Out of memory in Enq Repliction Server

[Former Member]

Hi Guys

We are facing issues related to the performace in High availability Portal system.

Windows, MSSQL, EP 700

Refer to the logs with Errors captured :

[Thr 348] EncIOWorker::Allocator (NIBUF-IN): failed to alloc 276 bytes from EnsMemObj (EnsMemObj: out of memory detected at ensutil.cpp:1012 Errorinfo: "EnsMemObj:getMemBlock")

[Thr 348] EncIOWorker::Allocator (NIBUF-IN): failed to alloc 75 bytes from EnsMemObj (EnsMemObj: out of memory detected at ensutil.cpp:1012 Errorinfo: "EnsMemObj:getMemBlock")

[Thr 348] EncIOWorker::Allocator (NIBUF-IN): failed to alloc 75 bytes from EnsMemObj

(EnsMemObj: out of memory detected at ensutil.cpp:1012 Errorinfo: "EnsMemObj:getMemBlock")

[Thr 8612] Tue Sep 15 08:30:41 2009

[Thr 8612] *** ERROR => IORequest::createNewResponseMem: caught exception

==>"EnsMemObj:getMemBlock"

while getting 54 bytes of new memory for the response area [enioreq.cpp 198]

[Thr 8612] *** ERROR => ReplicationThread::Loop(): exception caught!!!

==> "generic enqueue error detected at ensutil.cpp:1012 Errorinfo: "EnsMemObj:getMemBlock"" [eniothread.c 2120]

[Thr 8612] EncIOWorker::Allocator (NIBUF-IN): failed to alloc 91 bytes from EnsMemObj (EnsMemObj: out of memory detected at ensutil.cpp:1012 Errorinfo: "EnsMemObj:getMemBlock")

[Thr 8612] *** ERROR => ReplicationThread::Loop(): fatal error (read) on connection 000000000022B310 (hdl 5) [eniothread.c 1968]

JVM Settings for Dispatcher:

-XX:NewSize=85m

-XX:MaxNewSize=85m

-Xms256m

-verbose:gc

-Xloggc:GC.log

-Djava.security.policy=.\java.policy

-Djava.security.egd=file:/dev/urandom

-Djco.jarm=1

JVM Setting for Server node:

-Djco.jarm=1

-XX:MaxPermSize=512m

-XX:PermSize=512m

-Xms2048m

-XX:NewSize=340m

-XX:MaxNewSize=340m

-XX:+DisableExplicitGC

-verbose:gc

May I know which parameters we need to finetune.?

Regards

Rajesh

[former_member197561 · Accepted solution]

Hi Rajesh,

Please check SAP note 920979; this note refers to the stand-alone enqueue server and not specifically to replication, but the problem is the same.

The enqueue server is a kernel process and does not run under the java VM; the Java parameters are therefore not relevant in this situation.

Regards,

Mark

[Former Member]

Hi Rajesh,

Please check SAP note 920979; this note refers to the stand-alone enqueue server and not specifically to replication, but the problem is the same.

The enqueue server is a kernel process and does not run under the java VM; the Java parameters are therefore not relevant in this situation.

Regards,

Mark