On TOR scale setup, config processes take about an hour to get to stable state
Bug #1436230 reported by
Vedamurthy Joshi
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
Trunk |
Fix Committed
|
Medium
|
Hampapur Ajay |
Bug Description
R2.1 Build 43 Ubuntu 14.04 Multi-node setup
In this TOR scale setup, we have about 64K VMIs, 150 TORs, 1K Vmis per tor.
It is seen that on api-server restart or reinit, the config node takes about 30+ mins to get to a stable state(most notably ifmap-server, api-server and schema-transformer)
A majority of the time seems to be due to db_sync from cassandra to api-server.
While this is necessary, the delay seems to be too high in a production environment
Ajay, Prakash are aware of this issue.
To post a comment you must log in.
Fixed with control-node moving to rabbitmq instead of ifmap