Re: 2.6.28 ext4, xen and lvm volume becomes ro after snapshot

[Andreas Sundstrom <sunkan@zappa.cx>]

Theodore Tso wrote:
> Hmm... ok, to summarize, you are seeing this problem with 2.6.28 (have
> you tried earlier kernel versions) when an LVM volume has a read/only
> snapshot in existence and you try to mount the LVM volume using ext4.
> If you mount the same LVM volume using ext3, you don't see any
> problems.   Is that correct?
>   
Correct, and I have not been using barriers with any earlier kernel.
I stumbled upon this because barriers are enabled by default in ext4.
> Can you try mounting ext3 with barriers enabled?  Booting with the
> command linux option rootflags=barriers should do the trick.  If that
> fails, then it would indicate that trying to enable barriers in a Xen
> guest while the host OS has created a snapshot of volume causes an I/O
> error, thus leading errors which you are seeing.
>
> 						- Ted
>   
Yes, I mounted it with ext3 and barrier=1 and could reproduce the problem.
ext3 did not remount the fs ro though, it seems to only disable barriers:

[    7.681759] blkfront: xvda1: write barrier op failed
[    7.681776] blkfront: xvda1: barriers disabled
[    7.681785] end_request: I/O error, dev xvda1, sector 4584
[    7.681800] end_request: I/O error, dev xvda1, sector 4584
[    7.681886] JBD: barrier-based sync failed on xvda1 - disabling barriers

And then I tested with ext4 and barrier=0 and that also works.

I don't know if this is expected behaviour or not or if it's worth
looking into.
I just wanted to report back what I saw. For me personally it's no issue.
I'll be using ext4 without barriers for a while and see how it goes.

But I'm here if you want something tested or a patch verified or anything,
but I guess this might be a Xen issue rather than vanilla kernel stuff.

Thanks for helping out with the narrowing down of the issue

/Andreas