[SERVER-10731] Potential memory leak causing crash of mongod Created: 10/Sep/13  Updated: 11/Jul/16  Resolved: 13/Sep/13

Status: Closed
Project: Core Server
Component/s: Stability
Affects Version/s: 2.0.6
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Jeffrey Berger Assignee: James Wahlin
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Centos 6.3


Attachments: Text File mongodb.log    
Operating System: ALL
Participants:

 Description   

The mongod instance crashed, logs only show potential memleak warnings. Logs have been attached.



 Comments   
Comment by James Wahlin [ 13/Sep/13 ]

Thanks Jeff. Nothing else that comes to mind. Feel free to reopen this ticket if any further questions come up.

Comment by Jeffrey Berger [ 12/Sep/13 ]

James

Thanks for your comments, my organization has decided that we'll upgrade to 2.4.4 as planned as neither of those issues affects us. You can mark the ticket as closed unless there is anything else you can think of.

-Jeff

Comment by James Wahlin [ 11/Sep/13 ]

Hi Jeff,

mongod instances do have a 20K connection limit, regardless of ulimit settings so SERVER-10732 appears to be a separate issue. As for the upgrade, if your are using authentication I would definitely upgrade to at least 2.4.5 due to SERVER-10040. If you are sharded and have very large documents (16MB+) then I would upgrade to 2.4.6 due to SERVER-10478. Our general recommendation would be to upgrade to the latest stable version as it will have bug fixes that earlier versions do not.

thanks,
James

Comment by Jeffrey Berger [ 11/Sep/13 ]

Hi James

We experienced another issue and there is a JIRA ticket for that ( https://jira.mongodb.org/browse/SERVER-10732 ). We are planning on upgrading to 2.4.4 next week. Is there a reason to upgrade to 2.4.6 instead?

Thanks
Jeff

Comment by James Wahlin [ 11/Sep/13 ]

Hi Jeffrey,

The "could indicate a memory leak" are overly verbose and are logged when difference between the virtual address space and the mapped size is above an arbitrarily threshold. I would not be concerned by them as there is no substantial growth as they occur throughout your log. Unfortunately there is nothing else in the log that points to the reason for the server restart. Have you experienced other crashes recently?

I would recommend is an upgrade from MongoDB 2.0.6 to either 2.2.6 or 2.4.6. The version you are running is dated and there have been numerous improvements to MongoDB since then. If you are not in a position to go to 2.2 or 2.4, I would upgrade to 2.0.9 (the latest stable 2.0 release) at a minimum.

Thanks,
James

Generated at Thu Feb 08 03:23:55 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.