Thanks for addressing this issue, Josef!
On Thu, Sep 26, 2013 at 4:26 PM, Josef Bacik wrote:
> We can starve out the transaction commit with a bunch of caching threads all
> running at the same time. This is because we will only drop the
> extent_commit_sem if we need_resched(), which isn't like
Hey,
I'm seeing the deadlock below under a ceph-osd workload. There may be a
subtle problem with the async transaction sequence (since nobody but ceph
uses that that I know of), but not obvious to me why
create_pending_snapshots would get stuck on btrfs_tree_lock...
[ 602.217383] INFO: task
-- Forwarded message --
From: Chester
Date: Thu, Oct 17, 2013 at 3:51 PM
Subject: dmesg filled with btrfs messages
To: linux-btrfs
It's what looks like the same message repeated over and over again.
Here's my dmesg: Pasted below is an excerpt. The full netconsole log
is on paste
I asked the debian maintainer to add btrfs-zero-log to the initramfs since I
unfortunately have to use it from time to time.
It still fixes the problem for me when btrfs won't mount, but it errors at
the end.
See screenshot
Marc
--
"A mouse is a device used to point at the xterm you want to typ
On Thu, Oct 17, 2013 at 02:29:24PM -0700, Marc MERLIN wrote:
> I asked the debian maintainer to add btrfs-zero-log to the initramfs since I
> unfortunately have to use it from time to time.
>
> It still fixes the problem for me when btrfs won't mount, but it errors at
> the end.
>
> See screensho
