Many sessions in IDS 7.31 cause problem!!

Dear All... I've a IDS 7.31.UD8 running in RedHat Linux(uname -a showes 2.6.9-42 kernel), Every morning AM 8:00, it will be 270 sessions in it , and then at AM 8:30, there will be 380 sessions connect to IDS almost at the same time , and then the engine hang , I've to onmode -ky IDS ... These 380 sessions are for test only , after job done,they die !! that will only take for 2 to 3 minutes , unfortunately , some of them can not connect IDS because IDS die , I've noticed that if I stop running 190 sessions before AM 8:30 , I mean , only 80 sessions in IDS and the coming 380 sessions , the IDS will be fine ... Those stop running sessions do nothing at all at AM 8:30 , they just connect DB,and sleep , like following : if(strncmp(basetime_hms,"090100",6) < 0 ){ fprintf(fp,"[warmaker] System Time is ==> [%s]\\ ",basetime_hms); sleep(3); continue; } As you can see , those 190 stop running sessions do nothing at AM 8:30, just only connect IDS and then sleep until to AM 09:01 , stop these 190 sessions will help the test at AM 08:30 goes well , if not , the IDS hang ... NETTYPE ipcshm,1,20,CPU NETTYPE soctcp,16,50,NET and cpuvps = 20 , don't know why this happened , in my case , onstat -g ses | wc -l , it showes 605 and then I can not connect to IDS anymore .... no matter ap or dbaccess ..... Any comments , suggestions are welcome !! Mars

Hi, with the NETTYPE parameter you can adjust the number of poll threads in IDS, and with this you can (indirectly) determine how many concurrent sessions IDS can take. But concurrently attempted connects are a different story, because they need to go through the listener thread. In your case this seems to be the bottle neck. You could try to configure several listener threads listening on different port numbers (DBSERVERALIASES, sqlhosts, etc.). Though I'm not sure this will really help. Other than that you can try to make the application so that not all clients connect at the same time. Try to spread the time a bit, i.e. have them to connect over a period of 5 minutes at different time slots. This surely should help. And if they first sleep before starting to be active, this should not be too much of a problem - some will just sleep a bit less. :) Regards, Martin -- Martin Fuerderer IBM Informix Development Munich, Germany Information Management IBM Deutschland Research & Development GmbH Chairman of the Supervisory Board: Martin Jetter Board of Management: Erich Baier Corporate Seat: Boeblingen, Germany Reg.-Gericht: Amtsgericht Stuttgart, HRB 243294 ids-bounces@iiug.org wrote on 10/26/2009 03:39:54 AM: > > Dear All... > > I've a IDS 7.31.UD8 running in RedHat Linux(uname -a showes 2.6. > 9-42 kernel), > Every morning AM 8:00, it will be 270 sessions in it , and then atAM 8:30, > there will be 380 sessions connect to IDS almost at the same time , > and then the engine hang , I've to onmode -ky IDS ... > > These 380 sessions are for test only , after job done,they die !! > that will only take for 2 to 3 minutes , > unfortunately , some of them can not connect IDS because IDS die , > I've noticed that if I stop running 190 sessions before AM 8:30 , > I mean , only 80 sessions in IDS and the coming 380 sessions , > the IDS will be fine ... Those stop running sessions do nothing atall at AM > 8:30 , they just connect DB,and sleep , like following : > > if(strncmp(basetime_hms,"090100",6) < 0 ){ > fprintf(fp,"[warmaker] System Time is ==> [%s]\\ ",basetime_hms); > sleep(3); > continue; > } > > As you can see , those 190 stop running sessions do nothing at AM 8:30, > just only connect IDS and then sleep until to AM 09:01 , > stop these 190 sessions will help the test at AM 08:30 goes well , > if not , the IDS hang ... > > NETTYPE ipcshm,1,20,CPU > NETTYPE soctcp,16,50,NET> > and cpuvps = 20 , don't know why this happened , in my case , > onstat -g ses | wc -l , it showes 605 and then I can not connect to IDS > anymore .... no matter ap or dbaccess .....> > Any comments , suggestions are welcome !! > > Mars > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >

Dear All.... Thanks for all who gave me very helpful information ... After days of test , now I can figure out the real reason cause the problem .... I've a IDS called IF8 , another IDS called if12 ,both IDS are 7.31.UD8 running in RedHat , In my case , I've hundreds of esql/c ap which connect to IF8 and open cursor like the following in esql/c source : sprintf(StrSql,"%s%s%s%s%s%s%s%s", "select id,ptrade,puplim,pdownlim ", "from sit2@if12:quote ", "where id in ('",GlbWarid,"','",BasicStkid,"') ", "order by id "); EXEC SQL PREPARE pre_quote FROM $StrSql; EXEC SQL DECLARE cur_quote CURSOR FOR pre_quote; EXEC SQL OPEN cur_quote; while(1){ fetch ...fetch ...fetch... } EXEC SQL CLOSE cur_quote; EXEC SQL FREE cur_quote; EXEC SQL FREE pre_quote ; I've observed that if I don't execute this sql ,everything is fine , If I execute this , in IF8 and if12 , onstat -g opn|wc -l showes thousands of open cursors all the time , most of them are systables in both IDS engine , execute hundreds of such esql/c ap would hang the IF8, if12 keep alive .... if skip this sql , I can run 1000 such esql/c ap without problems !! I can use two connections in esql/c ap ,this might solve my problem , also I can move sit2:quote from if12 to IF8 , it works ,too . I wonder why this cause the problem I encountered ?! Any suggestions are welcome !!~~

Old bug in 7.31xD8 IIRC. Art Art S. Kagel Oninit (www.oninit.com) IIUG Board of Directors (art@iiug.org) Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Oninit, the IIUG, nor any other organization with which I am associated either explicitly or implicitly. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Mon, Oct 26, 2009 at 8:17 PM, MARS CHEN <mars@jsun.com> wrote: > Dear All.... > > Thanks for all who gave me very helpful information ... After days of test > , > now I can figure out the real reason cause the problem .... > > I've a IDS called IF8 , another IDS called if12 ,both IDS are 7.31.UD8 > running in RedHat , In my case , I've hundreds of esql/c ap which connect > to > IF8 and open cursor like the following in esql/c source : > > sprintf(StrSql,"%s%s%s%s%s%s%s%s", > > "select id,ptrade,puplim,pdownlim ", > > "from sit2@if12:quote ", > > "where id in ('",GlbWarid,"','",BasicStkid,"') ", > > "order by id "); > > EXEC SQL PREPARE pre_quote FROM $StrSql; > > EXEC SQL DECLARE cur_quote CURSOR FOR pre_quote; > > EXEC SQL OPEN cur_quote; > > while(1){ > > fetch ...fetch ...fetch... > > } > > EXEC SQL CLOSE cur_quote; > > EXEC SQL FREE cur_quote; > > EXEC SQL FREE pre_quote ; > > I've observed that if I don't execute this sql ,everything is fine , > If I execute this , in IF8 and if12 , onstat -g opn|wc -l showes > thousands of open cursors all the time , most of them are systables in > both IDS engine , execute hundreds of such esql/c ap would hang the IF8, > if12 keep alive .... if skip this sql , I can run 1000 such esql/c ap > without problems !! > > I can use two connections in esql/c ap ,this might solve my problem , > also I can move sit2:quote from if12 to IF8 , it works ,too . > I wonder why this cause the problem I encountered ?! > > Any suggestions are welcome !!~~ > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --001517473434ecc4b70476e19430

Thanks to Art .... I just reinstatement from 7.3 to 11.5 and purchase new IDS 11.5 recently , although I know IDS 7.X is out of support , but I need time to test the system and then upgrade IDS to 11.5 , I called vendor if I can still get the IDS 7.X version which fixed this bug , he said can not found a fixed 7.X version for this bug ... So I should consider IDS 11.5 seriously now ~~