Undercloud sometimes OOM kills things in CI
Bug #1536136 reported by
Steven Hardy
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
Invalid
|
Undecided
|
Unassigned |
Bug Description
See https:/
In this case we ran out of memory, and killed a heat-engine worker, with the predictable result that the CI test failed.
We need a better solution, even slowing things down a little by configuring swap is better than randomly killing stuff by default.
It's probably not what developers want to deal with from the default VM setup either.
To post a comment you must log in.
Note I think this is more likely to happen now we increased the CI vCPU count, as most projects use oslo processutils, which defaults to launching a worker process per core, thus using more memory.
Perhaps we need to increase the CI memory allowed a little more