We recently just upgraded to Juno from Icehouse. We used the upgrade_levels feature to get Nova off the ground, but ran into major issues around conductor being too slow to respond to messages and would cause a lot of our instances to not create. This would only happen at a particular high amount of load. A few vms being built were no problem, but get into anything over say 15, then the conductor queue would start showing a lot of unacknowledged messages and basically start causing a lot of instance creation failures.
I'm happy to provide any other details that are needed. --Joe On Wed, Sep 9, 2015 at 12:53 PM, Barrett, Carol L <carol.l.barr...@intel.com > wrote: > Ops – I’m looking for an example of the RPC compatibility issue that was > discussed in Palo Alto. Can someone help me out? > Thanks > Carol > > > _______________________________________________ > OpenStack-operators mailing list > OpenStack-operators@lists.openstack.org > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators > >
_______________________________________________ OpenStack-operators mailing list OpenStack-operators@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
