I'm attempting to move a 9.1E0426 system from a Solaris 10 1/06 OS to Solaris 10 10/09 and hit a shared memory issue.  The version of Solaris uses a new shared memory configuration but the the max-shm-memory parameter is set to 8GB.  The database servers will start with no errors after the server is booted.   However if the database servers are stopped and restarted  I start getting the following error:

09:05:05 BROKER   : No enough memory to lock shared memory segment in memory. (10015)
09:05:09 BROKER   : The data space of the process is not enough for the shm segment (1176)

If the database servers are stopped and started repeatly, the servers will fail to start at all:

09:10:12 BROKER   : No enough memory to lock shared memory segment in memory. (10015)
09:10:17 BROKER   : The data space of the process is not enough for the shm segment (1176)
09:10:17 BROKER   : Removed shared memory with segment_id: 33554472
09:10:17 BROKER   : Removed shared memory with segment_id: 33554456
09:10:17 BROKER   : Removed shared memory with segment_id: 33554457
09:10:17 BROKER   : Removed shared memory with segment_id: 33554458
09:10:17 BROKER   : Removed shared memory with segment_id: 33554459
09:10:17 BROKER   : Removed shared memory with segment_id: 33554460
09:10:17 BROKER   : Removed shared memory with segment_id: 33554461
09:10:17 BROKER   : Removed shared memory with segment_id: 33554462
09:10:17 BROKER   : Removed shared memory with segment_id: 33554463
09:10:17 BROKER   : Removed shared memory with segment_id: 33554465
09:10:17 BROKER   : Removed shared memory with segment_id: 33554466
09:10:17 BROKER   : ** This process terminated with exit code 2. (8619)

There is a Progress Knowledgebase entry for this issue (see below).

The patch 127127-11 referenced is integrated into this Solaris release so backing out the patch is not an option.  Has any fix or workaround been uncovered by anyone?  Please note that an upgrade to OE10 is not an option.

Regards,

Tom

ID: P137555
Title: "Getting error 1176 trying to restart database on Solaris and reboot is only way to get it to work"
Created: 11/20/2008Last Modified: 08/27/2009
Status: Verified


Symptoms: 
  • Getting error 1176 trying to restart database on Solaris and reboot is only way to get it to work
  • The data space of the process is not enough for the shm segment (1176)
  • Solaris patch 127127-11


Facts: 
  • OpenEdge 10.x
  • Solaris


Cause: 
  • Patch seems to be causing a problem


Fixes: 
Remove the Solaris Patch