snat health check generate too many api server updates
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
R2.20 |
Fix Committed
|
Critical
|
Rudra Rugge | |||
R2.20.x |
Fix Committed
|
Critical
|
Rudra Rugge | |||
R2.21.x |
Fix Committed
|
Critical
|
Rudra Rugge | |||
Trunk |
Fix Committed
|
Critical
|
Rudra Rugge |
Bug Description
SNAT and LBAAS are launched in a network namespace on a chosen vrouter. The vrouter selection happens based on the
UVE information from analytics. When service monitor detects that the vrouter is down then it tries to reschedule the
snat/lbaas netns to another vrouter.
This requires the virtual-machine placeholder object to be detached from the existing failed vrouter to a new vrouter. The
bug in the code misses the detach operation as information is not sent correctly to the api server. This causes a bunch of
messages/updates to the API server on a regular interval. The fix now correctly updates the virtual-machine and vrouter
references so that a new vrouter can be attached to the virtual-machine.
no longer affects: | juniperopenstack/trunk |
Changed in juniperopenstack: | |
milestone: | r2.20.1 → none |
no longer affects: | juniperopenstack/r3.0 |
information type: | Proprietary → Public |
Review in progress for https:/ /review. opencontrail. org/14217
Submitter: Hampapur Ajay (<email address hidden>)