Re: SFS Catalog Storage Pool
I am not just anyone. :-) I have fought my battles. Let someone who has skin in the game take up the banner. Regards, Richard Schuh > -Original Message- > From: The IBM z/VM Operating System > [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alan Altmark > Sent: Thursday, July 07, 2011 10:19 AM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: SFS Catalog Storage Pool > > On Thursday, 07/07/2011 at 12:35 EDT, "Schuh, Richard" > > wrote: > > Something ought to be done about the instructions regarding > MAXUSERS. > > A > few > > years ago, I had a small system that only had 64 users in > the directory. > > > It had performance problems (a complete freeze of any user > that tried > > to > use > > SFS) with MU of 100, 200, or 300. Upping it to 1000 fixed > the problem. > > > > And no, Alan or Mike, I will not open an incident or call > the Support > Center > > (Unless I call to say goodbye) :-) > > Anyone can send e-mail to mhvr...@us.ibm.com and comment on the pubs. > > Alan Altmark > > Senior Managing z/VM and Linux Consultant IBM System Lab > Services and Training ibm.com/systems/services/labservices > office: 607.429.3323 > mobile; 607.321.7556 > alan_altm...@us.ibm.com > IBM Endicott >
Re: SFS Catalog Storage Pool
On Thursday, 07/07/2011 at 12:35 EDT, "Schuh, Richard" wrote: > Something ought to be done about the instructions regarding MAXUSERS. A few > years ago, I had a small system that only had 64 users in the directory. > It had performance problems (a complete freeze of any user that tried to use > SFS) with MU of 100, 200, or 300. Upping it to 1000 fixed the problem. > > And no, Alan or Mike, I will not open an incident or call the Support Center > (Unless I call to say goodbye) :-) Anyone can send e-mail to mhvr...@us.ibm.com and comment on the pubs. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS Catalog Storage Pool
Something ought to be done about the instructions regarding MAXUSERS. A few years ago, I had a small system that only had 64 users in the directory. It had performance problems (a complete freeze of any user that tried to use SFS) with MU of 100, 200, or 300. Upping it to 1000 fixed the problem. And no, Alan or Mike, I will not open an incident or call the Support Center (Unless I call to say goodbye) :-) Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Kris Buelens Sent: Thursday, July 07, 2011 2:52 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS Catalog Storage Pool The physical space for the catalog is large enough, but SFS also has a notion of "logical space". SFS is derived from SQL/DS, remember DBEXTENTs and DBSPACEs). You need to run FILESERV REGENERATE and increase MAXUSERS and/or MAXDISKS, chapter 11 in the SFS admin manual. 2011/7/7 clifford jackson mailto:cliffordjackson...@msn.com>> I have a large SFS and monitoring my Catalog Space Information, I observed that my Percent used index blocks was at 97% I performed a control bata backup and done a FILESERV REORG. This in turn gave me 94% percent used index blocks. My Catalog storage pool 1 has the following specifications: MDISK STORAGE GROUP PERCENT Utilized 304 1 100% 310 1 20% 312 1 20% What do I need to do to get the percent used index blocks down and keep it to something manageable .. CliffJackson Senior Systems Programmer -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS Catalog Storage Pool
Thanks kris... Date: Thu, 7 Jul 2011 11:51:57 +0200 From: kris.buel...@gmail.com Subject: Re: SFS Catalog Storage Pool To: IBMVM@LISTSERV.UARK.EDU The physical space for the catalog is large enough, but SFS also has a notion of "logical space". SFS is derived from SQL/DS, remember DBEXTENTs and DBSPACEs). You need to run FILESERV REGENERATE and increase MAXUSERS and/or MAXDISKS, chapter 11 in the SFS admin manual. 2011/7/7 clifford jackson I have a large SFS and monitoring my Catalog Space Information, I observed that my Percent used index blocks was at 97% I performed a control bata backup and done a FILESERV REORG. This in turn gave me 94% percent used index blocks. My Catalog storage pool 1 has the following specifications: MDISK STORAGE GROUP PERCENT Utilized304 1 100%310 1 20%312 1 20% What do I need to do to get the percent used index blocks down and keep it to something manageable .. CliffJacksonSenior Systems Programmer -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS Catalog Storage Pool
The physical space for the catalog is large enough, but SFS also has a notion of "logical space". SFS is derived from SQL/DS, remember DBEXTENTs and DBSPACEs). You need to run FILESERV *RE*GENERATE and increase MAXUSERS and/or MAXDISKS, chapter 11 in the SFS admin manual. 2011/7/7 clifford jackson > I have a large SFS and monitoring my Catalog Space Information, I observed > that my Percent used index blocks was at 97% I performed a control bata > backup and done a FILESERV REORG. This in turn gave me 94% percent used > index blocks. My Catalog storage pool 1 has the following specifications: > > MDISK STORAGE GROUP PERCENT Utilized > 304 1 100% > 310 1 20% > 312 1 20% > > > What do I need to do to get the percent used index blocks down and keep it > to something manageable .. > > > > CliffJackson > Senior Systems Programmer > -- Kris Buelens, IBM Belgium, VM customer support
SFS Catalog Storage Pool
I have a large SFS and monitoring my Catalog Space Information, I observed that my Percent used index blocks was at 97% I performed a control bata backup and done a FILESERV REORG. This in turn gave me 94% percent used index blocks. My Catalog storage pool 1 has the following specifications: MDISK STORAGE GROUP PERCENT Utilized304 1 100%310 1 20%312 1 20% What do I need to do to get the percent used index blocks down and keep it to something manageable .. CliffJacksonSenior Systems Programmer
Re: SFS Possible problem
You have to redesign. The 16MB is indeed a restriction, and when you have really many files, an ACCESS of the dirid take ages. You could use a Directory Control Directory (DIRC) and place it in a dataspace, then the 16MB still applies, but you get the full 16MB (no space required for other resident stuff, like parts of the CMS nucleus). But, a DIRC can be accessed RW by only virtual machine at a time. If you place the DIRC in a dataspace, and if you assure always one user has it accessed, an ACCESS command is very fast, data and FST space are shared amongst all users. But, your users won't see new files until they reACCESS the DIRC, and if you have many updates throughout the day, it would cause many copies of the dataspace. For my former customer we had a structure like this: - a normal, small, DIR accessed as eg F, there updates during the day were placed - a big DIRC, with a dataspace accessed befind F. Each night, the files from the small DIR were moved into the DIRC. Then the service machine accessed it as RO, so we always had at least this one accessing the DIRC (and keeping the dataspace alive). Then users of the DIRC were recycled. We had some other service machine checking the number of dataspaces for that DIRC (using QUERY ACCESSORS dirid (DATASPACE and check the number of levels; if more than x, we sent out an email). A few months ago I redesigned a bad setup for another customer: they had two directories with batch job listings: 'this month' and 'previous month', with a monthly cleanup. But near the end of the month ACCESS became slow, sometimes impossible due to lack of storage. Now we have a dirid per day. Problems gone and much faster. To make the transition invisible, maybe the use of temporaly ALIASes can help (the code storing the files was changed at time x; the code reading the files still found its stuff but via ALIASes, and I had some extra day to fix the read code). 2011/4/28 clifford jackson > Hello All, I have a SFS data base with a Hierarchical directory structure > of File Control my tree structure currently has one top level directory, this > directory is used for report repository. This directory has grown very > large, I have an application that access this directory by customers (thru > worker machines) to view reports, I’ve run into a problem where it seems > that upon each request of the SFS data base for a report it seems the > complete top level directory is being loaded into the requesters virtual > machine storage and it seems there is a 16m line restriction. > > > > Is there a solution that I can implement without interfering with my > production environment too much or can I possibly redesign the SFS data > base structure and use Directory Control and possibly use Data Spaces. > > > > Cliff Jackson > > Senior Systems Programmer > -- Kris Buelens, IBM Belgium, VM customer support
SFS Possible problem
Hello All, I have a SFS data base with a Hierarchical directory structure of File Control my tree structure currently has one top level directory, this directory is used for report repository. This directory has grown very large, I have an application that access this directory by customers (thru worker machines) to view reports, I’ve run into a problem where it seems that upon each request of the SFS data base for a report it seems the complete top level directory is being loaded into the requesters virtual machine storage and it seems there is a 16m line restriction. Is there a solution that I can implement without interfering with my production environment too much or can I possibly redesign the SFS data base structure and use Directory Control and possibly use Data Spaces. Cliff Jackson Senior Systems Programmer
Re: SFS problem
Thank you, yes, that is correct... Apologies for mixing my terms. I should have been saying Alternate ID. John On Wed, Apr 20, 2011 at 12:33 PM, Schuh, Richard wrote: > Actually, SECUSER still exists and is something quite > different. SECUSER defines a secondary user for receiving console messages > and entering commands and responses. The entries are done as though they > came from, not on behalf of, the user. > > Regards, > Richard Schuh > > > > -- > *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On > Behalf Of *John Hall > *Sent:* Wednesday, April 20, 2011 7:04 AM > > *To:* IBMVM@LISTSERV.UARK.EDU > *Subject:* Re: SFS problem > > If I recall correctly, DIAG D4 is the one that manipulates the secondary > ID, aka "Alternate ID". (SECUSER is an old term). Diag 88 provides the > ability to link minidisks and perform userid/password validations (if the > issuer has appropriate authority). > > An interesting usage note for this is that DIAG D4 does not change the > userid associated with any existing (already active) SFS connections. This > is because it is a CP function and manipulates the VMDBK. Once set, future > connections (via APPC/VM Connect) will utilize the Alternate ID. ... This is > why severing all connections prior to setting the AltID will "fix" this type > of problem, because CMS will (re) connect and use the AltID. > > If this is Nora's problem, an easy work around would be wrap the job with > an exec that uses DMSGETWU and DMSPUSWU to set a new default work unit that > contains the appropriate user, then run the job from the exec, then finally > reset with DMSPOPWU. (If I'm remembering all of this correctly) > > John > > -- > John Hall > Safe Software, Inc. > 727-608-8799 > johnh...@safesoftware.com > > > > On Tue, Apr 19, 2011 at 11:48 AM, Schuh, Richard wrote: > >> Isn't that DIAG 88, instead of SECUSER? >> >> >> Regards, >> Richard Schuh >> >> >> >> >> -- >> *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On >> Behalf Of *John Hall >> *Sent:* Tuesday, April 19, 2011 6:41 AM >> >> *To:* IBMVM@LISTSERV.UARK.EDU >> *Subject:* Re: SFS problem >> >> Nora, >> Batch jobs normally run with the privileges of the "owner" of the job, >> using the SECUSER facility in z/VM. With SFS, this can lead to >> unexpected results when a prior batch job leaves the worker with a >> connection to the filepool under a different user's id. If the job >> ordering/selection of batch workers is somewhat random, you could see the >> outcome that you're experiencing (sometimes it works, sometimes it fails). >> >> >
Re: SFS problem
Actually, SECUSER still exists and is something quite different. SECUSER defines a secondary user for receiving console messages and entering commands and responses. The entries are done as though they came from, not on behalf of, the user. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Wednesday, April 20, 2011 7:04 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem If I recall correctly, DIAG D4 is the one that manipulates the secondary ID, aka "Alternate ID". (SECUSER is an old term). Diag 88 provides the ability to link minidisks and perform userid/password validations (if the issuer has appropriate authority). An interesting usage note for this is that DIAG D4 does not change the userid associated with any existing (already active) SFS connections. This is because it is a CP function and manipulates the VMDBK. Once set, future connections (via APPC/VM Connect) will utilize the Alternate ID. ... This is why severing all connections prior to setting the AltID will "fix" this type of problem, because CMS will (re) connect and use the AltID. If this is Nora's problem, an easy work around would be wrap the job with an exec that uses DMSGETWU and DMSPUSWU to set a new default work unit that contains the appropriate user, then run the job from the exec, then finally reset with DMSPOPWU. (If I'm remembering all of this correctly) John -- John Hall Safe Software, Inc. 727-608-8799 johnh...@safesoftware.com<mailto:johnh...@safesoftware.com> On Tue, Apr 19, 2011 at 11:48 AM, Schuh, Richard mailto:rsc...@visa.com>> wrote: Isn't that DIAG 88, instead of SECUSER? Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU<mailto:IBMVM@LISTSERV.UARK.EDU>] On Behalf Of John Hall Sent: Tuesday, April 19, 2011 6:41 AM To: IBMVM@LISTSERV.UARK.EDU<mailto:IBMVM@LISTSERV.UARK.EDU> Subject: Re: SFS problem Nora, Batch jobs normally run with the privileges of the "owner" of the job, using the SECUSER facility in z/VM. With SFS, this can lead to unexpected results when a prior batch job leaves the worker with a connection to the filepool under a different user's id. If the job ordering/selection of batch workers is somewhat random, you could see the outcome that you're experiencing (sometimes it works, sometimes it fails).
Re: SFS problem
Nora, Is there ever a case where the command (EXEC, or whatever) fails when it is NOT run in a VMBATCH worker? If they never fail outside of VMBATCH workers, John probably has a good handle on the diagnosis. You may have to read a little bit more in the VMBATCH documentation about how VMBATCH workers get initialized, end, and especially: how the post-job "cleanup" process works. It might be as simple as changing the program to release the SFS directory before ending (and perhaps examining the connection and looping until it has been cleared). BTW, the other common VMBATCH product is from CA, called VM:Batch. Many moons ago there were other publically available freeware VMBATCH solutions, too -- IIRC, from colleges. Mike Walter Aon Corporation The opinions expressed herein are mine alone, not my employer's. "Graves Nora E" Sent by: "The IBM z/VM Operating System" 04/20/2011 09:17 AM Please respond to "The IBM z/VM Operating System" To IBMVM@LISTSERV.UARK.EDU cc Subject Re: SFS problem I do not believe that this is the problem. I was giving you information on the failing job that I am most familiar with. Other jobs that fail are submitted to VMBatch by the ID that also owns the SFS directories. FYI, "VMBatch" is the IBM VM Batch Facility Version 2.2, I was not aware until recently that there are other products also known as "VMBatch". If this has caused confusion, I apologize. Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Wednesday, April 20, 2011 10:04 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem If I recall correctly, DIAG D4 is the one that manipulates the secondary ID, aka "Alternate ID". (SECUSER is an old term). Diag 88 provides the ability to link minidisks and perform userid/password validations (if the issuer has appropriate authority). An interesting usage note for this is that DIAG D4 does not change the userid associated with any existing (already active) SFS connections. This is because it is a CP function and manipulates the VMDBK. Once set, future connections (via APPC/VM Connect) will utilize the Alternate ID. ... This is why severing all connections prior to setting the AltID will "fix" this type of problem, because CMS will (re) connect and use the AltID. If this is Nora's problem, an easy work around would be wrap the job with an exec that uses DMSGETWU and DMSPUSWU to set a new default work unit that contains the appropriate user, then run the job from the exec, then finally reset with DMSPOPWU. (If I'm remembering all of this correctly) John -- John Hall Safe Software, Inc. 727-608-8799 johnh...@safesoftware.com On Tue, Apr 19, 2011 at 11:48 AM, Schuh, Richard wrote: Isn't that DIAG 88, instead of SECUSER? Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Tuesday, April 19, 2011 6:41 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Batch jobs normally run with the privileges of the "owner" of the job, using the SECUSER facility in z/VM. With SFS, this can lead to unexpected results when a prior batch job leaves the worker with a connection to the filepool under a different user's id. If the job ordering/selection of batch workers is somewhat random, you could see the outcome that you're experiencing (sometimes it works, sometimes it fails). The information contained in this e-mail and any accompanying documents may contain information that is confidential or otherwise protected from disclosure. If you are not the intended recipient of this message, or if this message has been addressed to you in error, please immediately alert the sender by reply e-mail and then delete this message, including any attachments. Any dissemination, distribution or other use of the contents of this message by anyone other than the intended recipient is strictly prohibited. All messages sent to and from this e-mail address may be monitored as permitted by applicable law and regulations to ensure compliance with our internal policies and to protect our business. E-mails are not secure and cannot be guaranteed to be error free as they can be intercepted, amended, lost or destroyed, or contain viruses. You are deemed to have accepted these risks if you communicate with us by e-mail.
Re: SFS problem
Interesting problem Nora. Nothing recorded in the SFS log either. Please post your POOLDEF statements and your DMSPARMS for the filepool having the problems. Also, what is the virtual storage size defined for the sfs server? Regards, Jim Hughes Consulting Systems Programmer Mainframe Technical Support Group Department of Information Technology State of New Hampshire 27 Hazen Drive Concord, NH 03301 603-271-5586Fax 603.271.1516 Statement of Confidentiality: The contents of this message are confidential. Any unauthorized disclosure, reproduction, use or dissemination (either whole or in part) is prohibited. If you are not the intended recipient of this message, please notify the sender immediately and delete the message from your system. From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Graves Nora E Sent: Wednesday, April 20, 2011 10:17 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem I do not believe that this is the problem. I was giving you information on the failing job that I am most familiar with. Other jobs that fail are submitted to VMBatch by the ID that also owns the SFS directories. FYI, "VMBatch" is the IBM VM Batch Facility Version 2.2, I was not aware until recently that there are other products also known as "VMBatch". If this has caused confusion, I apologize. Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Wednesday, April 20, 2011 10:04 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem If I recall correctly, DIAG D4 is the one that manipulates the secondary ID, aka "Alternate ID". (SECUSER is an old term). Diag 88 provides the ability to link minidisks and perform userid/password validations (if the issuer has appropriate authority). An interesting usage note for this is that DIAG D4 does not change the userid associated with any existing (already active) SFS connections. This is because it is a CP function and manipulates the VMDBK. Once set, future connections (via APPC/VM Connect) will utilize the Alternate ID. ... This is why severing all connections prior to setting the AltID will "fix" this type of problem, because CMS will (re) connect and use the AltID. If this is Nora's problem, an easy work around would be wrap the job with an exec that uses DMSGETWU and DMSPUSWU to set a new default work unit that contains the appropriate user, then run the job from the exec, then finally reset with DMSPOPWU. (If I'm remembering all of this correctly) John -- John Hall Safe Software, Inc. 727-608-8799 johnh...@safesoftware.com On Tue, Apr 19, 2011 at 11:48 AM, Schuh, Richard wrote: Isn't that DIAG 88, instead of SECUSER? Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Tuesday, April 19, 2011 6:41 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Batch jobs normally run with the privileges of the "owner" of the job, using the SECUSER facility in z/VM. With SFS, this can lead to unexpected results when a prior batch job leaves the worker with a connection to the filepool under a different user's id. If the job ordering/selection of batch workers is somewhat random, you could see the outcome that you're experiencing (sometimes it works, sometimes it fails).
Re: SFS problem
I do not believe that this is the problem. I was giving you information on the failing job that I am most familiar with. Other jobs that fail are submitted to VMBatch by the ID that also owns the SFS directories. FYI, "VMBatch" is the IBM VM Batch Facility Version 2.2, I was not aware until recently that there are other products also known as "VMBatch". If this has caused confusion, I apologize. Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Wednesday, April 20, 2011 10:04 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem If I recall correctly, DIAG D4 is the one that manipulates the secondary ID, aka "Alternate ID". (SECUSER is an old term). Diag 88 provides the ability to link minidisks and perform userid/password validations (if the issuer has appropriate authority). An interesting usage note for this is that DIAG D4 does not change the userid associated with any existing (already active) SFS connections. This is because it is a CP function and manipulates the VMDBK. Once set, future connections (via APPC/VM Connect) will utilize the Alternate ID. ... This is why severing all connections prior to setting the AltID will "fix" this type of problem, because CMS will (re) connect and use the AltID. If this is Nora's problem, an easy work around would be wrap the job with an exec that uses DMSGETWU and DMSPUSWU to set a new default work unit that contains the appropriate user, then run the job from the exec, then finally reset with DMSPOPWU. (If I'm remembering all of this correctly) John -- John Hall Safe Software, Inc. 727-608-8799 johnh...@safesoftware.com On Tue, Apr 19, 2011 at 11:48 AM, Schuh, Richard wrote: Isn't that DIAG 88, instead of SECUSER? Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Tuesday, April 19, 2011 6:41 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Batch jobs normally run with the privileges of the "owner" of the job, using the SECUSER facility in z/VM. With SFS, this can lead to unexpected results when a prior batch job leaves the worker with a connection to the filepool under a different user's id. If the job ordering/selection of batch workers is somewhat random, you could see the outcome that you're experiencing (sometimes it works, sometimes it fails).
Re: SFS problem
If I recall correctly, DIAG D4 is the one that manipulates the secondary ID, aka "Alternate ID". (SECUSER is an old term). Diag 88 provides the ability to link minidisks and perform userid/password validations (if the issuer has appropriate authority). An interesting usage note for this is that DIAG D4 does not change the userid associated with any existing (already active) SFS connections. This is because it is a CP function and manipulates the VMDBK. Once set, future connections (via APPC/VM Connect) will utilize the Alternate ID. ... This is why severing all connections prior to setting the AltID will "fix" this type of problem, because CMS will (re) connect and use the AltID. If this is Nora's problem, an easy work around would be wrap the job with an exec that uses DMSGETWU and DMSPUSWU to set a new default work unit that contains the appropriate user, then run the job from the exec, then finally reset with DMSPOPWU. (If I'm remembering all of this correctly) John -- John Hall Safe Software, Inc. 727-608-8799 johnh...@safesoftware.com On Tue, Apr 19, 2011 at 11:48 AM, Schuh, Richard wrote: > Isn't that DIAG 88, instead of SECUSER? > > > Regards, > Richard Schuh > > > > > -- > *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On > Behalf Of *John Hall > *Sent:* Tuesday, April 19, 2011 6:41 AM > > *To:* IBMVM@LISTSERV.UARK.EDU > *Subject:* Re: SFS problem > > Nora, > Batch jobs normally run with the privileges of the "owner" of the job, > using the SECUSER facility in z/VM. With SFS, this can lead to unexpected > results when a prior batch job leaves the worker with a connection to the > filepool under a different user's id. If the job ordering/selection of > batch workers is somewhat random, you could see the outcome that you're > experiencing (sometimes it works, sometimes it fails). > >
Re: SFS problem
Just to exhaust the possibilities in this line, we have had occasions when a job would fail for another reason, then be re-submitted by another userid without the required authority. On Tue, Apr 19, 2011 at 7:47 AM, Graves Nora E wrote: > Kris, > > The submitter is ABC, the directory owner is XYZ. The submitter has > WRITE/NEWRITE authority to the directory and files. > > Nora Graves > nora.e.gra...@irs.gov > Main IRS, Room 6531 > (202) 622-6735 > Fax (202) 622-3123 > SE:W:CAR:MP:D:KS:BRSI > > > -- > *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On > Behalf Of *Kris Buelens > *Sent:* Thursday, April 14, 2011 4:58 PM > > *To:* IBMVM@LISTSERV.UARK.EDU > *Subject:* Re: SFS problem > > Note: if you use VMBATCH, the worker machine connects to SFS with the > authority of the job submitter. > You say "This user is the owner of most of the directories" You mean: the > submitter is userid ABC, the dirids are all named "fpoolid:ABC.something"? > > > 2011/4/14 Graves Nora E > >> We are having an intermittent problem with SFS and I'm hoping someone >> may have some ideas of what to pursue next. >> >> We have several batch jobs that run under VMBATCH overnight. Sometimes >> they are not able to create a file in a directory, even though most times it >> is successful. The only differences in the executions are the file names; >> for many of these the File Type is the date. >> >> In the job I am most familiar with, these are the specifics. >> >> The job runs Monday-Saturday. This year, it has failed on January 4, >> January 12, February 9, March 18, March 25, and April 13. It has run >> successfully the other days. Other than the QUERY statements below, it has >> not changed. >> The job runs in a work machine, WORK7. >> The job is submitted by the User ID of the database owner. >> The SFS directories are owned by a 3rd user. Failures occur in many of >> the subdirectories, not just one subdirectory owned by this user. This user >> is the owner of most of the directories containing the data files we create >> in batch, so I don't think it's significant that it's the ID that has the >> problem. However, as far as I know, it is the only ID that does have the >> problem. >> This job uses VMLINK to acquire a write link to SFS directory. This >> always looks to be successful--no error is given. (Other jobs use GETFMADDR >> and ACCESS to acquire the write link to the directory. This always >> appears successful as well). >> Once the file is ready to be copied from the Work Machine's 191 disk to >> the SFS directory, the intermittent error appears. The vast majority of the >> time, the write is successful. However, sometimes, the job gets this error >> message: >> DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 >> >> The file is not large--last night's file was only 12 blocks. >> >> At the suggestion of our systems programmer, I've put in a lot of query >> statements. I've issued QUERY LIMITS for the job submitter; it's only used >> 84% of the allocation, with several thousand blocks available. The SFS >> directory owner has only used 76% of its allocation, with several thousand >> more blocks still available. The filepool is not full. >> >> I've issued QUERY FILEPOOL CONFLICT. There is no conflict. >> >> I've issued QUERY ACCESSED. The directory shows that is accessed R/W. >> >> When the write is unsuccessful, the program then loops through 5 tries of >> releasing the access, reacquiring the access, and attempting to write the >> file again. This has never been successful. I've issued both a COPYFILE >> and a PIPE to try to write the file; these do not work once there has been a >> failure. >> >> We've looked at the operator consoles to see if we can find any jobs >> running at the same time. We haven't found any that are accessing that >> directory structure. >> >> There aren't any dumps to look at--it looks perfectly successful other >> than the fact that it won't write the file. >> >> Does anyone have any suggestions of something to try next? >> >> >> Nora Graves >> nora.e.gra...@irs.gov >> >> > > > > -- > Kris Buelens, > IBM Belgium, VM customer support >
Re: SFS problem
Isn't that DIAG 88, instead of SECUSER? Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of John Hall Sent: Tuesday, April 19, 2011 6:41 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Batch jobs normally run with the privileges of the "owner" of the job, using the SECUSER facility in z/VM. With SFS, this can lead to unexpected results when a prior batch job leaves the worker with a connection to the filepool under a different user's id. If the job ordering/selection of batch workers is somewhat random, you could see the outcome that you're experiencing (sometimes it works, sometimes it fails).
Re: SFS problem
Nora, Batch jobs normally run with the privileges of the "owner" of the job, using the SECUSER facility in z/VM. With SFS, this can lead to unexpected results when a prior batch job leaves the worker with a connection to the filepool under a different user's id. If the job ordering/selection of batch workers is somewhat random, you could see the outcome that you're experiencing (sometimes it works, sometimes it fails). For example: If WORK7 first runs a job for "JOHN" and JOHN utilizes the same filepool, WORK7 could be left with a connection to the filepool that appears to be from the user "JOHN". If your failing job runs next, it would likely end up accessing the filepool as JOHN, not as the actual owner, which could cause the failure. If on another night, WORK7 is recycled before your failing job runs, then it would establish a connection to the file pool with the owner's ID (say OWNER) and then would run with the access rights of OWNER, and you'd see success. If this is your problem, there are many ways to solve it. The easiest would be to ensure that worker machines are recycled (logoff/xautolog or IPL CMS) prior to the job running. I wonder if there are VMBATCH options (system or job) that provide the ability to do a more advanced "clean up" of a worker prior to a job run. John -- John Hall Safe Software, Inc. 727-608-8799 johnh...@safesoftware.com On Thu, Apr 14, 2011 at 3:45 PM, Graves Nora E wrote: > We are having an intermittent problem with SFS and I'm hoping someone may > have some ideas of what to pursue next. > > We have several batch jobs that run under VMBATCH overnight. Sometimes > they are not able to create a file in a directory, even though most times it > is successful. The only differences in the executions are the file names; > for many of these the File Type is the date. > > In the job I am most familiar with, these are the specifics. > > The job runs Monday-Saturday. This year, it has failed on January 4, > January 12, February 9, March 18, March 25, and April 13. It has run > successfully the other days. Other than the QUERY statements below, it has > not changed. > The job runs in a work machine, WORK7. > The job is submitted by the User ID of the database owner. > The SFS directories are owned by a 3rd user. Failures occur in many of the > subdirectories, not just one subdirectory owned by this user. This user is > the owner of most of the directories containing the data files we create in > batch, so I don't think it's significant that it's the ID that has the > problem. However, as far as I know, it is the only ID that does have the > problem. > This job uses VMLINK to acquire a write link to SFS directory. This always > looks to be successful--no error is given. (Other jobs use GETFMADDR and > ACCESS to acquire the write link to the directory. This always > appears successful as well). > Once the file is ready to be copied from the Work Machine's 191 disk to the > SFS directory, the intermittent error appears. The vast majority of the > time, the write is successful. However, sometimes, the job gets this error > message: > DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 > > The file is not large--last night's file was only 12 blocks. > > At the suggestion of our systems programmer, I've put in a lot of query > statements. I've issued QUERY LIMITS for the job submitter; it's only used > 84% of the allocation, with several thousand blocks available. The SFS > directory owner has only used 76% of its allocation, with several thousand > more blocks still available. The filepool is not full. > > I've issued QUERY FILEPOOL CONFLICT. There is no conflict. > > I've issued QUERY ACCESSED. The directory shows that is accessed R/W. > > When the write is unsuccessful, the program then loops through 5 tries of > releasing the access, reacquiring the access, and attempting to write the > file again. This has never been successful. I've issued both a COPYFILE > and a PIPE to try to write the file; these do not work once there has been a > failure. > > We've looked at the operator consoles to see if we can find any jobs > running at the same time. We haven't found any that are accessing that > directory structure. > > There aren't any dumps to look at--it looks perfectly successful other than > the fact that it won't write the file. > > Does anyone have any suggestions of something to try next? > > > Nora Graves > nora.e.gra...@irs.gov > >
Re: SFS problem
Jim, There are basically no messages at all in the console for the SFS service machine. One of the failures was on April 13. This is the console for that day--no editing at all of the contents. HCPMID6001I TIME IS 00:00:00 EDT WEDNESDAY 04/13/11 00:00:00 HCPMID6001I TIME IS 00:00:00 EDT THURSDAY 04/14/11 00:00:00 Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Hughes, Jim Sent: Thursday, April 14, 2011 5:23 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Have you looked at the SFS service machine's console log for any odd messages? Could the SFS server being doing some sort of control backup? Jim Hughes Consulting Systems Programmer Mainframe Technical Support Group Department of Information Technology State of New Hampshire 27 Hazen Drive Concord, NH 03301 603-271-5586Fax 603.271.1516 Statement of Confidentiality: The contents of this message are confidential. Any unauthorized disclosure, reproduction, use or dissemination (either whole or in part) is prohibited. If you are not the intended recipient of this message, please notify the sender immediately and delete the message from your system.
Re: SFS problem
I've had similar issues before. There are really two choices you can make. The first is to open a PMR and do traces and such. Since it's a very intermittent failure, this can be difficult. The other is to make the user doing the writes a filepool administrator and get on with life. Mgmt chose the latter. -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Graves Nora E Sent: Tuesday, April 19, 2011 7:43 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Mike, Yes, I checked out the message. The thing is, the job runs nightly, and only fails intermittently. There is no reason to believe that the authority to access the directory is changing during that time period; there are only 2 people who regularly grant permissions to that user's directory. Neither of us has made a change. Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Mike Walter Sent: Thursday, April 14, 2011 4:13 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Did you issue: HELP DMSOPN1258E The response from the msg indicates an SFS authorization problem.: -- (c) Copyright IBM Corporation 1990, 2008 DMS1258E to write (to) file Explanation: You attempted to write to a file for which you do not have write authority, or you attempted to create a new file in a directory for which you do not have write authority. System Action: RC=12 or 28. Execution of the command is terminated. User Response: Ensure that you specified the correct file. If so, contact the owner to gain proper authorization to the file or directory. If the file specified is an alias, you may enter the QUERY ALIAS command to determine the owner of the base file. For a BFS file, enter the OPENVM LISTFILE command with the OWNER option for the file. This will tell you the owning user ID and group name for the file you wish to use. Refer to the z/VM: OpenExtensions Command Reference or enter HELP OPENVM for more information on the OPENVM commands. -- z/VM has a terrific HELP facility (including most messages), it's worth investing a bit of time to familiarize yourself with it. Were you using the VMLINK command in every case with the "(FORCERW" option? Regardless, check the syntax for your SFS authorizations. And if all those bazillion intricate SFS authorizations become overwhelming, consider checking out SafeSFS from safesfs.com. We're vary satisfied customers, it makes authorizations much MUCH simpler, saves huge amounts of effort and back time. Mike Walter Aon Corporation The opinions expressed herein are mine alone, not my employer's. "Graves Nora E" Sent by: "The IBM z/VM Operating System" 04/14/2011 02:45 PM Please respond to "The IBM z/VM Operating System" To IBMVM@LISTSERV.UARK.EDU cc Subject SFS problem We are having an intermittent problem with SFS and I'm hoping someone may have some ideas of what to pursue next. We have several batch jobs that run under VMBATCH overnight. Sometimes they are not able to create a file in a directory, even though most times it is successful. The only differences in the executions are the file names; for many of these the File Type is the date. In the job I am most familiar with, these are the specifics. The job runs Monday-Saturday. This year, it has failed on January 4, January 12, February 9, March 18, March 25, and April 13. It has run successfully the other days. Other than the QUERY statements below, it has not changed. The job runs in a work machine, WORK7. The job is submitted by the User ID of the database owner. The SFS directories are owned by a 3rd user. Failures occur in many of the subdirectories, not just one subdirectory owned by this user. This user is the owner of most of the directories containing the data files we create in batch, so I don't think it's significant that it's the ID that has the problem. However, as far as I know, it is the only ID that does have the problem. This job uses VMLINK to acquire a write link to SFS directory. This always looks to be successful--no error is given. (Other jobs use GETFMADDR and ACCESS to acquire the write link to the directory. This always appears successful as well). Once the file is ready to be copied from the Work Machine's 191 disk to the SFS directory, the intermittent error appears. The vast majority of the time, the write is successful. However, sometimes, the job gets this error message: DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 The file is not large--last night's file was only 12 blocks. At the suggestion of our systems programmer, I'
Re: SFS problem
Jim, I didn't think to look there, I will. Our regular backups are VMBACKUP, which is scheduled to run 90 minutes after this job runs. This job runs only 2 minutes maximum, so I know that isn't the problem. But the SFS console may have some good indications. I'll see what I can find there. Thanks, Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Hughes, Jim Sent: Thursday, April 14, 2011 5:23 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Have you looked at the SFS service machine's console log for any odd messages? Could the SFS server being doing some sort of control backup? Jim Hughes Consulting Systems Programmer Mainframe Technical Support Group Department of Information Technology State of New Hampshire 27 Hazen Drive Concord, NH 03301 603-271-5586Fax 603.271.1516 Statement of Confidentiality: The contents of this message are confidential. Any unauthorized disclosure, reproduction, use or dissemination (either whole or in part) is prohibited. If you are not the intended recipient of this message, please notify the sender immediately and delete the message from your system.
Re: SFS problem
Kris, The submitter is ABC, the directory owner is XYZ. The submitter has WRITE/NEWRITE authority to the directory and files. Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Kris Buelens Sent: Thursday, April 14, 2011 4:58 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Note: if you use VMBATCH, the worker machine connects to SFS with the authority of the job submitter. You say "This user is the owner of most of the directories" You mean: the submitter is userid ABC, the dirids are all named "fpoolid:ABC.something"? 2011/4/14 Graves Nora E We are having an intermittent problem with SFS and I'm hoping someone may have some ideas of what to pursue next. We have several batch jobs that run under VMBATCH overnight. Sometimes they are not able to create a file in a directory, even though most times it is successful. The only differences in the executions are the file names; for many of these the File Type is the date. In the job I am most familiar with, these are the specifics. The job runs Monday-Saturday. This year, it has failed on January 4, January 12, February 9, March 18, March 25, and April 13. It has run successfully the other days. Other than the QUERY statements below, it has not changed. The job runs in a work machine, WORK7. The job is submitted by the User ID of the database owner. The SFS directories are owned by a 3rd user. Failures occur in many of the subdirectories, not just one subdirectory owned by this user. This user is the owner of most of the directories containing the data files we create in batch, so I don't think it's significant that it's the ID that has the problem. However, as far as I know, it is the only ID that does have the problem. This job uses VMLINK to acquire a write link to SFS directory. This always looks to be successful--no error is given. (Other jobs use GETFMADDR and ACCESS to acquire the write link to the directory. This always appears successful as well). Once the file is ready to be copied from the Work Machine's 191 disk to the SFS directory, the intermittent error appears. The vast majority of the time, the write is successful. However, sometimes, the job gets this error message: DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 The file is not large--last night's file was only 12 blocks. At the suggestion of our systems programmer, I've put in a lot of query statements. I've issued QUERY LIMITS for the job submitter; it's only used 84% of the allocation, with several thousand blocks available. The SFS directory owner has only used 76% of its allocation, with several thousand more blocks still available. The filepool is not full. I've issued QUERY FILEPOOL CONFLICT. There is no conflict. I've issued QUERY ACCESSED. The directory shows that is accessed R/W. When the write is unsuccessful, the program then loops through 5 tries of releasing the access, reacquiring the access, and attempting to write the file again. This has never been successful. I've issued both a COPYFILE and a PIPE to try to write the file; these do not work once there has been a failure. We've looked at the operator consoles to see if we can find any jobs running at the same time. We haven't found any that are accessing that directory structure. There aren't any dumps to look at--it looks perfectly successful other than the fact that it won't write the file. Does anyone have any suggestions of something to try next? Nora Graves nora.e.gra...@irs.gov -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS problem
Alan, The submitter does have NEWRITE permissions. And nothing in that job erases the file. The job runs only once per day, and the File Type is the date, such as 20110419. However, I didn't know that authority to write the file disappeared with the file if NEWRITE wasn't granted, so I learned something! Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alan Altmark Sent: Thursday, April 14, 2011 4:52 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem On Thursday, 04/14/2011 at 03:47 EDT, Graves Nora E wrote: > DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 : > I've issued QUERY ACCESSED. The directory shows that is accessed R/W. > > When the write is unsuccessful, the program then loops through 5 tries of > releasing the access, reacquiring the access, and attempting to write the file > again. This has never been successful. I've issued both a COPYFILE and a > PIPE to try to write the file; these do not work once there has been a failure. > > We've looked at the operator consoles to see if we can find any jobs running > at the same time. We haven't found any that are accessing that directory > structure. > > There aren't any dumps to look at--it looks perfectly successful other than > the fact that it won't write the file. > > Does anyone have any suggestions of something to try next? As you can see, it's an authorization error. Does the worker/submittor have NEWRITE authority to the directory? If not, and some part of the job ERASEs the file from the directory and then tries to recreate it, it will fail, as the authority to the file is deleted if the file itself is deleted. In a FILECONTROL directory, R/W access to the directory does not imply R/W access to all the files in it! Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS problem
Mike, Yes, I checked out the message. The thing is, the job runs nightly, and only fails intermittently. There is no reason to believe that the authority to access the directory is changing during that time period; there are only 2 people who regularly grant permissions to that user's directory. Neither of us has made a change. Nora Graves nora.e.gra...@irs.gov Main IRS, Room 6531 (202) 622-6735 Fax (202) 622-3123 SE:W:CAR:MP:D:KS:BRSI -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Mike Walter Sent: Thursday, April 14, 2011 4:13 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS problem Nora, Did you issue: HELP DMSOPN1258E The response from the msg indicates an SFS authorization problem.: -- (c) Copyright IBM Corporation 1990, 2008 DMS1258E to write (to) file Explanation: You attempted to write to a file for which you do not have write authority, or you attempted to create a new file in a directory for which you do not have write authority. System Action: RC=12 or 28. Execution of the command is terminated. User Response: Ensure that you specified the correct file. If so, contact the owner to gain proper authorization to the file or directory. If the file specified is an alias, you may enter the QUERY ALIAS command to determine the owner of the base file. For a BFS file, enter the OPENVM LISTFILE command with the OWNER option for the file. This will tell you the owning user ID and group name for the file you wish to use. Refer to the z/VM: OpenExtensions Command Reference or enter HELP OPENVM for more information on the OPENVM commands. -- z/VM has a terrific HELP facility (including most messages), it's worth investing a bit of time to familiarize yourself with it. Were you using the VMLINK command in every case with the "(FORCERW" option? Regardless, check the syntax for your SFS authorizations. And if all those bazillion intricate SFS authorizations become overwhelming, consider checking out SafeSFS from safesfs.com. We're vary satisfied customers, it makes authorizations much MUCH simpler, saves huge amounts of effort and back time. Mike Walter Aon Corporation The opinions expressed herein are mine alone, not my employer's. "Graves Nora E" Sent by: "The IBM z/VM Operating System" 04/14/2011 02:45 PM Please respond to "The IBM z/VM Operating System" To IBMVM@LISTSERV.UARK.EDU cc Subject SFS problem We are having an intermittent problem with SFS and I'm hoping someone may have some ideas of what to pursue next. We have several batch jobs that run under VMBATCH overnight. Sometimes they are not able to create a file in a directory, even though most times it is successful. The only differences in the executions are the file names; for many of these the File Type is the date. In the job I am most familiar with, these are the specifics. The job runs Monday-Saturday. This year, it has failed on January 4, January 12, February 9, March 18, March 25, and April 13. It has run successfully the other days. Other than the QUERY statements below, it has not changed. The job runs in a work machine, WORK7. The job is submitted by the User ID of the database owner. The SFS directories are owned by a 3rd user. Failures occur in many of the subdirectories, not just one subdirectory owned by this user. This user is the owner of most of the directories containing the data files we create in batch, so I don't think it's significant that it's the ID that has the problem. However, as far as I know, it is the only ID that does have the problem. This job uses VMLINK to acquire a write link to SFS directory. This always looks to be successful--no error is given. (Other jobs use GETFMADDR and ACCESS to acquire the write link to the directory. This always appears successful as well). Once the file is ready to be copied from the Work Machine's 191 disk to the SFS directory, the intermittent error appears. The vast majority of the time, the write is successful. However, sometimes, the job gets this error message: DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 The file is not large--last night's file was only 12 blocks. At the suggestion of our systems programmer, I've put in a lot of query statements. I've issued QUERY LIMITS for the job submitter; it's only used 84% of the allocation, with several thousand blocks available. The SFS directory owner has only used 76% of its allocation, with several thousand more blocks still available. The filepool is not full. I've issued QUERY FILEPOOL CONFLICT. There is no conflict. I've issued QUERY ACCESSED. The directory shows that is accessed R/W. When the write is unsuccessful, the program then loops through 5 tr
Re: SFS problem
Nora, Have you looked at the SFS service machine's console log for any odd messages? Could the SFS server being doing some sort of control backup? Jim Hughes Consulting Systems Programmer Mainframe Technical Support Group Department of Information Technology State of New Hampshire 27 Hazen Drive Concord, NH 03301 603-271-5586Fax 603.271.1516 Statement of Confidentiality: The contents of this message are confidential. Any unauthorized disclosure, reproduction, use or dissemination (either whole or in part) is prohibited. If you are not the intended recipient of this message, please notify the sender immediately and delete the message from your system.
Re: SFS problem
Note: if you use VMBATCH, the worker machine connects to SFS with the authority of the job submitter. You say "This user is the owner of most of the directories" You mean: the submitter is userid ABC, the dirids are all named "fpoolid:ABC.something"? 2011/4/14 Graves Nora E > We are having an intermittent problem with SFS and I'm hoping someone may > have some ideas of what to pursue next. > > We have several batch jobs that run under VMBATCH overnight. Sometimes > they are not able to create a file in a directory, even though most times it > is successful. The only differences in the executions are the file names; > for many of these the File Type is the date. > > In the job I am most familiar with, these are the specifics. > > The job runs Monday-Saturday. This year, it has failed on January 4, > January 12, February 9, March 18, March 25, and April 13. It has run > successfully the other days. Other than the QUERY statements below, it has > not changed. > The job runs in a work machine, WORK7. > The job is submitted by the User ID of the database owner. > The SFS directories are owned by a 3rd user. Failures occur in many of the > subdirectories, not just one subdirectory owned by this user. This user is > the owner of most of the directories containing the data files we create in > batch, so I don't think it's significant that it's the ID that has the > problem. However, as far as I know, it is the only ID that does have the > problem. > This job uses VMLINK to acquire a write link to SFS directory. This always > looks to be successful--no error is given. (Other jobs use GETFMADDR and > ACCESS to acquire the write link to the directory. This always > appears successful as well). > Once the file is ready to be copied from the Work Machine's 191 disk to the > SFS directory, the intermittent error appears. The vast majority of the > time, the write is successful. However, sometimes, the job gets this error > message: > DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 > > The file is not large--last night's file was only 12 blocks. > > At the suggestion of our systems programmer, I've put in a lot of query > statements. I've issued QUERY LIMITS for the job submitter; it's only used > 84% of the allocation, with several thousand blocks available. The SFS > directory owner has only used 76% of its allocation, with several thousand > more blocks still available. The filepool is not full. > > I've issued QUERY FILEPOOL CONFLICT. There is no conflict. > > I've issued QUERY ACCESSED. The directory shows that is accessed R/W. > > When the write is unsuccessful, the program then loops through 5 tries of > releasing the access, reacquiring the access, and attempting to write the > file again. This has never been successful. I've issued both a COPYFILE > and a PIPE to try to write the file; these do not work once there has been a > failure. > > We've looked at the operator consoles to see if we can find any jobs > running at the same time. We haven't found any that are accessing that > directory structure. > > There aren't any dumps to look at--it looks perfectly successful other than > the fact that it won't write the file. > > Does anyone have any suggestions of something to try next? > > > Nora Graves > nora.e.gra...@irs.gov > > -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS problem
On Thursday, 04/14/2011 at 03:47 EDT, Graves Nora E wrote: > DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 : > I've issued QUERY ACCESSED. The directory shows that is accessed R/W. > > When the write is unsuccessful, the program then loops through 5 tries of > releasing the access, reacquiring the access, and attempting to write the file > again. This has never been successful. I've issued both a COPYFILE and a > PIPE to try to write the file; these do not work once there has been a failure. > > We've looked at the operator consoles to see if we can find any jobs running > at the same time. We haven't found any that are accessing that directory > structure. > > There aren't any dumps to look at--it looks perfectly successful other than > the fact that it won't write the file. > > Does anyone have any suggestions of something to try next? As you can see, it's an authorization error. Does the worker/submittor have NEWRITE authority to the directory? If not, and some part of the job ERASEs the file from the directory and then tries to recreate it, it will fail, as the authority to the file is deleted if the file itself is deleted. In a FILECONTROL directory, R/W access to the directory does not imply R/W access to all the files in it! Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS problem
Nora, Did you issue: HELP DMSOPN1258E The response from the msg indicates an SFS authorization problem.: -- (c) Copyright IBM Corporation 1990, 2008 DMS1258E to write (to) file Explanation: You attempted to write to a file for which you do not have write authority, or you attempted to create a new file in a directory for which you do not have write authority. System Action: RC=12 or 28. Execution of the command is terminated. User Response: Ensure that you specified the correct file. If so, contact the owner to gain proper authorization to the file or directory. If the file specified is an alias, you may enter the QUERY ALIAS command to determine the owner of the base file. For a BFS file, enter the OPENVM LISTFILE command with the OWNER option for the file. This will tell you the owning user ID and group name for the file you wish to use. Refer to the z/VM: OpenExtensions Command Reference or enter HELP OPENVM for more information on the OPENVM commands. -- z/VM has a terrific HELP facility (including most messages), it's worth investing a bit of time to familiarize yourself with it. Were you using the VMLINK command in every case with the "(FORCERW" option? Regardless, check the syntax for your SFS authorizations. And if all those bazillion intricate SFS authorizations become overwhelming, consider checking out SafeSFS from safesfs.com. We're vary satisfied customers, it makes authorizations much MUCH simpler, saves huge amounts of effort and back time. Mike Walter Aon Corporation The opinions expressed herein are mine alone, not my employer's. "Graves Nora E" Sent by: "The IBM z/VM Operating System" 04/14/2011 02:45 PM Please respond to "The IBM z/VM Operating System" To IBMVM@LISTSERV.UARK.EDU cc Subject SFS problem We are having an intermittent problem with SFS and I'm hoping someone may have some ideas of what to pursue next. We have several batch jobs that run under VMBATCH overnight. Sometimes they are not able to create a file in a directory, even though most times it is successful. The only differences in the executions are the file names; for many of these the File Type is the date. In the job I am most familiar with, these are the specifics. The job runs Monday-Saturday. This year, it has failed on January 4, January 12, February 9, March 18, March 25, and April 13. It has run successfully the other days. Other than the QUERY statements below, it has not changed. The job runs in a work machine, WORK7. The job is submitted by the User ID of the database owner. The SFS directories are owned by a 3rd user. Failures occur in many of the subdirectories, not just one subdirectory owned by this user. This user is the owner of most of the directories containing the data files we create in batch, so I don't think it's significant that it's the ID that has the problem. However, as far as I know, it is the only ID that does have the problem. This job uses VMLINK to acquire a write link to SFS directory. This always looks to be successful--no error is given. (Other jobs use GETFMADDR and ACCESS to acquire the write link to the directory. This always appears successful as well). Once the file is ready to be copied from the Work Machine's 191 disk to the SFS directory, the intermittent error appears. The vast majority of the time, the write is successful. However, sometimes, the job gets this error message: DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 The file is not large--last night's file was only 12 blocks. At the suggestion of our systems programmer, I've put in a lot of query statements. I've issued QUERY LIMITS for the job submitter; it's only used 84% of the allocation, with several thousand blocks available. The SFS directory owner has only used 76% of its allocation, with several thousand more blocks still available. The filepool is not full. I've issued QUERY FILEPOOL CONFLICT. There is no conflict. I've issued QUERY ACCESSED. The directory shows that is accessed R/W. When the write is unsuccessful, the program then loops through 5 tries of releasing the access, reacquiring the access, and attempting to write the file again. This has never been successful. I've issued both a COPYFILE and a PIPE to try to write the file; these do not work once there has been a failure. We've looked at the operator consoles to see if we can find any jobs running at the same time. We haven't found any that are accessing that directory structure. There aren't any dumps to look at--it looks perfectly successful other than the fact that it won't write the file. Does anyone have any suggestions of something to try next? Nora Graves nora.e.gra...@irs.gov The inform
Re: SFS problem
Would issuing a QUERY AUTHORITY at the time of the error give anything useful, I wonder. Maybe try for both the directory and the file? Peter -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Graves Nora E Sent: April 14, 2011 15:46 To: IBMVM@LISTSERV.UARK.EDU Subject: SFS problem We are having an intermittent problem with SFS and I'm hoping someone may have some ideas of what to pursue next. We have several batch jobs that run under VMBATCH overnight. Sometimes they are not able to create a file in a directory, even though most times it is successful. The only differences in the executions are the file names; for many of these the File Type is the date. In the job I am most familiar with, these are the specifics. The job runs Monday-Saturday. This year, it has failed on January 4, January 12, February 9, March 18, March 25, and April 13. It has run successfully the other days. Other than the QUERY statements below, it has not changed. The job runs in a work machine, WORK7. The job is submitted by the User ID of the database owner. The SFS directories are owned by a 3rd user. Failures occur in many of the subdirectories, not just one subdirectory owned by this user. This user is the owner of most of the directories containing the data files we create in batch, so I don't think it's significant that it's the ID that has the problem. However, as far as I know, it is the only ID that does have the problem. This job uses VMLINK to acquire a write link to SFS directory. This always looks to be successful--no error is given. (Other jobs use GETFMADDR and ACCESS to acquire the write link to the directory. This always appears successful as well). Once the file is ready to be copied from the Work Machine's 191 disk to the SFS directory, the intermittent error appears. The vast majority of the time, the write is successful. However, sometimes, the job gets this error message: DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 The file is not large--last night's file was only 12 blocks. At the suggestion of our systems programmer, I've put in a lot of query statements. I've issued QUERY LIMITS for the job submitter; it's only used 84% of the allocation, with several thousand blocks available. The SFS directory owner has only used 76% of its allocation, with several thousand more blocks still available. The filepool is not full. I've issued QUERY FILEPOOL CONFLICT. There is no conflict. I've issued QUERY ACCESSED. The directory shows that is accessed R/W. When the write is unsuccessful, the program then loops through 5 tries of releasing the access, reacquiring the access, and attempting to write the file again. This has never been successful. I've issued both a COPYFILE and a PIPE to try to write the file; these do not work once there has been a failure. We've looked at the operator consoles to see if we can find any jobs running at the same time. We haven't found any that are accessing that directory structure. There aren't any dumps to look at--it looks perfectly successful other than the fact that it won't write the file. Does anyone have any suggestions of something to try next? Nora Graves nora.e.gra...@irs.gov The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review retransmission dissemination or other use of or taking any action in reliance upon this information by persons or entities other than the intended recipient or delegate is strictly prohibited. If you received this in error please contact the sender and delete the material from any computer. The integrity and security of this message cannot be guaranteed on the Internet. The sender accepts no liability for the content of this e-mail or for the consequences of any actions taken on the basis of information provided. The recipient should check this e-mail and any attachments for the presence of viruses. The sender accepts no liability for any damage caused by any virus transmitted by this e-mail. This disclaimer is property of the TTC and must not be altered or circumvented in any manner.
SFS problem
We are having an intermittent problem with SFS and I'm hoping someone may have some ideas of what to pursue next. We have several batch jobs that run under VMBATCH overnight. Sometimes they are not able to create a file in a directory, even though most times it is successful. The only differences in the executions are the file names; for many of these the File Type is the date. In the job I am most familiar with, these are the specifics. The job runs Monday-Saturday. This year, it has failed on January 4, January 12, February 9, March 18, March 25, and April 13. It has run successfully the other days. Other than the QUERY statements below, it has not changed. The job runs in a work machine, WORK7. The job is submitted by the User ID of the database owner. The SFS directories are owned by a 3rd user. Failures occur in many of the subdirectories, not just one subdirectory owned by this user. This user is the owner of most of the directories containing the data files we create in batch, so I don't think it's significant that it's the ID that has the problem. However, as far as I know, it is the only ID that does have the problem. This job uses VMLINK to acquire a write link to SFS directory. This always looks to be successful--no error is given. (Other jobs use GETFMADDR and ACCESS to acquire the write link to the directory. This always appears successful as well). Once the file is ready to be copied from the Work Machine's 191 disk to the SFS directory, the intermittent error appears. The vast majority of the time, the write is successful. However, sometimes, the job gets this error message: DMSOPN1258E You are not authorized to write to file XX 20110413 Z1 The file is not large--last night's file was only 12 blocks. At the suggestion of our systems programmer, I've put in a lot of query statements. I've issued QUERY LIMITS for the job submitter; it's only used 84% of the allocation, with several thousand blocks available. The SFS directory owner has only used 76% of its allocation, with several thousand more blocks still available. The filepool is not full. I've issued QUERY FILEPOOL CONFLICT. There is no conflict. I've issued QUERY ACCESSED. The directory shows that is accessed R/W. When the write is unsuccessful, the program then loops through 5 tries of releasing the access, reacquiring the access, and attempting to write the file again. This has never been successful. I've issued both a COPYFILE and a PIPE to try to write the file; these do not work once there has been a failure. We've looked at the operator consoles to see if we can find any jobs running at the same time. We haven't found any that are accessing that directory structure. There aren't any dumps to look at--it looks perfectly successful other than the fact that it won't write the file. Does anyone have any suggestions of something to try next? Nora Graves nora.e.gra...@irs.gov
Re: SFS management question
Are you running the CRR server? I seem to remember that, even though you do not need it for its stated purpose, it is needed for performance reasons. Another p[possibility is that the log m-disks might be filling up. The control data backup is initiated at 80% full, a non-configurable but arbitrary size. If, while that backup is running, it reaches 95%, all activity is suspended until it is below a given percentage. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Kris Buelens Sent: Thursday, March 31, 2011 3:16 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS management question I some release it became less "required" to run SFS reorganisations. You can run a FILEPOOL REORG, that will only reorganise storage pool 1, that is, the SFS DB2-like catalog. There is no tool to reorganise the other storage groups (apart from a backup & restore). 2011/3/31 Tony Thigpen mailto:t...@vse2pdf.com>> It's been a long time since I managed a production SFS system so I need a little refresher. On our development system, I ran out of space on my SFS pool, but was able to reduce the usage down to just 45% by deleting a bunch of stuff we no longer needed. Now the SFS seems 'slow'. Do I need to run a job to 'compress' the SFS or re-org the directory? -- Tony Thigpen -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS management question
I some release it became less "required" to run SFS reorganisations. You can run a FILEPOOL REORG, that will only reorganise storage pool 1, that is, the SFS DB2-like catalog. There is no tool to reorganise the other storage groups (apart from a backup & restore). 2011/3/31 Tony Thigpen > It's been a long time since I managed a production SFS system so I need > a little refresher. > > On our development system, I ran out of space on my SFS pool, but was > able to reduce the usage down to just 45% by deleting a bunch of stuff > we no longer needed. Now the SFS seems 'slow'. Do I need to run a job to > 'compress' the SFS or re-org the directory? > > -- > > Tony Thigpen > -- Kris Buelens, IBM Belgium, VM customer support
SFS management question
It's been a long time since I managed a production SFS system so I need a little refresher. On our development system, I ran out of space on my SFS pool, but was able to reduce the usage down to just 45% by deleting a bunch of stuff we no longer needed. Now the SFS seems 'slow'. Do I need to run a job to 'compress' the SFS or re-org the directory? -- Tony Thigpen
Re: SFS
On Wednesday, 03/30/2011 at 06:00 EDT, clifford jackson wrote: > If I allocate x amount of storage to a SFS file pool and the tree structure is > such that I have x subdirectories under my top directory, will my > subdirectories have access to all allocated storage within that specific > storage pool, say for instance storage pool 3. Yes. All of the data within a file space is stored in the storage group assigned to the file space at ENROLL USER time. > Also will my subdirectories inherit all security controls from my top directory. No. Each directory and subdirectory must be managed separately. If you have ESM (e.g. RACF) protection for SFS, then the ESM can things like groups to make life easier. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS
If I allocate x amount of storage to a SFS file pool and the tree structure is such that I have x subdirectories under my top directory, will my subdirectories have access to all allocated storage within that specific storage pool, say for instance storage pool 3. Also will my subdirectories inherit all security controls from my top directory. Date: Wed, 23 Mar 2011 10:52:15 +0100 From: kris.buel...@gmail.com Subject: Re: SFS To: IBMVM@LISTSERV.UARK.EDU Yes indeed, I meant "are not copied". very bad eyes and bad typist, sorry. 2011/3/23 Les Koehler Spilling error below? Les Kris Buelens wrote: The RENAME and ERASE commands are indeed "atomic operations". You can't even issue a RENAME or ERASE when the current workunit has uncommited work. Then you can get around by getting a new workunit, making it the default, etc. There is no *need* to use other than the deafult workunit with DMSFILEC or DMSRENAM if you can afford committing all work that is pending in that workunit. Some other thing when using SFS: if you'd use the sequence COPY into tempfile, followed by a RENAME to install a new version of "MY FILE" note then that the auhtories granted to the orignal "MY FILE" or not copied to You meant: 'are not copied', right? the new version of "MY FILE". That pleads to use DMSFILEC with REPLACE. 2011/3/22 Alan Altmark On Tuesday, 03/22/2011 at 12:18 EDT, Les Koehler wrote: So with FILE CONTROL you can't replace co-dependent files in one atomic operation, as one would do with the 'copyfile two-step' (love that phrase!)? Yes, you can (and in DIRECTORY CONTROL directories, too), but you need to become familiar with workunits, DMSGETWU, DMSPUSWU and DMSPOPWU. See the discussion of workunits in Chapter 12 of the CMS Application Dev. Guide. If memory serves, you get a new workunit, push it on the stack, do all your file updates, then pop the workunit. But it's been a long time. Just don't expect RENAME and ERASE to obey the rules. I think they are atomic operations. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS
Yes indeed, I meant "are not copied". very bad eyes and bad typist, sorry. 2011/3/23 Les Koehler > Spilling error below? > Les > > > Kris Buelens wrote: > >> The RENAME and ERASE commands are indeed "atomic operations". >> You can't even issue a RENAME or ERASE when the current workunit has >> uncommited work. Then you can get around by getting a new workunit, >> making >> it the default, etc. There is no *need* to use other than the deafult >> workunit with DMSFILEC or DMSRENAM if you can afford committing all work >> that is pending in that workunit. >> >> Some other thing when using SFS: if you'd use the sequence COPY into >> tempfile, followed by a RENAME to install a new version of "MY FILE" note >> then that the auhtories granted to the orignal "MY FILE" or not copied to >> > You meant: 'are not copied', right? > > the new version of "MY FILE". That pleads to use DMSFILEC with REPLACE. >> >> 2011/3/22 Alan Altmark >> >> On Tuesday, 03/22/2011 at 12:18 EDT, Les Koehler >> > >>> wrote: >>> >>>> So with FILE CONTROL you can't replace co-dependent files in one atomic >>>> operation, as one would do with the 'copyfile two-step' (love that >>>> >>> phrase!)? >>> >>> Yes, you can (and in DIRECTORY CONTROL directories, too), but you need to >>> become familiar with workunits, DMSGETWU, DMSPUSWU and DMSPOPWU. See the >>> discussion of workunits in Chapter 12 of the CMS Application Dev. Guide. >>> >>> If memory serves, you get a new workunit, push it on the stack, do all >>> your file updates, then pop the workunit. But it's been a long time. >>> Just >>> don't expect RENAME and ERASE to obey the rules. I think they are atomic >>> operations. >>> >>> Alan Altmark >>> >>> z/VM and Linux on System z Consultant >>> IBM System Lab Services and Training >>> ibm.com/systems/services/labservices >>> office: 607.429.3323 >>> mobile; 607.321.7556 >>> alan_altm...@us.ibm.com >>> IBM Endicott >>> >>> >> >> >> -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS
Spilling error below? Les Kris Buelens wrote: The RENAME and ERASE commands are indeed "atomic operations". You can't even issue a RENAME or ERASE when the current workunit has uncommited work. Then you can get around by getting a new workunit, making it the default, etc. There is no *need* to use other than the deafult workunit with DMSFILEC or DMSRENAM if you can afford committing all work that is pending in that workunit. Some other thing when using SFS: if you'd use the sequence COPY into tempfile, followed by a RENAME to install a new version of "MY FILE" note then that the auhtories granted to the orignal "MY FILE" or not copied to You meant: 'are not copied', right? the new version of "MY FILE". That pleads to use DMSFILEC with REPLACE. 2011/3/22 Alan Altmark On Tuesday, 03/22/2011 at 12:18 EDT, Les Koehler wrote: So with FILE CONTROL you can't replace co-dependent files in one atomic operation, as one would do with the 'copyfile two-step' (love that phrase!)? Yes, you can (and in DIRECTORY CONTROL directories, too), but you need to become familiar with workunits, DMSGETWU, DMSPUSWU and DMSPOPWU. See the discussion of workunits in Chapter 12 of the CMS Application Dev. Guide. If memory serves, you get a new workunit, push it on the stack, do all your file updates, then pop the workunit. But it's been a long time. Just don't expect RENAME and ERASE to obey the rules. I think they are atomic operations. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS
The RENAME and ERASE commands are indeed "atomic operations". You can't even issue a RENAME or ERASE when the current workunit has uncommited work. Then you can get around by getting a new workunit, making it the default, etc. There is no *need* to use other than the deafult workunit with DMSFILEC or DMSRENAM if you can afford committing all work that is pending in that workunit. Some other thing when using SFS: if you'd use the sequence COPY into tempfile, followed by a RENAME to install a new version of "MY FILE" note then that the auhtories granted to the orignal "MY FILE" or not copied to the new version of "MY FILE". That pleads to use DMSFILEC with REPLACE. 2011/3/22 Alan Altmark > On Tuesday, 03/22/2011 at 12:18 EDT, Les Koehler > wrote: > > So with FILE CONTROL you can't replace co-dependent files in one atomic > > operation, as one would do with the 'copyfile two-step' (love that > phrase!)? > > Yes, you can (and in DIRECTORY CONTROL directories, too), but you need to > become familiar with workunits, DMSGETWU, DMSPUSWU and DMSPOPWU. See the > discussion of workunits in Chapter 12 of the CMS Application Dev. Guide. > > If memory serves, you get a new workunit, push it on the stack, do all > your file updates, then pop the workunit. But it's been a long time. Just > don't expect RENAME and ERASE to obey the rules. I think they are atomic > operations. > > Alan Altmark > > z/VM and Linux on System z Consultant > IBM System Lab Services and Training > ibm.com/systems/services/labservices > office: 607.429.3323 > mobile; 607.321.7556 > alan_altm...@us.ibm.com > IBM Endicott > -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS
On Tuesday, 03/22/2011 at 12:18 EDT, Les Koehler wrote: > So with FILE CONTROL you can't replace co-dependent files in one atomic > operation, as one would do with the 'copyfile two-step' (love that phrase!)? Yes, you can (and in DIRECTORY CONTROL directories, too), but you need to become familiar with workunits, DMSGETWU, DMSPUSWU and DMSPOPWU. See the discussion of workunits in Chapter 12 of the CMS Application Dev. Guide. If memory serves, you get a new workunit, push it on the stack, do all your file updates, then pop the workunit. But it's been a long time. Just don't expect RENAME and ERASE to obey the rules. I think they are atomic operations. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS
Kris, Thanks for correcting my memory! Back in my IGS days I had an EXEC that would do it all for my service machines, via MSG from authorized users. I just sent the DVM the files and a list of files identified by Change Request number. The EXEC did all the work, including sending notification of success/failure to a Notify List. It could even back of a change, since all files were RENAMEd. Les Kris Buelens wrote: Alan forgot to mention one restriction: with a DIRC, only one user can get R/W access at a given time. And, the benefit: a DIRC can be placed in a dataspace. As for the trick to update a few files in one atomic operation: you had it almost right, except that COPYFILE has no NOUPDIRT option, but that isn't needed. The use of temp files and the RENAMEs with NOUPDIRT that care of that. With SFS however, RENAME does not honor the NOUPDIRT option. I never tried but you can use the CSL routine DMSFILEC, it has a NOCOMMIT option. At the end of the DMSFILEC calls, you then call DMSCOMM to commit all copy operations (or DMSROLLB if you change your mind). There is also a DMSRENAM CSL routine, (with a NOCOMMIT option) but that is no longer required thanks to DMSFILEC NOCOMMIT. 2011/3/22 Les Koehler So with FILE CONTROL you can't replace co-dependent files in one atomic operation, as one would do with the 'copyfile two-step' (love that phrase!)? Les Alan Altmark wrote: On Monday, 03/21/2011 at 10:54 EDT, Les Koehler wrote: Does FILE CONTROL thus have the same exposure to a file being changed 'under the covers' as a minidisk has, possibly causing an error for the other users? Is there something similar to the minidisk trick of: copyfile new1 version a pseudo1 name k (noupdir copyfile new2 version a pseudo2 name k (noupdir rename old1 version k save1 version k (noupdir rename old2 version k save2 version k (noupdir rename pseudo1 name k old1 version k (noupdir rename pseudo2 name k old2 version k (hope I got that right!) to avoid impacting other users? You never get ERROR 3 READING FILE with SFS and you don't use the "copyfile twostep" with SFS. You just change the files. If someone accesses the directory while you're updating, then they will see changed files on for files that are closed when they access it. In that respect, DIRCONTROL directories are better than minidisks. FILECONTROL directories reveal file changes immediately. I.e. you can open a file and see one set of content, then close and re-open and see something new. But you never see a mix and you never see the file while it is open and being written to - you see the old version. And if you erase a file in a FILECONTROL directory, all authorizations are lost. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservicesoffice: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS
Alan forgot to mention one restriction: with a DIRC, only one user can get R/W access at a given time. And, the benefit: a DIRC can be placed in a dataspace. As for the trick to update a few files in one atomic operation: you had it almost right, except that COPYFILE has no NOUPDIRT option, but that isn't needed. The use of temp files and the RENAMEs with NOUPDIRT that care of that. With SFS however, RENAME does not honor the NOUPDIRT option. I never tried but you can use the CSL routine DMSFILEC, it has a NOCOMMIT option. At the end of the DMSFILEC calls, you then call DMSCOMM to commit all copy operations (or DMSROLLB if you change your mind). There is also a DMSRENAM CSL routine, (with a NOCOMMIT option) but that is no longer required thanks to DMSFILEC NOCOMMIT. 2011/3/22 Les Koehler > So with FILE CONTROL you can't replace co-dependent files in one atomic > operation, as one would do with the 'copyfile two-step' (love that phrase!)? > > Les > > > Alan Altmark wrote: > >> On Monday, 03/21/2011 at 10:54 EDT, Les Koehler >> wrote: >> >>> Does FILE CONTROL thus have the same exposure to a file being changed >>> >> 'under the >> >>> covers' as a minidisk has, possibly causing an error for the other >>> >> users? >> >>> Is there something similar to the minidisk trick of: >>> >>> copyfile new1 version a pseudo1 name k (noupdir >>> copyfile new2 version a pseudo2 name k (noupdir >>> rename old1 version k save1 version k (noupdir >>> rename old2 version k save2 version k (noupdir >>> rename pseudo1 name k old1 version k (noupdir >>> rename pseudo2 name k old2 version k >>> >>> (hope I got that right!) to avoid impacting other users? >>> >> >> You never get ERROR 3 READING FILE with SFS and you don't use the >> "copyfile twostep" with SFS. You just change the files. If someone >> accesses the directory while you're updating, then they will see changed >> files on for files that are closed when they access it. In that respect, >> DIRCONTROL directories are better than minidisks. >> >> FILECONTROL directories reveal file changes immediately. I.e. you can >> open a file and see one set of content, then close and re-open and see >> something new. But you never see a mix and you never see the file while it >> is open and being written to - you see the old version. And if you erase a >> file in a FILECONTROL directory, all authorizations are lost. >> >> Alan Altmark >> >> z/VM and Linux on System z Consultant >> IBM System Lab Services and Training >> ibm.com/systems/services/labservicesoffice: 607.429.3323 >> mobile; 607.321.7556 >> alan_altm...@us.ibm.com >> IBM Endicott >> >> -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS
So with FILE CONTROL you can't replace co-dependent files in one atomic operation, as one would do with the 'copyfile two-step' (love that phrase!)? Les Alan Altmark wrote: On Monday, 03/21/2011 at 10:54 EDT, Les Koehler wrote: Does FILE CONTROL thus have the same exposure to a file being changed 'under the covers' as a minidisk has, possibly causing an error for the other users? Is there something similar to the minidisk trick of: copyfile new1 version a pseudo1 name k (noupdir copyfile new2 version a pseudo2 name k (noupdir rename old1 version k save1 version k (noupdir rename old2 version k save2 version k (noupdir rename pseudo1 name k old1 version k (noupdir rename pseudo2 name k old2 version k (hope I got that right!) to avoid impacting other users? You never get ERROR 3 READING FILE with SFS and you don't use the "copyfile twostep" with SFS. You just change the files. If someone accesses the directory while you're updating, then they will see changed files on for files that are closed when they access it. In that respect, DIRCONTROL directories are better than minidisks. FILECONTROL directories reveal file changes immediately. I.e. you can open a file and see one set of content, then close and re-open and see something new. But you never see a mix and you never see the file while it is open and being written to - you see the old version. And if you erase a file in a FILECONTROL directory, all authorizations are lost. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS
On Monday, 03/21/2011 at 10:54 EDT, Les Koehler wrote: > Does FILE CONTROL thus have the same exposure to a file being changed 'under the > covers' as a minidisk has, possibly causing an error for the other users? > > Is there something similar to the minidisk trick of: > > copyfile new1 version a pseudo1 name k (noupdir > copyfile new2 version a pseudo2 name k (noupdir > rename old1 version k save1 version k (noupdir > rename old2 version k save2 version k (noupdir > rename pseudo1 name k old1 version k (noupdir > rename pseudo2 name k old2 version k > > (hope I got that right!) to avoid impacting other users? You never get ERROR 3 READING FILE with SFS and you don't use the "copyfile twostep" with SFS. You just change the files. If someone accesses the directory while you're updating, then they will see changed files on for files that are closed when they access it. In that respect, DIRCONTROL directories are better than minidisks. FILECONTROL directories reveal file changes immediately. I.e. you can open a file and see one set of content, then close and re-open and see something new. But you never see a mix and you never see the file while it is open and being written to - you see the old version. And if you erase a file in a FILECONTROL directory, all authorizations are lost. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS
Does FILE CONTROL thus have the same exposure to a file being changed 'under the covers' as a minidisk has, possibly causing an error for the other users? Is there something similar to the minidisk trick of: copyfile new1 version a pseudo1 name k (noupdir copyfile new2 version a pseudo2 name k (noupdir rename old1 version k save1 version k (noupdir rename old2 version k save2 version k (noupdir rename pseudo1 name k old1 version k (noupdir rename pseudo2 name k old2 version k (hope I got that right!) to avoid impacting other users? Les Alan Altmark wrote: On Monday, 03/21/2011 at 05:50 EDT, clifford jackson wrote: What are the major differences between filecontrol and directorycontrol in a SFS structure DIRECTORY CONTROL directories have two important attributes: 1. You don't seen any changes to any files until you re-ACCESS the directory. 2. Authority for the files resides at the directory level (DIRREAD, DIRWRITE). That is, all files have the same access rights, matching the access rights of the directory. You just grant access to the directory and the person has those rights to all the files. If there are new files, they automatically get access, but only after they re-access the directory. The above is called "Access-to-Release Consistency" and mimics the way a minidisk works. FILE CONTROL directories have separate file and directory access rights, and any changes to a file are visible to others as soon as the file is closed. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SFS
On Monday, 03/21/2011 at 05:50 EDT, clifford jackson wrote: > What are the major differences between filecontrol and directorycontrol in a > SFS structure DIRECTORY CONTROL directories have two important attributes: 1. You don't seen any changes to any files until you re-ACCESS the directory. 2. Authority for the files resides at the directory level (DIRREAD, DIRWRITE). That is, all files have the same access rights, matching the access rights of the directory. You just grant access to the directory and the person has those rights to all the files. If there are new files, they automatically get access, but only after they re-access the directory. The above is called "Access-to-Release Consistency" and mimics the way a minidisk works. FILE CONTROL directories have separate file and directory access rights, and any changes to a file are visible to others as soon as the file is closed. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: Temp SFS environment
The work space is not allocated the same way if control and logs have many blocks defined or not. I did several tests to suit my needs and i remarked that dasd pages could be used or (vs. Memory) not depending of what coded in the pooldef file. That seems normal anyway to me and glad to see that this resolves my need. Alain Envoyé de mon iPhone Le 17 mars 2011 à 20:23, "Gentry, Stephen" a écrit : > Wasn't the original poster trying to do this with virtual disk(s)? And > if that's the case it has a high potential of being in memory anyway? > Or did I miss something. > Steve > > -Original Message- > From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On > Behalf Of David Boyes > Sent: Thursday, March 17, 2011 2:51 PM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: Temp SFS environment > >> All that brings a subsidiary question : is it important to preserve > the machine xc >> in the VMSERVx directories (I should read the doc I think...) ? > > You need the XC mode setting if you want SFS to use VM dataspaces to map > parts of the data into memory (which would be consistent with the errors > you saw) for performance reasons. It works fine without it (as you've > seen), but performance will be impaired with very heavily loaded SFS > servers.
Re: Temp SFS environment
> Wasn't the original poster trying to do this with virtual disk(s)? And > if that's the case it has a high potential of being in memory anyway? > Or did I miss something. > Steve In this case, it probably doesn't matter at all, but that's the reason the XC mode setting is in the VMSERVx directory entries. There's probably some coding advantages to be able to just flip over to another address space and directly read or write a page, but I doubt it matters nearly as much as it used to.
Re: Temp SFS environment
With Dataspaces, the SFS end-user can directly read the CMS file data, without APPC communication to the SFS server. Only at ACCESS time, the end-user of a "dataspaced" DIRC must talk to the SFS server. 2011/3/17 Gentry, Stephen > Wasn't the original poster trying to do this with virtual disk(s)? And > if that's the case it has a high potential of being in memory anyway? > Or did I miss something. > Steve > > -Original Message- > From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On > Behalf Of David Boyes > Sent: Thursday, March 17, 2011 2:51 PM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: Temp SFS environment > > > All that brings a subsidiary question : is it important to preserve > the machine xc > > in the VMSERVx directories (I should read the doc I think...) ? > > You need the XC mode setting if you want SFS to use VM dataspaces to map > parts of the data into memory (which would be consistent with the errors > you saw) for performance reasons. It works fine without it (as you've > seen), but performance will be impaired with very heavily loaded SFS > servers. > -- Kris Buelens, IBM Belgium, VM customer support
Re: Temp SFS environment
Wasn't the original poster trying to do this with virtual disk(s)? And if that's the case it has a high potential of being in memory anyway? Or did I miss something. Steve -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of David Boyes Sent: Thursday, March 17, 2011 2:51 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Temp SFS environment > All that brings a subsidiary question : is it important to preserve the machine xc > in the VMSERVx directories (I should read the doc I think...) ? You need the XC mode setting if you want SFS to use VM dataspaces to map parts of the data into memory (which would be consistent with the errors you saw) for performance reasons. It works fine without it (as you've seen), but performance will be impaired with very heavily loaded SFS servers.
Re: Temp SFS environment
> All that brings a subsidiary question : is it important to preserve the > machine xc > in the VMSERVx directories (I should read the doc I think...) ? You need the XC mode setting if you want SFS to use VM dataspaces to map parts of the data into memory (which would be consistent with the errors you saw) for performance reasons. It works fine without it (as you've seen), but performance will be impaired with very heavily loaded SFS servers.
Re: Temp SFS environment
> Feedback: Reading Alan¹s answer, I thought to be unlucky,I changed my point of > view in making some tests by defining my MDISK 301 to 305 as 3390 T-DISK. > I fall in what Simon talked about : fileserv generate cmd didn¹t see the tdisk > and ask me to link them... More ugly than my previous tests. I decided to > freely supersede¹ Alan¹s directives to go back with my v-disk and check again > the directory... > and I got a idea. I had copied a VMSERVx directory to create my own SFS user. > VMSERVx come with a MACHINE XC definition. I removed that line. And it works > ! I am ready for communication¹. I don¹t say I will not have problems later > but I think to be on the good way. > All that brings a subsidiary question : is it important to preserve the > machine xc in the VMSERVx directories (I should read the doc I think...) ? > > Thanks to all > Alain >
Re: Temp SFS environment
V-disks stay around until the last user releases them, so could you have them linked read-only by a second user as well to keep them alive until you¹re done with them? -- Robert P. Nix Mayo Foundation.~. RO-OC-1-18 200 First Street SW/V\ 507-284-0844 Rochester, MN 55905 /( )\ -^^-^^ "In theory, theory and practice are the same, but in practice, theory and practice are different." On 3/17/11 8:43 AM, "Ivica Brodaric" wrote: > I don't know about detach, but re-link definitely happens. File pool minidisks > can be defined in the directory with linkmode R and they still become linked > R/W when the server starts and remain R/W after the server terminates. > > Ivica Brodaric > BNZ >
Re: Temp SFS environment
I don't know about detach, but re-link definitely happens. File pool minidisks can be defined in the directory with linkmode R and they still become linked R/W when the server starts and remain R/W after the server terminates. Ivica Brodaric BNZ
Re: Temp SFS environment
Thank you for sharing your reasoning. Makes sense to me. Thank you, Scott -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alain Benveniste Sent: Thursday, March 17, 2011 1:47 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Temp SFS environment Scott My VM is used as hypervisor for DR needs and defined with minimal dasd. I was looking for a way to transfer big files without adding perm or temp dasd space and avoid to increase time of backup... And we are waiting for a new dasd array coming in end of may. All the rest of free space is used to migrate 3380 to 3390. Alain Envoyé de mon iPhone Le 16 mars 2011 à 23:12, "Wandschneider, Scott" a écrit : > My curious mind wants me to ask. . . Why would you want a temporary SFS > Directory in memory? > > Thank you, > Scott > > > -Original Message- > From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On > Behalf Of Alain Benveniste > Sent: Wednesday, March 16, 2011 5:01 PM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Temp SFS environment > > Hi, > > I tried to create a Temp SFS environment in using FB-512 VDSK. > Part of my directory looks like this: > MDISK 0191 3390 11 1 VM0HVF MR > MDISK 0301 FB-512 V-DISK 64000 > MDISK 0302 FB-512 V-DISK 64000 > MDISK 0303 FB-512 V-DISK 64000 > MDISK 0304 FB-512 V-DISK 64000 > MDISK 0305 FB-512 V-DISK 64000 > > The fileserv generate command works perfectly but when I do a fileserv > st= > art > I get: > DMSWFV1117I FILESERV processing begun at 22:55:02 on 16 Mar 2011 > DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing > DMS4PD3400I Initializing begins for DDNAME = CONTROL > DMS4PD3400I Initializing ends for DDNAME = CONTROL > DMS4PD3400I Initializing begins for DDNAME = MDK1 > DMS4PD3400I Initializing ends for DDNAME = MDK1 > DMS4PD3400I Initializing begins for DDNAME = MDK2 > DMS4PD3400I Initializing ends for DDNAME = MDK2 > DMS4PG3404W File pool limit of 2 minidisks has been reached > DMS4PD3400I Initializing begins for DDNAME = LOG1 > DMS4PD3400I Initializing ends for DDNAME = LOG1 > DMS4PD3400I Initializing begins for DDNAME = LOG2 > DMS4PD3400I Initializing ends for DDNAME = LOG2 > DMS5FD3032I File pool server has terminated > DMSWFV1120I File SFS POOLDEF A1 created or replaced > DMSWFV1117I FILESERV processing ended at 22:55:52 on 16 Mar 2011 > DMSWFV1117I FILESERV processing begun at 22:55:52 on 16 Mar 2011 > DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing > DMSWFV1121I SFS POOLDEF A1 will be used for FILESERV processing > DMS5FE3040E File pool server system error occurred - DMS4FK 02 > > DMS3040E File pool server system error occurred - modulename nn > > Explanation: An internal error occurred within the file pool server. A > du= > mp is > taken according to the dump option chosen in the server startup > parameter= > s. > This is a system error. > > The modulename is the name of the module that detected the error. > > The nn is the error detection point within the module. > > The modulename nn is intended only for service personnel. > > > What modulename 02 means ? > > Fileserv works fine when the MDK are 'normal' mdisk but not when > they= > > are defined as temp... > > Is it just possible to do that ? > > Alain > > Confidentiality Note: This e-mail, including any attachment to it, may > contain material that is confidential, proprietary, privileged and/or > "Protected Health Information," within the meaning of the regulations under > the Health Insurance Portability & Accountability Act as amended. If it is > not clear that you are the intended recipient, you are hereby notified that > you have received this transmittal in error, and any review, dissemination, > distribution or copying of this e-mail, including any attachment to it, is > strictly prohibited. If you have received this e-mail in error, please > immediately return it to the sender and delete it from your system. Thank you. Confidentiality Note: This e-mail, including any attachment to it, may contain material that is confidential, proprietary, privileged and/or "Protected Health Information," within the meaning of the regulations under the Health Insurance Portability & Accountability Act as amended. If it is not clear that you are the intended recipient, you are hereby notified that you have received this transmittal in error, and any review, dissemination, distribution or copying of this e-mail, including any attachment to it, is strictly prohibited. If you have received this e-mail in error, please immediately return it to the sender and delete it from your system. Thank you.
Re: Temp SFS environment
Scott My VM is used as hypervisor for DR needs and defined with minimal dasd. I was looking for a way to transfer big files without adding perm or temp dasd space and avoid to increase time of backup... And we are waiting for a new dasd array coming in end of may. All the rest of free space is used to migrate 3380 to 3390. Alain Envoyé de mon iPhone Le 16 mars 2011 à 23:12, "Wandschneider, Scott" a écrit : > My curious mind wants me to ask. . . Why would you want a temporary SFS > Directory in memory? > > Thank you, > Scott > > > -Original Message- > From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On > Behalf Of Alain Benveniste > Sent: Wednesday, March 16, 2011 5:01 PM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Temp SFS environment > > Hi, > > I tried to create a Temp SFS environment in using FB-512 VDSK. > Part of my directory looks like this: > MDISK 0191 3390 11 1 VM0HVF MR > MDISK 0301 FB-512 V-DISK 64000 > MDISK 0302 FB-512 V-DISK 64000 > MDISK 0303 FB-512 V-DISK 64000 > MDISK 0304 FB-512 V-DISK 64000 > MDISK 0305 FB-512 V-DISK 64000 > > The fileserv generate command works perfectly but when I do a fileserv > st= > art > I get: > DMSWFV1117I FILESERV processing begun at 22:55:02 on 16 Mar 2011 > DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing > DMS4PD3400I Initializing begins for DDNAME = CONTROL > DMS4PD3400I Initializing ends for DDNAME = CONTROL > DMS4PD3400I Initializing begins for DDNAME = MDK1 > DMS4PD3400I Initializing ends for DDNAME = MDK1 > DMS4PD3400I Initializing begins for DDNAME = MDK2 > DMS4PD3400I Initializing ends for DDNAME = MDK2 > DMS4PG3404W File pool limit of 2 minidisks has been reached > DMS4PD3400I Initializing begins for DDNAME = LOG1 > DMS4PD3400I Initializing ends for DDNAME = LOG1 > DMS4PD3400I Initializing begins for DDNAME = LOG2 > DMS4PD3400I Initializing ends for DDNAME = LOG2 > DMS5FD3032I File pool server has terminated > DMSWFV1120I File SFS POOLDEF A1 created or replaced > DMSWFV1117I FILESERV processing ended at 22:55:52 on 16 Mar 2011 > DMSWFV1117I FILESERV processing begun at 22:55:52 on 16 Mar 2011 > DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing > DMSWFV1121I SFS POOLDEF A1 will be used for FILESERV processing > DMS5FE3040E File pool server system error occurred - DMS4FK 02 > > DMS3040E File pool server system error occurred - modulename nn > > Explanation: An internal error occurred within the file pool server. A > du= > mp is > taken according to the dump option chosen in the server startup > parameter= > s. > This is a system error. > > The modulename is the name of the module that detected the error. > > The nn is the error detection point within the module. > > The modulename nn is intended only for service personnel. > > > What modulename 02 means ? > > Fileserv works fine when the MDK are 'normal' mdisk but not when > they= > > are defined as temp... > > Is it just possible to do that ? > > Alain > > Confidentiality Note: This e-mail, including any attachment to it, may > contain material that is confidential, proprietary, privileged and/or > "Protected Health Information," within the meaning of the regulations under > the Health Insurance Portability & Accountability Act as amended. If it is > not clear that you are the intended recipient, you are hereby notified that > you have received this transmittal in error, and any review, dissemination, > distribution or copying of this e-mail, including any attachment to it, is > strictly prohibited. If you have received this e-mail in error, please > immediately return it to the sender and delete it from your system. Thank you.
Re: Temp SFS environment
I am not sure about this, but I think it might be that the mdisks get detached and re-linked, which would kill your SFS server if they were TDISK. See what response you get to: #cp q v 301-305 immediately after the failure. The reason I think I remember this is that years ago I once tried to define the mdisks as slightly different addresses than the ones in the CP directory (#CP DEF B301 301), and ran into problems. IIRC the server wanted to be able to LINK to the addresses I had given it. Shimon On Thu, Mar 17, 2011 at 12:01 AM, Alain Benveniste wrote: > Hi, > > I tried to create a Temp SFS environment in using FB-512 VDSK. > Part of my directory looks like this: > MDISK 0191 3390 11 1 VM0HVF MR > MDISK 0301 FB-512 V-DISK 64000 > MDISK 0302 FB-512 V-DISK 64000 > MDISK 0303 FB-512 V-DISK 64000 > MDISK 0304 FB-512 V-DISK 64000 > MDISK 0305 FB-512 V-DISK 64000 > > The fileserv generate command works perfectly but when I do a fileserv st > art > I get: > DMSWFV1117I FILESERV processing begun at 22:55:02 on 16 Mar 2011 > DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing > DMS4PD3400I Initializing begins for DDNAME = CONTROL > DMS4PD3400I Initializing ends for DDNAME = CONTROL > DMS4PD3400I Initializing begins for DDNAME = MDK1 > DMS4PD3400I Initializing ends for DDNAME = MDK1 > DMS4PD3400I Initializing begins for DDNAME = MDK2 > DMS4PD3400I Initializing ends for DDNAME = MDK2 > DMS4PG3404W File pool limit of 2 minidisks has been reached > DMS4PD3400I Initializing begins for DDNAME = LOG1 > DMS4PD3400I Initializing ends for DDNAME = LOG1 > DMS4PD3400I Initializing begins for DDNAME = LOG2 > DMS4PD3400I Initializing ends for DDNAME = LOG2 > DMS5FD3032I File pool server has terminated > DMSWFV1120I File SFS POOLDEF A1 created or replaced > DMSWFV1117I FILESERV processing ended at 22:55:52 on 16 Mar 2011 > DMSWFV1117I FILESERV processing begun at 22:55:52 on 16 Mar 2011 > DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing > DMSWFV1121I SFS POOLDEF A1 will be used for FILESERV processing > DMS5FE3040E File pool server system error occurred - DMS4FK 02 > > DMS3040E File pool server system error occurred - modulename nn > > Explanation: An internal error occurred within the file pool server. A du > mp is > taken according to the dump option chosen in the server startup parameter > s. > This is a system error. > > The modulename is the name of the module that detected the error. > > The nn is the error detection point within the module. > > The modulename nn is intended only for service personnel. > > > What modulename 02 means ? > > Fileserv works fine when the MDK are 'normal' mdisk but not when they > > are defined as temp... > > Is it just possible to do that ? > > Alain >
Re: Temp SFS environment
On Wednesday, 03/16/2011 at 06:01 EDT, Alain Benveniste wrote: > DMS5FE3040E File pool server system error occurred - DMS4FK 02 That error code indicates a failure of the MAPMDISK IDENTIFY function (see CP Programming Services), and the likely candidate error is "V-DISK not supported". I think you should be able to use T-disk, however. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: Temp SFS environment
My curious mind wants me to ask. . . Why would you want a temporary SFS Directory in memory? Thank you, Scott -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alain Benveniste Sent: Wednesday, March 16, 2011 5:01 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Temp SFS environment Hi, I tried to create a Temp SFS environment in using FB-512 VDSK. Part of my directory looks like this: MDISK 0191 3390 11 1 VM0HVF MR MDISK 0301 FB-512 V-DISK 64000 MDISK 0302 FB-512 V-DISK 64000 MDISK 0303 FB-512 V-DISK 64000 MDISK 0304 FB-512 V-DISK 64000 MDISK 0305 FB-512 V-DISK 64000 The fileserv generate command works perfectly but when I do a fileserv st= art I get: DMSWFV1117I FILESERV processing begun at 22:55:02 on 16 Mar 2011 DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing DMS4PD3400I Initializing begins for DDNAME = CONTROL DMS4PD3400I Initializing ends for DDNAME = CONTROL DMS4PD3400I Initializing begins for DDNAME = MDK1 DMS4PD3400I Initializing ends for DDNAME = MDK1 DMS4PD3400I Initializing begins for DDNAME = MDK2 DMS4PD3400I Initializing ends for DDNAME = MDK2 DMS4PG3404W File pool limit of 2 minidisks has been reached DMS4PD3400I Initializing begins for DDNAME = LOG1 DMS4PD3400I Initializing ends for DDNAME = LOG1 DMS4PD3400I Initializing begins for DDNAME = LOG2 DMS4PD3400I Initializing ends for DDNAME = LOG2 DMS5FD3032I File pool server has terminated DMSWFV1120I File SFS POOLDEF A1 created or replaced DMSWFV1117I FILESERV processing ended at 22:55:52 on 16 Mar 2011 DMSWFV1117I FILESERV processing begun at 22:55:52 on 16 Mar 2011 DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing DMSWFV1121I SFS POOLDEF A1 will be used for FILESERV processing DMS5FE3040E File pool server system error occurred - DMS4FK 02 DMS3040E File pool server system error occurred - modulename nn Explanation: An internal error occurred within the file pool server. A du= mp is taken according to the dump option chosen in the server startup parameter= s. This is a system error. The modulename is the name of the module that detected the error. The nn is the error detection point within the module. The modulename nn is intended only for service personnel. What modulename 02 means ? Fileserv works fine when the MDK are 'normal' mdisk but not when they= are defined as temp... Is it just possible to do that ? Alain Confidentiality Note: This e-mail, including any attachment to it, may contain material that is confidential, proprietary, privileged and/or "Protected Health Information," within the meaning of the regulations under the Health Insurance Portability & Accountability Act as amended. If it is not clear that you are the intended recipient, you are hereby notified that you have received this transmittal in error, and any review, dissemination, distribution or copying of this e-mail, including any attachment to it, is strictly prohibited. If you have received this e-mail in error, please immediately return it to the sender and delete it from your system. Thank you.
Temp SFS environment
Hi, I tried to create a Temp SFS environment in using FB-512 VDSK. Part of my directory looks like this: MDISK 0191 3390 11 1 VM0HVF MR MDISK 0301 FB-512 V-DISK 64000 MDISK 0302 FB-512 V-DISK 64000 MDISK 0303 FB-512 V-DISK 64000 MDISK 0304 FB-512 V-DISK 64000 MDISK 0305 FB-512 V-DISK 64000 The fileserv generate command works perfectly but when I do a fileserv st art I get: DMSWFV1117I FILESERV processing begun at 22:55:02 on 16 Mar 2011 DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing DMS4PD3400I Initializing begins for DDNAME = CONTROL DMS4PD3400I Initializing ends for DDNAME = CONTROL DMS4PD3400I Initializing begins for DDNAME = MDK1 DMS4PD3400I Initializing ends for DDNAME = MDK1 DMS4PD3400I Initializing begins for DDNAME = MDK2 DMS4PD3400I Initializing ends for DDNAME = MDK2 DMS4PG3404W File pool limit of 2 minidisks has been reached DMS4PD3400I Initializing begins for DDNAME = LOG1 DMS4PD3400I Initializing ends for DDNAME = LOG1 DMS4PD3400I Initializing begins for DDNAME = LOG2 DMS4PD3400I Initializing ends for DDNAME = LOG2 DMS5FD3032I File pool server has terminated DMSWFV1120I File SFS POOLDEF A1 created or replaced DMSWFV1117I FILESERV processing ended at 22:55:52 on 16 Mar 2011 DMSWFV1117I FILESERV processing begun at 22:55:52 on 16 Mar 2011 DMSWFV1121I SFS DMSPARMS A1 will be used for FILESERV processing DMSWFV1121I SFS POOLDEF A1 will be used for FILESERV processing DMS5FE3040E File pool server system error occurred - DMS4FK 02 DMS3040E File pool server system error occurred - modulename nn Explanation: An internal error occurred within the file pool server. A du mp is taken according to the dump option chosen in the server startup parameter s. This is a system error. The modulename is the name of the module that detected the error. The nn is the error detection point within the module. The modulename nn is intended only for service personnel. What modulename 02 means ? Fileserv works fine when the MDK are 'normal' mdisk but not when they are defined as temp... Is it just possible to do that ? Alain
Re: Old codger question: Can SFS be networked across systems?
You CAN put TSAF in between SFS and AVS+VTAM, but it's not necessary. Mark Wheeler UnitedHealth Group Date: Wed, 16 Mar 2011 13:13:28 -0400 From: rhamil...@cas.org Subject: Re: Old codger question: Can SFS be networked across systems? To: IBMVM@LISTSERV.UARK.EDU Yep; I used TSAF over AVS+VTAM and connected users+SFSs; that should work regardless of the size of your country. R; Rob Hamilton Chemical Abstracts Service
Re: Old codger question: Can SFS be networked across systems?
Les -- About the authentication nightmare part, the same criticism is frequently nailed on NFS in Unix/Linux land. The problem there is the same: You have placed a certain amount of admin trust in the file sharing peer systems. If you don't trust the sysadmin of an NFS client, don't export the content to his machines. (There are various ways people have tried to close the gap in NFS, wth varying results and varied satisfaction, all very confusing.) Similarly, if you don't trust the sysadmin of the other VM system(s), don't play SFS with them. Rob mentions username mapping with IPGATE. They do the same thing in NFS, but not quite as open-ended. IPGATE is kinda cool. -- R; <>< Rick Troth Velocity Software http://www.velocitysoftware.com/ On Wed, Mar 16, 2011 at 12:25, Les Koehler wrote: > A friend and I were (dis)cussing SFS and he thinks it can be networked > cross-country. Possible? I would guess that it would be an authentication > nightmare at the user level. Thoughts? > > Les >
Re: Old codger question: Can SFS be networked across systems?
Yep; I used TSAF over AVS+VTAM and connected users+SFSs; that should work regardless of the size of your country. R; Rob Hamilton Chemical Abstracts Service From: The IBM z/VM Operating System [mailto:IBMVM@listserv.uark.edu] On Behalf Of Mark Wheeler Sent: Wednesday, March 16, 2011 12:57 PM To: IBMVM@listserv.uark.edu Subject: Re: Old codger question: Can SFS be networked across systems? In the unlikely situation where you run VTAM on your z/VM systems, you can config your AVS gateways and connect the two SFS's via LU6.2. Mark Wheeler UnitedHealth Group > Date: Wed, 16 Mar 2011 17:38:33 +0100 > From: rvdh...@gmail.com > Subject: Re: Old codger question: Can SFS be networked across systems? > To: IBMVM@LISTSERV.UARK.EDU > > On Wed, Mar 16, 2011 at 5:25 PM, Les Koehler wrote: > > A friend and I were (dis)cussing SFS and he thinks it can be networked > > cross-country. Possible? I would guess that it would be an authentication > > nightmare at the user level. Thoughts? > > If you share via ISFC that means both sides must be in the same > administrative domain. But given the length of the FICON connection, > it will only work for small countries ;-) > IPGATE run over TCP/IP connections and does long distance. Latency can > make it less fun for regular use, but IMHO you really should not have > critical data on a remote system. You define user access and mapping > in the IPGATE configuration files. Provided there is some > understanding between userid management on both sides, you could even > use dummy userids as the target of the mapping (eg map MAINT at SYSA > to user MAINT-A at SYSB). > > Rob Confidentiality Notice: This electronic message transmission, including any attachment(s), may contain confidential, proprietary, or privileged information from Chemical Abstracts Service ("CAS"), a division of the American Chemical Society ("ACS"). If you have received this transmission in error, be advised that any disclosure, copying, distribution, or use of the contents of this information is strictly prohibited. Please destroy all copies of the message and contact the sender immediately by either replying to this message or calling 614-447-3600.
Re: Old codger question: Can SFS be networked across systems?
In the unlikely situation where you run VTAM on your z/VM systems, you can config your AVS gateways and connect the two SFS's via LU6.2. Mark Wheeler UnitedHealth Group > Date: Wed, 16 Mar 2011 17:38:33 +0100 > From: rvdh...@gmail.com > Subject: Re: Old codger question: Can SFS be networked across systems? > To: IBMVM@LISTSERV.UARK.EDU > > On Wed, Mar 16, 2011 at 5:25 PM, Les Koehler wrote: > > A friend and I were (dis)cussing SFS and he thinks it can be networked > > cross-country. Possible? I would guess that it would be an authentication > > nightmare at the user level. Thoughts? > > If you share via ISFC that means both sides must be in the same > administrative domain. But given the length of the FICON connection, > it will only work for small countries ;-) > IPGATE run over TCP/IP connections and does long distance. Latency can > make it less fun for regular use, but IMHO you really should not have > critical data on a remote system. You define user access and mapping > in the IPGATE configuration files. Provided there is some > understanding between userid management on both sides, you could even > use dummy userids as the target of the mapping (eg map MAINT at SYSA > to user MAINT-A at SYSB). > > Rob
Re: Old codger question: Can SFS be networked across systems?
On Wed, Mar 16, 2011 at 5:25 PM, Les Koehler wrote: > A friend and I were (dis)cussing SFS and he thinks it can be networked > cross-country. Possible? I would guess that it would be an authentication > nightmare at the user level. Thoughts? If you share via ISFC that means both sides must be in the same administrative domain. But given the length of the FICON connection, it will only work for small countries ;-) IPGATE run over TCP/IP connections and does long distance. Latency can make it less fun for regular use, but IMHO you really should not have critical data on a remote system. You define user access and mapping in the IPGATE configuration files. Provided there is some understanding between userid management on both sides, you could even use dummy userids as the target of the mapping (eg map MAINT at SYSA to user MAINT-A at SYSB). Rob
Old codger question: Can SFS be networked across systems?
A friend and I were (dis)cussing SFS and he thinks it can be networked cross-country. Possible? I would guess that it would be an authentication nightmare at the user level. Thoughts? Les
Re: SFS question
Ivica and others. Thank you for the help. Increasing the size of the control disk fixed the problem. Steve From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Ivica Brodaric Sent: Saturday, March 12, 2011 8:12 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS question Stephen, Do I understand the formula and definitions correctly? You do, but that's just to get you started. It really all depends on how many objects are going to be created in that file pool. So, take that formula as an attempt to predict the future. MAXUSERS determines the logical size of the catalog and that formula is MAXUSERS * 85 (and that's a real formula). Half of that goes to data blocks and half to index blocks. The physical size of the catalog (number of blocks in group 1) can initially be much smaller than that. That smaller space would again be divided equally to data and index blocks. It is much easier to add another minidisk to group 1 (when you hit physical limit) than to do regenerate file pool (when you hit logical limit). Therefore, allocate MAXUSERS generously, so that you don't have to do this again soon. In your case, you have 640 cylinders in group 1 (16 * 40 cyl), which equates to 115200 blocks. You may have over-allocated group 1, but to decrease that now would be a pain, so leave it. To be able to use all that space, the logical size of the catalog has to be at least that much, so MAXUSERS must be at least 1356, which is 115200 / 85, rounded up. I would suggest that you use at least 5000 if not 1 for MAXUSERS. Over-allocating MAXUSERS is not harmful, it just takes a bit of space in control minidisk. Size of control minidisk (or of the rest of it) determines the number of potentially addressable blocks in the file pool (maximum size to which the file pool can grow to). You can see that number if you do QUERY FILEPOOL OVERVIEW, but roughly, it is a bit less than one million addressable blocks for each control disk cylinder. There is no formula for this, because other things go to control disk as well (MAXDISKS has a hand in it too), but for example (roughly again), with a 50-cylinder control disk you will be left with at least 45 million potentially addressable blocks, which equates to around 75 3390 mod3's. That's the maximum size of disk space that you will ever be able to have in the file pool before having to regenerate it. While regenerating file pool, you will have to increase the size of the control disk. You have to do this to allow for the increase of MAXUSERS in order to maintain at least the same number of potentially addressable blocks. Increase it by 25-33% to be sure, but do *not* over-allocate control disk. Whenever the file pool server performs control data backup (and it does it automatically when log disks fill up more than 80%), the file pool will be unavailable to end users for the duration of the backup. You want that backup to be quick especially if your log disks are small and the backup is run often. Control data backup backs up *entire* control disk and *used* blocks in group 1 (catalog). Therefore, over-allocating group 1 is OK, but over-allocating control disk is not. Follow the procedure to regenerate a repository file pool to which you have been pointed to already and you'll be fine. One final word of advice though: before doing anything like this, make sure that you also have a reliable backup of the whole file pool (storage groups 2 and above) handy. Ivica Brodaric BNZ
Re: SFS question
Stephen, Do I understand the formula and definitions correctly? > > You do, but that's just to get you started. It really all depends on how many objects are going to be created in that file pool. So, take that formula as an attempt to predict the future. MAXUSERS determines the logical size of the catalog and that formula is MAXUSERS * 85 (and that's a real formula). Half of that goes to data blocks and half to index blocks. The physical size of the catalog (number of blocks in group 1) can initially be much smaller than that. That smaller space would again be divided equally to data and index blocks. It is much easier to add another minidisk to group 1 (when you hit physical limit) than to do regenerate file pool (when you hit logical limit). Therefore, allocate MAXUSERS generously, so that you don't have to do this again soon. In your case, you have 640 cylinders in group 1 (16 * 40 cyl), which equates to 115200 blocks. You may have over-allocated group 1, but to decrease that now would be a pain, so leave it. To be able to use all that space, the logical size of the catalog has to be at least that much, so MAXUSERS must be at least 1356, which is 115200 / 85, rounded up. I would suggest that you use at least 5000 if not 1 for MAXUSERS. Over-allocating MAXUSERS is not harmful, it just takes a bit of space in control minidisk. Size of control minidisk (or of the rest of it) determines the number of potentially addressable blocks in the file pool (maximum size to which the file pool can grow to). You can see that number if you do QUERY FILEPOOL OVERVIEW, but roughly, it is a bit less than one million addressable blocks for each control disk cylinder. There is no formula for this, because other things go to control disk as well (MAXDISKS has a hand in it too), but for example (roughly again), with a 50-cylinder control disk you will be left with at least 45 million potentially addressable blocks, which equates to around 75 3390 mod3's. That's the maximum size of disk space that you will ever be able to have in the file pool before having to regenerate it. While regenerating file pool, you will have to increase the size of the control disk. You have to do this to allow for the increase of MAXUSERS in order to maintain at least the same number of potentially addressable blocks. Increase it by 25-33% to be sure, but do *not* over-allocate control disk. Whenever the file pool server performs control data backup (and it does it automatically when log disks fill up more than 80%), the file pool will be unavailable to end users for the duration of the backup. You want that backup to be quick especially if your log disks are small and the backup is run often. Control data backup backs up *entire* control disk and *used* blocks in group 1 (catalog). Therefore, over-allocating group 1 is OK, but over-allocating control disk is not. Follow the procedure to regenerate a repository file pool to which you have been pointed to already and you'll be fine. One final word of advice though: before doing anything like this, make sure that you also have a reliable backup of the whole file pool (storage groups 2 and above) handy. Ivica Brodaric BNZ
Re: SFS question
Yup, that's correct, Steve.you need to set MAXUSERS to at least 1500. DJ On 03/11/2011 08:16 AM, Gentry, Stephen wrote: > So far, thanks to all who have replied. > I'm going over various numbers and need some clarification. In the CMS > Pool Planning guide, Estimate Max Pool Size, the formula is: maximum > enrolled users = 300 * (# system defined users / # system active users) > If I understand the definitions correctly #system defined users is > number of users in USER DIRECT file. For us that number is 750 (but to > CMA, I'm going round up to 1000), #system active users is the number of > users logged on. On an average we have about 200 users logged on. So, > plugging in the values to the formula above: > 1500= 300 * ( 1000 / 200 ) > So the MAXUSER statement should be 1500 ? > If so, I'm way low, currently 300. > Do I understand the formula and definitions correctly? > > Steve > > -Original Message- > From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On > Behalf Of Sue Farrell > Sent: Friday, March 11, 2011 8:56 AM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: SFS question > > Steve, > > You need to increase your MAXUSERS setting by doing a FILESERV > REGENERATE > > for the file pool. > > Although it's buried, the 51010 reason code is mentioned in Chapter 5 of > > > the CMS File Pool Planning, Administration, and Operation manual: > > What Happens When the Limit is Reached: Logical catalog space is > reserved > > during file pool generation. FILESERV GENERATE processing uses the > MAXUSERS value to estimate and set the maximum logical catalog space. > Whe > n > the server runs out of logical space, it displays a warning message on > it > s > console > and continues processing. Depending on their use of the file pool, users > > > may > receive error messages (DMS1146E) and error return codes (with reason > cod > es > 51010 or 51020). When the logical catalog space is exhausted, you need > to > > increase the MAXUSERS value for the file pool. Follow the instructions > in > > Chapter 11, "Regenerating a Repository File Pool," on page xxx. > > Thanks, > Sue > -- Dave Jones V/Soft Software www.vsoft-software.com Houston, TX 281.578.7544
Re: SFS question
So far, thanks to all who have replied. I'm going over various numbers and need some clarification. In the CMS Pool Planning guide, Estimate Max Pool Size, the formula is: maximum enrolled users = 300 * (# system defined users / # system active users) If I understand the definitions correctly #system defined users is number of users in USER DIRECT file. For us that number is 750 (but to CMA, I'm going round up to 1000), #system active users is the number of users logged on. On an average we have about 200 users logged on. So, plugging in the values to the formula above: 1500= 300 * ( 1000 / 200 ) So the MAXUSER statement should be 1500 ? If so, I'm way low, currently 300. Do I understand the formula and definitions correctly? Steve -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Sue Farrell Sent: Friday, March 11, 2011 8:56 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS question Steve, You need to increase your MAXUSERS setting by doing a FILESERV REGENERATE for the file pool. Although it's buried, the 51010 reason code is mentioned in Chapter 5 of the CMS File Pool Planning, Administration, and Operation manual: What Happens When the Limit is Reached: Logical catalog space is reserved during file pool generation. FILESERV GENERATE processing uses the MAXUSERS value to estimate and set the maximum logical catalog space. Whe n the server runs out of logical space, it displays a warning message on it s console and continues processing. Depending on their use of the file pool, users may receive error messages (DMS1146E) and error return codes (with reason cod es 51010 or 51020). When the logical catalog space is exhausted, you need to increase the MAXUSERS value for the file pool. Follow the instructions in Chapter 11, "Regenerating a Repository File Pool," on page xxx. Thanks, Sue
Re: SFS question
Steve, You need to increase your MAXUSERS setting by doing a FILESERV REGENERATE for the file pool. Although it's buried, the 51010 reason code is mentioned in Chapter 5 of the CMS File Pool Planning, Administration, and Operation manual: What Happens When the Limit is Reached: Logical catalog space is reserved during file pool generation. FILESERV GENERATE processing uses the MAXUSERS value to estimate and set the maximum logical catalog space. Whe n the server runs out of logical space, it displays a warning message on it s console and continues processing. Depending on their use of the file pool, users may receive error messages (DMS1146E) and error return codes (with reason cod es 51010 or 51020). When the logical catalog space is exhausted, you need to increase the MAXUSERS value for the file pool. Follow the instructions in Chapter 11, “Regenerating a Repository File Pool,” on page xxx. Thanks, Sue
Re: SFS question
How large is the catalog? -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Dave Jones Sent: Thursday, March 10, 2011 3:40 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS question Hi, Steve. Yes, you need to expand your storage group 1 size by adding more DASD space DJ On 03/10/2011 03:33 PM, Gentry, Stephen wrote: > I'm getting the following error on a DMSCLOSE: > > 51010 - No space for data left in catalog space. > > I am writing a group of files, 733 of them, to an SFS pool > > Does the error message mean that I don't have enough room in storage > group 1? > > TIA > > Steve > -- Dave Jones V/Soft Software www.vsoft-software.com Houston, TX 281.578.7544 == This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to which they are addressed. If you have received this email in error please notify the system manager. This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately by e-mail if you have received this e-mail by mistake and delete this e-mail from your system. If you are not the intended recipient you are notified that disclosing, copying, distributing or taking any action in reliance on the contents of this information is strictly prohibited.
Re: SFS question
Do QUERY FILEPOOL CATALOG to see the amount of catalog data blocks and catalog index blocks used. Total number of catalog space blocks (data+index) is MAXUSERS*85. Maybe your MAXUSERS value is too small? Ivica Brodaric BNZ
Re: SFS question
Here is the results from the QUERY: q filepool storgrp vmsyse: VMSYSE File Pool Storage Groups Start-up Date 03/02/11 Query Date 03/10/11 Start-up Time 23:39:25 Query Time 16:57:49 STORAGE GROUP INFORMATION Storage4K Blocks 4K Blocks Group No.In-Use Free 1 17973 - 16% 96971 2 8712106 - 67%4278828 The total number of files is 733, occupying approx. 2022 4k blocks. So, I don't understand why I'm getting this error. I have defined storage group 1 with 40 cylinders (7184 - 4k blocks) across 16 3390 mod3's. To clarify, that's 40 cylinders on each mod3. I'm wondering if there isn't enough room on one of the mod3's and thus I get this message. Due to the nature of SFS, I assume it would spread the load across the 16 drives. Steve From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Kris Buelens Sent: Thursday, March 10, 2011 4:43 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS question Thta's what I'd conclude too. Try Q FILEPOOL STORGRP (or use SFSULIST, it shows the summary too) 2011/3/10 Gentry, Stephen I'm getting the following error on a DMSCLOSE: 51010 - No space for data left in catalog space. I am writing a group of files, 733 of them, to an SFS pool Does the error message mean that I don't have enough room in storage group 1? TIA Steve -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS question
Yes. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Gentry, Stephen Sent: Thursday, March 10, 2011 1:33 PM To: IBMVM@LISTSERV.UARK.EDU Subject: SFS question I'm getting the following error on a DMSCLOSE: 51010 - No space for data left in catalog space. I am writing a group of files, 733 of them, to an SFS pool Does the error message mean that I don't have enough room in storage group 1? TIA Steve
Re: SFS question
Thta's what I'd conclude too. Try Q FILEPOOL STORGRP (or use SFSULIST, it shows the summary too) 2011/3/10 Gentry, Stephen > I’m getting the following error on a DMSCLOSE: > > 51010 - No space for data left in catalog space. > > I am writing a group of files, 733 of them, to an SFS pool > > Does the error message mean that I don’t have enough room in storage group > 1? > > TIA > > Steve > -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS question
Hi, Steve. Yes, you need to expand your storage group 1 size by adding more DASD space DJ On 03/10/2011 03:33 PM, Gentry, Stephen wrote: > I’m getting the following error on a DMSCLOSE: > > 51010 - No space for data left in catalog space. > > I am writing a group of files, 733 of them, to an SFS pool > > Does the error message mean that I don’t have enough room in storage > group 1? > > TIA > > Steve > -- Dave Jones V/Soft Software www.vsoft-software.com Houston, TX 281.578.7544
SFS question
I'm getting the following error on a DMSCLOSE: 51010 - No space for data left in catalog space. I am writing a group of files, 733 of them, to an SFS pool Does the error message mean that I don't have enough room in storage group 1? TIA Steve
Re: CMS SFS Question
I don't believe that I said DELETE USER RICHARD. I certainly did not intend to imply that, nor did I intend for someone to infer it. I should have stated it better. Regards, Richard Schuh > -Original Message- > From: The IBM z/VM Operating System > [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler > Sent: Tuesday, March 01, 2011 2:07 PM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: CMS SFS Question > > That's NOT the scenario you gave in your original note! You > wrote about deleting Richard when you wrote: > > It is possible for one user to grant access to other users > >>>> who are not enrolled. DELETE USER does not clean up > these >>>> permissions. > > I don't see *any* indication that would trigger a DELETE USER > Les (using your scenario, which was reversed from mine, > further confusing the issue). > > Les > > Schuh, Richard wrote: > > It is permissions granted to users who are not enrolled > that is the issue. Here is the scenario: > > > > User Richard is enrolled > > User Les is not enrolled > > Richard grants Les some SFS authorities. > > DELETE USER LES is issued without enrolling LES (or no > DELETE USER is > > issued for LES) The authorities granted to LES by RICHARD > are left hanging and will be applied to any newly created LES > regardless of the identity of the owner. > > > > If LES is enrolled before the DELETE USER, those > authorities granted to LES by others are removed. By doing > the ENROLL for 0 blocks for any userid that is to be deleted, > no ghost authorities are given to new users. The userids are > unconditionally enrolled. If the user has already been > enrolled and owns a file space, the enroll will fail. Because > all I care about is that the user be enrolled, I ignore that failure. > > > > > > Regards, > > Richard Schuh > > > > > > > >> -Original Message- > >> From: The IBM z/VM Operating System > >> [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler > >> Sent: Tuesday, March 01, 2011 1:24 PM > >> To: IBMVM@LISTSERV.UARK.EDU > >> Subject: Re: CMS SFS Question > >> > >> I guess there's something implied there that I don't get. > >> Scenario, from your note: > >> > >> Your task is to delete LES, who is enrolled, from the SFS > system LES > >> has granted rights to RICHARD but RICHARD is not enrolled > >> > >> How does enrolling LES for 0 blocks do anything about the granted > >> rights that RICHARD has? > >> > >> Les > >> > >> Schuh, Richard wrote: > >>> I simply enroll any user to be deleted for 0 blocks. The > >> alternative is to scan the sfs directories and files > looking for such > >> users. It is much easier to attempt the enroll. If it fails, it is > >> because the user is already enrolled. > >>> Regards, > >>> Richard Schuh > >>> > >>> > >>> > >>>> -Original Message- > >>>> From: The IBM z/VM Operating System > >>>> [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler > >>>> Sent: Tuesday, March 01, 2011 12:22 PM > >>>> To: IBMVM@LISTSERV.UARK.EDU > >>>> Subject: Re: CMS SFS Question > >>>> > >>>> I'm curious: How do you find the user who is not enrolled, but > >>>> granted rights to the target user to be deleted? > >>>> > >>>> Les > >>>> > >>>> Schuh, Richard wrote: > >>>>> The Pipe is the easiest. > >>>>> > >>>>> PIPE < user list | spec /delete user/ 1 w1 nw | cms | > > >> delete log a > >>>>> Note, however, that if you have an SFS that has a lot of > >>>> files and permissions, each DELETE USER can take a long > >> time, so you > >>>> do not want to do this on an id that you might need soon > after you > >>>> enter the PIPE command. In our shop, an individual > DELETE USER can > >>>> take upwards of 10 minutes. > >>>>> Cleaning up SFS when a userid is deleted is important from > >>>> a security standpoint. If the same id should be given to a > >> different > >>>> person, it would automatically inherit permissions from > the prior > >>>> owner. You should be doing a DELETE USER every time that a > >> userid is > >&
Re: CMS SFS Question
That's NOT the scenario you gave in your original note! You wrote about deleting Richard when you wrote: It is possible for one user to grant access to other users >>>> who are not enrolled. DELETE USER does not clean up these >>>> permissions. I don't see *any* indication that would trigger a DELETE USER Les (using your scenario, which was reversed from mine, further confusing the issue). Les Schuh, Richard wrote: It is permissions granted to users who are not enrolled that is the issue. Here is the scenario: User Richard is enrolled User Les is not enrolled Richard grants Les some SFS authorities. DELETE USER LES is issued without enrolling LES (or no DELETE USER is issued for LES) The authorities granted to LES by RICHARD are left hanging and will be applied to any newly created LES regardless of the identity of the owner. If LES is enrolled before the DELETE USER, those authorities granted to LES by others are removed. By doing the ENROLL for 0 blocks for any userid that is to be deleted, no ghost authorities are given to new users. The userids are unconditionally enrolled. If the user has already been enrolled and owns a file space, the enroll will fail. Because all I care about is that the user be enrolled, I ignore that failure. Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler Sent: Tuesday, March 01, 2011 1:24 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: CMS SFS Question I guess there's something implied there that I don't get. Scenario, from your note: Your task is to delete LES, who is enrolled, from the SFS system LES has granted rights to RICHARD but RICHARD is not enrolled How does enrolling LES for 0 blocks do anything about the granted rights that RICHARD has? Les Schuh, Richard wrote: I simply enroll any user to be deleted for 0 blocks. The alternative is to scan the sfs directories and files looking for such users. It is much easier to attempt the enroll. If it fails, it is because the user is already enrolled. Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler Sent: Tuesday, March 01, 2011 12:22 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: CMS SFS Question I'm curious: How do you find the user who is not enrolled, but granted rights to the target user to be deleted? Les Schuh, Richard wrote: The Pipe is the easiest. PIPE < user list | spec /delete user/ 1 w1 nw | cms | > delete log a Note, however, that if you have an SFS that has a lot of files and permissions, each DELETE USER can take a long time, so you do not want to do this on an id that you might need soon after you enter the PIPE command. In our shop, an individual DELETE USER can take upwards of 10 minutes. Cleaning up SFS when a userid is deleted is important from a security standpoint. If the same id should be given to a different person, it would automatically inherit permissions from the prior owner. You should be doing a DELETE USER every time that a userid is deleted from the directory. It is possible for one user to grant access to other users who are not enrolled. DELETE USER does not clean up these permissions. To get rid of them, you have to first enroll the user in the pool even if it is for 0 blocks. To solve this in our automated process, each user to be deleted is enrolled for 0 blocks, ignoring the return code. We don't care if the user is already enrolled, the attempt does no harm. After the enroll, the deletion will clean out all permissions granted to or by the user being deleted. Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Rick Troth Sent: Tuesday, March 01, 2011 10:54 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: CMS SFS Question Nahh ... even easier ... Pipes. I'm thinking two pipes. One to gather the Q ENROLL output then a second to actually perform the deletes. In between shove that Q ENROLL output into a file, manually edit for confirmation, then feed the selected content into DELETE USER. -- R; Rick Troth Velocity Software http://www.velocitysoftware.com/ On Tue, 1 Mar 2011, Rich Smrcina wrote: REXX? On 03/01/2011 12:35 PM, Wandschneider, Scott wrote: Is there a way to delete multiple users at once or create a "batch" job to delete multiple users that are enrolled in SFS? Thank you, Scott R Wandschneider Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 Miracle Hills Drive, Omaha, NE, 68154-4457|| ': 402.963.8905 || Ë:847.849.7223 || : scott.wandschnei...@infocrossing.com **Think Green - Please print responsibly** Confidentiality Note: This e-mail, including any attachment to it, may contain material that is confidential, proprieta
Re: CMS SFS Question
It is permissions granted to users who are not enrolled that is the issue. Here is the scenario: User Richard is enrolled User Les is not enrolled Richard grants Les some SFS authorities. DELETE USER LES is issued without enrolling LES (or no DELETE USER is issued for LES) The authorities granted to LES by RICHARD are left hanging and will be applied to any newly created LES regardless of the identity of the owner. If LES is enrolled before the DELETE USER, those authorities granted to LES by others are removed. By doing the ENROLL for 0 blocks for any userid that is to be deleted, no ghost authorities are given to new users. The userids are unconditionally enrolled. If the user has already been enrolled and owns a file space, the enroll will fail. Because all I care about is that the user be enrolled, I ignore that failure. Regards, Richard Schuh > -Original Message- > From: The IBM z/VM Operating System > [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler > Sent: Tuesday, March 01, 2011 1:24 PM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: CMS SFS Question > > I guess there's something implied there that I don't get. > Scenario, from your note: > > Your task is to delete LES, who is enrolled, from the SFS > system LES has granted rights to RICHARD but RICHARD is not enrolled > > How does enrolling LES for 0 blocks do anything about the > granted rights that RICHARD has? > > Les > > Schuh, Richard wrote: > > I simply enroll any user to be deleted for 0 blocks. The > alternative is to scan the sfs directories and files looking > for such users. It is much easier to attempt the enroll. If > it fails, it is because the user is already enrolled. > > > > Regards, > > Richard Schuh > > > > > > > >> -Original Message- > >> From: The IBM z/VM Operating System > >> [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler > >> Sent: Tuesday, March 01, 2011 12:22 PM > >> To: IBMVM@LISTSERV.UARK.EDU > >> Subject: Re: CMS SFS Question > >> > >> I'm curious: How do you find the user who is not enrolled, but > >> granted rights to the target user to be deleted? > >> > >> Les > >> > >> Schuh, Richard wrote: > >>> The Pipe is the easiest. > >>> > >>> PIPE < user list | spec /delete user/ 1 w1 nw | cms | > > delete log a > >>> > >>> Note, however, that if you have an SFS that has a lot of > >> files and permissions, each DELETE USER can take a long > time, so you > >> do not want to do this on an id that you might need soon after you > >> enter the PIPE command. In our shop, an individual DELETE USER can > >> take upwards of 10 minutes. > >>> Cleaning up SFS when a userid is deleted is important from > >> a security standpoint. If the same id should be given to a > different > >> person, it would automatically inherit permissions from the prior > >> owner. You should be doing a DELETE USER every time that a > userid is > >> deleted from the directory. > >>> It is possible for one user to grant access to other users > >> who are not enrolled. DELETE USER does not clean up these > >> permissions. To get rid of them, you have to first enroll > the user in > >> the pool even if it is for 0 blocks. To solve this in our > automated > >> process, each user to be deleted is enrolled for 0 blocks, > ignoring > >> the return code. We don't care if the user is already > enrolled, the > >> attempt does no harm. After the enroll, the deletion will > clean out > >> all permissions granted to or by the user being deleted. > >>> > >>> Regards, > >>> Richard Schuh > >>> > >>> > >>> > >>>> -Original Message- > >>>> From: The IBM z/VM Operating System > >>>> [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Rick Troth > >>>> Sent: Tuesday, March 01, 2011 10:54 AM > >>>> To: IBMVM@LISTSERV.UARK.EDU > >>>> Subject: Re: CMS SFS Question > >>>> > >>>> Nahh ... even easier ... Pipes. > >>>> I'm thinking two pipes. One to gather the Q ENROLL > output then a > >>>> second to actually perform the deletes. In between shove that Q > >>>> ENROLL output into a file, manually edit for confirmation, > >> then feed > >>>> the selected content into DELETE USER. > >>>> > >>>>
Re: CMS SFS Question
I guess there's something implied there that I don't get. Scenario, from your note: Your task is to delete LES, who is enrolled, from the SFS system LES has granted rights to RICHARD but RICHARD is not enrolled How does enrolling LES for 0 blocks do anything about the granted rights that RICHARD has? Les Schuh, Richard wrote: I simply enroll any user to be deleted for 0 blocks. The alternative is to scan the sfs directories and files looking for such users. It is much easier to attempt the enroll. If it fails, it is because the user is already enrolled. Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler Sent: Tuesday, March 01, 2011 12:22 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: CMS SFS Question I'm curious: How do you find the user who is not enrolled, but granted rights to the target user to be deleted? Les Schuh, Richard wrote: The Pipe is the easiest. PIPE < user list | spec /delete user/ 1 w1 nw | cms | > delete log a Note, however, that if you have an SFS that has a lot of files and permissions, each DELETE USER can take a long time, so you do not want to do this on an id that you might need soon after you enter the PIPE command. In our shop, an individual DELETE USER can take upwards of 10 minutes. Cleaning up SFS when a userid is deleted is important from a security standpoint. If the same id should be given to a different person, it would automatically inherit permissions from the prior owner. You should be doing a DELETE USER every time that a userid is deleted from the directory. It is possible for one user to grant access to other users who are not enrolled. DELETE USER does not clean up these permissions. To get rid of them, you have to first enroll the user in the pool even if it is for 0 blocks. To solve this in our automated process, each user to be deleted is enrolled for 0 blocks, ignoring the return code. We don't care if the user is already enrolled, the attempt does no harm. After the enroll, the deletion will clean out all permissions granted to or by the user being deleted. Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Rick Troth Sent: Tuesday, March 01, 2011 10:54 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: CMS SFS Question Nahh ... even easier ... Pipes. I'm thinking two pipes. One to gather the Q ENROLL output then a second to actually perform the deletes. In between shove that Q ENROLL output into a file, manually edit for confirmation, then feed the selected content into DELETE USER. -- R; Rick Troth Velocity Software http://www.velocitysoftware.com/ On Tue, 1 Mar 2011, Rich Smrcina wrote: REXX? On 03/01/2011 12:35 PM, Wandschneider, Scott wrote: Is there a way to delete multiple users at once or create a "batch" job to delete multiple users that are enrolled in SFS? Thank you, Scott R Wandschneider Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 Miracle Hills Drive, Omaha, NE, 68154-4457|| ': 402.963.8905 || Ë:847.849.7223 || : scott.wandschnei...@infocrossing.com **Think Green - Please print responsibly** Confidentiality Note: This e-mail, including any attachment to it, may contain material that is confidential, proprietary, privileged and/or "Protected Health Information," within the meaning of the regulations under the Health Insurance Portability& Accountability Act as amended. If it is not clear that you are the intended recipient, you are hereby notified that you have received this transmittal in error, and any review, dissemination, distribution or copying of this e-mail, including any attachment to it, is strictly prohibited. If you have received this e-mail in error, please immediately return it to the sender and delete it from your system. Thank you. -- Rich Smrcina Velocity Software, Inc. http://www.velocitysoftware.com Catch the WAVV! http://www.wavv.org WAVV 2011 - April 15-19, 2011 Colorado Springs, CO
Re: CMS SFS Question
I simply enroll any user to be deleted for 0 blocks. The alternative is to scan the sfs directories and files looking for such users. It is much easier to attempt the enroll. If it fails, it is because the user is already enrolled. Regards, Richard Schuh > -Original Message- > From: The IBM z/VM Operating System > [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Les Koehler > Sent: Tuesday, March 01, 2011 12:22 PM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: CMS SFS Question > > I'm curious: How do you find the user who is not enrolled, > but granted rights to the target user to be deleted? > > Les > > Schuh, Richard wrote: > > The Pipe is the easiest. > > > > PIPE < user list | spec /delete user/ 1 w1 nw | cms | > delete log a > > > > Note, however, that if you have an SFS that has a lot of > files and permissions, each DELETE USER can take a long time, > so you do not want to do this on an id that you might need > soon after you enter the PIPE command. In our shop, an > individual DELETE USER can take upwards of 10 minutes. > > > > Cleaning up SFS when a userid is deleted is important from > a security standpoint. If the same id should be given to a > different person, it would automatically inherit permissions > from the prior owner. You should be doing a DELETE USER every > time that a userid is deleted from the directory. > > > > It is possible for one user to grant access to other users > who are not enrolled. DELETE USER does not clean up these > permissions. To get rid of them, you have to first enroll the > user in the pool even if it is for 0 blocks. To solve this in > our automated process, each user to be deleted is enrolled > for 0 blocks, ignoring the return code. We don't care if the > user is already enrolled, the attempt does no harm. After the > enroll, the deletion will clean out all permissions granted > to or by the user being deleted. > > > > > > Regards, > > Richard Schuh > > > > > > > >> -Original Message----- > >> From: The IBM z/VM Operating System > >> [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Rick Troth > >> Sent: Tuesday, March 01, 2011 10:54 AM > >> To: IBMVM@LISTSERV.UARK.EDU > >> Subject: Re: CMS SFS Question > >> > >> Nahh ... even easier ... Pipes. > >> I'm thinking two pipes. One to gather the Q ENROLL output then a > >> second to actually perform the deletes. In between shove that Q > >> ENROLL output into a file, manually edit for confirmation, > then feed > >> the selected content into DELETE USER. > >> > >> -- R; > >> Rick Troth > >> Velocity Software > >> http://www.velocitysoftware.com/ > >> > >> > >> > >> On Tue, 1 Mar 2011, Rich Smrcina wrote: > >> > >>> REXX? > >>> > >>> On 03/01/2011 12:35 PM, Wandschneider, Scott wrote: > >>>> Is there a way to delete multiple users at once or create > >> a "batch" job to delete multiple users that are enrolled in SFS? > >>>> Thank you, > >>>> Scott R Wandschneider > >>>> Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 > >>>> Miracle Hills Drive, Omaha, NE, 68154-4457|| ': 402.963.8905 || > >>>> Ë:847.849.7223 || : > >> scott.wandschnei...@infocrossing.com **Think > >>>> Green - Please print responsibly** > >>>> > >>>> > >>>> > >>>> Confidentiality Note: This e-mail, including any > >> attachment to it, may contain material that is confidential, > >> proprietary, privileged and/or "Protected Health > Information," within > >> the meaning of the regulations under the Health Insurance > >> Portability& Accountability Act as amended. > >> If it is not clear that you are the intended recipient, you are > >> hereby notified that you have received this transmittal in > error, and > >> any review, dissemination, distribution or copying of this e-mail, > >> including any attachment to it, is strictly prohibited. If > you have > >> received this e-mail in error, please immediately return it to the > >> sender and delete it from your system. Thank you. > >>> > >>> -- > >>> Rich Smrcina > >>> Velocity Software, Inc. > >>> http://www.velocitysoftware.com > >>> > >>> Catch the WAVV! http://www.wavv.org > >>> WAVV 2011 - April 15-19, 2011 Colorado Springs, CO > >>> > >>> >
Re: CMS SFS Question
I'm curious: How do you find the user who is not enrolled, but granted rights to the target user to be deleted? Les Schuh, Richard wrote: The Pipe is the easiest. PIPE < user list | spec /delete user/ 1 w1 nw | cms | > delete log a Note, however, that if you have an SFS that has a lot of files and permissions, each DELETE USER can take a long time, so you do not want to do this on an id that you might need soon after you enter the PIPE command. In our shop, an individual DELETE USER can take upwards of 10 minutes. Cleaning up SFS when a userid is deleted is important from a security standpoint. If the same id should be given to a different person, it would automatically inherit permissions from the prior owner. You should be doing a DELETE USER every time that a userid is deleted from the directory. It is possible for one user to grant access to other users who are not enrolled. DELETE USER does not clean up these permissions. To get rid of them, you have to first enroll the user in the pool even if it is for 0 blocks. To solve this in our automated process, each user to be deleted is enrolled for 0 blocks, ignoring the return code. We don't care if the user is already enrolled, the attempt does no harm. After the enroll, the deletion will clean out all permissions granted to or by the user being deleted. Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Rick Troth Sent: Tuesday, March 01, 2011 10:54 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: CMS SFS Question Nahh ... even easier ... Pipes. I'm thinking two pipes. One to gather the Q ENROLL output then a second to actually perform the deletes. In between shove that Q ENROLL output into a file, manually edit for confirmation, then feed the selected content into DELETE USER. -- R; Rick Troth Velocity Software http://www.velocitysoftware.com/ On Tue, 1 Mar 2011, Rich Smrcina wrote: REXX? On 03/01/2011 12:35 PM, Wandschneider, Scott wrote: Is there a way to delete multiple users at once or create a "batch" job to delete multiple users that are enrolled in SFS? Thank you, Scott R Wandschneider Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 Miracle Hills Drive, Omaha, NE, 68154-4457|| ': 402.963.8905 || Ë:847.849.7223 || : scott.wandschnei...@infocrossing.com **Think Green - Please print responsibly** Confidentiality Note: This e-mail, including any attachment to it, may contain material that is confidential, proprietary, privileged and/or "Protected Health Information," within the meaning of the regulations under the Health Insurance Portability& Accountability Act as amended. If it is not clear that you are the intended recipient, you are hereby notified that you have received this transmittal in error, and any review, dissemination, distribution or copying of this e-mail, including any attachment to it, is strictly prohibited. If you have received this e-mail in error, please immediately return it to the sender and delete it from your system. Thank you. -- Rich Smrcina Velocity Software, Inc. http://www.velocitysoftware.com Catch the WAVV! http://www.wavv.org WAVV 2011 - April 15-19, 2011 Colorado Springs, CO
Re: CMS SFS Question
The Pipe is the easiest. PIPE < user list | spec /delete user/ 1 w1 nw | cms | > delete log a Note, however, that if you have an SFS that has a lot of files and permissions, each DELETE USER can take a long time, so you do not want to do this on an id that you might need soon after you enter the PIPE command. In our shop, an individual DELETE USER can take upwards of 10 minutes. Cleaning up SFS when a userid is deleted is important from a security standpoint. If the same id should be given to a different person, it would automatically inherit permissions from the prior owner. You should be doing a DELETE USER every time that a userid is deleted from the directory. It is possible for one user to grant access to other users who are not enrolled. DELETE USER does not clean up these permissions. To get rid of them, you have to first enroll the user in the pool even if it is for 0 blocks. To solve this in our automated process, each user to be deleted is enrolled for 0 blocks, ignoring the return code. We don't care if the user is already enrolled, the attempt does no harm. After the enroll, the deletion will clean out all permissions granted to or by the user being deleted. Regards, Richard Schuh > -Original Message- > From: The IBM z/VM Operating System > [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Rick Troth > Sent: Tuesday, March 01, 2011 10:54 AM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: CMS SFS Question > > Nahh ... even easier ... Pipes. > I'm thinking two pipes. One to gather the Q ENROLL output > then a second to actually perform the deletes. In between > shove that Q ENROLL output into a file, manually edit for > confirmation, then feed the selected content into DELETE USER. > > -- R; > Rick Troth > Velocity Software > http://www.velocitysoftware.com/ > > > > On Tue, 1 Mar 2011, Rich Smrcina wrote: > > > REXX? > > > > On 03/01/2011 12:35 PM, Wandschneider, Scott wrote: > > > Is there a way to delete multiple users at once or create > a "batch" job to delete multiple users that are enrolled in SFS? > > > > > > Thank you, > > > Scott R Wandschneider > > > Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 > > > Miracle Hills Drive, Omaha, NE, 68154-4457|| ': 402.963.8905 || > > > Ë:847.849.7223 || : > scott.wandschnei...@infocrossing.com **Think > > > Green - Please print responsibly** > > > > > > > > > > > > Confidentiality Note: This e-mail, including any > attachment to it, may contain material that is confidential, > proprietary, privileged and/or "Protected Health > Information," within the meaning of the regulations under the > Health Insurance Portability& Accountability Act as amended. > If it is not clear that you are the intended recipient, you > are hereby notified that you have received this transmittal > in error, and any review, dissemination, distribution or > copying of this e-mail, including any attachment to it, is > strictly prohibited. If you have received this e-mail in > error, please immediately return it to the sender and delete > it from your system. Thank you. > > > > > > -- > > Rich Smrcina > > Velocity Software, Inc. > > http://www.velocitysoftware.com > > > > Catch the WAVV! http://www.wavv.org > > WAVV 2011 - April 15-19, 2011 Colorado Springs, CO > > > > >
Re: CMS SFS Question
Nahh ... even easier ... Pipes. I'm thinking two pipes. One to gather the Q ENROLL output then a second to actually perform the deletes. In between shove that Q ENROLL output into a file, manually edit for confirmation, then feed the selected content into DELETE USER. -- R; Rick Troth Velocity Software http://www.velocitysoftware.com/ On Tue, 1 Mar 2011, Rich Smrcina wrote: > REXX? > > On 03/01/2011 12:35 PM, Wandschneider, Scott wrote: > > Is there a way to delete multiple users at once or create a "batch" job to > > delete multiple users that are enrolled in SFS? > > > > Thank you, > > Scott R Wandschneider > > Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 Miracle Hills > > Drive, Omaha, NE, 68154-4457|| : 402.963.8905 || :847.849.7223 || : > > scott.wandschnei...@infocrossing.com **Think Green - Please print > > responsibly** > > > > > > > > Confidentiality Note: This e-mail, including any attachment to it, may > > contain material that is confidential, proprietary, privileged and/or > > "Protected Health Information," within the meaning of the regulations under > > the Health Insurance Portability& Accountability Act as amended. If it is > > not clear that you are the intended recipient, you are hereby notified that > > you have received this transmittal in error, and any review, dissemination, > > distribution or copying of this e-mail, including any attachment to it, is > > strictly prohibited. If you have received this e-mail in error, please > > immediately return it to the sender and delete it from your system. Thank > > you. > > > -- > Rich Smrcina > Velocity Software, Inc. > http://www.velocitysoftware.com > > Catch the WAVV! http://www.wavv.org > WAVV 2011 - April 15-19, 2011 Colorado Springs, CO > >
Re: CMS SFS Question
REXX? On 03/01/2011 12:35 PM, Wandschneider, Scott wrote: Is there a way to delete multiple users at once or create a "batch" job to delete multiple users that are enrolled in SFS? Thank you, Scott R Wandschneider Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 Miracle Hills Drive, Omaha, NE, 68154-4457|| : 402.963.8905 || :847.849.7223 || : scott.wandschnei...@infocrossing.com **Think Green - Please print responsibly** Confidentiality Note: This e-mail, including any attachment to it, may contain material that is confidential, proprietary, privileged and/or "Protected Health Information," within the meaning of the regulations under the Health Insurance Portability& Accountability Act as amended. If it is not clear that you are the intended recipient, you are hereby notified that you have received this transmittal in error, and any review, dissemination, distribution or copying of this e-mail, including any attachment to it, is strictly prohibited. If you have received this e-mail in error, please immediately return it to the sender and delete it from your system. Thank you. -- Rich Smrcina Velocity Software, Inc. http://www.velocitysoftware.com Catch the WAVV! http://www.wavv.org WAVV 2011 - April 15-19, 2011 Colorado Springs, CO
CMS SFS Question
Is there a way to delete multiple users at once or create a "batch" job to delete multiple users that are enrolled in SFS? Thank you, Scott R Wandschneider Systems Programmer 3|| Infocrossing, a Wipro Company || 11707 Miracle Hills Drive, Omaha, NE, 68154-4457|| : 402.963.8905 || :847.849.7223 || : scott.wandschnei...@infocrossing.com **Think Green - Please print responsibly**
Re: SFS question
Not really true Dave: the SFSULIST panel is complete when not being SFS admin. Only the Catalog stats are not available to non-admins (upper right corner). But, yes, using PF11 to start DIRLIST of the user pointed too will only list the subdirs one is authorized to. 2011/1/4 Dave Jones > Steve, note that your user id must be enrolled in the file pool as an > administrator for SFSULIST to return all of the menaingful results. > > DJ > > On 01/04/2011 08:25 AM, Gentry, Stephen wrote: > > Thanks Dave. I was sure there was something out there that would > > provide this info. I had SFSULIST already downloaded just didn't know > > that's what it would/could do. > > Steve > > > > -Original Message- > > From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On > > Behalf Of Dave Jones > > Sent: Tuesday, January 04, 2011 9:20 AM > > To: IBMVM@LISTSERV.UARK.EDU > > Subject: Re: SFS question > > > > Steve, grab Kris Buelen's SFSULIST tool from the VM download page; it > > does exactly what you are looking for. > > > > Have a good one. > > > > On 01/04/2011 08:13 AM, Gentry, Stephen wrote: > >> How can I tell what users and/or what files are in a storage group? > > I'd > >> prefer to know users but can trace it back if I know what files. > >> > >> Thanks, > >> > >> Steve > >> > >> > > > > -- > Dave Jones > V/Soft Software > www.vsoft-software.com > Houston, TX > 281.578.7544 > -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS question
Steve, note that your user id must be enrolled in the file pool as an administrator for SFSULIST to return all of the menaingful results. DJ On 01/04/2011 08:25 AM, Gentry, Stephen wrote: > Thanks Dave. I was sure there was something out there that would > provide this info. I had SFSULIST already downloaded just didn't know > that's what it would/could do. > Steve > > -Original Message- > From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On > Behalf Of Dave Jones > Sent: Tuesday, January 04, 2011 9:20 AM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: SFS question > > Steve, grab Kris Buelen's SFSULIST tool from the VM download page; it > does exactly what you are looking for. > > Have a good one. > > On 01/04/2011 08:13 AM, Gentry, Stephen wrote: >> How can I tell what users and/or what files are in a storage group? > I'd >> prefer to know users but can trace it back if I know what files. >> >> Thanks, >> >> Steve >> >> > -- Dave Jones V/Soft Software www.vsoft-software.com Houston, TX 281.578.7544
Re: SFS question
SFSULIST is an interactive tool. If you need this information in a REXX exec: Q LIMITS is your friend. If you also need to know the files: after Q LIMITS, you use LISTDIR; ACCESS and LISTFILE... 2011/1/4 Gentry, Stephen > Thanks Dave. I was sure there was something out there that would > provide this info. I had SFSULIST already downloaded just didn't know > that's what it would/could do. > Steve > > -Original Message- > From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On > Behalf Of Dave Jones > Sent: Tuesday, January 04, 2011 9:20 AM > To: IBMVM@LISTSERV.UARK.EDU > Subject: Re: SFS question > > Steve, grab Kris Buelen's SFSULIST tool from the VM download page; it > does exactly what you are looking for. > > Have a good one. > > On 01/04/2011 08:13 AM, Gentry, Stephen wrote: > > How can I tell what users and/or what files are in a storage group? > I'd > > prefer to know users but can trace it back if I know what files. > > > > Thanks, > > > > Steve > > > > > > -- > Dave Jones > V/Soft Software > www.vsoft-software.com > Houston, TX > 281.578.7544 > -- Kris Buelens, IBM Belgium, VM customer support
Re: SFS question
Thanks Dave. I was sure there was something out there that would provide this info. I had SFSULIST already downloaded just didn't know that's what it would/could do. Steve -Original Message- From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Dave Jones Sent: Tuesday, January 04, 2011 9:20 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SFS question Steve, grab Kris Buelen's SFSULIST tool from the VM download page; it does exactly what you are looking for. Have a good one. On 01/04/2011 08:13 AM, Gentry, Stephen wrote: > How can I tell what users and/or what files are in a storage group? I'd > prefer to know users but can trace it back if I know what files. > > Thanks, > > Steve > > -- Dave Jones V/Soft Software www.vsoft-software.com Houston, TX 281.578.7544
Re: SFS question
Steve, grab Kris Buelen's SFSULIST tool from the VM download page; it does exactly what you are looking for. Have a good one. On 01/04/2011 08:13 AM, Gentry, Stephen wrote: > How can I tell what users and/or what files are in a storage group? I'd > prefer to know users but can trace it back if I know what files. > > Thanks, > > Steve > > -- Dave Jones V/Soft Software www.vsoft-software.com Houston, TX 281.578.7544
SFS question
How can I tell what users and/or what files are in a storage group? I'd prefer to know users but can trace it back if I know what files. Thanks, Steve
Re: Time out an SFS command
On Fri, 24 Dec 2010 18:44:04 +0100, Berry van Sleeuwen wrote: >Would PIPSTOP cancel the SFS command (or any CP/CMS command for that >matter)? If so then you could use PIPSTOP to end the pipeline. Just a >theory, not tested in any way... As John Hartmann explained: >> From: "John P. Hartmann" >> Subject: Re: Time out a CMS command >> >> The pipeline is dead in the water while the CMS command is executing; >> no way it can force a timeout. If you have two virtual machines to >> play with, the situation is different. Then you can send the command >> and wait for a response with STARMSG and send HX when it times out. >> Whether CMS reacts to the HX is another matter.
Re: Time out an SFS command
We think the problem is IPGATE, which IBM does not support, right? In one case when my PROFILE EXEC in Richmond, VA issues ACCESS SFSKC:ACKERMAN.TOOLS C I can fix the problem via #CP IPL ACC (NOPROF FORCE IPGATE AUTOLOG IPGATE In the case I reported here, I don't even get an error message until the RC = 55, 18 hours after the hang started. At that point the problem is gone. On Fri, 24 Dec 2010 00:20:52 -0500, Alan Altmark wrote: >John was right, you can't set a timeout for sfs commands. Having anothe r >ID hx your ID after a too-long wait is about it. That said, you're bett er >off trying to find out why your workunit is hanging and solve the *real* >problem. Maybe the server is hung up on a backup. > >Depending on what's going on, you may have grounds for a PMR. > >Alan Altmark >IBM Lab Services > >Merry Christmas, everyone! > = == ==
Re: Time out an SFS command
If anyone is interested, I have a TIMEBOMB module that lets a CMS user set the CP command(s) to be executed and the amount of time to wait before the timebomb 'explodes'. Drop me a note off list if anyone would like a copy. Merry Christmas to all DJ On 12/24/2010 07:44 AM, Ronald van der Laan wrote: > Or use the CP timebomb to issue a timed CP command within your own userid. > > Ronald van der Laan > -- Dave Jones V/Soft Software www.vsoft-software.com Houston, TX 281.578.7544
Re: Time out an SFS command
Would PIPSTOP cancel the SFS command (or any CP/CMS command for that matter)? If so then you could use PIPSTOP to end the pipeline. Just a theory, not tested in any way... 'PIPE (end \) command LISTDIR ...', /* execute command */ '| stem',/* store */ '\ literal +60', /* 60 seconds */ '| delay', /* wait for it */ '| pipstop' /* cancel command */ Regards, Berry. Op 24-12-10 01:09, Alan Ackerman schreef: > Moving this question from CMS-PIPELINES to IBMVM, since I am assured by > John Hartmann that it cannot be done with CMS Pipelines. Anyone figured > out how to get it to time out? > > Sender: CMSTSO Pipelines Discussion List pipeli...@vm.marist.edu> > From: "Ackerman, Alan" > Subject: Time out a CMS command > > Is there a PIPELINE idiom to force a CMS command to timeout after a > certain length of time? > > I am issuing: > > 'PIPE command LISTDIR SFSLNB00:SFSADMIN.GETSMTP | stem emsg.' > > Occasionally (about once a week), this hangs for a long time (18 hours > the last time) and then returns with RC = 55. > > I can live with RC=55, but not with my virtual machine being tied up for > 18 hours. (There are other VMSCHED jobs it should be running.) > > From: "Schuh, Richard" > Subject: Re: Time out a CMS command > > Look at the beat stage. It may work for this problem. > > From: "John P. Hartmann" > Subject: Re: Time out a CMS command > > The pipeline is dead in the water while the CMS command is executing; > no way it can force a timeout. If you have two virtual machines to > play with, the situation is different. Then you can send the command > and wait for a response with STARMSG and send HX when it times out. > Whether CMS reacts to the HX is another matter. > > From: Rob van der Heij > Subject: Re: Time out a CMS command > > Some SFS commands have the ability to hang forever and prevent any > recovery. I recall we did some check right before these, just to make > sure the remote file pool is actually alive. But I don't remember > which check it was, maybe Bruce or Rod can fill in the blanks... > > From: "Ackerman, Alan" > Subject: Re: Time out a CMS command > > I suppose I can just CP FORCE the hung userid. I'm not sure what state > that would leave SFS in, though. > > Date: Tue, 21 Dec 2010 20:16:34 -0500 > From: Rich Greenberg > Subject: Re: Time out a CMS command > > If you don't find a fix, instead of running it from VMSCHED, have > VMSCHED autolog another service machine for the actual LISTDIR. It > wouldn't be a problem if the extra SVM gets hung. > > Date: Wed, 22 Dec 2010 12:09:49 +0100 > From: Rod Furey > Subject: Re: Time out a CMS command > > Methinks Holger did it this way: > > start up a thread or two via multitasking CMS > set up a timer on one thread > set up the access (or whatever) on the other thread > if the access completes before the time ticks, kill the timer thread > if the timer thread ticks before the access completes, kill the access > thread > > Don't ask me about the ramifications of doing this and what happens > about cleanup. Multitasking CMS was never an area I hit before I went > in search of other work. > > I do recall that the dev group did discover some problems in the mtCMS > area at the time. > I would hope that they've been fixed by now. > > Date: Wed, 22 Dec 2010 08:07:34 -0500 > From: Bruce Hayden > Subject: Re: Time out a CMS command > > I haven't looked at this in years! But, looking at the code, you have > it essentially correct. The only sticky part is that the code doesn't > attempt to do an access, but tries to simulate some kind of SFS > communication via APPC, which I'm sure is not a documented interface > (it is coded in the exec as a long hex string.) > > Anyway, the best approach for the rest of us to this problem would be > to use 2 userids, as was already suggested. I doubt it would have any > affect on SFS. If this happened to my own id, I just enter #CP IPL > CMS to cancel the APPC wait and recover and there was never a problem > in SFS after the communication was fixed or reset. > > >
Re: Time out an SFS command
On 12/24/2010 8:44 AM, Ronald van der Laan wrote: Or use the CP timebomb to issue a timed CP command within your own userid. Ronald van der Laan What about doing an async dmsconn csl call, and then if you get no response in several iterations of dmscheck, then close the workunit ??? John
Re: Time out an SFS command
Or use the CP timebomb to issue a timed CP command within your own userid. Ronald van der Laan
Re: Time out an SFS command
John was right, you can't set a timeout for sfs commands. Having another ID hx your ID after a too-long wait is about it. That said, you're better off trying to find out why your workunit is hanging and solve the *real* problem. Maybe the server is hung up on a backup. Depending on what's going on, you may have grounds for a PMR. Alan Altmark IBM Lab Services Merry Christmas, everyone!
Re: Time out an SFS command
John was right, you can't cause an sfs command to time out. - Original Message - From: Alan Ackerman [alan.acker...@bankofamerica.com] Sent: 12/23/2010 06:09 PM CST To: IBMVM@LISTSERV.UARK.EDU Subject: [IBMVM] Time out an SFS command Moving this question from CMS-PIPELINES to IBMVM, since I am assured by John Hartmann that it cannot be done with CMS Pipelines. Anyone figured = out how to get it to time out? Sender: CMSTSO Pipelines Discussion List From: "Ackerman, Alan" Subject: Time out a CMS command Is there a PIPELINE idiom to force a CMS command to timeout after a certain length of time? I am issuing: 'PIPE command LISTDIR SFSLNB00:SFSADMIN.GETSMTP | stem emsg.' Occasionally (about once a week), this hangs for a long time (18 hours the last time) and then returns with RC = 55. I can live with RC=55, but not with my virtual machine being tied up fo= r 18 hours. (There are other VMSCHED jobs it should be running.) From: "Schuh, Richard" Subject: Re: Time out a CMS command Look at the beat stage. It may work for this problem. From: "John P. Hartmann" Subject: Re: Time out a CMS command The pipeline is dead in the water while the CMS command is executing; no way it can force a timeout. If you have two virtual machines to play with, the situation is different. Then you can send the command and wait for a response with STARMSG and send HX when it times out. Whether CMS reacts to the HX is another matter. From: Rob van der Heij Subject: Re: Time out a CMS command Some SFS commands have the ability to hang forever and prevent any recovery. I recall we did some check right before these, just to make sure the remote file pool is actually alive. But I don't remember which check it was, maybe Bruce or Rod can fill in the blanks... From: "Ackerman, Alan" Subject: Re: Time out a CMS command I suppose I can just CP FORCE the hung userid. I'm not sure what state that would leave SFS in, though. Date: Tue, 21 Dec 2010 20:16:34 -0500 From: Rich Greenberg Subject: Re: Time out a CMS command If you don't find a fix, instead of running it from VMSCHED, have VMSCHED autolog another service machine for the actual LISTDIR. It wouldn't be a problem if the extra SVM gets hung. Date: Wed, 22 Dec 2010 12:09:49 +0100 From: Rod Furey Subject: Re: Time out a CMS command Methinks Holger did it this way: start up a thread or two via multitasking CMS set up a timer on one thread set up the access (or whatever) on the other thread if the access completes before the time ticks, kill the timer thread if the timer thread ticks before the access completes, kill the access thread Don't ask me about the ramifications of doing this and what happens about cleanup. Multitasking CMS was never an area I hit before I went in search of other work. I do recall that the dev group did discover some problems in the mtCMS area at the time. I would hope that they've been fixed by now. Date: Wed, 22 Dec 2010 08:07:34 -0500 From: Bruce Hayden Subject: Re: Time out a CMS command I haven't looked at this in years! But, looking at the code, you have it essentially correct. The only sticky part is that the code doesn't attempt to do an access, but tries to simulate some kind of SFS communication via APPC, which I'm sure is not a documented interface (it is coded in the exec as a long hex string.) Anyway, the best approach for the rest of us to this problem would be to use 2 userids, as was already suggested. I doubt it would have any affect on SFS. If this happened to my own id, I just enter #CP IPL CMS to cancel the APPC wait and recover and there was never a problem in SFS after the communication was fixed or reset.