shared memory full

Hi, Today, memory of our informix instance increased from 2.5GB to 8GB within 9 minutes. That instance is running with 2.5GB shared memory for around last 4 months. 8GB is defined as shmtotal so further requests for memory chunks overturned by the instance. Usually we have session with a few KB memory to around 10MB memory. Sorting the sessions on memory we got a session aquiring 1.83GB as can be seen below: onstat -g ses 950863 appuser - -1 fmcpdbsp 1 9551872 9526088 off 955301 appuser - -1 fmcpappp 1 9908224 9822320 off 954997 appuser - -1 fmcpappp 1 9928704 9846984 off 945865 appuser - -1 fmcpappp 1 10002432 9965144 off 950909 appuser - -1 fmcpappp 1 10113024 10069920 off 947189 appuser - -1 fmcpappp 1 10407936 10399920 off 955492 appuser - -1 CASUN185 1 1971294208 1961083904 off -bash-3.2$ onstat -g ses 955492 IBM Informix Dynamic Server Version 11.50.FC8W3 -- On-Line (Prim) -- Up 97 days 09:11:52 -- 8259584 Kbytes session effective #RSAM total used dynamic id user user tty pid hostname threads memory memory explain 955492 appuser - - -1 CASUN185 1 1981358080 1970916600 off tid name rstcb flags curstk status 954290 sqlexec 15d3efb78 ---P--- 7855 running- Memory pools count 2 name class addr totalsize freesize #allocfrag #freefrag 955492 V 14804b040 1982713856 10469440 2729374 37206 955492*O0 V 17568e040 4096 808 1 1 name free used name free used overhead 0 6576 mtmisc 0 72 scb 0 144 opentable 0 238856560 filetable 0 6072 ru 0 600 misc 0 80 log 0 16536 temprec 0 21664 keys 0 736 ralloc 0 1524405528 gentcb 0 1712 ostcb 0 2816 sqscb 0 9639672 sql 0 72 rdahead 0 352 hashfiletab 0 552 osenv 0 2992 buft_buffer 0 2168 sqtcb 0 3177936 fragman 0 196105192 udr 0 168 sqscb info scb sqscb optofc pdqpriority optcompind directives 16c105b88 133144028 0 0 2 1 Sess SQL Current Iso Lock SQL ISAM F.E. Id Stmt type Database Lvl Mode ERR ERR Vers Explain 955492 SELECT inventory CR Wait 5 0 0 9.28 Off Current Role : app_role Current SQL statement : Last parsed SQL statement : SELECT 1 FROM SYSTABLES WHERE TABID = 1 Operating system: Sun Solaris The second largest session w.r.t. memory was holding 9.9 MB. We are unable to find or guess where the rest of memory gone. Also: how a session can consume so much memory? if cummulative memory of session (through onstat -g ses) is consuming only 20% of total instance memory, why we receing the messages in online log that no more memory chunks are allowed as shmtotal limit reached

If one of your sessions allocated too much locks you may have seen a lock table growth. This will consume memory (sometimes a lot) but will leave evidences in online.log Your session may have got a lot of prepared statements that were not freed. Most times bad Java coding is the cause. An onstat -g stm SID may show you this. Regards. On Mon, Jan 13, 2014 at 6:27 PM, KAMRAN HAQ <khaq@i2cinc.com> wrote: > Hi, > Today, memory of our informix instance increased from 2.5GB to 8GB within 9 > minutes. > That instance is running with 2.5GB shared memory for around last 4 months. > 8GB is defined as shmtotal so further requests for memory chunks > overturned by > the instance. > Usually we have session with a few KB memory to around 10MB memory. > Sorting the sessions on memory we got a session aquiring 1.83GB as can be > seen > below: > > onstat -g ses> > 950863 appuser - -1 fmcpdbsp 1 9551872 9526088 off > 955301 appuser - -1 fmcpappp 1 9908224 9822320 off > 954997 appuser - -1 fmcpappp 1 9928704 9846984 off > 945865 appuser - -1 fmcpappp 1 10002432 9965144 off > 950909 appuser - -1 fmcpappp 1 10113024 10069920 off > 947189 appuser - -1 fmcpappp 1 10407936 10399920 off > 955492 appuser - -1 CASUN185 1 1971294208 1961083904 off > -bash-3.2$ onstat -g ses 955492> > IBM Informix Dynamic Server Version 11.50.FC8W3 -- On-Line (Prim) -- Up 97 > days 09:11:52 -- 8259584 Kbytes> > session effective #RSAM total used dynamic > id user user tty pid hostname threads memory memory explain > 955492 appuser - - -1 CASUN185 1 1981358080 1970916600 off > > tid name rstcb flags curstk status > 954290 sqlexec 15d3efb78 ---P--- 7855 running- > > Memory pools count 2 > name class addr totalsize freesize #allocfrag #freefrag > 955492 V 14804b040 1982713856 10469440 2729374 37206 > 955492*O0 V 17568e040 4096 808 1 1 > > name free used name free used > overhead 0 6576 mtmisc 0 72 > scb 0 144 opentable 0 238856560 > filetable 0 6072 ru 0 600 > misc 0 80 log 0 16536 > temprec 0 21664 keys 0 736 > ralloc 0 1524405528 gentcb 0 1712 > ostcb 0 2816 sqscb 0 9639672 > sql 0 72 rdahead 0 352 > hashfiletab 0 552 osenv 0 2992 > buft_buffer 0 2168 sqtcb 0 3177936 > fragman 0 196105192 udr 0 168 > > sqscb info > scb sqscb optofc pdqpriority optcompind directives > 16c105b88 133144028 0 0 2 1 > > Sess SQL Current Iso Lock SQL ISAM F.E. > Id Stmt type Database Lvl Mode ERR ERR Vers Explain > 955492 SELECT inventory CR Wait 5 0 0 9.28 Off > > Current Role : app_role > > Current SQL statement : > > Last parsed SQL statement : > SELECT 1 FROM SYSTABLES> > WHERE TABID = 1 > > Operating system: Sun Solaris > > The second largest session w.r.t. memory was holding 9.9 MB. We are unable > to > find or guess > where the rest of memory gone. Also: > how a session can consume so much memory? > if cummulative memory of session (through onstat -g ses) is consuming only > 20% > of total instance memory, > why we receing the messages in online log that no more memory chunks are > allowed as shmtotal limit reached > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --089e01176cc3d14dc704efde5697