Details
-
Bug
-
Resolution: Unresolved
-
Major
-
7.2.5, 7.6.1
-
Untriaged
-
0
-
Unknown
Description
Related to this AV-77598. In Capella App Services tests, we try to deploy 500 App Services(SGW) collections. Each SGW collection creates 5 indexes, so we want to create 2500 indexes in total.
Looking at this comment, we can see that SGW timed out while waiting for an index to be created. These indexes normally take 1 minute or less to be ready, while SGW timeout is 15 minutes, and the slowest indexes can take 1 hour or more until they are ready.
We only noticed this issue when running on capella, both on 7.2.5 and 7.6.1. However, it is intermittent, and it only happens in tests where we want to deploy 500 SGW collections (it happens in about a third of the runs), while it didn't occur in any run with 250 collections or less. The cluster in which the issue occured cannot be accessed anymore, so we can't obtain the server logs from it. However, I am currently rerunning the test and will add the logs here as soon as we hit the issue again. In the meantime, any feeback would be greatly appreciated. This is a run where the issue happened again: https://perf.jenkins.couchbase.com/job/SGW-Capella-2/81/console. This is the capella server cluster: perf-cluster-c05790
Attachments
Issue Links
- relates to
-
AV-77598 Loading...