VRM Service is Stopped - Replication Appliance Root Partition Full

[SJDewar]

Hi,

We've recently had an upgrade to vsphere 5.5 and a replication appliance setup (appliance version: 5.8.0.1).

Everything was okay to start with but now the VRM service is stopped and won't start.

On investigation we found that the root partition of the appliance was full.  We tracked this down to the a directory called 16384 in the /var/lib/vrmsdb/base directory.

I'm assuming this is why the VRM service won't start.

Can you tell me what what's filling up this directory and can the size of this directory be controlled or do we just have to add more disk space?

Thanks in advance.

S.

[mvalkanov · Accepted solution]

Hi,

Please open a SR to get assistance with adding additional disk space and moving the embedded vPostgres DB to a partition with additional space.

Once the DB is up and running, the DB tables can be analyzed to see the root cause for the full DB.

In hms-configuration.xml file there is a setting hms-eventlog-maxage that can be lowered (for example to 604800, which is 7 days), in order to limit old data kept in IncomingEventLogEntity and OutgoingEventLogEntity tables. Without dump of the DB it would be hard to know what is the cause of the full DB, but there have been reported issues with highly loaded environment and occasional VR server disconnects that have led to large number of notifications between VRM servers and that filled the IncomingEventLogEntity and OutgoingEventLogEntity tables.

Regards,

Martin

[SJDewar]

Thanks for the quick reply and advice Martin, much appreciated, cheers.