[SERVER-8111] Exception in replicated delete caused server crash Created: 08/Jan/13  Updated: 11/Jul/16  Resolved: 25/Jan/13

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: 2.2.0
Fix Version/s: None

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

Linux x86_64 CentOS 6.3


Attachments: Text File mongo_backtrace.txt    
Operating System: Linux
Participants:

 Description   

Just had a mongod crash, looking at the trace (attached) it's during replication of deletes.



 Comments   
Comment by Carl Johnstone [ 25/Jan/13 ]

Hi,

Yes we've upgraded and it's been running fine without any problems so far.

That said, we'd been running 2.2.0 pretty much straight from release in production (we wanted the aggregation framework) and had no problems with it until 8th Jan. Not being exactly sure what triggered the crash, obviously I can't be entirely sure that we're now good.

As I said in the previous comment, I'm happy to bow down to the better knowledge of the internals you guys have, and can reopen the ticket if I get a re-occurrence.

Carl

Comment by James Wahlin [ 24/Jan/13 ]

Hi Carl,

Just checking in to see whether you have upgraded to 2.2.2 yet. If so how did the upgrade go?

Thanks,
James

Comment by James Wahlin [ 15/Jan/13 ]

Hi Carl,

Reviewing SEVER-7238, it appears likely to be the same issue. That said, let's keep this open until we have confirmation on the fix. Can you let us know the results once you have upgraded to 2.2.2 (or higher) and have tested?

Thanks,
James

Comment by Carl Johnstone [ 08/Jan/13 ]

We're using TTL connections, so therefore powerOf2Sizes. I wanted to check as I wouldn't expect to be doing a bucket allocation for a large document on a replicated delete operation. However I'm happy to accept it as a dupe and am already planning on a 2.2.2 upgrade.

I also meant to come back and add that this was followed by a failure on the primary server around 2 hours later. It didn't crash - but it did allocate all available disk space to the database. Again if this is further indication of https://jira.mongodb.org/browse/SERVER-7238 I'm a happy user.

Comment by Tad Marshall [ 08/Jan/13 ]

This seems to be SERVER-7238, fixed in version 2.2.1. Are you using TTL collections or powerOf2Sizes?

Generated at Thu Feb 08 03:16:34 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.