Process Error 131:Process is out of memory

[balakumar86]

Hi,

Currently some of the index engines are returing this error and we are unable to run indexing. I have seen in a discussion in KC that it's due to the value set in otadmin.cfg, i.e. the amount of memory that UD is allowed to use. Anyone having any ideas?

Regards,
Bala

 

[ggriffit]

Bala,
This is usually due to the amount of RAM required by Livelink and the Partitions, if you have LES 971 you can amend the amount of Meta Data allocated to RAM versus Disk. As you mentioned you can also update the otadmin.cfg file to amend the amount of RAM allocated to Java.

Greg Griffiths
Livelink Certified Developer & ECM Global Star Champion 2005 & 2006

http://www.greggriffiths.org/livelink/

 

[LiveLinkPRO]

No - don't mess around with the otadmin.cfg unless you know what you're doing, you can't just increase the memory values and hope it will work. Especially not on Windows due to the Java Memory limitations.

Make sure your server is not running out of physical memory (figure 2.4 GB per 50 GB partition plus OS plus whatnot) and that not being the case, check why the error comes up. In particular, check the checkpoint file for each partition and if the metadata memory settings have been manipulated.

Jack

 

[ggriffit]

I agree with LivelinkPro's comments about editing any configuration files without knowing what you were doing etc, I mentioned it because I've done it on a system to get around the issue and the posts in the KC that are referenced in the original post all indicate the same fix.

The issue is usually with the amount of Physical RAM available, as both LivelinkPro and I mentioned, so either adding more RAM, spreading the load across more than one Admin server or moving some metadata to disk will improve performance.

Greg Griffiths
Livelink Certified Developer & ECM Global Star Champion 2005 & 2006

http://www.greggriffiths.org/livelink/

 

[balakumar86]

The issue is now solved.

I have found one entry in the message file which was too big and occupying 3/4 of the whole message file. I've removed that entry from the message file and restarted indexing. Unluckily, there was some lock by the windows explorer on the message file and it returned Process Error 174. I've closed all explorer process from task manager and tried again. But, got the same error. Then I've cleared all pending messages from the dataflow, resynchronized and restarted indexing. Now, Indexing is running fine and no issues.

Now, My question is, will the docs that were cleared off the dataflow get indexed or not. I think, only if the doc gets edited, an entry will be added to dtreenotify and will come in the indexing processing. otherwise the document will not be searchable. If I have to add these docs manually in the indexing process, then please help me with that.

Thanks for all your help.

 

[ggriffit]

if you are having those issues then you MUST call OT support who will be able to review the data and provide you with methods to reindex any content you remove etc.

Greg Griffiths
Livelink Certified Developer & ECM Global Star Champion 2005 & 2006

http://www.greggriffiths.org/livelink/

 

[LiveLinkPRO]

If you hack stuff out of the dataflow, you will have to reindex the deleted objects manually (one ipool message contains up to 500 objects).

This can be done by inserting all objects into the dtreenotify that match a certain time frame (where ModifyDate).

Consult OT Support to walk you through this exercise, it is not a big deal.

Also, as you already noticed, do not browse the dataflow as the processes are running. This will cause all sorts of process errors due to file locks. Suspend the dataflow first.

Jack