Bug 437798
Summary: | GFS2: xen virtual machine crash, gfs2 kernel bug on glock | ||
---|---|---|---|
Product: | Red Hat Enterprise Linux 5 | Reporter: | Maurizio Rottin <maurizio.rottin> |
Component: | kernel | Assignee: | Red Hat Kernel Manager <kernel-mgr> |
Status: | CLOSED INSUFFICIENT_DATA | QA Contact: | Red Hat Kernel QE team <kernel-qe> |
Severity: | medium | Docs Contact: | |
Priority: | low | ||
Version: | 5.1 | CC: | cluster-maint, edamato, rpeterso, rwheeler, swhiteho |
Target Milestone: | rc | ||
Target Release: | --- | ||
Hardware: | i686 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2008-12-03 10:01:51 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Maurizio Rottin
2008-03-17 14:30:40 UTC
I'm using the same bits that you're using (only that I've got rhel and I'm not using xen). I've tried a few things (including leaving the fs alone), but haven't been able to reproduce this problem. a) Is it possible for you to hit this on a freshly gfs2.mkfs'ed filesystem? Can you hit it on a standalone (mkfs'ed with -p lock_nolock) gfs2? b) What operations do you run until you hit this BUG? The second stacktrace you posted starts with an unlink operation, but the first one comes from gfs2_quotad. (These may to two similar bugs) I'll try to reproduce this and figure out what's going on, but it'd be ideal if you can hit it reliably and share the steps. Thanks! Hi, i went back to gfs(even if slower), and i have no more spare server to try now. Anyway, can the problem be xen? i never tried without xen, but i saw this bug trace a lot of times during one week. Can that be related to xen or some hardware i am using? a) the second trace is from a freshly made gfs2.mkfs, but never tried with lock_nolock. I remember that this happened often and also with only one node mounting the filesystem. The first trace is from a filesystem mounted with quota=on, but never initialized by me. b) actually i did nothing. I believed the problem was the clusterfs.sh script, so i mounted the fs manually, but the problem was still there. The glock code for RHEL 5.2 and up has been extensively cleaned up, so that it's very unlikely that this still applies to more recent kernels. Please upsgrade and let us know if its still a problem. i'm very busy right now and i must setup the same environment, i believe i'll be able to recreate the scenario in the next week, please wait until then. thank you. Its not a problem if you can't look at this right away, but we need to leave the bug in NEEDINFO in the mean time so that it doesn't mess up our stats. This was originally reported against Centos 5.1, since that code is very old indeed, I'd be very surprised if this bug still exists. I've not seen anything similar reported recently. If we don't hear any more information in the next few weeks, we'll close this bug as cannot reproduce. |