[SERVER-4880] frequent journal warnings with Win2008 Created: 06/Feb/12  Updated: 11/Jul/16  Resolved: 28/Feb/12

Status: Closed
Project: Core Server
Component/s: Storage
Affects Version/s: 2.1.0
Fix Version/s: 2.1.1

Type: Bug Priority: Minor - P4
Reporter: Eric Milkie Assignee: Eric Milkie
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Windows 2008


Issue Links:
Related
related to SERVER-5080 journal rotates while dbMutex is held Closed
Operating System: Windows
Participants:

 Description   

By running "mongod.exe" with no args, and then simply inserting a lot of data, you get a log like this:

Mon Feb 06 10:47:02 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:07 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:12 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:17 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:22 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:27 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:32 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:37 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:42 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:47 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:52 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare
Mon Feb 06 10:47:57 [journal] info journal _rotate called insider dbMutex - ok but should be somewhat rare

This could be related to the new SRW locks.



 Comments   
Comment by auto [ 28/Feb/12 ]

Author:

{u'login': u'milkie', u'email': u'milkie@10gen.com', u'name': u'Eric Milkie'}

Message: SERVER-4880 log only one message every 20 seconds instead of every 5
Branch: master
https://github.com/mongodb/mongo/commit/73f34a28286f3594bc71811b1e6e1fe0db84114d

Comment by Eric Milkie [ 24/Feb/12 ]

Filed SERVER-5080 to look at the journal rotate stuff.
For this one, I am going to make LOGSOME log less often.

Comment by Aaron Staple [ 18/Feb/12 ]

Just want to comment that I'm seeing this on osx as well (potentially it occurs on all oses).

Comment by Dwight Merriman [ 06/Feb/12 ]

not sure why it is being called so much inside the lock. rotating in the lock would be a little bad as the world would be stopped while we unlink a 2GB file which might take a little time. but it's a rare event.

the logging is happening before the if statement that checks if we are due to rotate – it's good it's before the rotate as rotates are so rare we'd never notice it otherwise.

so we should analyze and see what is up but severity is not high

Comment by Eliot Horowitz (Inactive) [ 06/Feb/12 ]

We're getting now because i fixed the LOGSOME macro

Generated at Thu Feb 08 03:07:15 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.