[U2] Help -- UD System hang...
I recently upgraded a client to UniData 6.1.12 running on SCO OpenServer 5.0.7. The installation (fresh installs of both) went smoothly, and everything seemed fine after the upgrade. This morning, however, once all of the users logged in, they are intermittently experiencing hangs (i.e. no response from their terminals). Unix is still up and running fine, but the Unidata sessions can't be killed using stopudt pid. They CAN be killed using kill -9, but a 'listuser' shows them as still logged in to Unidata. The hang affects ALL Unidata users. If I attempt to begin a new Unidata session, it starts, but immediately hangs. I'm not seeing any obvious problems in the Unidata log files. The hang may or may not be related to a couple of cron jobs that we have running once per hour. After the last two, I noticed that there were PHANTOM users in the listuser output. There were no errors in the _PH_ output files or the log files that the PHANTOM jobs created. All of the kernel tunable parameters are set at or above recommended settings (I can provide the config settings if needed). Any ideas? Larry Hiscock Western Computer Services http://www.wcs-corp.com --- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
Re: [U2] Help -- UD System hang...
ran out of message headers? kernel MSGTQL $UDTBIN/ipcstat -qonwill just show message queues with non-zero CBYTES or QNUM or 'ipcs -qop' - look at QNUM total and compare to kernel MSGTQL Wally Terhune U2 Support Architect IBM Information Management 4700 South Syracuse Street, Denver, CO 80237 Tel: 303.773.7969 Fax: 303.773.5915 [EMAIL PROTECTED] Larry Hiscock [EMAIL PROTECTED] com To Sent by: u2-users@listserver.u2ug.org [EMAIL PROTECTED] cc stserver.u2ug.org Subject [U2] Help -- UD System hang... 10/23/2006 10:20 AM Please respond to [EMAIL PROTECTED] er.u2ug.org I recently upgraded a client to UniData 6.1.12 running on SCO OpenServer 5.0.7. The installation (fresh installs of both) went smoothly, and everything seemed fine after the upgrade. This morning, however, once all of the users logged in, they are intermittently experiencing hangs (i.e. no response from their terminals). Unix is still up and running fine, but the Unidata sessions can't be killed using stopudt pid. They CAN be killed using kill -9, but a 'listuser' shows them as still logged in to Unidata. The hang affects ALL Unidata users. If I attempt to begin a new Unidata session, it starts, but immediately hangs. I'm not seeing any obvious problems in the Unidata log files. The hang may or may not be related to a couple of cron jobs that we have running once per hour. After the last two, I noticed that there were PHANTOM users in the listuser output. There were no errors in the _PH_ output files or the log files that the PHANTOM jobs created. All of the kernel tunable parameters are set at or above recommended settings (I can provide the config settings if needed). Any ideas? Larry Hiscock Western Computer Services http://www.wcs-corp.com --- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/ [demime 1.01d removed an attachment of type image/gif which had a name of graycol.gif] [demime 1.01d removed an attachment of type image/gif which had a name of pic11604.gif] [demime 1.01d removed an attachment of type image/gif which had a name of ecblank.gif] --- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
RE: [U2] Help -- UD System hang...
Hi Wally, Thanks for the suggestion. Here's what I've currently got configured: MSGTQL is set to 1024. ipcstat -qon shows nothing ipcs -qop shows 5 queues (0-4) all with 0 in the CBYTES and QNUM columns I ran udtconf allowed it to autoconf, then rebooted the system. So far (knock on wood) we haven't had another hang, BUT I have all of the cron jobs that start Unidata phantoms temporarily disabled. Larry Hiscock Western Computer Services -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Wally Terhune Sent: Monday, October 23, 2006 10:11 AM To: u2-users@listserver.u2ug.org Subject: Re: [U2] Help -- UD System hang... ran out of message headers? kernel MSGTQL $UDTBIN/ipcstat -qonwill just show message queues with non-zero CBYTES or QNUM or 'ipcs -qop' - look at QNUM total and compare to kernel MSGTQL Wally Terhune U2 Support Architect IBM Information Management 4700 South Syracuse Street, Denver, CO 80237 Tel: 303.773.7969 Fax: 303.773.5915 [EMAIL PROTECTED] Larry Hiscock [EMAIL PROTECTED] com To Sent by: u2-users@listserver.u2ug.org [EMAIL PROTECTED] cc stserver.u2ug.org Subject [U2] Help -- UD System hang... 10/23/2006 10:20 AM Please respond to [EMAIL PROTECTED] er.u2ug.org I recently upgraded a client to UniData 6.1.12 running on SCO OpenServer 5.0.7. The installation (fresh installs of both) went smoothly, and everything seemed fine after the upgrade. This morning, however, once all of the users logged in, they are intermittently experiencing hangs (i.e. no response from their terminals). Unix is still up and running fine, but the Unidata sessions can't be killed using stopudt pid. They CAN be killed using kill -9, but a 'listuser' shows them as still logged in to Unidata. The hang affects ALL Unidata users. If I attempt to begin a new Unidata session, it starts, but immediately hangs. I'm not seeing any obvious problems in the Unidata log files. The hang may or may not be related to a couple of cron jobs that we have running once per hour. After the last two, I noticed that there were PHANTOM users in the listuser output. There were no errors in the _PH_ output files or the log files that the PHANTOM jobs created. All of the kernel tunable parameters are set at or above recommended settings (I can provide the config settings if needed). Any ideas? Larry Hiscock Western Computer Services http://www.wcs-corp.com --- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/ [demime 1.01d removed an attachment of type image/gif which had a name of graycol.gif] [demime 1.01d removed an attachment of type image/gif which had a name of pic11604.gif] [demime 1.01d removed an attachment of type image/gif which had a name of ecblank.gif] --- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/ --- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
RE: [U2] Help -- UD System hang...
So - generate a udtdiag dump when the problem is occuring and open a support case with your UniData support provider. (would be my next suggestion) 1) Log in to your system as 'root' 2) In the UNIX shell, cd to a directory that has as much as 200MB of available disk space. Note: For systems without the Recoverable File System (RFS) enabled, you may only need 50MB or so. The size of the dump depends on udtdconfig settings and the size of your logs and errlogs in $UDTBIN. 3) Decide on a name for the dump directory you want to create. This directory must not already exist in the directory you are now positioned in. Folks often construct a unique name involving the current date (e.g. udtdiagdump_080105). If you already have an IBM U2 Support case opened, you may want to use the case number. The choice is yours. The rest of the instructions use this example dump directory name. 4) Run the udtdiag script (using the example above): $UDTBIN/udtdiag udtdiagdump_080105 The script runs a series of commands that copy logs and capture the current state of UniData and your system environment. 5) Use the UNIX tar command to make a relative path tar image of the dump directory. For example: tar cvf udtdiagdump_080105.tar udtdiagdump_080105 6) Compress the tar image: compress udtdiagdump_080105.tar 7) Use ftp it to send this compressed tar image to IBM U2 Support: ftp testcase.boulder.ibm.com login as anonymous, use your email address as password cd /software/toibm/u2 bin put udtdiagdump_080105.tar.Z quit Note: If you move the compressed tar file to another system in order to access testcase.boulder.ibm.com, be sure to treat it as a binary file at every step. 8) Contact the appropriate IBM U2 support office (for your geography) where you have opened a support case and advise them of the name of the compressed tar file you just posted. Note: If you are creating a diagnostic dump because you are having a problem with UniData Replication, please create dumps on all systems involved in your replication setup - the publishing system and the subscribing system(s). Wally Terhune U2 Support Architect IBM Information Management 4700 South Syracuse Street, Denver, CO 80237 Tel: 303.773.7969 Fax: 303.773.5915 [EMAIL PROTECTED] Larry Hiscock [EMAIL PROTECTED] com To Sent by: u2-users@listserver.u2ug.org [EMAIL PROTECTED] cc stserver.u2ug.org Subject RE: [U2] Help -- UD System hang... 10/23/2006 11:56 AM Please respond to [EMAIL PROTECTED] er.u2ug.org Hi Wally, Thanks for the suggestion. Here's what I've currently got configured: MSGTQL is set to 1024. ipcstat -qon shows nothing ipcs -qop shows 5 queues (0-4) all with 0 in the CBYTES and QNUM columns I ran udtconf allowed it to autoconf, then rebooted the system. So far (knock on wood) we haven't had another hang, BUT I have all of the cron jobs that start Unidata phantoms temporarily disabled. Larry Hiscock Western Computer Services [demime 1.01d removed an attachment of type image/gif which had a name of graycol.gif] [demime 1.01d removed an attachment of type image/gif which had a name of pic25008.gif] [demime 1.01d removed an attachment of type image/gif which had a name of ecblank.gif] --- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/