[SERVER-47284] Log oplog entries from all nodes after a test failure Created: 02/Apr/20 Updated: 06/Dec/22 Resolved: 21/Apr/20 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Replication, Testing Infrastructure |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Samyukta Lanka | Assignee: | Backlog - Server Tooling and Methods (STM) (Inactive) |
| Resolution: | Won't Fix | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Assigned Teams: |
Server Tooling & Methods
|
| Participants: |
| Description |
|
There are many test failures where it would help the replication team to diagnose what is going on if we could see the contents of each node's oplog. One options is that we could log the last 100 oplog entries for each node when a test fails, similar to what happens with a dbhash mismatch. |
| Comments |
| Comment by Samyukta Lanka [ 21/Apr/20 ] |
|
I don't have specific BFs in mind at the moment. I'm going to close this ticket for now, but I'll reopen it when we come across a BF where having access to the contents of the oplog could have helped us diagnose it. |
| Comment by Brooke Miller [ 14/Apr/20 ] |
|
Hey samy.lanka, any update on the above question? Are there any BFs that we can add to this to show as an example so that we can investigate this? |
| Comment by Max Hirschhorn [ 02/Apr/20 ] |
samy.lanka, are these test failures concentrated in particular test files or test suites where it'd be possible to enable capturing data files? Trying to run more JavaScript code within a mongo shell which spawned a replica set using ReplSetTest seems cumbersome given that a test failure is by definition an uncaught JavaScript exception. |