[SERVER-84005] Streams: prevent SP process from being unschedulable after encountering the first OOM signal Created: 08/Dec/23  Updated: 11/Dec/23  Resolved: 11/Dec/23

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

Type: Bug Priority: Major - P3
Reporter: Aadesh Patel (Inactive) Assignee: Aadesh Patel (Inactive)
Resolution: Fixed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
is related to SERVER-81129 Detect high memory usage in stream pr... Closed
Assigned Teams:
Atlas Streams
Backwards Compatibility: Fully Compatible
Sprint: Sprint 37, Sprint 38
Participants:

 Description   

When the mongostream process runs out of memory, the memory usage monitor flips on the "exceeded memory limit" signal which will cause all current stream processors to error out on that process, but that signal is never reseted afterwards so that process/pod is rendered useless.



 Comments   
Comment by Githook User [ 11/Dec/23 ]

Author:

{'name': 'Aadesh Patel', 'email': 'aadesh.patel@mongodb.com', 'username': 'Aadeshp'}

Message: SERVER-84005: make ConcurrentMemoryAggregator take shared_ptr instead of unique_ptr for callback

GitOrigin-RevId: 5d33da40f361b9c8fde41bad9d028df68cde7200
Branch: master
https://github.com/mongodb/mongo/commit/542387a9f2f22724c366cc3277a21943d0bfc76c

Generated at Thu Feb 08 06:53:47 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.