Bug 607245 - snapshot of root lv hangs
Summary: snapshot of root lv hangs
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: rawhide
Hardware: All
OS: Linux
low
medium
Target Milestone: ---
Assignee: Eric Sandeen
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Keywords:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-06-23 15:41 UTC by Zing
Modified: 2010-07-01 20:48 UTC (History)
18 users (show)

(edit)
Clone Of:
(edit)
Last Closed: 2010-07-01 20:48:23 UTC


Attachments (Terms of Use)

Description Zing 2010-06-23 15:41:49 UTC
Description of problem:
when I try to snapshot my root lv it hangs here:

==========================================================================
[root@RAWHIDE ~]# lvcreate -v -s -n snaproot -L 100M /dev/mainvg/rootlv
    Setting chunksize to 8 sectors.
    Setting logging type to disk
    Finding volume group "mainvg"
    Archiving volume group "mainvg" metadata (seqno 114).
    Creating logical volume snaproot
    Creating volume group backup "/etc/lvm/backup/mainvg" (seqno 115).
    Found volume group "mainvg"
    Creating mainvg-snaproot
    Loading mainvg-snaproot table (253:2)
    Resuming mainvg-snaproot (253:2)
    Clearing start of logical volume "snaproot"
    Creating logical volume snapshot0
    Found volume group "mainvg"
    Found volume group "mainvg"
    Creating mainvg-rootlv-real
    Loading mainvg-rootlv-real table (253:3)
    Resuming mainvg-rootlv-real (253:3)
    Loading mainvg-rootlv table (253:0)
    Creating mainvg-snaproot-cow
    Loading mainvg-snaproot-cow table (253:4)
    Resuming mainvg-snaproot-cow (253:4)
    Loading mainvg-snaproot table (253:2)
    Suspending mainvg-rootlv (253:0) with filesystem sync with device flush
=========================================================================

I/O on the system is then blocked and reboot needed.

lv's other than the root lv I can successfully snapshot.

Version-Release number of selected component (if applicable):
kernel-2.6.35-0.2.rc3.git0.fc14.x86_64
lvm2-2.02.67-1.fc14.x86_64
device-mapper-1.02.49-1.fc14.x86_64

Additional info:
If I revert back to kernel-2.6.34-20.fc14.x86_64, snapshots of my root lv are successful again.

This install is running in a qemu-kvm if that matters.

Comment 1 Milan Broz 2010-06-23 15:57:05 UTC
This is probably problem described here
https://www.redhat.com/archives/dm-devel/2010-June/msg00084.html

Reassigning to kernel, according the thread, reverting 6b0310fbf087ad6 fixes the problem.

Comment 2 Eric Sandeen 2010-06-23 16:01:05 UTC
Yeah, this one's mine.

It doesn't have to be the root, I think, any volume w/ active IO may hit this, sadly.

Comment 3 Eric Sandeen 2010-06-23 20:45:20 UTC
I sent a patch upstream (linux-fsdevel & linux-ext4) for this, there seem to be other deadlocks lurking but this should fix up the egregious one I introduced.

[PATCH] ext4: fix freeze deadlock under IO

I'll toss it into rawhide for testing.

Comment 4 Eric Sandeen 2010-07-01 20:48:23 UTC
s/b fixed in 2.6.35-0.15.rc3.git3

pls test & re-open if you still have trouble.

Thanks,
-Eric


Note You need to log in before you can comment on or make changes to this bug.