"Onmode -F " hangs engine

Hello for all. Sorry for my English. We have SAP R/3 4.6C on Informix 7.31.TD1X2 (W2K). Engine eats too mach memory for 2-3 weeks and system began swapping. I read somewhere about "onmode -F". And I had shedule this command at every 4 hours. This works fine on NT 4.0, but since upgrade to W2K, we sometimes has a big problem: "onmode -F" is standstill for very large time (more than 30 minutes). After this, oninit consumpt 50% CPU time, and Informix does not respond to any connections. Even, I can't stop engine with "onmode -yuk". I must to reboot the system. This problem appear only 3 times for a year, when engine has a peak of memory consumption (more than 1 GB). This is some parameters of our onconfig: RESIDENT 0 # Forced residency flag (Yes = 1, No = 0) MULTIPROCESSOR 1 # 0 for single-processor, 1 for multi-processor NUMCPUVPS 2 # Number of user (cpu) vps SINGLE_CPU_VP 0 # If non-zero, limit number of cpu vps toone BUFFERS 100000 # Maximum number of shared buffers NUMAIOVPS 1 # Number of IO vps SHMBASE 0xC000000 # Shared memory base address SHMVIRTSIZE 327680 # initial virtual shared memory segment size SHMADD 32768 # Size of new shared memory segments (Kbytes) SHMTOTAL 0 # Total shared memory (Kbytes). 0=>unlimited Has anybody the same situation in the past? Who know what is the problem, please answer. With Best Regards, Yury Borinsky Basis Administrator and DBA of JSC "Dneproenergo" (Ukrane)

Hi. Given problem can take place when sesions which allocate a lot of memory have being worked a long time. Perhaps they do a hard DSS queries. Can you drop and restart this connections from time to time. onmode -F don't release memory while a sesions which use that memoryisn't terminated. You also can set SHMTOTAL parameter on the onconfig. In this case all sessions which need for additional memory will be rolled back when server already have took SHMTOTAL amount of memory. Server will be work, but some transactions will be roolled back by server. Best regards, Konstantin. p.s. I'm from Dnipropetrovsk too. ----- Original Message ----- From: "Yury Borinsky" <bor@gc.dnepr.energy.gov.ua> To: <ids@iiug.org> Sent: Wednesday, January 19, 2005 10:35 AM Subject: "Onmode -F " hangs engine [4042] > Hello for all. > Sorry for my English. > > We have SAP R/3 4.6C on Informix 7.31.TD1X2 (W2K). > Engine eats too mach memory for 2-3 weeks and system began swapping. > I read somewhere about "onmode -F". And I had shedule this command at > every > 4 hours. > This works fine on NT 4.0, but since upgrade to W2K, we sometimes has a > big > problem: > "onmode -F" is standstill for very large time (more than 30 minutes). > After > this, > oninit consumpt 50% CPU time, and Informix does not respond to any > connections. > Even, I can't stop engine with "onmode -yuk". > I must to reboot the system. This problem appear only 3 times for a year, > when engine > has a peak of memory consumption (more than 1 GB). > > This is some parameters of our onconfig: > RESIDENT 0 # Forced residency flag (Yes = 1,> No > = 0) > MULTIPROCESSOR 1 # 0 for single-processor, 1 for> multi-processor > NUMCPUVPS 2 # Number of user (cpu) vps > SINGLE_CPU_VP 0 # If non-zero, limit number of cpu vps to> one > BUFFERS 100000 # Maximum number of shared buffers > NUMAIOVPS 1 # Number of IO vps > SHMBASE 0xC000000 # Shared memory base address > SHMVIRTSIZE 327680 # initial virtual shared memory segment> size > SHMADD 32768 # Size of new shared memory segments > (Kbytes) > SHMTOTAL 0 # Total shared memory (Kbytes). > 0=>unlimited> > Has anybody the same situation in the past? > Who know what is the problem, please answer. > > With Best Regards, Yury Borinsky > Basis Administrator and DBA of JSC "Dneproenergo" (Ukrane) > > > > > >

Hi Yury, We used to have that problem, and probably still do, but it's not noticable much anymore, we are on unix. We upgraded our 7.31.UC... Informix engine to 7.31.FC... (64 bit), this increased the grand memory level for informix past the 3.? GB limit. Our informix engine starts up with nearly 6GB of memory now. >>7.31.TD1X2 (yours)... Is that 64 bit? Upgrading to 64 bit is one option. Another is as Konstantin said, but killing the session inside SAP, not at the informix level. To be clean, you have to track the informix session id back into SAP and find which sapr3 thread is on that session id, then go to SAP trans (forgot which one), and kill the sap session when no process is on it. The SAP thread will respawn it self accordingly to your SAP settings. Another option is that there is an SAP note (forgot which one, it's been a couple of years since we looked at it) that talks about setting SAP parameters that will automatically kill/respawn SAP sessions after a time or usage interval (I don't remember the specifics). We did not implement it as our manager at the time was afraid to. So we worked around it, but perhaps you can find that OSS note and try it. (easiest to enter an OSS ticket and ask.. Sometimes those OSS searches are fruitless for a long time before you find something!). Good luck. Norma Jean -----Original Message----- From: forum.subscriber@iiug.org [mailto:forum.subscriber@iiug.org] On Behalf Of Konstantin .... Sent: Wednesday, January 19, 2005 4:24 AM To: ids@iiug.org Subject: Re: "Onmode -F " hangs engine [4043] Hi. Given problem can take place when sesions which allocate a lot of memory have being worked a long time. Perhaps they do a hard DSS queries. Can you drop and restart this connections from time to time. onmode -F don't release memory while a sesions which use that memoryisn't terminated. You also can set SHMTOTAL parameter on the onconfig. In this case all sessions which need for additional memory will be rolled back when server already have took SHMTOTAL amount of memory. Server will be work, but some transactions will be roolled back by server. Best regards, Konstantin. p.s. I'm from Dnipropetrovsk too. ----- Original Message ----- From: "Yury Borinsky" <bor@gc.dnepr.energy.gov.ua> To: <ids@iiug.org> Sent: Wednesday, January 19, 2005 10:35 AM Subject: "Onmode -F " hangs engine [4042] > Hello for all. > Sorry for my English. > > We have SAP R/3 4.6C on Informix 7.31.TD1X2 (W2K). > Engine eats too mach memory for 2-3 weeks and system began swapping. > I read somewhere about "onmode -F". And I had shedule this command at > every > 4 hours. > This works fine on NT 4.0, but since upgrade to W2K, we sometimes has a > big > problem: > "onmode -F" is standstill for very large time (more than 30 minutes). > After > this, > oninit consumpt 50% CPU time, and Informix does not respond to any > connections. > Even, I can't stop engine with "onmode -yuk". > I must to reboot the system. This problem appear only 3 times for a year, > when engine > has a peak of memory consumption (more than 1 GB). > > This is some parameters of our onconfig: > RESIDENT 0 # Forced residency flag (Yes = 1,> No > = 0) > MULTIPROCESSOR 1 # 0 for single-processor, 1 for> multi-processor > NUMCPUVPS 2 # Number of user (cpu) vps > SINGLE_CPU_VP 0 # If non-zero, limit number of cpu vpsto > one > BUFFERS 100000 # Maximum number of shared buffers > NUMAIOVPS 1 # Number of IO vps > SHMBASE 0xC000000 # Shared memory base address > SHMVIRTSIZE 327680 # initial virtual shared memorysegment > size > SHMADD 32768 # Size of new shared memorysegments > (Kbytes) > SHMTOTAL 0 # Total shared memory (Kbytes). > 0=>unlimited> > Has anybody the same situation in the past? > Who know what is the problem, please answer. > > With Best Regards, Yury Borinsky > Basis Administrator and DBA of JSC "Dneproenergo" (Ukrane) > > > > > > ============================================================ The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs ============================================================

Konstantin, Per default for SAP systems the onconfig parameter SHMTOTAL is actually set to value 0. If you set SHMTOTAL to other value in similiar situation without any other configuration changes you will run into same problem but earlier than before. If the database system requires more shared memory than is allowed by the parameter SHMTOTAL, this should be set to a higher value or to 0 (=unrestricted). If, however, the shared memory that has already been used is right on the limit of the shared memory that is made available by the operating system, then you should consider measures to reduce the shared memory consumption; the removal of other applications to other machines or a storage extension. The parameter SHMTOTAL should only be set to a value not equal to 0 if the available shared memory has to be shared online with other applications. For SAP systems SAP note 38307 might be helpful... See below what I wrote to Yuri in the meantime: Dear Yury, Currently used 7.31.TD1X2 is quite old. The problem sounds like product defect IBM Informix PTS 149742 which is fixed in IDS releases >= 7.31.TD2. 149742 ONMODE -F CAN HANG SERVER,... For SAP system you should upgrade to 7.31.TD2X7 in minimum, alternatively to latest available release 7.31.TD6 according SAP note 156549. For large SHM configurations you may consider IDS release notes and SAP recommendations as well. See also SAP note 41360. If the problem still persists please open a SAP case. Please see SAP note 113047 for needed information. For other questions please feel free to contact me again using mail address robert.seifert@sap.com By the way, do you know that there exists a SAP list @ IUUG as well? You may subscribe sapmix@iiug.org. HTH Best regards, Robert Robert Seifert _________________________________________________________________ IBM Deutschland GmbH, Data Management Solutions, Software, Informix Products, Advanced Support Elite VAR support engineer Development Support for IBM Informix Products on-site @ SAP c/o SAP AG, -TechDev05-,Neurottstrasse 16, D-69190 Walldorf, Germany Phone +49-6227-7-45713, Fax +49-6227-7-47147 mailto:robert.seifert@de.ibm.com or mailto:robert.seifert@sap.com http://www.ibm.com/software/data/informix "Konstantin ...." <cvk@pbcards.dp.u a> To Sent by: ids@iiug.org forum.subscriber@ cc iiug.org Subject Re: "Onmode -F " hangs engine 19.01.2005 11:24 [4043] Hi. Given problem can take place when sesions which allocate a lot of memory have being worked a long time. Perhaps they do a hard DSS queries. Can you drop and restart this connections from time to time. onmode -F don't release memory while a sesions which use that memoryisn't terminated. You also can set SHMTOTAL parameter on the onconfig. In this case all sessions which need for additional memory will be rolled back when server already have took SHMTOTAL amount of memory. Server will be work, but some transactions will be roolled back by server. Best regards, Konstantin. p.s. I'm from Dnipropetrovsk too. ----- Original Message ----- From: "Yury Borinsky" <bor@gc.dnepr.energy.gov.ua> To: <ids@iiug.org> Sent: Wednesday, January 19, 2005 10:35 AM Subject: "Onmode -F " hangs engine [4042] > Hello for all. > Sorry for my English. > > We have SAP R/3 4.6C on Informix 7.31.TD1X2 (W2K). > Engine eats too mach memory for 2-3 weeks and system began swapping. > I read somewhere about "onmode -F". And I had shedule this command at > every > 4 hours. > This works fine on NT 4.0, but since upgrade to W2K, we sometimes has a > big > problem: > "onmode -F" is standstill for very large time (more than 30 minutes). > After > this, > oninit consumpt 50% CPU time, and Informix does not respond to any > connections. > Even, I can't stop engine with "onmode -yuk". > I must to reboot the system. This problem appear only 3 times for a year, > when engine > has a peak of memory consumption (more than 1 GB). > > This is some parameters of our onconfig: > RESIDENT 0 # Forced residency flag (Yes = 1,> No > = 0) > MULTIPROCESSOR 1 # 0 for single-processor, 1 for> multi-processor > NUMCPUVPS 2 # Number of user (cpu) vps > SINGLE_CPU_VP 0 # If non-zero, limit number of cpu vps to> one > BUFFERS 100000 # Maximum number of shared buffers > NUMAIOVPS 1 # Number of IO vps > SHMBASE 0xC000000 # Shared memory base address > SHMVIRTSIZE 327680 # initial virtual shared memory segment> size > SHMADD 32768 # Size of new shared memory segments > (Kbytes) > SHMTOTAL 0 # Total shared memory (Kbytes). > 0=>unlimited> > Has anybody the same situation in the past? > Who know what is the problem, please answer. > > With Best Regards, Yury Borinsky > Basis Administrator and DBA of JSC "Dneproenergo" (Ukrane) > > > > > >