HPL failpoint: 15902 while attempting hplunload

Hi, We do nightly unloads via hplunload. In the script that runs, we drop yesterday's project, then create the project and jobs for each table. This works fine the majority of the time, but sometimes we get the following error (which repeats for each table): Failed to determine shared memory base address. Set PLOAD_SHMBASE to appropriate value for PLOAD to work. HPL failpoint: 15902 After the unloads, we also do a level 0 backup, which also fails: /usr/informix/bin/ontape -v -s -L 0 -d -t STDIO 1> /dbbackup/backupdb/ontape/fullsize/backup.l0.new 23:56:56 shmat: [22]: operating system error 23:56:56 Client could not attach server shared memory segment, use IFX_XFER_SHMBASE. Server is in an incompatible state or user authentication failed. The failures happen about once a week. I have neither PLOAD_SHMBASE nor IFX_XFER_SHMBASE in my onconfig or environment variables, so I'm confused as to why it would complain about this sometimes but not others. For reference, this database is running Informix 11.1 on AIX 5.3 Thanks in advance, -Justin

I suggest performing an onstat -g seg to see how many virtual (V) memory segments you have each day and how many you have when the issue occurs. I suspect you will find multiple segments on the latter and perhaps the other times as well. If multiple segments exist, you may want to increase your SHMVIRTSIZE to encompass these segments and/or increase SHMADD to make obtained segments larger. You also might want to schedule several onmode -F commands during non peak times to release no longer needed segments. Lastly, I would, at a minimum, upgrade to the latest version of your current engine. Hope this helps. Clifton On Thu, Jun 6, 2013 at 5:02 PM, Justin Killen < jkillen@allamericanasphalt.com> wrote: > Hi, > > We do nightly unloads via hplunload. In the script that runs, we drop > yesterday's project, then create the project and jobs for each table. This > works fine the majority of the time, but sometimes we get the following > error > (which repeats for each table): > > Failed to determine shared memory base address. > Set PLOAD_SHMBASE to appropriate value for PLOAD to work. > HPL failpoint: 15902 > > After the unloads, we also do a level 0 backup, which also fails: > > /usr/informix/bin/ontape -v -s -L 0 -d -t STDIO 1> > /dbbackup/backupdb/ontape/fullsize/backup.l0.new > 23:56:56 shmat: [22]: operating system error > 23:56:56 Client could not attach server shared memory segment, use > IFX_XFER_SHMBASE. > Server is in an incompatible state or user authentication failed. > > The failures happen about once a week. I have neither PLOAD_SHMBASE nor > IFX_XFER_SHMBASE in my onconfig or environment variables, so I'm confused > as > to why it would complain about this sometimes but not others. > > For reference, this database is running Informix 11.1 on AIX 5.3 > > Thanks in advance, > -Justin > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Take care. Clifton This communication may contain privileged or other confidential information. If you have received it in error, please advise the sender by reply email and immediately delete the message and any attachments without copying or disclosing the contents. Thank you. Consider the environment. Please don't print this e-mail unless you need to. --001a11c25b1addb63f04de7f20e1

Hello. You should always use PLOAD_SHMBASE, as recommended in machine notes, for using HPL. (It might be trying to access some memory range that the engine is using). For your backup issue, check this technote: http://www-01.ibm.com/support/docview.wss?uid=swg21623474 Regards. Alexandre Marini IBM Informix Certified Professional v10 / v11.50 / v11.70 IBM Information Management Informix Technical Professional IBM Infosphere DataStage Technical Professional Informix Senior DBA - Orizon Brasil BRIUG website administrator Informix independent consultant > To: ids@iiug.org > From: jkillen@allamericanasphalt.com > Subject: HPL failpoint: 15902 while attempting hplunload [30443] > Date: Thu, 6 Jun 2013 12:02:05 -0400 > > Hi, > > We do nightly unloads via hplunload. In the script that runs, we drop > yesterday's project, then create the project and jobs for each table. This > works fine the majority of the time, but sometimes we get the following error > (which repeats for each table): > > Failed to determine shared memory base address. > Set PLOAD_SHMBASE to appropriate value for PLOAD to work. > HPL failpoint: 15902 > > After the unloads, we also do a level 0 backup, which also fails: > > /usr/informix/bin/ontape -v -s -L 0 -d -t STDIO 1> > /dbbackup/backupdb/ontape/fullsize/backup.l0.new > 23:56:56 shmat: [22]: operating system error > 23:56:56 Client could not attach server shared memory segment, use > IFX_XFER_SHMBASE. > Server is in an incompatible state or user authentication failed. > > The failures happen about once a week. I have neither PLOAD_SHMBASE nor > IFX_XFER_SHMBASE in my onconfig or environment variables, so I'm confused as > to why it would complain about this sometimes but not others. > > For reference, this database is running Informix 11.1 on AIX 5.3 > > Thanks in advance, > -Justin > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >

Thanks - I'll add the onstat -g seg to the backup script and start tracking the segments. Hopefully after a few weeks I can confirm the correlation between the segment usage and the failures, then be able to take corrective action. An upgrade to 12.1 is in the works already for later this year (migrating over to new hardware, etc), we're just trying to keep the old system up and running until migration day. -Justin -----Original Message----- From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of Clifton Bean Sent: Thursday, June 06, 2013 9:52 AM To: ids@iiug.org Subject: Re: HPL failpoint: 15902 while attempting hplu.... [30444] I suggest performing an onstat -g seg to see how many virtual (V) memory segments you have each day and how many you have when the issue occurs. I suspect you will find multiple segments on the latter and perhaps the other times as well. If multiple segments exist, you may want to increase your SHMVIRTSIZE to encompass these segments and/or increase SHMADD to make obtained segments larger. You also might want to schedule several onmode -F commands during non peak times to release no longer needed segments. Lastly, I would, at a minimum, upgrade to the latest version of your current engine. Hope this helps. Clifton On Thu, Jun 6, 2013 at 5:02 PM, Justin Killen < jkillen@allamericanasphalt.com> wrote: > Hi, > > We do nightly unloads via hplunload. In the script that runs, we drop > yesterday's project, then create the project and jobs for each table. This > works fine the majority of the time, but sometimes we get the following > error > (which repeats for each table): > > Failed to determine shared memory base address. > Set PLOAD_SHMBASE to appropriate value for PLOAD to work. > HPL failpoint: 15902 > > After the unloads, we also do a level 0 backup, which also fails: > > /usr/informix/bin/ontape -v -s -L 0 -d -t STDIO 1> > /dbbackup/backupdb/ontape/fullsize/backup.l0.new > 23:56:56 shmat: [22]: operating system error > 23:56:56 Client could not attach server shared memory segment, use > IFX_XFER_SHMBASE. > Server is in an incompatible state or user authentication failed. > > The failures happen about once a week. I have neither PLOAD_SHMBASE nor > IFX_XFER_SHMBASE in my onconfig or environment variables, so I'm confused > as > to why it would complain about this sometimes but not others. > > For reference, this database is running Informix 11.1 on AIX 5.3 > > Thanks in advance, > -Justin > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Take care. Clifton This communication may contain privileged or other confidential information. If you have received it in error, please advise the sender by reply email and immediately delete the message and any attachments without copying or disclosing the contents. Thank you. Consider the environment. Please don't print this e-mail unless you need to. --001a11c25b1addb63f04de7f20e1 ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.

Okay, so it finally happened again. No luck on the 'onstat -g seg' - here is the output: >/usr/informix/bin/onstat -g seg 22:18:03 shmat: [22]: operating system error 22:18:03 Client could not attach server shared memory segment, use IFX_XFER_SHMBASE. -Justin -----Original Message----- From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of Justin Killen Sent: Thursday, June 06, 2013 1:51 PM To: ids@iiug.org Subject: RE: HPL failpoint: 15902 while attempting hplu.... [30446] Thanks - I'll add the onstat -g seg to the backup script and start tracking the segments. Hopefully after a few weeks I can confirm the correlation between the segment usage and the failures, then be able to take corrective action. An upgrade to 12.1 is in the works already for later this year (migrating over to new hardware, etc), we're just trying to keep the old system up and running until migration day. -Justin -----Original Message----- From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of Clifton Bean Sent: Thursday, June 06, 2013 9:52 AM To: ids@iiug.org Subject: Re: HPL failpoint: 15902 while attempting hplu.... [30444] I suggest performing an onstat -g seg to see how many virtual (V) memory segments you have each day and how many you have when the issue occurs. I suspect you will find multiple segments on the latter and perhaps the other times as well. If multiple segments exist, you may want to increase your SHMVIRTSIZE to encompass these segments and/or increase SHMADD to make obtained segments larger. You also might want to schedule several onmode -F commands during non peak times to release no longer needed segments. Lastly, I would, at a minimum, upgrade to the latest version of your current engine. Hope this helps. Clifton On Thu, Jun 6, 2013 at 5:02 PM, Justin Killen < jkillen@allamericanasphalt.com> wrote: > Hi, > > We do nightly unloads via hplunload. In the script that runs, we drop > yesterday's project, then create the project and jobs for each table. This > works fine the majority of the time, but sometimes we get the following > error > (which repeats for each table): > > Failed to determine shared memory base address. > Set PLOAD_SHMBASE to appropriate value for PLOAD to work. > HPL failpoint: 15902 > > After the unloads, we also do a level 0 backup, which also fails: > > /usr/informix/bin/ontape -v -s -L 0 -d -t STDIO 1> > /dbbackup/backupdb/ontape/fullsize/backup.l0.new > 23:56:56 shmat: [22]: operating system error > 23:56:56 Client could not attach server shared memory segment, use > IFX_XFER_SHMBASE. > Server is in an incompatible state or user authentication failed. > > The failures happen about once a week. I have neither PLOAD_SHMBASE nor > IFX_XFER_SHMBASE in my onconfig or environment variables, so I'm confused > as > to why it would complain about this sometimes but not others. > > For reference, this database is running Informix 11.1 on AIX 5.3 > > Thanks in advance, > -Justin > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Take care. Clifton This communication may contain privileged or other confidential information. If you have received it in error, please advise the sender by reply email and immediately delete the message and any attachments without copying or disclosing the contents. Thank you. Consider the environment. Please don't print this e-mail unless you need to. --001a11c25b1addb63f04de7f20e1 ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum. ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.