Hello,

me again, sorry... We're looking into https://jira.opendaylight.org/browse/
NETVIRT-985, and it reminds me of the https://jira.opendaylight.
org/browse/CONTROLLER-1756 from two months ago:

Again seeing a HUGE ShardDataTree, and a lot of messages in the log which
seem to indicate something may be wrong re. re. transaction chain usage,
again from openflowplugin like before, but with slightly different messages
(likely because the fix for the first OOM two months ago changed the
related code?). Do any openflowplugin experts know what they would have to
do to fix this?

Or is a huge ShardDataTree just perfectly normal, because that's the
in-memory database? Must we just bump up our Xmx of 2 GB to much more for
this to be able to work at scale? (But then why was it not normal 2 months
ago, and whatever openflowplugin did at the time in
https://jira.opendaylight.org/browse/OPNFLWPLUG-933 significantly reduced
the size of the ShardDataTree there?)

Thanks a lot for any insight, proposed next steps, workarounds and the like!

M.
--
Michael Vorburger, Red Hat
vorbur...@redhat.com | IRC: vorburger @freenode | ~ = http://vorburger.ch
_______________________________________________
controller-dev mailing list
controller-dev@lists.opendaylight.org
https://lists.opendaylight.org/mailman/listinfo/controller-dev

Reply via email to