neutron api_workers/rpc_workers

[Matt Thompson]

When running tempest on my multi-node cloud servers install running master, I'm hitting the following issue:

https://bugs.launchpad.net/neutron/+bug/1382064

When I lower neutron's api_workers/rpc_workers to 0 and restart neutron-server the failing tests all pass.

The neutron defaults for api_workers/rpc_workers is 0 -- should we temporarily lower these values in the neutron.conf template until this bug is fixed?

Further, we probably do need to make variables for these two settings so they can be overriden. Also, the defaults should probably be automatically determined based on CPU count.

Does anyone, perhaps QE specifically, have any input here?

--Matt

[Matt Thompson (mattt416)]

We could probably do w/ scanning all service templates to see if workers, etc. are hard-coded and create variables w/ calculated defaults if they are.