Re: Duplicate VOLID's
But if the normally lower one does not respond ... Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of David Boyes Sent: Friday, August 27, 2010 5:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Who's using it is a different issue, and with the size of systems we're starting to see, the name of the disk isn't going to tell you anything worthwhile anyway. 6 characters just isn't enough to overload with any useful meaning beyond a fairly small environment. I manage that problem in the change management system - OTRS is really quite nice for that. 8-) On 8/27/10 7:21 PM, Schuh, Richard rsc...@visa.com wrote: Order of acquisition will stay static. Is V04371 a spare or is it used by one of the n LPARS? If the latter, which one?
Duplicate VOLID's
There is my post of Thu, 20 Nov 2003 07:34:22 -0500 Richard Corak
Re: Duplicate VOLID's
That is absolutely the wrong thing to do. Is it so black and white? Our team has been using this convention for a number of years and it has worked fine for us. However, I will admit we have never been asked to migrate to a new disk array. I believe the main argument for not using the rdevs in the volser is that after migration, the volsers will be wrong (e.g. migrate VM at rdev to a new DASD at rdev and the volser convention is now wrong - the DASD at now has a volser of VM). I understand that changing the volsers throughout the (perhaps) many z/VM systems would be labor-intensive and perhaps error prone (e.g. clipping the VM to VM and then making sure that all references to VM are updated correctly). I agree that allowing this would break the convention and then things would become very confusing. But, is there not a model whereby the new DASD can have the same rdev range as the old DASD? For example: 1) Put the new DASD online assigning it a temporary range of rdevs (e.g. ) 2) FLASHCOPY/DDR the old DASD to the new DASD (e.g. = , and every other DASD in a z/VM LPAR) 3) Reassign the new DASD the old DASD's address ranges, and take the old DASD offline 4) Restart the LPAR. This should work on paper, but I may be missing something obvious. Yes, it would require more interaction with the hardware guys. Has anyone used this model? I'm just trying to understand and appreciate the wisdom of the community, where many have much more experience as sysadmins than I. Thanks. Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
I often bring up a second level VM system with the same volsers: 540RES, 540WK1, 540WK2, ... I always use a higher address for the duplicated volume when I DDR copy. In other words if 540WK1 is say 209, the DDR copied volume would be something like 847. This way, if 1st level VM happens to go down, before I finish my testing, it should find the lower addressed volumes (production volumes) and use these to IPL with. VM will start looking at each address with lower addresses prior to higher addresses. The only way this would fail, would be if VM fails on the read of the production VM volume or volumes. Once testing is completed, I ICKDSF these copied volumes back to their original state/volid. Been doing this for 30 years and haven't got burned yet, yet, yet. Ray From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 9:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.commailto:tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 4:56 PM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Michael MacIsaac Sent: Thursday, August 26, 2010 10:55 AM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.commailto:mike...@us.ibm.com (845) 433-7061 NOTICE: This e-mail is intended solely for the use of the individual to whom it is addressed and may contain information that is privileged, confidential or otherwise exempt from disclosure. If the reader of this e-mail is not the intended
Re: Duplicate VOLID's
OK Ray, that is the best tip yet. But there are flaws you are assuming that you are the one in control. Picture a situation where there is a pool of DASD that is available for 'anyone' to use, for anything they want to use them for. They could have z/VM system volumes, z/OS, z/LINUX, z/VSE, or just application data. At some point they all get reinitialized but not until the user is done which could be weeks. Also consider that there are several z/VM 'production' systems in different LPARS or CEC's They can't all be at the lowest address.. Some others have suggested labeling schemes ie VMrdev. As for labeling the volumes with the rdev ie VM1234, yes I use that scheme when returning the packs to the user pool. Labeling the packs as VMrdev doesn't address the problem, I am not interested tracking those volid's I only care about xxxRES, xxxWK1, xxxWK2, etc On Fri, Aug 27, 2010 at 7:47 AM, Ray Waters ray.wat...@opensolutions.comwrote: I often bring up a second level VM system with the same volsers: 540RES, 540WK1, 540WK2, … I always use a higher address for the duplicated volume when I DDR copy. In other words if 540WK1 is say 209, the DDR copied volume would be something like 847. This way, if 1st level VM happens to go down, before I finish my testing, it should find the lower addressed volumes (production volumes) and use these to IPL with. VM will start looking at each address with lower addresses prior to higher addresses. The only way this would fail, would be if VM fails on the read of the production VM volume or volumes. Once testing is completed, I ICKDSF these copied volumes back to their original state/volid. Been doing this for 30 years and haven’t got burned yet, yet, yet. Ray *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Scott Rohling *Sent:* Thursday, August 26, 2010 9:45 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Scott Rohling *Sent:* Thursday, August 26, 2010 4:56 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody
Re: Duplicate VOLID's
Yes -- that's a good strategy too ... I usually pay attention to the addresses - but this was our test/play system, so didn't give it the careful thought I normally would. Higher addresses can also prevent an issue with duplicates Scott Rohling On Fri, Aug 27, 2010 at 6:47 AM, Ray Waters ray.wat...@opensolutions.comwrote: I often bring up a second level VM system with the same volsers: 540RES, 540WK1, 540WK2, … I always use a higher address for the duplicated volume when I DDR copy. In other words if 540WK1 is say 209, the DDR copied volume would be something like 847. This way, if 1st level VM happens to go down, before I finish my testing, it should find the lower addressed volumes (production volumes) and use these to IPL with. VM will start looking at each address with lower addresses prior to higher addresses. The only way this would fail, would be if VM fails on the read of the production VM volume or volumes. Once testing is completed, I ICKDSF these copied volumes back to their original state/volid. Been doing this for 30 years and haven’t got burned yet, yet, yet. Ray *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Scott Rohling *Sent:* Thursday, August 26, 2010 9:45 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Scott Rohling *Sent:* Thursday, August 26, 2010 4:56 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Michael MacIsaac *Sent:* Thursday, August 26, 2010 10:55 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061 -- NOTICE: This e-mail
Re: Duplicate VOLID's
Very True Tom. I was just giving a situation with one second level guest. I realize, some shops cannot control how many duplicated volsers and who is creating these duplicates (such as a school). My situation is a controlled environment, where only Tech Support does this kind of stuff, and we communicate well. Ray From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Tom Huegel Sent: Friday, August 27, 2010 8:22 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's OK Ray, that is the best tip yet. But there are flaws you are assuming that you are the one in control. Picture a situation where there is a pool of DASD that is available for 'anyone' to use, for anything they want to use them for. They could have z/VM system volumes, z/OS, z/LINUX, z/VSE, or just application data. At some point they all get reinitialized but not until the user is done which could be weeks. Also consider that there are several z/VM 'production' systems in different LPARS or CEC's They can't all be at the lowest address.. Some others have suggested labeling schemes ie VMrdev. As for labeling the volumes with the rdev ie VM1234, yes I use that scheme when returning the packs to the user pool. Labeling the packs as VMrdev doesn't address the problem, I am not interested tracking those volid's I only care about xxxRES, xxxWK1, xxxWK2, etc On Fri, Aug 27, 2010 at 7:47 AM, Ray Waters ray.wat...@opensolutions.commailto:ray.wat...@opensolutions.com wrote: I often bring up a second level VM system with the same volsers: 540RES, 540WK1, 540WK2, ... I always use a higher address for the duplicated volume when I DDR copy. In other words if 540WK1 is say 209, the DDR copied volume would be something like 847. This way, if 1st level VM happens to go down, before I finish my testing, it should find the lower addressed volumes (production volumes) and use these to IPL with. VM will start looking at each address with lower addresses prior to higher addresses. The only way this would fail, would be if VM fails on the read of the production VM volume or volumes. Once testing is completed, I ICKDSF these copied volumes back to their original state/volid. Been doing this for 30 years and haven't got burned yet, yet, yet. Ray From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 9:45 PM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.commailto:tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 4:56 PM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: That is absolutely
Re: Duplicate VOLID's
Also, if you have dasd that is EMIFd to n LPARS, how do you know (a) whether a disk is spare or in use if all of the dasd have volsers ZL, and (b) if in use, by which LPAR? IO can see using VM as an indicator that a volume has not yet been deployed, but once it has, it needs to be labeled using another convention. Since we seem to replace our DASD farm every 2 or 3 years, device number has no place in the scheme other than to denote a spare volume. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Michael MacIsaac Sent: Friday, August 27, 2010 5:31 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's That is absolutely the wrong thing to do. Is it so black and white? Our team has been using this convention for a number of years and it has worked fine for us. However, I will admit we have never been asked to migrate to a new disk array. I believe the main argument for not using the rdevs in the volser is that after migration, the volsers will be wrong (e.g. migrate VM at rdev to a new DASD at rdev and the volser convention is now wrong - the DASD at now has a volser of VM). I understand that changing the volsers throughout the (perhaps) many z/VM systems would be labor-intensive and perhaps error prone (e.g. clipping the VM to VM and then making sure that all references to VM are updated correctly). I agree that allowing this would break the convention and then things would become very confusing. But, is there not a model whereby the new DASD can have the same rdev range as the old DASD? For example: 1) Put the new DASD online assigning it a temporary range of rdevs (e.g. ) 2) FLASHCOPY/DDR the old DASD to the new DASD (e.g. = , and every other DASD in a z/VM LPAR) 3) Reassign the new DASD the old DASD's address ranges, and take the old DASD offline 4) Restart the LPAR. This should work on paper, but I may be missing something obvious. Yes, it would require more interaction with the hardware guys. Has anyone used this model? I'm just trying to understand and appreciate the wisdom of the community, where many have much more experience as sysadmins than I. Thanks. Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
Tom, Been following this with interest. It sounds like your environment is a lot different from most of the responders, so it has been educational looking at what everyone else does. You have said that you know the volumes addresses of the production VM volumes. If so, why not code the SYSTEM CONFIG to only bring those online at IPL? Leave all the rest offline until you are up, then use an exec (perhaps on AUTOLOG1) to bring the rest of the volumes online? In ancient days, before there was a SYSTEM CONFIG, my AUTOLOG1 always brought the volumes up and attached them to the system. There have been two approaches for multiple VM res volumes discussed on the list in the past. One is to make the production volumes have different volsers than everyone else. You only need to do that for the floor system and let your many testers sort through whose 540RES is this? If you have to do that for many LPAR, then you have the fun of coming up with many unique names. Personally, my test VM system uses volumes that are copies of the floor system but one cylinder less, so I can leave the real volser untouched and give my testbed mdisk volumes that start at cylinder 1. IBM still makes RES, W01 and W02 stop 1 cylinder before full (something that they were doing when I put in my first VM system on 3310s in 1979) and I think that is so you can create a complete copy on a second volume but not need to touch cyl 0 of the second volume. These volumes are not IPLable on different LPAR, or different machines sharing the DASD subsystem, but it works okay on the same LPAR. It works for me because I have one machine and no LPAR. Ron On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments?
Re: Duplicate VOLID's
It will only take one use of an incorrect spool volume to really burn you. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Ray Waters Sent: Friday, August 27, 2010 5:47 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's I often bring up a second level VM system with the same volsers: 540RES, 540WK1, 540WK2, ... I always use a higher address for the duplicated volume when I DDR copy. In other words if 540WK1 is say 209, the DDR copied volume would be something like 847. This way, if 1st level VM happens to go down, before I finish my testing, it should find the lower addressed volumes (production volumes) and use these to IPL with. VM will start looking at each address with lower addresses prior to higher addresses. The only way this would fail, would be if VM fails on the read of the production VM volume or volumes. Once testing is completed, I ICKDSF these copied volumes back to their original state/volid. Been doing this for 30 years and haven't got burned yet, yet, yet. Ray From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 9:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.commailto:tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 4:56 PM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Michael MacIsaac Sent: Thursday, August 26, 2010 10:55 AM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike
Re: Duplicate VOLID's
Do higher addresses really protect you? I do not think that there is a guarantee that they do. And if IBM does not guarantee it, depending on something that seems to be is a risk that has no place in a production system. The volsers belonging to the production system should not be used for disks belonging to any second level system that can be ipled on the bare iron or whose disks begin at cylinder 0. Yes, you have been doing it for 30 years. That is kind of like cheating on your spouse for 30 years and not getting caught. You only need to get caught once. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Scott Rohling Sent: Friday, August 27, 2010 6:25 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Yes -- that's a good strategy too ... I usually pay attention to the addresses - but this was our test/play system, so didn't give it the careful thought I normally would. Higher addresses can also prevent an issue with duplicates Scott Rohling On Fri, Aug 27, 2010 at 6:47 AM, Ray Waters ray.wat...@opensolutions.commailto:ray.wat...@opensolutions.com wrote: I often bring up a second level VM system with the same volsers: 540RES, 540WK1, 540WK2, ... I always use a higher address for the duplicated volume when I DDR copy. In other words if 540WK1 is say 209, the DDR copied volume would be something like 847. This way, if 1st level VM happens to go down, before I finish my testing, it should find the lower addressed volumes (production volumes) and use these to IPL with. VM will start looking at each address with lower addresses prior to higher addresses. The only way this would fail, would be if VM fails on the read of the production VM volume or volumes. Once testing is completed, I ICKDSF these copied volumes back to their original state/volid. Been doing this for 30 years and haven't got burned yet, yet, yet. Ray From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 9:45 PM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.commailto:tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 4:56 PM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk
Re: Duplicate VOLID's
Richard is right on this as far as i remember to have read on the online help that order of chpids in the config is the way the rdev are known by the system. We have 4 clones of our production + 2 ranges of contigus addresses cloned on the same array... all the possibilities to IPl is driven by a specific config file. At salpl, the ipled dasd on thé Hmc is the filename of the config file. So all these files are propagated at clone time. Never had a problem that way. Alain Envoyé de mon iPhone Le 27 août 2010 à 17:55, Schuh, Richard rsc...@visa.com a écrit : Do higher addresses really protect you? I do not think that there is a guarantee that they do. And if IBM does not guarantee it, depending on something that seems to be is a risk that has no place in a production system. The volsers belonging to the production system should not be used for disks belonging to any second level system that can be ipled on the bare iron or whose disks begin at cylinder 0. Yes, you have been doing it for 30 years. That is kind of like cheating on your spouse for 30 years and not getting caught. You only need to get caught once. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Scott Rohling Sent: Friday, August 27, 2010 6:25 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Yes -- that's a good strategy too ... I usually pay attention to the addresses - but this was our test/play system, so didn't give it the careful thought I normally would. Higher addresses can also prevent an issue with duplicates Scott Rohling On Fri, Aug 27, 2010 at 6:47 AM, Ray Waters ray.wat...@opensolutions.com wrote: I often bring up a second level VM system with the same volsers: 540RES, 540WK1, 540WK2, … I always use a higher address for the duplicated volume when I DDR copy. In other words if 540WK1 is say 209, the DDR copied volume would be something like 847. This way, if 1st level VM happens to go down, before I finish my testing, it should find the lower addressed volumes (production volumes) and use these to IPL with. VM will start looking at each address with lower addresses prior to higher addresses. The only way this would fail, would be if VM fails on the read of the production VM volume or volumes. Once testing is completed, I ICKDSF these copied volumes back to their original state/volid. Been doing this for 30 years and haven’t got burned yet, yet, yet. Ray From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 9:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 4:56 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - anddoesn't care about volsers the mindisks are on.The processfor DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update
Re: Duplicate VOLID's
On Thursday, 08/26/2010 at 09:58 EDT, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? Rich Corak has explained on previous occasions how volser recognition works. If CP detects a duplicate volser for a device to be attached to SYSTEM at IPL, you will get HCP954I DASD rdev1 VOLID volid IS A DUPLICATE OF DASD rdev2 and you are responsible to ensure that rdev2 is the one you intended to be attached to the system. Out of all the volumes on the system with that volid, the lowest *device number* (address) will win, without regard to who responds first. As has been suggested, this is no guarantee since you can find yourself in trouble if you have a dasd problem or someone dinks with the I/O config and makes a mistake. Being able to place the RDEV on the CP_owned and user_volume_ statements is one way to help mitigate the problem. There is another way to identify devices: by their architected node element descriptor (NED). Or in the lingua franca, universally unique identifiers (UUIDs). When Single System Image hits the streets, you will find UUIDs very much in evidence, though not in the context of IPL or SYSTEM CONFIG. But I could imagine something like this in SYSTEM CONFIG: CP_Owned Slot 1 6X0RES ID 2105.000.IBM.13.3737504EE.0D0A CP_owned Slot 2 6X0TD1 ID 2107.900.IBM.13.29839621A.0D0A CP_owned Slot 3 6X0PG1 ID 2107.900.IBM.13.4924295DC.0D0A, (yes, there are two IDs for slot 3) 2107.900.IBM.13.0358113AA.0D0A (..choose in the order given) so that you wouldn't have to worry about the RDEV. However, there is a Heisenberg-esque trade-off to be made between flexibility and security. Good news - the above syntax ensures you won't have any bogus volumes. Bad news - it won't tolerate any copies or change in DASD, so it's generally hostile to a dynamic DR environment, and you can't get the NED until you IPL. Hmmmorwhen IPL has finished, CP could write the found UUIDs to the warm start area and use those in preference to any other on a subsequent IPL. For PPRC pairs, write both UUIDs and allow either. If a volume with the needed UUID is not available, then (based on configuration) ask which RDEV to use a la z/OS, or just take the lowest-numbered one available. In either case, update the UUID in the warm start area. h. Alan Altmark z/VM Development IBM Endicott
Re: Duplicate VOLID's
Yes, and it's specified in examples in some documentation (Redbook Paper Linux on IBM eserver zSeries and S/390: VSWITCH and VLAN Features of z/VM 4.4). Yeah, it's old! :) Frank M. Ramaekers Jr. -Original Message- From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Alan Altmark Sent: Friday, August 27, 2010 1:38 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's On Thursday, 08/26/2010 at 09:58 EDT, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? Rich Corak has explained on previous occasions how volser recognition works. If CP detects a duplicate volser for a device to be attached to SYSTEM at IPL, you will get HCP954I DASD rdev1 VOLID volid IS A DUPLICATE OF DASD rdev2 and you are responsible to ensure that rdev2 is the one you intended to be attached to the system. Out of all the volumes on the system with that volid, the lowest *device number* (address) will win, without regard to who responds first. As has been suggested, this is no guarantee since you can find yourself in trouble if you have a dasd problem or someone dinks with the I/O config and makes a mistake. Being able to place the RDEV on the CP_owned and user_volume_ statements is one way to help mitigate the problem. There is another way to identify devices: by their architected node element descriptor (NED). Or in the lingua franca, universally unique identifiers (UUIDs). When Single System Image hits the streets, you will find UUIDs very much in evidence, though not in the context of IPL or SYSTEM CONFIG. But I could imagine something like this in SYSTEM CONFIG: CP_Owned Slot 1 6X0RES ID 2105.000.IBM.13.3737504EE.0D0A CP_owned Slot 2 6X0TD1 ID 2107.900.IBM.13.29839621A.0D0A CP_owned Slot 3 6X0PG1 ID 2107.900.IBM.13.4924295DC.0D0A, (yes, there are two IDs for slot 3) 2107.900.IBM.13.0358113AA.0D0A (..choose in the order given) so that you wouldn't have to worry about the RDEV. However, there is a Heisenberg-esque trade-off to be made between flexibility and security. Good news - the above syntax ensures you won't have any bogus volumes. Bad news - it won't tolerate any copies or change in DASD, so it's generally hostile to a dynamic DR environment, and you can't get the NED until you IPL. Hmmmorwhen IPL has finished, CP could write the found UUIDs to the warm start area and use those in preference to any other on a subsequent IPL. For PPRC pairs, write both UUIDs and allow either. If a volume with the needed UUID is not available, then (based on configuration) ask which RDEV to use a la z/OS, or just take the lowest-numbered one available. In either case, update the UUID in the warm start area. h. Alan Altmark z/VM Development IBM Endicott _ This message contains information which is privileged and confidential and is solely for the use of the intended recipient. If you are not the intended recipient, be aware that any review, disclosure, copying, distribution, or use of the contents of this message is strictly prohibited. If you have received this in error, please destroy it immediately and notify us at privacy...@ailife.com.
Re: Duplicate VOLID's
BINGO!! Problem solved. On Fri, Aug 27, 2010 at 11:38 AM, Alan Altmark alan_altm...@us.ibm.comwrote: On Thursday, 08/26/2010 at 09:58 EDT, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? Rich Corak has explained on previous occasions how volser recognition works. If CP detects a duplicate volser for a device to be attached to SYSTEM at IPL, you will get HCP954I DASD rdev1 VOLID volid IS A DUPLICATE OF DASD rdev2 and you are responsible to ensure that rdev2 is the one you intended to be attached to the system. Out of all the volumes on the system with that volid, the lowest *device number* (address) will win, without regard to who responds first. As has been suggested, this is no guarantee since you can find yourself in trouble if you have a dasd problem or someone dinks with the I/O config and makes a mistake. Being able to place the RDEV on the CP_owned and user_volume_ statements is one way to help mitigate the problem. There is another way to identify devices: by their architected node element descriptor (NED). Or in the lingua franca, universally unique identifiers (UUIDs). When Single System Image hits the streets, you will find UUIDs very much in evidence, though not in the context of IPL or SYSTEM CONFIG. But I could imagine something like this in SYSTEM CONFIG: CP_Owned Slot 1 6X0RES ID 2105.000.IBM.13.3737504EE.0D0A CP_owned Slot 2 6X0TD1 ID 2107.900.IBM.13.29839621A.0D0A CP_owned Slot 3 6X0PG1 ID 2107.900.IBM.13.4924295DC.0D0A, (yes, there are two IDs for slot 3) 2107.900.IBM.13.0358113AA.0D0A (..choose in the order given) so that you wouldn't have to worry about the RDEV. However, there is a Heisenberg-esque trade-off to be made between flexibility and security. Good news - the above syntax ensures you won't have any bogus volumes. Bad news - it won't tolerate any copies or change in DASD, so it's generally hostile to a dynamic DR environment, and you can't get the NED until you IPL. Hmmmorwhen IPL has finished, CP could write the found UUIDs to the warm start area and use those in preference to any other on a subsequent IPL. For PPRC pairs, write both UUIDs and allow either. If a volume with the needed UUID is not available, then (based on configuration) ask which RDEV to use a la z/OS, or just take the lowest-numbered one available. In either case, update the UUID in the warm start area. h. Alan Altmark z/VM Development IBM Endicott
Re: Duplicate VOLID's
On Friday, 08/27/2010 at 12:22 EDT, Mike Walter mike.wal...@hewitt.com wrote: There have been previous reliable posts (which I can't seem to find right now) relating that the order in which real devices (RDEVs) come online during IPL is NOT GUARANTEED to be in serially ascending order. Certainly not since the XA subsystem was implemented. Perhaps not even before then, some ancient, little-used synapses are whispering contingent connection in my ear (along with all those other people who just whisper in my mind). That's how it worked ages ago, but it was specifically re-engineered to work in ascending RDEV order, regardless of variation in DASD response times or subchannel assignments. Alan Altmark z/VM Development IBM Endicott
Re: Duplicate VOLID's
Well, there you go, Alan. There's a reason I (and you) remember it from ages ago. It's hard to break some habits, and even harder to unlearn some painfully learned knowledge. BTW, I *did* search the list many different ways for Rich Corak's posts (his name was one of the first search arguments I tried). No joy. Mike (grey-beard) Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. Alan Altmark alan_altm...@us.ibm.com Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 08/27/2010 04:09 PM Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject Re: Duplicate VOLID's On Friday, 08/27/2010 at 12:22 EDT, Mike Walter mike.wal...@hewitt.com wrote: There have been previous reliable posts (which I can't seem to find right now) relating that the order in which real devices (RDEVs) come online during IPL is NOT GUARANTEED to be in serially ascending order. Certainly not since the XA subsystem was implemented. Perhaps not even before then, some ancient, little-used synapses are whispering contingent connection in my ear (along with all those other people who just whisper in my mind). That's how it worked ages ago, but it was specifically re-engineered to work in ascending RDEV order, regardless of variation in DASD response times or subchannel assignments. Alan Altmark z/VM Development IBM Endicott 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: Duplicate VOLID's
The order it searches is guaranteed, as is the order of non-responsive devices. If a spool volume does not respond but there is a duplicate volser in the configuration, you still have problems, especially if the operator says GO. Having a guaranteed order does lessen the threat considerably, but it does not completely eliminate it. There have only been two ways to eliminate it mentioned in this thread. 1. Insure that each device has a unique volser, or 2. have the duplicates on devices that are Offline_at_IPL and wait until the system initialization completes before varying them on. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Mike Walter Sent: Friday, August 27, 2010 2:14 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Well, there you go, Alan. There's a reason I (and you) remember it from ages ago. It's hard to break some habits, and even harder to unlearn some painfully learned knowledge. BTW, I *did* search the list many different ways for Rich Corak's posts (his name was one of the first search arguments I tried). No joy. Mike (grey-beard) Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. Alan Altmark alan_altm...@us.ibm.com Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 08/27/2010 04:09 PM Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject Re: Duplicate VOLID's On Friday, 08/27/2010 at 12:22 EDT, Mike Walter mike.wal...@hewitt.com wrote: There have been previous reliable posts (which I can't seem to find right now) relating that the order in which real devices (RDEVs) come online during IPL is NOT GUARANTEED to be in serially ascending order. Certainly not since the XA subsystem was implemented. Perhaps not even before then, some ancient, little-used synapses are whispering contingent connection in my ear (along with all those other people who just whisper in my mind). That's how it worked ages ago, but it was specifically re-engineered to work in ascending RDEV order, regardless of variation in DASD response times or subchannel assignments. Alan Altmark z/VM Development IBM Endicott 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: Duplicate VOLID's
Mike I understand your pain. Bill Bitner just this week had to tell me my thoughts on processor utiliization was OUTDATED. So sad. Even grayer beard Bill Munson - Original Message - From: Mike Walter [mike.wal...@hewitt.com] Sent: 08/27/2010 04:14 PM EST To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Well, there you go, Alan. There's a reason I (and you) remember it from ages ago. It's hard to break some habits, and even harder to unlearn some painfully learned knowledge. BTW, I *did* search the list many different ways for Rich Corak's posts (his name was one of the first search arguments I tried). No joy. Mike (grey-beard) Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. Alan Altmark alan_altm...@us.ibm.com Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 08/27/2010 04:09 PM Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject Re: Duplicate VOLID's On Friday, 08/27/2010 at 12:22 EDT, Mike Walter mike.wal...@hewitt.com wrote: There have been previous reliable posts (which I can't seem to find right now) relating that the order in which real devices (RDEVs) come online during IPL is NOT GUARANTEED to be in serially ascending order. Certainly not since the XA subsystem was implemented. Perhaps not even before then, some ancient, little-used synapses are whispering contingent connection in my ear (along with all those other people who just whisper in my mind). That's how it worked ages ago, but it was specifically re-engineered to work in ascending RDEV order, regardless of variation in DASD response times or subchannel assignments. Alan Altmark z/VM Development IBM Endicott 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. *** IMPORTANT NOTE*-- The opinions expressed in this message and/or any attachments are those of the author and not necessarily those of Brown Brothers Harriman Co., its subsidiaries and affiliates (BBH). There is no guarantee that this message is either private or confidential, and it may have been altered by unauthorized sources without your or our knowledge. Nothing in the message is capable or intended to create any legally binding obligations on either party and it is not intended to provide legal advice. BBH accepts no responsibility for loss or damage from its use, including damage from virus.
Re: Duplicate VOLID's
I’ve always liked Vn, where n is the order of acquisition in your organization, eg V1 for the first volume acquired, V2 for the second, etc. The V is there to not confuse commands that need a volser as input into thinking they’re dealing with a real address (once had that problem with a system that had all numeric userids --- ick), and I’ve never rolled over within the same complex during my time there. If you stick to decimal numbering most people don’t confuse them with addresses, and if you go through a lot of disks (as I assume Visa does), you’re still unlikely to replace 99,999 at a time. Works with LPARs, VM, z/OS, pretty much any situation you come to. On 8/27/10 10:45 AM, Schuh, Richard rsc...@visa.com wrote: Also, if you have dasd that is EMIFd to n LPARS, how do you know (a) whether a disk is spare or in use if all of the dasd have volsers ZL, and (b) if in use, by which LPAR? IO can see using VM as an indicator that a volume has not yet been deployed, but once it has, it needs to be labeled using another convention. Since we seem to replace our DASD farm every 2 or 3 years, device number has no place in the scheme other than to denote a spare volume. Regards, Richard Schuh
Re: Duplicate VOLID's
On 8/27/10 1:38 PM, Alan Altmark alan_altm...@us.ibm.com wrote: But I could imagine something like this in SYSTEM CONFIG: CP_Owned Slot 1 6X0RES ID 2105.000.IBM.13.3737504EE.0D0A CP_owned Slot 2 6X0TD1 ID 2107.900.IBM.13.29839621A.0D0A CP_owned Slot 3 6X0PG1 ID 2107.900.IBM.13.4924295DC.0D0A, (yes, there are two IDs for slot 3) 2107.900.IBM.13.0358113AA.0D0A (..choose in the order given) Meh. Kinda has the same problem that ATM addresses had -- too much detail stored as context. Hmmmorwhen IPL has finished, CP could write the found UUIDs to the warm start area and use those in preference to any other on a subsequent IPL. For PPRC pairs, write both UUIDs and allow either. If a volume with the needed UUID is not available, then (based on configuration) ask which RDEV to use a la z/OS, or just take the lowest-numbered one available. In either case, update the UUID in the warm start area. h. Much better.
Re: Duplicate VOLID's
Who wants to write the requirement? On Fri, Aug 27, 2010 at 4:04 PM, David Boyes dbo...@sinenomine.net wrote: On 8/27/10 1:38 PM, Alan Altmark alan_altm...@us.ibm.com wrote: But I could imagine something like this in SYSTEM CONFIG: CP_Owned Slot 1 6X0RES ID 2105.000.IBM.13.3737504EE.0D0A CP_owned Slot 2 6X0TD1 ID 2107.900.IBM.13.29839621A.0D0A CP_owned Slot 3 6X0PG1 ID 2107.900.IBM.13.4924295DC.0D0A, (yes, there are two IDs for slot 3) 2107.900.IBM.13.0358113AA.0D0A (..choose in the order given) Meh. Kinda has the same problem that ATM addresses had -- too much detail stored as context. Hmmmorwhen IPL has finished, CP could write the found UUIDs to the warm start area and use those in preference to any other on a subsequent IPL. For PPRC pairs, write both UUIDs and allow either. If a volume with the needed UUID is not available, then (based on configuration) ask which RDEV to use a la z/OS, or just take the lowest-numbered one available. In either case, update the UUID in the warm start area. h. Much better.
Re: Duplicate VOLID's
Order of acquisition will stay static. Is V04371 a spare or is it used by one of the n LPARS? If the latter, which one? Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of David Boyes Sent: Friday, August 27, 2010 4:00 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's I've always liked Vn, where n is the order of acquisition in your organization, eg V1 for the first volume acquired, V2 for the second, etc. The V is there to not confuse commands that need a volser as input into thinking they're dealing with a real address (once had that problem with a system that had all numeric userids --- ick), and I've never rolled over within the same complex during my time there. If you stick to decimal numbering most people don't confuse them with addresses, and if you go through a lot of disks (as I assume Visa does), you're still unlikely to replace 99,999 at a time. Works with LPARs, VM, z/OS, pretty much any situation you come to. On 8/27/10 10:45 AM, Schuh, Richard rsc...@visa.com wrote: Also, if you have dasd that is EMIFd to n LPARS, how do you know (a) whether a disk is spare or in use if all of the dasd have volsers ZL, and (b) if in use, by which LPAR? IO can see using VM as an indicator that a volume has not yet been deployed, but once it has, it needs to be labeled using another convention. Since we seem to replace our DASD farm every 2 or 3 years, device number has no place in the scheme other than to denote a spare volume. Regards, Richard Schuh
Re: Duplicate VOLID's
Just submit Alan's suggestion as-is and make sure that he gets full credit for it. :-) Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Tom Huegel Sent: Friday, August 27, 2010 5:12 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Who wants to write the requirement? On Fri, Aug 27, 2010 at 4:04 PM, David Boyes dbo...@sinenomine.netmailto:dbo...@sinenomine.net wrote: On 8/27/10 1:38 PM, Alan Altmark alan_altm...@us.ibm.commailto:alan_altm...@us.ibm.com wrote: But I could imagine something like this in SYSTEM CONFIG: CP_Owned Slot 1 6X0RES ID 2105.000.IBM.13.3737504EE.0D0A CP_owned Slot 2 6X0TD1 ID 2107.900.IBM.13.29839621A.0D0A CP_owned Slot 3 6X0PG1 ID 2107.900.IBM.13.4924295DC.0D0A, (yes, there are two IDs for slot 3) 2107.900.IBM.13.0358113AA.0D0A (..choose in the order given) Meh. Kinda has the same problem that ATM addresses had -- too much detail stored as context. Hmmmorwhen IPL has finished, CP could write the found UUIDs to the warm start area and use those in preference to any other on a subsequent IPL. For PPRC pairs, write both UUIDs and allow either. If a volume with the needed UUID is not available, then (based on configuration) ask which RDEV to use a la z/OS, or just take the lowest-numbered one available. In either case, update the UUID in the warm start area. h. Much better.
Re: Duplicate VOLID's
Who’s using it is a different issue, and with the size of systems we’re starting to see, the name of the disk isn’t going to tell you anything worthwhile anyway. 6 characters just isn’t enough to overload with any useful meaning beyond a fairly small environment. I manage that problem in the change management system — OTRS is really quite nice for that. 8-) On 8/27/10 7:21 PM, Schuh, Richard rsc...@visa.com wrote: Order of acquisition will stay static. Is V04371 a spare or is it used by one of the n LPARS? If the latter, which one?
Re: Duplicate VOLID's
On Wed, 25 Aug 2010 23:07:04 -0500, Tom Huegel tehue...@gmail.com wrote : In a normal production environment this is not such an insurmountable problem. The problem is this is a test lab, and I don't necessarily know what happens to the different disk volumes. I do know what addresses my system disks are on, but there may be copies that someone was testing with floating around. It may have been a in a second level machine, or a first level test.. It's only a problem at IPL .. The suggestions for using ONLINE_AT_IPL and OFFLINE_AT_IPL are great when you are in configuration where you know what DASD devices addresses are available and your systems disks have been restored to the appropriate addresses. Examples would be a 2nd level guest or a known 1st level LPAR . If you want the flexibility to be able to use any LPAR configuration with any DASD device addresses that happen to be configured to it then you nee d to take an extra step. Build a 1 pack recovery system that allows all devices to be online at IPL time. You restore and IPL the recovery syste m which won't care what volids are on any DASD address except itself. Then you have multiple options: A) You could relabel all (or just the problem subset of) the online DASD, thus solving your problem, or B) You could restore your main system packs and use the recovery system to update the SYSTEM CONFIG on your main RES volume with the appropriate ONLINE_AT_IPL and OFFLINE_AT_IPL addresses. Brian Nielsen
Re: Duplicate VOLID's
I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
Alas, that's about the worst possible convention to adopt if you ever plan to replace your DASD farm. Mark Wheeler UnitedHealth Group Date: Thu, 26 Aug 2010 13:55:05 -0400 From: mike...@us.ibm.com Subject: Re: Duplicate VOLID's To: IBMVM@LISTSERV.UARK.EDU I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
The obvious argument against using the rdev in the volser is when you end up needing to move the data to a new volume, or restore the pack after a physical problem, then you no longer have a match between the volser and the rdev, and it becomes very confusing from there. There really isn¹t one ideal way to do it. -- 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 8/26/10 12:55 PM, Michael MacIsaac mike...@us.ibm.com wrote: I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Michael MacIsaac Sent: Thursday, August 26, 2010 10:55 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Michael MacIsaac *Sent:* Thursday, August 26, 2010 10:55 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Scott Rohling Sent: Thursday, August 26, 2010 4:56 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.commailto:rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Michael MacIsaac Sent: Thursday, August 26, 2010 10:55 AM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.commailto:mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Scott Rohling *Sent:* Thursday, August 26, 2010 4:56 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Michael MacIsaac *Sent:* Thursday, August 26, 2010 10:55 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
Then it sounds like changing volids isn't such a big deal? ;-) Automation can really help simplify the annoying stuff.. Scott Rohling On Thu, Aug 26, 2010 at 6:15 PM, Schuh, Richard rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Scott Rohling *Sent:* Thursday, August 26, 2010 4:56 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Michael MacIsaac *Sent:* Thursday, August 26, 2010 10:55 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling On Thu, Aug 26, 2010 at 7:58 PM, Tom Huegel tehue...@gmail.com wrote: So I guess there is no way to absolutly protect z/VM from using the wrong pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? On Thu, Aug 26, 2010 at 5:15 PM, Schuh, Richard rsc...@visa.com wrote: DIRMAINT is just a directory manager. It is similar to the directory manager component of VM:Secure. DIRMAINT does have the capability to do mass updates of the directory. VM:Secure does not. I have my own form of mass updater. I create code to perform the update of a generic single user and temporarily EXECLOAD it as PROFILE XEDIT. Then I run a pipe that looks something like this: 'PIPE id list a | spec /vmsecure edit/ 1 w1 nw | cms | log file a' It usually runs quickly because our directory has fewer than 2000 userids in it. It might not be acceptable on a system with 1+ userids. Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Scott Rohling *Sent:* Thursday, August 26, 2010 4:56 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's Hmm.. RACF isn't really related as it's protecting minidisks on z/VM, at least - and doesn't care about volsers the mindisks are on.The process for DIRMAINT is probably similar to the things that need doing on VM:Secure to do the directory changes: - Make a 'monolithic' copy of the directory and run a PIPE to change all volsers.. then initialize DIRMAINT using the new directory (USER INPUT) - Put the directory online (DIRM DIRECT) - Change EXTENT CONTROL similarly and do a DIRM RLDE I'm in favor of labels using the rdev - unless you really have frequent changes of DASD - to me, the benefits outweigh the occassional need to update the directory. YMMV, as this thread indicates. Scott Rohling On Thu, Aug 26, 2010 at 3:33 PM, Schuh, Richard rsc...@visa.com wrote: That is absolutely the wrong thing to do. I am now suffering because someone else did that to dasd that is EMFFd to 3 LPARS. (It was all ZLccuu). It requires meticulous record keeping and is very error prone. I did wipe out a disk needed by one system because the records I received were not complete Fortunately, it was a disk that was to be used by a new system and had not been updated; it was easy to restore. Also, it is a huge headache if you ever replace your DASD. I don't know about RACF, but there is no mechanism built into VM:Secure for easily doing a mass update of volsers ( I know, you can change the volser with one command - if it is a VM:Secure .controlled disk and nobody is linked to it. The latter is hard to achieve around here.) Regards, Richard Schuh -- *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Michael MacIsaac *Sent:* Thursday, August 26, 2010 10:55 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: Duplicate VOLID's I do know what addresses my system disks are on, Ah! - an argument for the convention of using the RDEV as the last four characters of the volser :)) Mike MacIsaac mike...@us.ibm.com (845) 433-7061
Re: Duplicate VOLID's
On Thu, 26 Aug 2010 18:58:08 -0700, Tom Huegel tehue...@gmail.com wrote : So I guess there is no way to absolutly protect z/VM from using the wron g pack at IPL.. Maybe a requirement? In SYSTEM CONFIG allow optional rdev on the SLOT deffinations. comments? I would rather see z/VM issue prompt messages the way z/OS does for duplicate volsers/duplicate SYSRES packs: IEA213A DUPLICATE VOLUME volname FOUND ON DEVICES dev1 AND dev2. REPLY DEVICE NUMBER WHICH IS TO REMAIN OFFLINE IEA214A DUPLICATE SYSRES volname FOUND ON DEVICE dev. VERIFY THAT CORRECT DEVICE WAS USED FOR IPL. DUPLICATE DEVICE WILL REMAIN OFFLINE. REPLY CONT TO CONTINUE IPL This gives you/operations the chance to pick which duplicate volume should be offline, (or which one should be online), instead of z/VM picking one for you and probably using the wrong one. This is one of the few things that z/OS does better then z/VM! :-) -- Dale R. Smith
Re: Duplicate VOLID's
How much 'protection' is really necessary here? If something like this is implemented *AND* the wrong device number is used in the slot definition, your system could be rendered non-IPLable because the volid on the slot doesn't match the one on the volume. What takes precedence? Should the IPL stop and CP go through a dialog asking a bunch of silly questions about which volume you 'really' intended to use? The volumes should be labeled properly in the first place. I'll vote strongly negative for a requirement to change this process for the reasons indicated so far. On 08/26/2010 09:45 PM, Scott Rohling wrote: Would definitely agree an rdev specification on the SLOT def would be very useful. I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling -- Rich Smrcina Phone: 414-491-6001 http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2011 - April 15-19, 2011 Colorado Springs, CO
Re: Duplicate VOLID's
One can't always be sure the packs are labeled properly.. Besides why depend on people when software can do it better. I suspect the reasons for VM just choosing by volid goes back to when the disk packs were removable, put your second level testRES on the shelf when you were done. Putting an OPTIONAL rdev on the SLOT makes sense, similiar to a DEDICATE statement with both rdev and volid. If there is no rdev on the SLOT and there are dups, or the voild on the rdev doesn't match, then ask questions. Maybe be able to run a mini editor from SAPL to correct SYSTEM CONFIG errors. On Thu, Aug 26, 2010 at 10:09 PM, Rich Smrcina r...@velocitysoftware.comwrote: How much 'protection' is really necessary here? If something like this is implemented *AND* the wrong device number is used in the slot definition, your system could be rendered non-IPLable because the volid on the slot doesn't match the one on the volume. What takes precedence? Should the IPL stop and CP go through a dialog asking a bunch of silly questions about which volume you 'really' intended to use? The volumes should be labeled properly in the first place. I'll vote strongly negative for a requirement to change this process for the reasons indicated so far. On 08/26/2010 09:45 PM, Scott Rohling wrote: Would definitely agree an rdev specification on the SLOT def would be very useful.I just recently built a 2nd level guest and neglected to relabel the volumes before they IPL'd the 1st level system ... ugly. Wouldn't have happened if the real address was specified... good idea! Scott Rohling -- Rich Smrcina Phone: 414-491-6001 http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2011 - April 15-19, 2011 Colorado Springs, CO
Duplicate VOLID's
I am looking for a way to verify at IPL time that z/VM is using the volumes I intended. It is possible that there are more than one volume with a volid of xxxRES xxxWK1 xxxWK2 etc. I could put something in the AUTOLOG1 PROFILE EXEC to do Q DASD and verify that xxxRES is at address x'100' and xxxWK1 is at x'101' and xxxWK2 is at x'102'.. etc. but that can be messy and diffacult to maintain. Then I thought I could do something in the SYSTEM CONFIG file. SYSTEM_IDENTIFIER_2094_123ABC_GOODSYS SYSTEM_IDENTIFER_DEFAULT_BADSYS IMBED -SYSTEM- CONFIG In GOODSYS CONFIG I would have all of the normal stuff. In BADSYS CONFIG maybe just a bunch of SAY 'WRONG SYSTEM' statements.. This doesn't verify anything beyond that this is the correct xxxRES for this LPAR. If the volume was cloned (DDR) it would pass this test anway. Just courious as how others handle this, if at all. Thanks for any thoughts.
Re: Duplicate VOLID's
You could change the SAIPL defaults on iplable volumes so that it would be obvious to everyone which disk was IPLed. That is assuming that whoever ipls will look at the comments before hitting PF10. Also, make sure that no two of the systems have a matching volser in the spool configuration. That could be very bad. (On the other hand, if you do include a few duplicates in the spool, the operators will also get messages about spool errors and have to give permission for the start-up to continue. Just hope they are wise enough to stop the process and call you.). Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Tom Huegel Sent: Wednesday, August 25, 2010 4:49 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Duplicate VOLID's I am looking for a way to verify at IPL time that z/VM is using the volumes I intended. It is possible that there are more than one volume with a volid of xxxRES xxxWK1 xxxWK2 etc. I could put something in the AUTOLOG1 PROFILE EXEC to do Q DASD and verify that xxxRES is at address x'100' and xxxWK1 is at x'101' and xxxWK2 is at x'102'.. etc. but that can be messy and diffacult to maintain. Then I thought I could do something in the SYSTEM CONFIG file. SYSTEM_IDENTIFIER_2094_123ABC_GOODSYS SYSTEM_IDENTIFER_DEFAULT_BADSYS IMBED -SYSTEM- CONFIG In GOODSYS CONFIG I would have all of the normal stuff. In BADSYS CONFIG maybe just a bunch of SAY 'WRONG SYSTEM' statements.. This doesn't verify anything beyond that this is the correct xxxRES for this LPAR. If the volume was cloned (DDR) it would pass this test anway. Just courious as how others handle this, if at all. Thanks for any thoughts.
Re: Duplicate VOLID's
Cloned systems immediately are assigned a new set of volid's. You may want to consider assigning new sets of volids. An alternative is to take the other systems devices offline at IPL, but if they move for whatever reason that is one more thing to keep track of and maintain. On 08/25/2010 06:49 PM, Tom Huegel wrote: I am looking for a way to verify at IPL time that z/VM is using the volumes I intended. It is possible that there are more than one volume with a volid of xxxRES xxxWK1 xxxWK2 etc. I could put something in the AUTOLOG1 PROFILE EXEC to do Q DASD and verify that xxxRES is at address x'100' and xxxWK1 is at x'101' and xxxWK2 is at x'102'.. etc. but that can be messy and diffacult to maintain. Then I thought I could do something in the SYSTEM CONFIG file. SYSTEM_IDENTIFIER_2094_123ABC_GOODSYS SYSTEM_IDENTIFER_DEFAULT_BADSYS IMBED -SYSTEM- CONFIG In GOODSYS CONFIG I would have all of the normal stuff. In BADSYS CONFIG maybe just a bunch of SAY 'WRONG SYSTEM' statements.. This doesn't verify anything beyond that this is the correct xxxRES for this LPAR. If the volume was cloned (DDR) it would pass this test anway. Just courious as how others handle this, if at all. Thanks for any thoughts. -- Rich Smrcina Phone: 414-491-6001 http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2011 - April 15-19, 2011 Colorado Springs, CO
Re: Duplicate VOLID's
We code stuff in the SYSTEM CONFIG to control which lpar sees which volumes at IPL time. It does require some work to set up and to maintain it. Paul Feller AIT Mainframe Technical Support From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Tom Huegel Sent: Wednesday, August 25, 2010 6:49 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Duplicate VOLID's I am looking for a way to verify at IPL time that z/VM is using the volumes I intended. It is possible that there are more than one volume with a volid of xxxRES xxxWK1 xxxWK2 etc. I could put something in the AUTOLOG1 PROFILE EXEC to do Q DASD and verify that xxxRES is at address x'100' and xxxWK1 is at x'101' and xxxWK2 is at x'102'.. etc. but that can be messy and diffacult to maintain. Then I thought I could do something in the SYSTEM CONFIG file. SYSTEM_IDENTIFIER_2094_123ABC_GOODSYS SYSTEM_IDENTIFER_DEFAULT_BADSYS IMBED -SYSTEM- CONFIG In GOODSYS CONFIG I would have all of the normal stuff. In BADSYS CONFIG maybe just a bunch of SAY 'WRONG SYSTEM' statements.. This doesn't verify anything beyond that this is the correct xxxRES for this LPAR. If the volume was cloned (DDR) it would pass this test anway. Just courious as how others handle this, if at all. Thanks for any thoughts.
Re: Duplicate VOLID's
On Wednesday, 08/25/2010 at 07:49 EDT, Tom Huegel tehue...@gmail.com wrote: I am looking for a way to verify at IPL time that z/VM is using the volumes I intended. It is possible that there are more than one volume with a volid of xxxRES xxxWK1 xxxWK2 etc. I could put something in the AUTOLOG1 PROFILE EXEC to do Q DASD and verify that xxxRES is at address x'100' and xxxWK1 is at x'101' and xxxWK2 is at x'102'.. etc. but that can be messy and diffacult to maintain. Then I thought I could do something in the SYSTEM CONFIG file. SYSTEM_IDENTIFIER_2094_123ABC_GOODSYS SYSTEM_IDENTIFER_DEFAULT_BADSYS IMBED -SYSTEM- CONFIG In GOODSYS CONFIG I would have all of the normal stuff. In BADSYS CONFIG maybe just a bunch of SAY 'WRONG SYSTEM' statements.. This doesn't verify anything beyond that this is the correct xxxRES for this LPAR. If the volume was cloned (DDR) it would pass this test anway. Just courious as how others handle this, if at all. Thanks for any thoughts. You need to use the DEVICES OFFLINE_AT_IPL and ONLINE_AT_IPL statements in SYSTEM CONFIG that I discussed in a July post. http://listserv.uark.edu/scripts/wa.exe?A2=ind1007L=IBMVMP=R1084. Checking AFTER the system has IPLed is too late. Alan Altmark z/VM Development IBM Endicott
Re: Duplicate VOLID's
In a normal production environment this is not such an insurmountable problem. The problem is this is a test lab, and I don't necessarily know what happens to the different disk volumes. I do know what addresses my system disks are on, but there may be copies that someone was testing with floating around. It may have been a in a second level machine, or a first level test.. It's only a problem at IPL .. On Wed, Aug 25, 2010 at 10:08 PM, Feller, Paul pfel...@aegonusa.com wrote: We code stuff in the SYSTEM CONFIG to control which lpar sees which volumes at IPL time. It does require some work to set up and to maintain it. *Paul Feller* *AIT Mainframe Technical Support* *From:* The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] *On Behalf Of *Tom Huegel *Sent:* Wednesday, August 25, 2010 6:49 PM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Duplicate VOLID's I am looking for a way to verify at IPL time that z/VM is using the volumes I intended. It is possible that there are more than one volume with a volid of xxxRES xxxWK1 xxxWK2 etc. I could put something in the AUTOLOG1 PROFILE EXEC to do Q DASD and verify that xxxRES is at address x'100' and xxxWK1 is at x'101' and xxxWK2 is at x'102'.. etc. but that can be messy and diffacult to maintain. Then I thought I could do something in the SYSTEM CONFIG file. SYSTEM_IDENTIFIER_2094_123ABC_GOODSYS SYSTEM_IDENTIFER_DEFAULT_BADSYS IMBED -SYSTEM- CONFIG In GOODSYS CONFIG I would have all of the normal stuff. In BADSYS CONFIG maybe just a bunch of SAY 'WRONG SYSTEM' statements.. This doesn't verify anything beyond that this is the correct xxxRES for this LPAR. If the volume was cloned (DDR) it would pass this test anway. Just courious as how others handle this, if at all. Thanks for any thoughts.