[SERVER-61432] symbolizer should not change global logging config Created: 12/Nov/21  Updated: 29/Oct/23  Resolved: 24/Oct/22

Status: Closed
Project: Core Server
Component/s: Testing Infrastructure
Affects Version/s: None
Fix Version/s: 6.2.0-rc0

Type: Bug Priority: Major - P3
Reporter: Robert Guo (Inactive) Assignee: Jasur Nurboev (Inactive)
Resolution: Fixed Votes: 0
Labels: evergreen-yaml
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
related to SERVER-71402 Hang analyzer continues to modify glo... Closed
Backwards Compatibility: Fully Compatible
Sprint: DAG 2022-10-31
Participants:
Story Points: 1

 Description   

Timebox: 1 day



 Comments   
Comment by Githook User [ 24/Oct/22 ]

Author:

{'name': 'jasurbeknurboyev', 'email': '998946972365'}

Message: SERVER-61432 symbolizer should not change global logging config
Branch: master
https://github.com/mongodb/mongo/commit/a80bb2bae4458cfb570b3fb48994bdaa7044cecc

Comment by Robert Guo (Inactive) [ 17/Oct/22 ]

iryna.zhuravlova@mongodb.com jasur.nurboev@mongodb.com yep. Here's the place that we're setting the global logging config. The productionized approach is to create a dedicated logger. Here's how we do it in the hang analyzer

Comment by Iryna Zhuravlova [ 17/Oct/22 ]

hey robert.guo@mongodb.com can you help us (jasur.nurboev@mongodb.com) to identify where this might be coming from? 

Comment by Max Hirschhorn [ 17/Aug/22 ]

Re-opening this ticket because it is still incorrect to use logging.basicConfig() to globally change the log settings for the current resmoke Python process. It causes the test output to end up showing up in the task logs.

[2022/08/10 09:29:29.248] [resmoke] 05:29:29.248Z Aborting process util_concurrency_test with pid 48705
[2022/08/10 09:29:29.248] [2022-08-10 05:29:29,248 - resmoke - INFO] Aborting process util_concurrency_test with pid 48705
[2022/08/10 09:29:29.249] [2022-08-10 05:29:29,249 - cpp_unit_test:util_concurrency_test - INFO] | 2022-08-10T09:29:29.248Z F  CONTROL  6384300 [thread1] "Writing fatal message","attr":{"message":"\n"}
[2022/08/10 09:29:29.249] [2022-08-10 05:29:29,249 - cpp_unit_test:util_concurrency_test - INFO] | 2022-08-10T09:29:29.248Z F  CONTROL  6384300 [thread1] "Writing fatal message","attr":{"message":"Got signal: 6 (Abort trap: 6).\n"}

https://evergreen.mongodb.com/lobster/evergreen/task/mongodb_mongo_master_nightly_macos_arm64_run_unittests_1ad5ef42f5161542550262647ab79e85cc6a7cf2_22_08_10_01_16_26/0/task#bookmarks=0%2C1340&l=1&shareLine=1071

Comment by Robert Guo (Inactive) [ 03/Jun/22 ]

Thanks for the reminder sviatlana.zuiko@mongodb.com; I had missed the connection between the two issues. Checking the code just now, Misha refactored the code 2 weeks ago here to call out to the db-contrib-tool subprocess, so logging config should no longer impact anything else in resmoke. Could you let me know if you are still seeing the issue? It might be a bug elsewhere.

Comment by Sviatlana Zuiko [ 31/May/22 ]

robert.guo@mongodb.com, do I understand correctly that this issue caused test logs being merged in task logs?
Actually, I'm facing such a behavior in recent runs https://evergreen.mongodb.com/task_log_raw/mongodb_mongo_master_enterprise_rhel_80_64_bit_dynamic_required_tenant_migration_kill_primary_jscore_passthrough_1_enterprise_2a7a609a78eb12dd5a237933b49b5f6bd03043f2_22_05_31_07_33_10/0?type=T
Is this represent a different issue?

Comment by Robert Guo (Inactive) [ 27/May/22 ]

No longer a priority after db-contrib-tool was refactored

Comment by Robert Guo (Inactive) [ 26/Apr/22 ]

Could we park this under a label for tech debt? Not urgent IMO

Generated at Thu Feb 08 05:52:24 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.