The failures only happen in some runs. Is this still a blocker?

I believe this job should not be part of daily stable job, as it's been not stable job. Irrespective of that we need to fix this job, so marking this high. ariel.adam ecelgp Please let me know if you guys think otherwise.

I agree and job is already out of the stable list.

shague ecelgp guys, if we know this job is prone to fail let's remove it from the stable view for now and remove this bug.

Arunprakash any update on this ?

Initial analysis:

The failures might seem to be not related to bundle based resync.

In the failed TC,

a) static flows were added to the config DS

b) the presence of the same is verified in operational DS 

c) then the resync is triggered

But the TCs had failed sometimes at the b) step wherein they failed and didnt proceed to step c)

Also though, there is some wait and retries while checking operational DS while checking flows, in the failure cases there were no retries. 

The operational DS was completely empty in these instances so the reconciliation might nt ve completed. We might have to add a larger wait before checking operational DS?

The wait for the flow to show in operational is 30 sec which should be more than enough, I think the problem may be in a previous step like when pushing the group the flow depends on. Unfortunately the test is weak and it does not check the groups for example are actually pushed to the OVS before adding the flow, so I believe adding this check will mitigate the issue. I would recommend to neglect/disable the test until someone has cycles to rework it.

Lowering down priority as this is test issue not controller.