Re: Problem that is a blast from the past...
Note: by coding an explicit CONS option on WAKEP, it will stop with RC=6 when there is something in the stack when WAKEUP is started. May that be the problem? At the other hand: I don't see a TIME option on the WAKEUP command in the first append, so WAKEUP would not stack 3 lines but only 2. 2009/9/10 Cal c...@the-fishers.com Hi Martha Where did this exec come from? The way that wakeup works is it always stacks the next line from the times file. Actually it stacks 3 lines 1. Current date and time 2. Line from Wakeup Times file 3. SPM, VMCF, SMSG, IUCV message, IO or externat interrupt data. So if you wrote your own exec you are using the stack the line that you are really intersted in is the last line on the stack. If you pull the line from the times file and execute it you will leave something on the stack and wakeup will exit. The 300 secs come from the +5 Cal Fisher MVMUA website http://www2.marist.edu/~mvmua/http://www2.marist.edu/%7Emvmua/ My Navy memoirs http://www.the-fishers.com/cal/Navy - Original Message - From: Martha McConaghy u...@vm.marist.edu To: IBMVM@LISTSERV.UARK.EDU Sent: Wednesday, September 09, 2009 5:58 PM Subject: Re: Problem that is a blast from the past... That's the strange part, there is nothing. This is happening on VM systems with very little going on, so there isn't any noise. Here's what the console looks like when it happens: DMSCYW2246I 15:06:26 WAKEUP in (299 sec). DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS DMSCYW2246I* 00067 ==/==/== +5 15:11:26 EXEC HOBVMCPU DMSCYW2246I* 00068 ==/==/== +5 15:11:26 EXEC HOBVMPRC Number of VMs: 19 DMSCYW2246I* 00070 ==/==/== +5 15:11:26 EXEC HOBVMDSK DMSCYW2246I* 00071 ==/==/== +5 15:11:26 EXEC HOBVMFLE DMSERS002E File HOBVM700 CLIENT A not found DMSCYW2246I* 00072 ==/==/== +5 15:11:26 EXEC HOBVMPOR DMSCYW2246I* 00073 ==/==/== +5 15:11:26 EXEC HOBVMIFC DMSCYW2246I* 00077 ==/==/== +5 15:11:26 EXEC HOBVMCD DMSCYW2246I 15:11:26 WAKEUP in (300 sec). --300 secs always shows up DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --This isn't right Console interrupt... queue: 2 Queue data: * 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --my diags Queue data: * 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --my diags The sequence is to run HOBVARS, HOBVMCPU, HOBVMPRC, HOBVMDSK, HOBVMFLE, HOBVMPOR, HOBVMIFC and then HOBVMCD. It sleeps and then starts over. Whenever I see the WAKEUP in (300 sec) I know it is going to fail. If the time is anything less than 300 sec, then it will be OK. It happens too consistently to be a coincidence. When it fails, HOBVARS always shows up twice. I think that maybe what is being interpreted as a console interrupt, i.e. someone typing on the console. I can't see any reason why that happens. HOBVARS never gets run at that point. I've put traces on it and it doesn't get executed. Its almost like WAKEUP is getting confused. Could there be something on the program stack that is getting it messed up? Is there any way to trace what WAKEUP is doing? Martha On Wed, 9 Sep 2009 23:50:38 +0200 Alan Altmark said: On Wednesday, 09/09/2009 at 05:26 EDT, Martha McConaghy u...@vm.marist.edu wrote: WAKEUP +5 ( CONS EXT SMSG FILE(HOBBIT TIMES *) Sometimes, it will run through a sequence and then exit, sometimes it will run for several days before it happens. This is happening on different systems to, not just on one VM system. I suspect that some silly thing is not set correctly, but I have no idea what. I finally did a CP TRACE EXT on one of them and found that it is getting an external interrupt code 1004. According to my trusty old reference book, that is a clock comparator interrupt. That is what is causing WAKEUP to stop with RC=6. While it's true that EXT 1004 is a timer pop, RC=6 from WAKEUP indicates it detected a console I/O interrupt. I am wondering if some sort of automation sequence (CP SEND) is bothering the virtual machine. Since there's no QUIET option, the reason for the wakeup should be in the console. Alan Altmark z/VM Development IBM Endicott -- Kris Buelens, IBM Belgium, VM customer support
How much memory?
Hi all, I'm not a VM expert, so forgive me if this seems a newbie quetion. I have a z800 running z/VM, which has two production z/OS guests, and the usual collection of CMS guests (TCPIP / OPERATOR etc). We discovered recently that the z800 has an IFL (long story - don't ask!) and would like to steal some of the 8GB currently allocated to VM, to run up a VM / Linux LPAR for a POC. The two z/OS guests are defined at 256MB 512MB (Bill Gates, are you watching??), and don't page. In your most esteemed opinion, what could I steal from the VM LPAR withou t hurting it's and the guests performance / making it page? Personally, I think 2GB would be enough. Comments and opinions welcome.
Re: How much memory?
It will depend on what your POC encompasses.. how many Linux guests? What software/middleware?. You might want to consider a 2nd level guest rather than building a separate LPAR -- that way you can manage things under a single zVM and dynamically change things (give your 2nd level zVM more storage, for example) much more easily... Scott On Mon, Sep 14, 2009 at 7:52 AM, Vince Getgood vincent.getg...@xchanging.com wrote: Hi all, I'm not a VM expert, so forgive me if this seems a newbie quetion. I have a z800 running z/VM, which has two production z/OS guests, and the usual collection of CMS guests (TCPIP / OPERATOR etc). We discovered recently that the z800 has an IFL (long story - don't ask!) and would like to steal some of the 8GB currently allocated to VM, to run up a VM / Linux LPAR for a POC. The two z/OS guests are defined at 256MB 512MB (Bill Gates, are you watching??), and don't page. In your most esteemed opinion, what could I steal from the VM LPAR without hurting it's and the guests performance / making it page? Personally, I think 2GB would be enough. Comments and opinions welcome.
Re: How much memory?
On Mon, Sep 14, 2009 at 3:52 PM, Vince Getgood vincent.getg...@xchanging.com wrote: Hi all, I'm not a VM expert, so forgive me if this seems a newbie quetion. I have a z800 running z/VM, which has two production z/OS guests, and the usual collection of CMS guests (TCPIP / OPERATOR etc). We discovered recently that the z800 has an IFL (long story - don't ask!) and would like to steal some of the 8GB currently allocated to VM, to run up a VM / Linux LPAR for a POC. The two z/OS guests are defined at 256MB 512MB (Bill Gates, are you watching??), and don't page. In your most esteemed opinion, what could I steal from the VM LPAR without hurting it's and the guests performance / making it page? Personally, I think 2GB would be enough. The real answer will be in z/VM monitor data. But as long as z/VM has enough memory it is not easy to tell what would happen when it had less. You probably can do with 1G. You should not expect z/VM to overcommit memory for the two z/OS guests, so you need the memory available for that. The remaining service machines will probably do fine in 256 MB real memory. And you should make sure you do have some paging space to make up for the difference... Your Linux workload can probably use all the memory you can give it. And I do hope you don't have too high expectations of the z800 IFL (they're a bit slow for today's standards with Linux). Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Problem that is a blast from the past...
Yes, I aware of that. The issue is why the console interrupt is happening in the first place. I haven't been able to track down a cause. So, I've just taken CONS out at this point. Its not worth spending so much time on. Martha On Mon, 14 Sep 2009 11:15:05 +0200 Kris Buelens said: Note: by coding an explicit CONS option on WAKEP, it will stop with RC=6 when there is something in the stack when WAKEUP is started. May that be the problem? At the other hand: I don't see a TIME option on the WAKEUP command in the first append, so WAKEUP would not stack 3 lines but only 2. 2009/9/10 Cal c...@the-fishers.com Hi Martha Where did this exec come from? The way that wakeup works is it always stacks the next line from the times file. Actually it stacks 3 lines 1. Current date and time 2. Line from Wakeup Times file 3. SPM, VMCF, SMSG, IUCV message, IO or externat interrupt data. So if you wrote your own exec you are using the stack the line that you are really intersted in is the last line on the stack. If you pull the line from the times file and execute it you will leave something on the stack and wakeup will exit. The 300 secs come from the +5 Cal Fisher MVMUA website http://www2.marist.edu/~mvmua/http://www2.marist.edu/%7Emvmua/ My Navy memoirs http://www.the-fishers.com/cal/Navy - Original Message - From: Martha McConaghy u...@vm.marist.edu To: IBMVM@LISTSERV.UARK.EDU Sent: Wednesday, September 09, 2009 5:58 PM Subject: Re: Problem that is a blast from the past... That's the strange part, there is nothing. This is happening on VM systems with very little going on, so there isn't any noise. Here's what the console looks like when it happens: DMSCYW2246I 15:06:26 WAKEUP in (299 sec). DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS DMSCYW2246I* 00067 ==/==/== +5 15:11:26 EXEC HOBVMCPU DMSCYW2246I* 00068 ==/==/== +5 15:11:26 EXEC HOBVMPRC Number of VMs: 19 DMSCYW2246I* 00070 ==/==/== +5 15:11:26 EXEC HOBVMDSK DMSCYW2246I* 00071 ==/==/== +5 15:11:26 EXEC HOBVMFLE DMSERS002E File HOBVM700 CLIENT A not found DMSCYW2246I* 00072 ==/==/== +5 15:11:26 EXEC HOBVMPOR DMSCYW2246I* 00073 ==/==/== +5 15:11:26 EXEC HOBVMIFC DMSCYW2246I* 00077 ==/==/== +5 15:11:26 EXEC HOBVMCD DMSCYW2246I 15:11:26 WAKEUP in (300 sec). --300 secs always shows up DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --This isn't right Console interrupt... queue: 2 Queue data: * 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --my diags Queue data: * 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --my diags The sequence is to run HOBVARS, HOBVMCPU, HOBVMPRC, HOBVMDSK, HOBVMFLE, HOBVMPOR, HOBVMIFC and then HOBVMCD. It sleeps and then starts over. Whenever I see the WAKEUP in (300 sec) I know it is going to fail. If the time is anything less than 300 sec, then it will be OK. It happens too consistently to be a coincidence. When it fails, HOBVARS always shows up twice. I think that maybe what is being interpreted as a console interrupt, i.e. someone typing on the console. I can't see any reason why that happens. HOBVARS never gets run at that point. I've put traces on it and it doesn't get executed. Its almost like WAKEUP is getting confused. Could there be something on the program stack that is getting it messed up? Is there any way to trace what WAKEUP is doing? Martha On Wed, 9 Sep 2009 23:50:38 +0200 Alan Altmark said: On Wednesday, 09/09/2009 at 05:26 EDT, Martha McConaghy u...@vm.marist.edu wrote: WAKEUP +5 ( CONS EXT SMSG FILE(HOBBIT TIMES *) Sometimes, it will run through a sequence and then exit, sometimes it will run for several days before it happens. This is happening on different systems to, not just on one VM system. I suspect that some silly thing is not set correctly, but I have no idea what. I finally did a CP TRACE EXT on one of them and found that it is getting an external interrupt code 1004. According to my trusty old reference book, that is a clock comparator interrupt. That is what is causing WAKEUP to stop with RC=6. While it's true that EXT 1004 is a timer pop, RC=6 from WAKEUP indicates it detected a console I/O interrupt. I am wondering if some sort of automation sequence (CP SEND) is bothering the virtual machine. Since there's no QUIET option, the reason for the wakeup should be in the console. Alan Altmark z/VM Development IBM Endicott -- Kris Buelens, IBM Belgium, VM customer support
Re: How much memory?
On a 2G z/VM system, we run one OS/390 machine (128M), two z/OS machines (512M 1024M), three z/VSE machines (128M each) and four Linux machines (3 at 2048M and one at 256M/512M). But our configuration is not a production environment. On Mon, Sep 14, 2009 at 7:02 AM, Gentry, Stephen stephen.gen...@lafayettelife.com wrote: 2gb would be a good starting point. When we did a POC, we had a little over 7gb available and we could run 3 Linux guests, running DB2, comfortably. -Original Message- From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Vince Getgood Sent: Monday, September 14, 2009 9:53 AM To: IBMVM@LISTSERV.UARK.EDU Subject: How much memory? Hi all, I'm not a VM expert, so forgive me if this seems a newbie quetion. I have a z800 running z/VM, which has two production z/OS guests, and the usual collection of CMS guests (TCPIP / OPERATOR etc). We discovered recently that the z800 has an IFL (long story - don't ask!) and would like to steal some of the 8GB currently allocated to VM, to run up a VM / Linux LPAR for a POC. The two z/OS guests are defined at 256MB 512MB (Bill Gates, are you watching??), and don't page. In your most esteemed opinion, what could I steal from the VM LPAR withou t hurting it's and the guests performance / making it page? Personally, I think 2GB would be enough. Comments and opinions welcome. -- Daniel Allen | Serena Software, Inc. | Senior Systems Programmer - Mainframe Services Phone: 1-800-457-3736x11241
Re: How to find out what user holds dasd
Q SYSTEM 16E0 will tell you 2009/9/14 Bauer, Bobby (NIH/CIT) [E] baue...@mail.nih.gov I'm trying detach and take a couple of volumes offline but seems some user has them and I can't identify which one. Is there a way to force a detach or to display who has it. Normally the devices are used for a server named UPST We are running z/VM 5.4 at 0901 q 16e0 DASD 16E0 CP SYSTEM UPST01 1 det 16e0 system HCPDTS124E DASD 16E0 in use by 1 users Ready(00124); T=0.01/0.01 11:45:07 det 16e0 upst HCPDTC045E UPST not logged on Ready(00045); T=0.01/0.01 11:45:12 Bobby Bauer Center for Information Technology National Institutes of Health Bethesda, MD 20892-5628 301-594-7474 -- Kris Buelens, IBM Belgium, VM customer support
Re: How much memory?
2gb would be a good starting point. When we did a POC, we had a little over 7gb available and we could run 3 Linux guests, running DB2, comfortably. -Original Message- From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Vince Getgood Sent: Monday, September 14, 2009 9:53 AM To: IBMVM@LISTSERV.UARK.EDU Subject: How much memory? Hi all, I'm not a VM expert, so forgive me if this seems a newbie quetion. I have a z800 running z/VM, which has two production z/OS guests, and the usual collection of CMS guests (TCPIP / OPERATOR etc). We discovered recently that the z800 has an IFL (long story - don't ask!) and would like to steal some of the 8GB currently allocated to VM, to run up a VM / Linux LPAR for a POC. The two z/OS guests are defined at 256MB 512MB (Bill Gates, are you watching??), and don't page. In your most esteemed opinion, what could I steal from the VM LPAR withou t hurting it's and the guests performance / making it page? Personally, I think 2GB would be enough. Comments and opinions welcome.
Re: How to find out what user holds dasd
Was upst a keying error...wouldn't it be upst01?? mace -Original Message- From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Bauer, Bobby (NIH/CIT) [E] Sent: Monday, September 14, 2009 11:47 AM To: IBMVM@LISTSERV.UARK.EDU Subject: How to find out what user holds dasd I'm trying detach and take a couple of volumes offline but seems some user has them and I can't identify which one. Is there a way to force a detach or to display who has it. Normally the devices are used for a server named UPST We are running z/VM 5.4 at 0901 q 16e0 DASD 16E0 CP SYSTEM UPST01 1 det 16e0 system HCPDTS124E DASD 16E0 in use by 1 users Ready(00124); T=0.01/0.01 11:45:07 det 16e0 upst HCPDTC045E UPST not logged on Ready(00045); T=0.01/0.01 11:45:12 Bobby Bauer Center for Information Technology National Institutes of Health Bethesda, MD 20892-5628 301-594-7474 - The information transmitted is intended solely for the individual or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of or taking action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you have received this email in error please contact the sender and delete the material from any computer.
Re: How to find out what user holds dasd
You're looking for the command: CP Query SYSTEM rdev In your example: CP Q SYSTEM 16E0 Mike Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. Bauer, Bobby (NIH/CIT) [E] baue...@mail.nih.gov Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 09/14/2009 10:47 AM Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject How to find out what user holds dasd I'm trying detach and take a couple of volumes offline but seems some user has them and I can't identify which one. Is there a way to force a detach or to display who has it. Normally the devices are used for a server named UPST We are running z/VM 5.4 at 0901 q 16e0 DASD 16E0 CP SYSTEM UPST01 1 det 16e0 system HCPDTS124E DASD 16E0 in use by 1 users Ready(00124); T=0.01/0.01 11:45:07 det 16e0 upst HCPDTC045E UPST not logged on Ready(00045); T=0.01/0.01 11:45:12 Bobby Bauer Center for Information Technology National Institutes of Health Bethesda, MD 20892-5628 301-594-7474 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: How to find out what user holds dasd
Thank you, that is exactly what I was looking for. Bobby Bauer Center for Information Technology National Institutes of Health Bethesda, MD 20892-5628 301-594-7474 -Original Message- From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Mike Walter Sent: Monday, September 14, 2009 11:55 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: How to find out what user holds dasd You're looking for the command: CP Query SYSTEM rdev In your example: CP Q SYSTEM 16E0 Mike Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. Bauer, Bobby (NIH/CIT) [E] baue...@mail.nih.gov Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 09/14/2009 10:47 AM Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject How to find out what user holds dasd I'm trying detach and take a couple of volumes offline but seems some user has them and I can't identify which one. Is there a way to force a detach or to display who has it. Normally the devices are used for a server named UPST We are running z/VM 5.4 at 0901 q 16e0 DASD 16E0 CP SYSTEM UPST01 1 det 16e0 system HCPDTS124E DASD 16E0 in use by 1 users Ready(00124); T=0.01/0.01 11:45:07 det 16e0 upst HCPDTC045E UPST not logged on Ready(00045); T=0.01/0.01 11:45:12 Bobby Bauer Center for Information Technology National Institutes of Health Bethesda, MD 20892-5628 301-594-7474 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: How much memory?
My initial guess is that you need about 1 GB for your current workload. Some things to check: 1. Do a QSYSOWN qsysown ** Summary Information: Total-Pages TypeAllocd In-Use %-Used SPOL 1201680 348517 29.0 PAGE 1201680 172559 14.4 Look at the %-Used for PAGE. If it is near zero, where it should be if you never paged, great. If it is not near zero, then you did/have been paging. 2. Find how much storage you are using. Do an IND USER for each guest: ind user stlesa2 USERID=STLESA2 MACH=ESA STOR=700M VIRT=V XSTORE=NONE IPLSYS=DEV 0120 DEVNUM=00073 PAGES: RES=00021448 WS=00021444 LOCKEDREAL=0004 RESVD= NPREF=00011685 PREF= READS=00015515 WRITES=00051555 XSTORE=000646 READS=003233 WRITES=025769 MIGRATES=021694 CPU 00: CTIME=92:37 VTIME=721:27 TTIME=839:31 IO=335062 RDR=015839 PRT=790969 PCH=078332 Look at the RES figure. This is the number of pages that machine is using now. Sum up all the pages for all the currently logged on machines, multiply by 4K an that is the storage your guests are using. 3. Do a Q FRAMES: q frames All Frames: Configured=163839 Real=163839 Usable=163839 Offline=0 Pageable=148743 NotInitialized=0 GlobalClearedAvail=32 LocalClearedAvail=32 LocalUnclearedAvail=31 The Configured minus Pageablejust say VM overhead. Sum up 2 and 3 and that is the amount of storage you need at your current paging rate (which should be near zero). My guess is you have 6 GB left. 1 GB for your system. 1 GB taken for microcode. 6 GB remaining. Taking 2 GB for the zLinux LPAR is a good start. You will have a little more than a GB for Linux stuff. On the Linux LPAR, make sure you have 4+ full paging packs, as you are going to page here. Perhaps give .5 GB for expanded storage, also, as you are going to page here. Don't forget to implement vdisk support so you have vdisk packs for Linux swap areas. Good time to also bring up vswitch/guest LAN. Of course if you have a performance monitor, getting the right amount of storage is a lot easier, but it is also doable manually. How often can you reconfigure your production LPAR? If rare, push for a performance monitor as it will save you Power On Resets as you keep reconfiguring your production LPAR. If you can do it weekly/monthly, well, take 2 GB, monitor for a while, take another 1 GB, monitor for a while, take another GB, perhaps give back 512 MB. You know the drill. Tom Duerbusch THD Consulting Vince Getgood vincent.getg...@xchanging.com 9/14/2009 8:52 AM Hi all, I'm not a VM expert, so forgive me if this seems a newbie quetion. I have a z800 running z/VM, which has two production z/OS guests, and the usual collection of CMS guests (TCPIP / OPERATOR etc). We discovered recently that the z800 has an IFL (long story - don't ask!) and would like to steal some of the 8GB currently allocated to VM, to run up a VM / Linux LPAR for a POC. The two z/OS guests are defined at 256MB 512MB (Bill Gates, are you watching??), and don't page. In your most esteemed opinion, what could I steal from the VM LPAR without hurting it's and the guests performance / making it page? Personally, I think 2GB would be enough. Comments and opinions welcome.
Re: Problem that is a blast from the past...
Is there any chance of some other SVM issuing a 'CP SEND' or 'CP FOR' command to the server running WAKEUP and experiencing the unexpected interrupt? Of course, in such a case of one disconnected SVM waking another up in that manner, one might expect to hear the faint strains of Dueling Banjos playing softly in the background! ;-) Mike Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. Martha McConaghy u...@vm.marist.edu Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 09/14/2009 09:08 AM Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject Re: Problem that is a blast from the past... Yes, I aware of that. The issue is why the console interrupt is happening in the first place. I haven't been able to track down a cause. So, I've just taken CONS out at this point. Its not worth spending so much time on. Martha On Mon, 14 Sep 2009 11:15:05 +0200 Kris Buelens said: Note: by coding an explicit CONS option on WAKEP, it will stop with RC=6 when there is something in the stack when WAKEUP is started. May that be the problem? At the other hand: I don't see a TIME option on the WAKEUP command in the first append, so WAKEUP would not stack 3 lines but only 2. 2009/9/10 Cal c...@the-fishers.com Hi Martha Where did this exec come from? The way that wakeup works is it always stacks the next line from the times file. Actually it stacks 3 lines 1. Current date and time 2. Line from Wakeup Times file 3. SPM, VMCF, SMSG, IUCV message, IO or externat interrupt data. So if you wrote your own exec you are using the stack the line that you are really intersted in is the last line on the stack. If you pull the line from the times file and execute it you will leave something on the stack and wakeup will exit. The 300 secs come from the +5 Cal Fisher MVMUA website http://www2.marist.edu/~mvmua/http://www2.marist.edu/%7Emvmua/ My Navy memoirs http://www.the-fishers.com/cal/Navy - Original Message - From: Martha McConaghy u...@vm.marist.edu To: IBMVM@LISTSERV.UARK.EDU Sent: Wednesday, September 09, 2009 5:58 PM Subject: Re: Problem that is a blast from the past... That's the strange part, there is nothing. This is happening on VM systems with very little going on, so there isn't any noise. Here's what the console looks like when it happens: DMSCYW2246I 15:06:26 WAKEUP in (299 sec). DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS DMSCYW2246I* 00067 ==/==/== +5 15:11:26 EXEC HOBVMCPU DMSCYW2246I* 00068 ==/==/== +5 15:11:26 EXEC HOBVMPRC Number of VMs: 19 DMSCYW2246I* 00070 ==/==/== +5 15:11:26 EXEC HOBVMDSK DMSCYW2246I* 00071 ==/==/== +5 15:11:26 EXEC HOBVMFLE DMSERS002E File HOBVM700 CLIENT A not found DMSCYW2246I* 00072 ==/==/== +5 15:11:26 EXEC HOBVMPOR DMSCYW2246I* 00073 ==/==/== +5 15:11:26 EXEC HOBVMIFC DMSCYW2246I* 00077 ==/==/== +5 15:11:26 EXEC HOBVMCD DMSCYW2246I 15:11:26 WAKEUP in (300 sec). --300 secs always shows up DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS DMSCYW2246I* 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --This isn't right Console interrupt... queue: 2 Queue data: * 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --my diags Queue data: * 00066 ==/==/== +5 15:11:26 EXEC HOBVARS --my diags The sequence is to run HOBVARS, HOBVMCPU, HOBVMPRC, HOBVMDSK, HOBVMFLE, HOBVMPOR, HOBVMIFC and then HOBVMCD. It sleeps and then starts over. Whenever I see the WAKEUP in (300 sec) I know it is going to fail. If the time is anything less than 300 sec, then it will be OK. It happens too consistently to be a coincidence. When it fails, HOBVARS always shows up twice. I think that maybe what is being interpreted as a console interrupt, i.e. someone typing on the console. I can't see any reason why that happens. HOBVARS never gets run at that point. I've put traces on it and it doesn't get executed. Its almost like WAKEUP is getting confused. Could there be something on the program stack that is getting it messed up? Is there any way to trace what WAKEUP is doing? Martha On Wed, 9 Sep 2009 23:50:38 +0200 Alan Altmark said: On Wednesday, 09/09/2009 at 05:26 EDT, Martha McConaghy u...@vm.marist.edu wrote: WAKEUP +5 ( CONS EXT SMSG FILE(HOBBIT TIMES *) Sometimes, it will run through a sequence and then exit, sometimes it will run for several days before it happens. This is happening on different systems to, not just on one VM system. I suspect that some silly thing is not set correctly, but I have no idea what. I finally did a CP TRACE EXT on one of them and found that it is getting an external interrupt code 1004. According to my trusty old reference book, that is a clock comparator interrupt. That is what is causing WAKEUP to stop with RC=6. While it's true that EXT 1004 is a timer pop, RC=6 from WAKEUP indicates it detected a console I/O
Re: RELABEL SYSRES was: Re: Backup RES Labeling Question
Shimon, Thanks for the corrections! They have all been included in V1R1. Before sending V1R1 to the list, ... Oops! Note to self: check the To: address very carefully before replying! :-( Mike Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. 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.
Mark Workman/IS/Shelter is out of the office.
I will be out of the office starting 09/14/2009 and will not return until 10/01/2009. I will respond to your message when I return. This e-mail is intended only for its addressee and may contain information that is privileged, confidential, or otherwise protected from disclosure. If you have received this communication in error, please notify us immediately by e-mailing postmas...@shelterinsurance.com; then delete the original message.
Re: Problem that is a blast from the past...
I really think Kris's first response about the CONS option is the correct one. You don't want to use this option unless you have some specific need. WAKEUP will wake and give a rc 6 when you hit enter on the console without this option. I have done traces before and 'watched' the rc 6 occur because of something being put temporarily on the stack by CMS. In those cases WAKEUP is too efficient and catches what you don't want caught. Another quirky thing with WAKEUP is using DESBUF without CONWAIT. DROPBUF works much better with WAKEUP and isn't as finicky about whether or not CONWAIT is used. (It has been too long since I chased some of these things. Memory fades ...) Colleen M Brown IBM z/VM and Related Products Development and Service Kris Buelens kris.buel...@gmail.com Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 09/14/2009 04:03 PM Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject Re: Problem that is a blast from the past... As far as I know: CP FOR can only be used to execute CP commands on behalf of the target user, it does not generate console interrupts as opposed to CP SEND. 2009/9/14 Mike Walter mike.wal...@hewitt.com Is there any chance of some other SVM issuing a 'CP SEND' or 'CP FOR' command to the server running WAKEUP and experiencing the unexpected interrupt? Of course, in such a case of one disconnected SVM waking another up in that manner, one might expect to hear the faint strains of Dueling Banjos playing softly in the background! ;-) Mike Walter Hewitt Associates The opinions expressed herein are mine alone, not my employer's. -- Kris Buelens, IBM Belgium, VM customer support