My Informix Never free up memory!

Hi, group My ESQL/C program run a insert/update loop to load about 90,000 records from data file into tables. During run, I foud the share memory size kept growing from it's initial size to 750M! My box only got 128M physical RAM, I think it is a big performance problem for Informix will cause a large disk swapping, won't it? Can anyone please tell me why Informix dont' free memory after every excuttion of insert/update sql? I think I already free every sql id after prepare & execute it. In fact, even after exit of the program, the share memory size still dont go down. What it means? Below is some run time info, can anyone please checks and gives me and advice. 1, Tranaction logs is turned off for databases I operated on. 2, Shared Memory Parameters LOCKS 80000 # Maximum number of locks BUFFERS 500 # Maximum number of shared buffers NUMAIOVPS # Number of IO vps PHYSBUFF 2048 # Physical log buffer size (Kbytes) LOGBUFF 2048 # Logical log buffer size (Kbytes)LOGSMAX 20 # Maximum number of logical log files CLEANERS 1 # Number of buffer cleaner processes SHMBASE 0x82000000 # Shared memory base address SHMVIRTSIZE 81920 # initial virtual shared memorysegment size SHMADD 20480 # Size of new shared memory segments (Kbytes) SHMTOTAL 0 # Total shared memory (Kbytes). 0=>unlimited CKPTINTVL 1800 # Check point interval (in sec) LRUS 8 # Number of LRU queues LRU_MAX_DIRTY 60 # LRU percent dirty begin cleaninglimit LRU_MIN_DIRTY 50 # LRU percent dirty end cleaning limit LTXHWM 50 # Long transaction high water markpercentage LTXEHWM 60 # Long transaction high water mark (exclusive) TXTIMEOUT 0x12c # Transaction timeout (in sec) STACKSIZE 64 # Stack size (Kbytes) 3, onstat -b INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 15:13:00 -- 754408 Kbytes Buffers address userthread flgs pagenum memaddr nslots pgflgs xflgs owner waitlist 825a5a2c 0 23 395cdf 82fb9800 3 1 10 0 0 825a86dc 0 23 39bada 82ff0800 3 1 10 0 0 825ad4dc 0 23 395cdd 83050800 3 1 10 0 0 278 modified, 500 total, 512 hash buckets, 2048 buffer size 4, onstat -p INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 15:19:48 -- 754408 Kbytes Profile dskreads pagreads bufreads %cached dskwrits pagwrits bufwrits %cached 103425 104107 30929193 99.67 49203 92680 837271 94.12 isamtot open start read write rewrite delete commit rollbk 19291051 1633683 2082423 11202078 203016 122600 0 0 0 ovlock ovuserthread ovbuff usercpu syscpu numckpts flushes 0 0 0 1855.32 200.83 4 62 bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress seqscans 39330 0 24964352 0 0 2 0 29 ixda-RA idx-RA da-RA RA-pgsused lchwaits 50464 0 132 50556 300 5, onstat -g ses INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 15:20:24 -- 754408 Kbytes session #RSAM total used id user tty pid hostname threads memory memory 15 informix - 0 - 0 8192 4588 14 informix ttyp0 2712 lczmdbsv 1 422092800 422053416 8 informix - 0 - 0 8192 4588 6 informix - 0 - 0 8192 4588 5 informix - 0 - 0 16384 12224 4 informix - 0 - 0 16384 12224 3 informix - 0 - 0 8192 4588 2 informix - 0 - 0 8192 4588

steven wrote: > Hi, group > > My ESQL/C program run a insert/update loop to load about 90,000 > records from data file into tables. During run, I foud the share > memory size kept growing from it's initial size to 750M! My box only > got 128M physical RAM, I think it is a big performance problem for > Informix will cause a large disk swapping, won't it? > > Can anyone please tell me why Informix dont' free memory after every > excuttion of insert/update sql? I think I already free every sql id > after prepare & execute it. In fact, even after exit of the program, > the share memory size still dont go down. What it means? > > Below is some run time info, can anyone please checks and gives me and > advice. > > 1, Tranaction logs is turned off for databases I operated on. > > 2, Shared Memory Parameters > > LOCKS 80000 # Maximum number of locks > BUFFERS 500 # Maximum number of shared buffers > NUMAIOVPS # Number of IO vps > PHYSBUFF 2048 # Physical log buffer size (Kbytes) > LOGBUFF 2048 # Logical log buffer size (Kbytes) Your buffers are vastly underconfigured - 50000 would be a starting point, and go up from there. However, your ESQL/C program is going to more relevant to debugging the memory leak. There's nothing in the ONCONFIG file that says "leak memory" or "do not free memory". Gut feel, you should be looking at the free statements in your program. -- Jonathan Leffler #include <disclaimer.h> Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/

You need to review your code, you appear to have a memory leak and/or you are not freeing resources as you should. You could throw clint at the code, it's very good at spotting problems steven wrote: > > Hi, group > > My ESQL/C program run a insert/update loop to load about 90,000 > records from data file into tables. During run, I foud the share > memory size kept growing from it's initial size to 750M! My box only > got 128M physical RAM, I think it is a big performance problem for > Informix will cause a large disk swapping, won't it? > > Can anyone please tell me why Informix dont' free memory after every > excuttion of insert/update sql? I think I already free every sql id > after prepare & execute it. In fact, even after exit of the program, > the share memory size still dont go down. What it means? > > Below is some run time info, can anyone please checks and gives me and > advice. > > 1, Tranaction logs is turned off for databases I operated on. > > 2, Shared Memory Parameters > > LOCKS 80000 # Maximum number of locks > BUFFERS 500 # Maximum number of shared buffers > NUMAIOVPS # Number of IO vps > PHYSBUFF 2048 # Physical log buffer size (Kbytes) > LOGBUFF 2048 # Logical log buffer size (Kbytes)> LOGSMAX 20 # Maximum number of logical log files > CLEANERS 1 # Number of buffer cleaner processes > SHMBASE 0x82000000 # Shared memory base address > SHMVIRTSIZE 81920 # initial virtual shared memory> segment size > SHMADD 20480 # Size of new shared memory segments > (Kbytes) > SHMTOTAL 0 # Total shared memory (Kbytes). > 0=>unlimited > CKPTINTVL 1800 # Check point interval (in sec) > LRUS 8 # Number of LRU queues > LRU_MAX_DIRTY 60 # LRU percent dirty begin cleaning> limit > LRU_MIN_DIRTY 50 # LRU percent dirty end cleaning limit > LTXHWM 50 # Long transaction high water mark> percentage > LTXEHWM 60 # Long transaction high water mark > (exclusive) > TXTIMEOUT 0x12c # Transaction timeout (in sec) > STACKSIZE 64 # Stack size (Kbytes)> > 3, onstat -b > INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 15:13:00 -- 754408 > Kbytes > > Buffers > address userthread flgs pagenum memaddr nslots pgflgs xflgs owner > waitlist > 825a5a2c 0 23 395cdf 82fb9800 3 1 10 0 > 0 > 825a86dc 0 23 39bada 82ff0800 3 1 10 0 > 0 > 825ad4dc 0 23 395cdd 83050800 3 1 10 0 > 0 > 278 modified, 500 total, 512 hash buckets, 2048 buffer size > > 4, onstat -p > INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 15:19:48 -- 754408 > Kbytes > Profile > dskreads pagreads bufreads %cached dskwrits pagwrits bufwrits %cached > 103425 104107 30929193 99.67 49203 92680 837271 94.12 > isamtot open start read write rewrite delete commit > rollbk > 19291051 1633683 2082423 11202078 203016 122600 0 0 > 0 > ovlock ovuserthread ovbuff usercpu syscpu numckpts flushes > 0 0 0 1855.32 200.83 4 62 > bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress > seqscans > 39330 0 24964352 0 0 2 0 29 > ixda-RA idx-RA da-RA RA-pgsused lchwaits > 50464 0 132 50556 300 > > 5, onstat -g ses > INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 15:20:24 -- 754408 > Kbytes > session #RSAM total used > id user tty pid hostname threads memory memory > 15 informix - 0 - 0 8192 4588 > 14 informix ttyp0 2712 lczmdbsv 1 422092800 > 422053416 > 8 informix - 0 - 0 8192 4588 > 6 informix - 0 - 0 8192 4588 > 5 informix - 0 - 0 16384 12224 > 4 informix - 0 - 0 16384 12224 > 3 informix - 0 - 0 8192 4588 > 2 informix - 0 - 0 8192 4588 -- Paul Watson # Oninit Ltd # Growing old is mandatory Tel: +44 1436 672201 # Growing up is optional Fax: +44 1436 678693 # Mob: +44 7818 003457 # www.oninit.com #

Paul Watson <paul@oninit.com> wrote in message news:<3F6432B8.E225C65A@oninit.com>... > You need to review your code, you appear to have a memory leak > and/or you are not freeing resources as you should. You could throw > clint at the code, it's very good at spotting problems what you mean 'throw clint'? I dont understand, could you tell me more? thanks! -- steven@lczmsoft.com

Jonathan Leffler <jleffler@earthlink.net> wrote in message news:<84U8b.2850$BS5.1636@newsread4.news.pas.earthlink.net>... > Your buffers are vastly underconfigured - 50000 would be a starting > point, and go up from there. > I skimed through the Informix Performance Guide, it said, that the BUFFERS should be about 20% - 25% of physical memory size. For my host equipment with 128M physical memory and page size if 2k, so I think 50000 is too large to me. Should I set it to about 15000? > However, your ESQL/C program is going to more relevant to debugging > the memory leak. There's nothing in the ONCONFIG file that says "leak > memory" or "do not free memory". Gut feel, you should be looking at > the free statements in your program. I still dont find any leak in my code :-(, is there any quite easy tool?

Do a web search for clint - its a c-code checker - it's very good steven wrote: > > Paul Watson <paul@oninit.com> wrote in message news:<3F6432B8.E225C65A@oninit.com>... > > You need to review your code, you appear to have a memory leak > > and/or you are not freeing resources as you should. You could throw > > clint at the code, it's very good at spotting problems > > what you mean 'throw clint'? I dont understand, could you tell me more? thanks! > > -- > steven@lczmsoft.com -- Paul Watson # Oninit Ltd # Growing old is mandatory Tel: +44 1436 672201 # Growing up is optional Fax: +44 1436 678693 # Mob: +44 7818 003457 # www.oninit.com #

steven@steven4u.net (steven) writes: > Hi, group > > My ESQL/C program run a insert/update loop to load about 90,000 > records from data file into tables. During run, I foud the share > memory size kept growing from it's initial size to 750M! My box only > got 128M physical RAM, I think it is a big performance problem for > Informix will cause a large disk swapping, won't it? > > Can anyone please tell me why Informix dont' free memory after every > excuttion of insert/update sql? I think I already free every sql id > after prepare & execute it. In fact, even after exit of the program, > the share memory size still dont go down. What it means? > > Below is some run time info, can anyone please checks and gives me and > advice. Can you pleasy clarify if this supposed memory leakage happens in the client or in the server? From your post I suspect you're talking about the server, but the answers here have all been talking about the client. 128M physical RAM is not very much these days, to put it mildly. -- 'yvind Gjerstad Systems dept Tollpost-Globe AS N-6301 'ndalsnes/Norway E-mail: ogj@tollpost.no Phone: +47 7122 6663 Fax: +47 7122 6694