[SERVER-7417] Invalid access at address: 0 from thread: journal Created: 18/Oct/12  Updated: 08/Mar/13  Resolved: 09/Nov/12

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

Type: Bug Priority: Major - P3
Reporter: Bobby J Assignee: Eric Milkie
Resolution: Cannot Reproduce Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Ubuntu 12.04


Operating System: Linux
Participants:

 Description   

Saving new entries on multiple python processes (via pymongo 2.3), and get the error below. Has happened repeatedly, but not reproducibly.

Thu Oct 18 13:56:50 [initandlisten] connection accepted from 127.0.0.1:46942 #18413
(36 connections now open)
Thu Oct 18 13:56:50 [conn18413] end connection 127.0.0.1:46942 (35 connections now open)
Thu Oct 18 13:56:50 Invalid access at address: 0 from thread: journal

Thu Oct 18 13:56:50 Got signal: 11 (Segmentation fault).

Thu Oct 18 13:56:50 Backtrace:
0xade6e1 0x5582d9 0x558862 0x7fc0a4db1cb0 0x7fc0a411b0d0 0x583c30 0x5847f8 0x584b8c 0x6228ae 0x622c12 0xa0a30b 0xa0a70c 0xa0ae24 0x7c3659 0x7fc0a4da9e9a 0x7fc0a40bf4bd
mongod(_ZN5mongo15printStackTraceERSo+0x21) [0xade6e1]
mongod(_ZN5mongo10abruptQuitEi+0x399) [0x5582d9]
mongod(_ZN5mongo24abruptQuitWithAddrSignalEiP7siginfoPv+0x262) [0x558862]
/lib/x86_64-linux-gnu/libpthread.so.0(+0xfcb0) [0x7fc0a4db1cb0]
/lib/x86_64-linux-gnu/libc.so.6(+0x14e0d0) [0x7fc0a411b0d0]
mongod(_ZN6snappy8internal16CompressFragmentEPKcmPcPti+0x300) [0x583c30]
mongod(_ZN6snappy8CompressEPNS_6SourceEPNS_4SinkE+0x158) [0x5847f8]
mongod(_ZN6snappy11RawCompressEPKcmPcPm+0x5c) [0x584b8c]
mongod(_ZN5mongo3dur7Journal7journalERKNS0_11JSectHeaderERKNS_14AlignedBuilderE+0xae) [0x6228ae]
mongod(_ZN5mongo3dur14WRITETOJOURNALENS0_11JSectHeaderERNS_14AlignedBuilderE+0x32) [0x622c12]
mongod() [0xa0a30b]
mongod() [0xa0a70c]
mongod(_ZN5mongo3dur9durThreadEv+0x364) [0xa0ae24]
mongod() [0x7c3659]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x7e9a) [0x7fc0a4da9e9a]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7fc0a40bf4bd]



 Comments   
Comment by Bobby J [ 13/Nov/12 ]

Realized that this occurred during an (unrelated) catastrophic failure of the disk it was using. Given this unique event, I don't think we'll be able to reproduce. Thanks for looking into it.

Comment by Eric Milkie [ 23/Oct/12 ]

Hi Bobby. Thanks for reporting this issue.
I'd like to try to reproduce this on my own machine. How often do you experience a crash like this? Did you have something working okay with a prior version of the server or driver? Is it possible to get a copy of your Python code used to trigger this problem?
Can you attach a full log (beginning with server startup) from a crash? Is there anything special or unusual about your system setup?

The crash seems to be happening because we're using a null buffer for the temporary storage that the journaling system uses to compress its redo log. Since this buffer is static and protected by a mutex, I'm not yet sure how it could be seen as null during normal program execution.

Generated at Thu Feb 08 03:14:28 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.