Comment 3 for bug 1796976

Hi,
Just found out that the docs are only partially corrected.
I was setting up the compute node following the self-service network model.
When trying to lauch an instance in the provider network it remained in build status until timeout 300 seconds was reached and then went to error status.
Digging in the logs i found that this was happening because nova never received notification that the virtual interface was plugged in with no apparent cause.
I also found in the neutron logs the same depicted before in this bug poiting to a missing lock_path parameter.
As set ip up as in the controller node and then the instance started with no issues.
Looking into the documentation web pages i found no all of them are updated. For example:
- https://docs.openstack.org/neutron/latest/install/compute-install-rdo.html#configure-the-common-component properly states the need for setting up the lock_path variable in the [oslo_concurrency] section
- https://docs.openstack.org/neutron/latest/install/compute-install-ubuntu.html which is the sama page but for ubuntu install still lacks that info...which leads to the behavior described above.
Could you please update it?

Thanks a lot in advance.