Make dhcp agent not recycle ports in binding_failed state
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
neutron |
Fix Released
|
Medium
|
Rossella Sblendido |
Bug Description
We just happened to have broken dhcp because the dhcp ports were in binding_failed state.
We tried to disable/enable dhcp again, but apparently (not 100% sure, unfortunately), disabling didn't remove the ports (assuming that it was because they were in binding_failed state) and re-enabling just ended up with ports in binding_failed states again (likely the same ports because dhcp agent tries to recycle ports whenever possible). Unfortunately, I don't have the logs / traces of this.
So disabling / re-enabling didn't fix anything while, I believe, users would try that first to fix the situation. If we could make that just work, then that would make it easier for people to get out of this "no dhcp" problem when ports are failing to bind.
In the end, we had to remove the ports to have new ports created when dhcp was disable / re-enabled.
Changed in neutron: | |
assignee: | nobody → Rossella Sblendido (rossella-o) |
tags: | added: dhc |
tags: |
added: l3-ipam-dhcp removed: dhc |
Changed in neutron: | |
importance: | Undecided → Medium |
Changed in neutron: | |
status: | Fix Committed → Fix Released |
Would https:/ /review. openstack. org/#/c/ 162260/ solve this?