From: "Bruno Prémont" <bonbons@linux-vserver.org>
To: Dave Chinner <david@fromorbit.com>
Cc: xfs@oss.sgi.com
Subject: Re: XFS crashing system with general protection fault
Date: Mon, 29 Dec 2014 08:44:52 +0100 [thread overview]
Message-ID: <20141229084452.615e1900@pluto.restena.lu> (raw)
In-Reply-To: <20141228115127.GN24183@dastard>
Hi Dave,
On Sun, 28 Dec 2014 22:51:27 +1100 Dave Chinner wrote:
> On Wed, Dec 24, 2014 at 11:14:03AM +0100, Bruno Prémont wrote:
> > On a server I've got the following traces, the first on Monday, the second
> > one today. On Monday kernel was 3.14.17 and 3.14.27 for today (both captured
> > via netconsole).
> >
> > Is that fixed in a newer kernel?
> >
> > I've xfs_repaired one of the two XFS partitions on the server though it
> > found nothing to complain about. The other partition, containing /, has
> > not been explicitly checked yet.
> >
> > If there is some information I should gather before xfs_repairing, please
> > tell as soon as possible!
> >
> >
> > Thanks,
> > Bruno
> >
> > [6149136.014757] general protection fault: 0000 [#1] SMP
> > [6149136.022825] Modules linked in: netconsole configfs
> > [6149136.028996] CPU: 4 PID: 151 Comm: kworker/4:1H Not tainted 3.14.18-x86_64 #1
> > [6149136.040750] Hardware name: HP ProLiant DL360 G6, BIOS P64 07/02/2013
> > [6149136.048936] Workqueue: xfslogd xfs_buf_iodone_work
> > [6149136.056836] task: ffff880212c67500 ti: ffff8800def3c000 task.ti: ffff8800def3c000
> > [6149136.067023] RIP: 0010:[<ffffffff81255b67>] [<ffffffff81255b67>] xfs_trans_ail_delete_bulk+0x87/0x1a0
> > [6149136.080940] RSP: 0018:ffff8800def3dce8 EFLAGS: 00010202
> > [6149136.088889] RAX: dead000000100100 RBX: ffff88000211bd10 RCX: ffff88010e23fbb1
> > [6149136.098962] RDX: 6b6b6b6b6b6b6b6b RSI: 6b6b6b6b6b6b6b6b RDI: ffff88000211bd10
> > [6149136.110787] RBP: ffff8800def3dd38 R08: 6b6b6b6b6b6b6b6b R09: 2900000000000000
>
> You have memory poisoning turned on?
>
> #define POISON_FREE 0x6b /* for use-after-free poisoning */
Yes, I do.
> Did this occur at unmount? Can you reproduce it on a 3.18 kernel?
No, it happens at runtime (apparently triggered/made likely by backup
daemon reading through the filesystem, but not each time).
Though that server is always busy writing to the disks (so backup
makes it even more busy).
It has two XFS partitions, one root partition including /var/
and a second data partition, both being written to (the data partition
more aggressively that the root one - root partition receives some deal
of logging).
Thanks,
Bruno
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2014-12-29 7:45 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-12-24 10:14 XFS crashing system with general protection fault Bruno Prémont
2014-12-28 11:51 ` Dave Chinner
2014-12-29 7:44 ` Bruno Prémont [this message]
2015-01-13 7:17 ` Bruno Prémont
2015-02-05 14:10 ` Bruno Prémont
2015-02-05 22:15 ` Dave Chinner
2015-02-09 8:47 ` Bruno Prémont
2015-02-09 21:24 ` Dave Chinner
2015-02-10 7:05 ` Bruno Prémont
2015-02-23 7:56 ` Bruno Prémont
2015-03-12 14:15 ` Bruno Prémont
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20141229084452.615e1900@pluto.restena.lu \
--to=bonbons@linux-vserver.org \
--cc=david@fromorbit.com \
--cc=xfs@oss.sgi.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox