[SERVER-38870] IndexCatalogImpl::createIndexOnEmptyCollection() doesn't roll back properly Created: 07/Jan/19 Updated: 27/Oct/23 Resolved: 17/Apr/20 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Storage |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Major - P3 |
| Reporter: | Gregory Wlodarek | Assignee: | Maria van Keulen |
| Resolution: | Gone away | Votes: | 0 |
| Labels: | neweng | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Operating System: | ALL | ||||||||
| Sprint: | Execution Team 2020-04-20, Execution Team 2020-05-04 | ||||||||
| Participants: | |||||||||
| Description |
|
If we call createIndexOnEmptyCollection() and roll back by not committing the WriteUnitOfWork that it was called in, then the second time we try to call this function, we hit an invariant:
The following pseudocode can be used to reproduce this:
|
| Comments |
| Comment by Maria van Keulen [ 17/Apr/20 ] |
|
I added code similar to the pseudocode under a failpoint and did not reproduce the failure. gregory.wlodarek also wrote a unit test similar to the pseudocode, which passes, so it seems like this is no longer an issue. Per a discussion with Gregory, the issue in this ticket most likely came from a unit test that wasn't committed. I don't think it's necessary to add a regression test for the issue at this time, so I'm closing this ticket as Gone Away. |