[SERVER-35964] Delete UBSan concurrency_replication* tasks Created: 05/Jul/18  Updated: 01/Jan/24  Resolved: 21/Dec/23

Status: Closed
Project: Core Server
Component/s: Testing Infrastructure
Affects Version/s: 4.1.1
Fix Version/s: 7.3.0-rc0

Type: Bug Priority: Major - P3
Reporter: Robert Guo (Inactive) Assignee: Mikhail Shchatko
Resolution: Fixed Votes: 0
Labels: open_todo_in_code, tig-concurrency
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Backports
Depends
Related
related to SERVER-35037 Create new concurrency suite that run... Closed
Assigned Teams:
Correctness
Backwards Compatibility: Fully Compatible
Operating System: ALL
Backport Requested:
v4.0, v3.6
Participants:
Linked BF Score: 30

 Description   

concurrency_replication* suites have high memory usage on UBSan, so we run them differently for now while the underlying cause is still being investigated



 Comments   
Comment by Githook User [ 21/Dec/23 ]

Author:

{'name': 'Mikhail Shchatko', 'email': 'mikhail.shchatko@mongodb.com', 'username': 'MikhailShchatko'}

Message: SERVER-35964 Delete UBSan concurrency_replication* tasks (#17716)

GitOrigin-RevId: 39fb7ee8fb6ecb7e5a6d6252a41c5d3a410477f1
Branch: master
https://github.com/mongodb/mongo/commit/38a98af97aa04a31f5fda75d9f6c570f12d84026

Comment by Steven Vannelli [ 10/May/22 ]

Moving this ticket to the Backlog and removing the "Backlog" fixVersion as per our latest policy for using fixVersions.

Comment by Max Hirschhorn [ 25/Feb/19 ]

Max Hirschhorn is the relation we've seen between the system allocator and memory usage sufficient to say we've "understood" the cause? If so, SERVER-26300 has been pushed which may mean this is ready for a look.

daniel.gottlieb, I think we can see whether the behavior is also observed when using --sanitize=undefined --allocator=tcmalloc, and if so, then set heapProfilingEnable=1 to see where the allocations are come from that don't appear to released until the process exits.

Comment by Daniel Gottlieb (Inactive) [ 25/Feb/19 ]

max.hirschhorn is the relation we've seen between the system allocator and memory usage sufficient to say we've "understood" the cause? If so, SERVER-26300 has been pushed which may mean this is ready for a look.

Comment by Max Hirschhorn [ 31/Oct/18 ]

Does this ticket still need to be in investigating status? What do we need to figure out to schedule it?

april.schoffer, we need to figure out the root cause behind the memory pressure when running with UBSan (undefined behavior sanitizer) enabled. The work that has been done on this ticket was to stop the bleeding in the linked BF ticket, but we need to understand the actual problem in order to revert those changes.

Comment by April Schoffer [ 31/Oct/18 ]

Does this ticket still need to be in investigating status? What do we need to figure out to schedule it?

Comment by Max Hirschhorn [ 06/Sep/18 ]

Do you want to file a separate ticket for blacklisting that suite?

No - I wasn't sure if it would give a clue as to where the memory pressure is coming from.

Comment by Robert Guo (Inactive) [ 06/Sep/18 ]

max.hirschhorn Nope, that was an oversight as the simultaneous suite didn't fail in the BF ticket. Do you want to file a separate ticket for blacklisting that suite?

Comment by Max Hirschhorn [ 06/Sep/18 ]

robert.guo, I forget - did you already look into why the concurrency_simultaneous_replication task doesn't have this issue on the UBSan builder?

Comment by Max Hirschhorn [ 05/Sep/18 ]

The concurrency_replication_multi_stmt_txn_ubsan.yml test suite introduced by the changes from 806b9c3 as part of SERVER-35037 should also be reverted once the memory pressure issue is understood and fixed.

Comment by Max Hirschhorn [ 11/Jul/18 ]

Bumping this out far into the future. It's something we should definitely continue to look into but isn't the most pressing thing now that the UBSan builder has returned to being green.

https://evergreen.mongodb.com/build/mongodb_mongo_master_ubuntu1604_debug_ubsan_80c7c825a44cf99b17e81f4233445c7ab1927706_18_07_11_01_45_09

Comment by Githook User [ 06/Jul/18 ]

Author:

{'username': 'guoyr', 'name': 'Robert Guo', 'email': 'robert.guo@10gen.com'}

Message: SERVER-35964 create UBSan concurrency_replication* tasks

(cherry picked from commit a6e2c10af2dcfa862c6afa0fb1504a5d2b092d09)
Branch: v4.0
https://github.com/mongodb/mongo/commit/a73913de26efe8e7b15d7d058ee8bccb6534a7c6

Comment by Githook User [ 05/Jul/18 ]

Author:

{'username': 'guoyr', 'name': 'Robert Guo', 'email': 'robert.guo@10gen.com'}

Message: SERVER-35964 create UBSan concurrency_replication* tasks
Branch: master
https://github.com/mongodb/mongo/commit/a6e2c10af2dcfa862c6afa0fb1504a5d2b092d09

Generated at Thu Feb 08 04:41:36 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.