[openstack-dev] [neutron-lbaas][octavia]Octavia request poll interval not respected

Michael Johnson johnsomor at gmail.com
Thu Feb 1 18:18:31 UTC 2018


Hi Mihaela,

The polling logic that the neutron-lbaas octavia driver uses to update
the neutron database is as follows:

Once a Create/Update/Delete action is executed against a load balancer
using the Octavia driver a polling thread is created.
On every request_poll_interval the thread queries the Octavia v1 API
to check the status of the object modified.
It will save the updated state in the neutron databse and exit if the
objects provisioning status becomes on of: "ACTIVE", "DELETED", or
"ERROR".
It will repeat this polling until one of those provisioning statuses
is met, or the request_poll_timeout is exceeded.

My suspicion is the GET requests you are seeing for those objects is
occurring from another source.
You can test this by running neutron-lbaas in debug mode.  I will then
log a debug message for every polling interval.

The code for this thread is located here:
https://github.com/openstack/neutron-lbaas/blob/stable/ocata/neutron_lbaas/drivers/octavia/driver.py#L66

Michael



More information about the OpenStack-dev mailing list