[SERVER-30008] Core files not being archived Created: 06/Jul/17  Updated: 18/Jul/17  Resolved: 18/Jul/17

Status: Closed
Project: Core Server
Component/s: Testing Infrastructure
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Jonathan Abrahams Assignee: DO NOT USE - Backlog - Test Infrastructure Group (TIG)
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Duplicate
Related
Operating System: ALL
Participants:
Linked BF Score: 0

 Description   

If the core file that's being moved already resides in the directory, this can cause an errexit:

[2017/07/05 17:58:52.110] mv: cannot stat `../src/dump_mongo.25629.core': No such file or directory
[2017/07/05 17:58:52.111] Command failed: exit status 1



 Comments   
Comment by Max Hirschhorn [ 18/Jul/17 ]

Core dumps are being archived in Evergreen again, so I'm closing this ticket as a duplicate of EVG-1828.

See this task as an example of where there's a "Core Dumps - Execution 0" tarball since a test intentionally aborts.

Comment by Max Hirschhorn [ 09/Jul/17 ]

Does "post" step enable errexit?

No, the Evergreen agent runs the "post" phase as a list of commands the same it does for the task definition; it doesn't implicitly enable errexit. Could it just be that the for-loop ended and the script exited with the last mv statement's exit code since dump_mongo.25629.core was the only core file seen to exist because the others had already been deleted?

Comment by Jonathan Abrahams [ 07/Jul/17 ]

There shouldn't be an error exit either. Does "post" step enable errexit?

Comment by Max Hirschhorn [ 07/Jul/17 ]

sam.kleinman, could this be a result of EVG-1828 where the core files are being deleted before they can be packaged up during the "post" phase and uploaded to S3?

Generated at Thu Feb 08 04:22:24 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.