[SERVER-23602] Resmoke should delete all job directories regardless of -j setting Created: 07/Apr/16 Updated: 06/Dec/22 Resolved: 19/Nov/21 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Testing Infrastructure |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Mathias Stearn | Assignee: | Backlog - Server Tooling and Methods (STM) (Inactive) |
| Resolution: | Won't Fix | Votes: | 0 |
| Labels: | stm, tig-resmoke | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Assigned Teams: |
Server Tooling & Methods
|
| Participants: |
| Description |
|
Currently it only deletes the job directories that it will use in its current run. This has the unfortunate effect that if you run out of disk space from using too high of a -j setting, using a lower -j, even 1, won't solve the problem. Today, this requires manually cleaning up the directories. This may be simpler if resmoke always used a single subdirectory in --dbpathPrefix, or if the default was something like /data/db/resmoke/. |
| Comments |
| Comment by Brooke Miller [ 19/Nov/21 ] |
|
There hasn't been a recent ask for this, so we're closing as won't fix. |
| Comment by Max Hirschhorn [ 26/Jan/18 ] |
|
I think there are cases (e.g. for the fuzzer) where it may be useful to run a sequence of tests and preserve the dbpath so that it is possible to later run a single test starting from pre-existing dbpath and reproduce a failure. Whenever we do this ticket, we should ensure there is a way to prevent the data files from being cleaned up after the test succeeds. It may be enough to say if you're running with --jobs=1 and didn't specify --continueOnFailure, then the data directory is left intact. |