[MXS-4403] Memory leak in 6.4.2 Created: 2022-11-16  Updated: 2023-12-07

Status: Needs Feedback
Project: MariaDB MaxScale
Component/s: maxbase
Affects Version/s: 6.4.2, 6.4.4
Fix Version/s: 6.4

Type: Bug Priority: Major
Reporter: Pon Suresh Pandian (Inactive) Assignee: Niclas Antti
Resolution: Unresolved Votes: 4
Labels: None
Environment:

Debian OS 11.3


Attachments: PNG File 11.08-11.16 (1).png     PNG File 11.14-11.16 (1).png     PNG File Bildschirm_foto 2022-11-14 um 15.21.29 (1).png     File maxscale.cnf    
Sprint: MXS-SPRINT-171, MXS-SPRINT-172, MXS-SPRINT-173, MXS-SPRINT-174, MXS-SPRINT-179, MXS-SPRINT-180, MXS-SPRINT-181, MXS-SPRINT-182, MXS-SPRINT-183

 Description   

Hi Team,

We see a linear increase of ram usage, finally crashing maxscale at 100% ram usage..

The maxscale process was killed by the oom_killer.. Here I have attached the maxsclae config file and memory usage graphs.. Please check it..



 Comments   
Comment by Naresh Chandra [ 2022-11-17 ]

Even, I have seen the same issue in the Maxscale 22.08 version as well. I have faced the same OOM error.

Comment by markus makela [ 2023-05-22 ]

naresh.chandra@copart.com Can you try if adding transaction_replay_max_size=1Mi into the configuration fixes the problem for you?

Comment by Naresh Chandra [ 2023-06-05 ]

markus makela,

Actually, we have disabled the transaction_replay settings and we are using the schema router service.

Comment by markus makela [ 2023-06-06 ]

Ah, in that case you might be seeing a different problem. naresh.chandra@copart.com can you open a new bug report for the increase in memory usage with schemarouter? I think it would be better to track it separately from this issue which focuses on readwritesplit.

Comment by markus makela [ 2023-08-21 ]

This could be explained by MXS-4726 if the client application used a connection pool to a readconnroute service.

Generated at Thu Feb 08 04:28:23 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.