Kernel crash

Kernel crash

[Guido Winkelmann]

Hi,

I just saw a kernel crash on one of my machines. It had the cephfs from the 
ceph cluster mounted using the in-kernel client:

[522247.751071]  [<ffffffff814d3383>] ? release_sock+0xe3/0x110
[522247.751182]  [<ffffffff815ea438>] __bad_area_nosemaphore+0x1d1/0x1f0
[522247.751290]  [<ffffffff815ea46a>] bad_area_nosemaphore+0x13/0x15
[522247.751397]  [<ffffffff815f7b76>] do_page_fault+0x416/0x4f0
[522247.751503]  [<ffffffff814ce5dd>] ? sock_recvmsg+0x11d/0x140
[522247.751611]  [<ffffffff812c0ea6>] ? cpumask_next_and+0x36/0x50
[522247.751718]  [<ffffffff815f4475>] page_fault+0x25/0x30
[522247.751828]  [<ffffffffa03ccba4>] ? ceph_x_destroy_authorizer+0x14/0x40 
[libceph]
[522247.751995]  [<ffffffffa040f9be>] get_authorizer+0x6e/0x140 [ceph]
[522247.752104]  [<ffffffff814ce646>] ? kernel_recvmsg+0x46/0x60
[522247.752213]  [<ffffffffa03b969a>] prepare_write_connect+0x17a/0x270 
[libceph]
[522247.752378]  [<ffffffffa03bba75>] con_work+0x755/0x2c40 [libceph]
[522247.752486]  [<ffffffff810876a3>] ? update_rq_clock+0x43/0x1b0
[522247.752598]  [<ffffffffa03bb320>] ? ceph_msg_new+0x2d0/0x2d0 [libceph]
[522247.752707]  [<ffffffff810747ae>] process_one_work+0x11e/0x470
[522247.752815]  [<ffffffff810755bf>] worker_thread+0x15f/0x360
[522247.752925]  [<ffffffff81075460>] ? manage_workers+0x230/0x230
[522247.753032]  [<ffffffff81079da3>] kthread+0x93/0xa0
[522247.753137]  [<ffffffff815fd2a4>] kernel_thread_helper+0x4/0x10
[522247.753245]  [<ffffffff81079d10>] ? 
kthread_freezable_should_stop+0x70/0x70
[522247.753355]  [<ffffffff815fd2a0>] ? gs_change+0x13/0x13
[522247.753459] ---[ end trace b9ba686594d99f89 ]---

These lines are all that I could still read on the screen. (Good thing there's 
Opens Source OCR programs out there...) I do not know how to extract more 
information about that crash (scrolling up does not work), but I'm leaving the 
machine like that over night in case someone can tell me.

Kernel version was 3.3.6-3.fc16.x86_64, Ceph cluster is version 0.47.2. The 
crash happened after I issued an rbd command. Another thing that might be 
related is that I stopped and restarted the entire cluster twice since 
mounting the cephfs. The first time, I disabled cephx, the second time I 
enabled it again.

Regards,
	Guido

Re: Kernel crash

[Tommi Virtanen]

On Wed, May 30, 2012 at 10:35 AM, Guido Winkelmann
<guido-ceph@thisisnotatest.de> wrote:
> I just saw a kernel crash on one of my machines. It had the cephfs from the
> ceph cluster mounted using the in-kernel client:
...
> [522247.751290]  [<ffffffff815ea46a>] bad_area_nosemaphore+0x13/0x15
> [522247.751397]  [<ffffffff815f7b76>] do_page_fault+0x416/0x4f0
> [522247.751503]  [<ffffffff814ce5dd>] ? sock_recvmsg+0x11d/0x140
> [522247.751611]  [<ffffffff812c0ea6>] ? cpumask_next_and+0x36/0x50
> [522247.751718]  [<ffffffff815f4475>] page_fault+0x25/0x30
> [522247.751828]  [<ffffffffa03ccba4>] ? ceph_x_destroy_authorizer+0x14/0x40
> [libceph]
> [522247.751995]  [<ffffffffa040f9be>] get_authorizer+0x6e/0x140 [ceph]
> [522247.752104]  [<ffffffff814ce646>] ? kernel_recvmsg+0x46/0x60

Hi. I can't find any other reports of kernel crashes related to
ceph_x_destroy_authorizer either.

If this keeps happening, please file a bug report at
http://tracker.newdream.net/ with details on the circumstance in which
it triggers, and we'll get back to it once we re-focus on the Ceph
Distributed File System. Right now, we are focusing our efforts in the
RADOS, RBD and radosgw functionality, in an effort to stabilize and
optimize the core object store of the product even more.
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html