Re: Performance problem Linux under Zvm
I made the mistake of believing what I was told. cp q storage 05:57:21 STORAGE = 2G Ready; T=0.01/0.01 05:57:21 cp q virtual storage 05:57:35 STORAGE = 128M Ready; T=0.01/0.01 05:57:35 cp q srm 06:01:42 IABIAS : INTENSITY=90%; DURATION=2 06:01:42 LDUBUF : Q1=100% Q2=75% Q3=60% 06:01:42 STORBUF: Q1=300% Q2=300% Q3=300% 06:01:42 DSPBUF : Q1=32767 Q2=32767 Q3=32767 06:01:42 DISPATCHING MINOR TIMESLICE = 5 MS 06:01:42 MAXWSS : LIMIT=% 06:01:42 .. : PAGES=99 06:01:42 XSTORE : 0% Any recommendation for LDUBUF? I just raised it to 100 100 100. Please advise if that change was counter indicated. Thank you, Dave O'Brien National Institutes of Health From: Gentry, Stephen [Sent: Thu 3/20/2008 2:51 PM Subject: Re: Performance problem Linux under Zvm What command did you use to determine that you had 768m central storage? QUERY STOREAGE? QUERY VIRTUAL STORAGE? Steve G. -Original Message- Subject: Re: Performance problem Linux under Zvm Thanks John cp q srm IABIAS : INTENSITY=90%; DURATION=2 LDUBUF : Q1=100% Q2=75% Q3=60% STORBUF: Q1=300% Q2=200% Q3=200% DSPBUF : Q1=32767 Q2=32767 Q3=32767 DISPATCHING MINOR TIMESLICE = 5 MS MAXWSS : LIMIT=% .. : PAGES=99 XSTORE : 0% Just got the following from one of the other techs (non-VM) We were able to diagnose the problem and make the necessary correction. The problem was z/VM has a total 768m of central available. The Linux guests (3 total) each had 768m of central allocated, therefore contention. The Linux guests are over allocated and are storage constrained with 768m of central. Understanding the Linux guests would be in contention with each other for this storage VM time sliced what it could for each guest, therefore the symptoms we experienced. My question to this group - Does a Linux quest really require 768MB of Central? Regards, Dave O'Brien From: Romanowski, John (OFT) Subject: Re: Performance problem Linux under Zvm If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
Issue CP IND USER linuxXyz Then we can see how big your Linux guest is. The Q V STOR is followed by a Ready message, which tells us it is executed under CMS, hence probably not in the Linux guest, but e.g. in MAINT, and your command would reveal how big MAINT is defined. 2008/3/21, O'Brien, David W. (NIH/CIT) [C] [EMAIL PROTECTED]: I made the mistake of believing what I was told. cp q storage 05:57:21 STORAGE = 2G Ready; T=0.01/0.01 05:57:21 cp q virtual storage 05:57:35 STORAGE = 128M Ready; T=0.01/0.01 05:57:35 cp q srm 06:01:42 IABIAS : INTENSITY=90%; DURATION=2 06:01:42 LDUBUF : Q1=100% Q2=75% Q3=60% 06:01:42 STORBUF: Q1=300% Q2=300% Q3=300% 06:01:42 DSPBUF : Q1=32767 Q2=32767 Q3=32767 06:01:42 DISPATCHING MINOR TIMESLICE = 5 MS 06:01:42 MAXWSS : LIMIT=% 06:01:42 .. : PAGES=99 06:01:42 XSTORE : 0% Any recommendation for LDUBUF? I just raised it to 100 100 100. Please advise if that change was counter indicated. Thank you, Dave O'Brien National Institutes of Health From: Gentry, Stephen [Sent: Thu 3/20/2008 2:51 PM Subject: Re: Performance problem Linux under Zvm What command did you use to determine that you had 768m central storage? QUERY STOREAGE? QUERY VIRTUAL STORAGE? Steve G. -Original Message- Subject: Re: Performance problem Linux under Zvm Thanks John cp q srm IABIAS : INTENSITY=90%; DURATION=2 LDUBUF : Q1=100% Q2=75% Q3=60% STORBUF: Q1=300% Q2=200% Q3=200% DSPBUF : Q1=32767 Q2=32767 Q3=32767 DISPATCHING MINOR TIMESLICE = 5 MS MAXWSS : LIMIT=% .. : PAGES=99 XSTORE : 0% Just got the following from one of the other techs (non-VM) We were able to diagnose the problem and make the necessary correction. The problem was z/VM has a total 768m of central available. The Linux guests (3 total) each had 768m of central allocated, therefore contention. The Linux guests are over allocated and are storage constrained with 768m of central. Understanding the Linux guests would be in contention with each other for this storage VM time sliced what it could for each guest, therefore the symptoms we experienced. My question to this group - Does a Linux quest really require 768MB of Central? Regards, Dave O'Brien From: Romanowski, John (OFT) Subject: Re: Performance problem Linux under Zvm If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health -- Kris Buelens, IBM Belgium, VM customer support
Re: Performance problem Linux under Zvm
On Fri, Mar 21, 2008 at 11:08 AM, O'Brien, David W. (NIH/CIT) [C] [EMAIL PROTECTED] wrote: I made the mistake of believing what I was told. One man's virtual is another man's real. We probably don't want to change CP after 40 years, but it is sometimes confusing that Q STOR returns either virtual or real depending on the CP privileges... It is easy to get confused. :story type=sad. One of my colleagues who managed DIRMAINT and thus could give himself class ABCDEFGH (sigh) opened a problem ticket about DEFINE STORAGE not working. Even after he did DEF STOR 8M and IPL CMS, Q STOR would still show him 256M (yes, long ago). :estory. So now we know that your Linux virtual machine were 768M and your z/VM is 2GB. Which means we can not have all Linux virtual machines completely resident (you also need some for z/VM itself and other users running). When your Linux virtual machines don't drop nicely from queue. 06:01:42 LDUBUF : Q1=100% Q2=75% Q3=60% When your virtual machines do not drop from queue, they end up as Q3 and CP will only want them to take 60% of storage. Any recommendation for LDUBUF? I just raised it to 100 100 100. Please advise if that change was counter indicated. It may appear to address your problem, but it really does not. It does make you more likely to thrash. It will tell CP to use all storage for Q3 users when needed. It's probably not enough for the full Linux servers, but may be enough when they have just restarted and did not gobble up enough data to cache and bother z/VM. No doubt someone will suggest to raise this even way over 100. That will make your system thrash even harder (and your configuration would get into that). Don't. It does not help to tell CP to imagine it had more than it really does. The real approach is to reduce the working set of the Linux virtual machines to fit in what you have available, instead of telling CP to think that it has 2 or 3 times as much. One way to reduce the working set size is to change the virtual machines into 512M plus 256M VDISK for cache. It is almost as fast (because VDISK is data-in-memory) but discourages Linux to use it as cache. It could also be that the Linux server did not really need 768M but someone just invented the number (or used this to run the GUI installer). If the 512M + 256M does not swap, you can probably lower the 512M and see where it goes. With defaults, MDC will probably gobble up quote some of your main storage as well (check with Q MDC). Considering your system, you probably want to disable that completely. Your z/VM is rather small to current standards. If you do find you page with the workload you want to run, it might make sense to add some expanded storage (or if you have nothing left, change some main into expanded). With that, and with decent paging subsystem you might be able to challenge CP a bit and tweak the SRM settings a little. Rob -- Rob van der Heij Velocity Software GmbH http://velocitysoftware.com/
Re: Performance problem Linux under Zvm
Thank you to all who have responded. What follows is the latest feedback from the person tasked with fixing this problem: Today we were able to re-allocate 512m of central storage to the Linux guests. This will place us comfortably below the 2gig available for VM. We performed this by a shutdown of the Linux guest, logging on as the user of the guest and specifying the new central storage value of (512m). We then re-ipled the Linux guest and noted 512m was now the new storage value. This however will not persist between IPL's. The z/VM USER DIRECT file in VM's directory must be edited and the associated values for USER changed to the new storage values. The question remains whether DIRMAINT will be used to make this change or xedit. We will wait for Bobby on Monday to determine what has been used to date. From: Mary Anne Matyaz [ Subject: Re: Performance problem Linux under Zvm The real approach is to reduce the working set of the Linux virtual machines to fit in what you have available, instead of telling CP to think that it has 2 or 3 times as much. One way to reduce the working set size is to change the virtual machines into 512M plus 256M VDISK for cache. It is almost as fast (because VDISK is data-in-memory) but discourages Linux to use it as cache. It could also be that the Linux server did not really need 768M but someone just invented the number (or used this to run the GUI installer). If the 512M + 256M does not swap, you can probably lower the 512M and see where it goes. Rob, Say you inherited a few linuxes and they are webservers, file servers, etc, that are set at 768M, 1024M, even 1536M. How do you tell what the working set size is? Is the wss indicated on the perfkit paging screen a valid indicator? I can't necessarily get in to these linuxes either. :) Thanks, Mary Anne
Re: Performance problem Linux under Zvm
The guideline for LDUBUF is to LOWER it from default, NEVER raise it unless you like to re-IPL z/VM. O'Brien, David W. (NIH/CIT) [C] wrote: I made the mistake of believing what I was told. cp q storage 05:57:21 STORAGE = 2G Ready; T=0.01/0.01 05:57:21 cp q virtual storage 05:57:35 STORAGE = 128M Ready; T=0.01/0.01 05:57:35 cp q srm 06:01:42 IABIAS : INTENSITY=90%; DURATION=2 06:01:42 LDUBUF : Q1=100% Q2=75% Q3=60% 06:01:42 STORBUF: Q1=300% Q2=300% Q3=300% 06:01:42 DSPBUF : Q1=32767 Q2=32767 Q3=32767 06:01:42 DISPATCHING MINOR TIMESLICE = 5 MS 06:01:42 MAXWSS : LIMIT=% 06:01:42 .. : PAGES=99 06:01:42 XSTORE : 0% Any recommendation for LDUBUF? I just raised it to 100 100 100. Please advise if that change was counter indicated. Thank you, Dave O'Brien National Institutes of Health From: Gentry, Stephen [Sent: Thu 3/20/2008 2:51 PM Subject: Re: Performance problem Linux under Zvm What command did you use to determine that you had 768m central storage? QUERY STOREAGE? QUERY VIRTUAL STORAGE? Steve G. -Original Message- Subject: Re: Performance problem Linux under Zvm Thanks John cp q srm IABIAS : INTENSITY=90%; DURATION=2 LDUBUF : Q1=100% Q2=75% Q3=60% STORBUF: Q1=300% Q2=200% Q3=200% DSPBUF : Q1=32767 Q2=32767 Q3=32767 DISPATCHING MINOR TIMESLICE = 5 MS MAXWSS : LIMIT=% .. : PAGES=99 XSTORE : 0% Just got the following from one of the other techs (non-VM) We were able to diagnose the problem and make the necessary correction. The problem was z/VM has a total 768m of central available. The Linux guests (3 total) each had 768m of central allocated, therefore contention. The Linux guests are over allocated and are storage constrained with 768m of central. Understanding the Linux guests would be in contention with each other for this storage VM time sliced what it could for each guest, therefore the symptoms we experienced. My question to this group - Does a Linux quest really require 768MB of Central? Regards, Dave O'Brien From: Romanowski, John (OFT) Subject: Re: Performance problem Linux under Zvm If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
Not enough information: How many Linux images on what size box (IFLs, memory). Also, is this all users in one or more specific Linux images, or certain users within an image? Could you show some numbers, like maybe an IND LOAD on the z/VM box? -- Robert P. Nix Mayo Foundation.~. RO-OE-5-55 200 First Street SW/V\ 507-284-0844 Rochester, MN 55905 /( )\ -^^-^^ In theory, theory and practice are the same, but in practice, theory and practice are different. On 3/20/08 12:36 PM, O'Brien, David W. (NIH/CIT) [C] [EMAIL PROTECTED] wrote: Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 1:37 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
Robert, Thank you for your response. This is what I have from the user coordinator: We have 3 virtual servers defined, but only 2 of them at a time seem to be running. We have only 1 ifl. There are only about 5 users (all except one in DCSS) defined to the servers. I'll try to get additional info. From: RPN01 [mailto:[EMAIL PROTECTED] Sent: Thu 3/20/2008 1:43 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Performance problem Linux under Zvm Not enough information: How many Linux images on what size box (IFLs, memory). Also, is this all users in one or more specific Linux images, or certain users within an image? Could you show some numbers, like maybe an IND LOAD on the z/VM box? -- Robert P. Nix Mayo Foundation.~. RO-OE-5-55 200 First Street SW/V\ 507-284-0844 Rochester, MN 55905 /( )\ -^^-^^ In theory, theory and practice are the same, but in practice, theory and practice are different. On 3/20/08 12:36 PM, O'Brien, David W. (NIH/CIT) [C] [EMAIL PROTECTED] wrote: Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
You might want to search the IBMVM archives for LDUBUF/STORBUF and SRM settings. Sounds like you may have guests in the 'E3' queue. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 1:37 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
Thanks John cp q srm IABIAS : INTENSITY=90%; DURATION=2 LDUBUF : Q1=100% Q2=75% Q3=60% STORBUF: Q1=300% Q2=200% Q3=200% DSPBUF : Q1=32767 Q2=32767 Q3=32767 DISPATCHING MINOR TIMESLICE = 5 MS MAXWSS : LIMIT=% .. : PAGES=99 XSTORE : 0% Just got the following from one of the other techs (non-VM) We were able to diagnose the problem and make the necessary correction. The problem was z/VM has a total 768m of central available. The Linux guests (3 total) each had 768m of central allocated, therefore contention. The Linux guests are over allocated and are storage constrained with 768m of central. Understanding the Linux guests would be in contention with each other for this storage VM time sliced what it could for each guest, therefore the symptoms we experienced. My question to this group - Does a Linux quest really require 768MB of Central? Regards, Dave O'Brien From: Romanowski, John (OFT) [mailto:[EMAIL PROTECTED] Sent: Thu 3/20/2008 1:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Performance problem Linux under Zvm If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 1:37 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
I've put up my popular presentation configuring z/VM and Linux for Performance at velocitysoftware.com/present/config. This presentation looks at how to ensure your performance for Linux and z/VM is optimal, and provides the best practices. I'm still working on the notes. YOu will want to look at the SET SRM STORBUF setting towards the end of the presentation. O'Brien, David W. (NIH/CIT) [C] wrote: Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
The problem was z/VM has a total 768m of central available. Really? That's so little - does someone want your POC to fail? Or was that an oversite? Verify: From a CLASS B privileged ID (try maint), issue Q STOR and Q XSTOR Does a Linux quest really require 768MB of Central? Depends on what it is running... WebSphere AppServer 6.1 - much more! Apache, way less! DB2 -- you might be OK there. How many VM paging volumes do you have? Issue Q ALLOC PAGE Keep %full 50, prefereably 35 or so. Marcy Cortes This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose, or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 11:12 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: [IBMVM] Performance problem Linux under Zvm The problem was z/VM has a total 768m of central available. The Linux guests (3 total) each had 768m of central allocated, therefore contention. The Linux guests are over allocated and are storage constrained with 768m of central. Understanding the Linux guests would be in contention with each other for this storage VM time sliced what it could for each guest, therefore the symptoms we experienced. My question to this group - Does a Linux quest really require 768MB of Central? Regards, Dave O'Brien From: Romanowski, John (OFT) [mailto:[EMAIL PROTECTED] Sent: Thu 3/20/2008 1:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Performance problem Linux under Zvm If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 1:37 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
Hello Dave, In either case, if a linux (zlinux or intel) gets 768M it will eventually allocate all storage it can get. If not for applications it will use the 'free' memory for datacache. So in that repect, yes, the linux will use the storage you assign to the guest. But as for if the image really needs the storage, that will depend on the application you are running in the guest. In some cases you really need the storage but some applications offer the requirements to be set. Oracle for example can set buffersizes. I always try to convince the linux and/or application owners to request less storage for their linux image. Try, but not always successfull. Our zLinux machines run from 200M up to about 6000M. The first are small machines such as installation and SSL, the latter are large oracle machines. We run the linux images in two VM's of 10G each. At least you're not the only one who has seen this. We also have found the situation where we over allocated the VM image. That's not bad, in fact it is our unique selling point, but if SRM is not set accordingly you'll find guests in the E-list and start to wonder why a guest will stop processing. Evaluate your LDUBUF and STORBUF and also make sure you will have enough page space to hold all of your (vitual) storage needs. Regards, Berry. O'Brien, David W. (NIH/CIT) [C] schreef: Thanks John cp q srm IABIAS : INTENSITY=90%; DURATION=2 LDUBUF : Q1=100% Q2=75% Q3=60% STORBUF: Q1=300% Q2=200% Q3=200% DSPBUF : Q1=32767 Q2=32767 Q3=32767 DISPATCHING MINOR TIMESLICE = 5 MS MAXWSS : LIMIT=% .. : PAGES=99 XSTORE : 0% Just got the following from one of the other techs (non-VM) We were able to diagnose the problem and make the necessary correction. The problem was z/VM has a total 768m of central available. The Linux guests (3 total) each had 768m of central allocated, therefore contention. The Linux guests are over allocated and are storage constrained with 768m of central. Understanding the Linux guests would be in contention with each other for this storage VM time sliced what it could for each guest, therefore the symptoms we experienced. My question to this group - Does a Linux quest really require 768MB of Central? Regards, Dave O'Brien From: Romanowski, John (OFT) [mailto:[EMAIL PROTECTED] Sent: Thu 3/20/2008 1:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Performance problem Linux under Zvm If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 1:37 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
Re: Performance problem Linux under Zvm
What command did you use to determine that you had 768m central storage? QUERY STOREAGE? QUERY VIRTUAL STORAGE? Steve G. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 2:12 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Performance problem Linux under Zvm Thanks John cp q srm IABIAS : INTENSITY=90%; DURATION=2 LDUBUF : Q1=100% Q2=75% Q3=60% STORBUF: Q1=300% Q2=200% Q3=200% DSPBUF : Q1=32767 Q2=32767 Q3=32767 DISPATCHING MINOR TIMESLICE = 5 MS MAXWSS : LIMIT=% .. : PAGES=99 XSTORE : 0% Just got the following from one of the other techs (non-VM) We were able to diagnose the problem and make the necessary correction. The problem was z/VM has a total 768m of central available. The Linux guests (3 total) each had 768m of central allocated, therefore contention. The Linux guests are over allocated and are storage constrained with 768m of central. Understanding the Linux guests would be in contention with each other for this storage VM time sliced what it could for each guest, therefore the symptoms we experienced. My question to this group - Does a Linux quest really require 768MB of Central? Regards, Dave O'Brien From: Romanowski, John (OFT) [mailto:[EMAIL PROTECTED] Sent: Thu 3/20/2008 1:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Performance problem Linux under Zvm If CP INDICATE QUEUES shows an En (like E3) in the 2nd column for one or more userids try CP QUERY SRM (write down response for reviewing ) and do this quick fix CP SET SRM STORBUF 300% 300% 300% This e-mail, including any attachments, may be confidential, privileged or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of O'Brien, David W. (NIH/CIT) [C] Sent: Thursday, March 20, 2008 1:37 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Performance problem Linux under Zvm Our shop is new to Zvm and Linux. We have a very small number of Linux users who are reporting significant response time problems. It almost seems as if each stops running for a period of time and is then re-dispatched. Is there a VM parameter that we might have taken the default on that needs tweaking? Any help or advice appreciated as this is a proof of concept endeavour and we would like not to turn off prospective users from the start. Thank you, Dave O'Brien National Institutes of Health
