Hi,
I have been testing this a while now, and just finished testing your
untested patch. The rbd caching problem still persists.
The system I am testing on has the following characteristics:
Dom0:
- Linux xen-001 3.2.0-4-amd64 #1 SMP Debian 3.2.46-1 x86_64
- Most recent git checkout of blktap rbd branch
DomU:
- Same kernel as dom0
- Root (xvda1) is a logical volume on dom0
- xvda2 is a Rados Block Device format 1
Let me start by saying that the errors only occur with RBD client
caching ON.
I will give the error messages of both dom0 and domU before and after I
applied the patch.
Actions in domU to trigger errors:
~# mkfs.xfs -f /dev/xvda2
~# mount /dev/xvda2 /mnt
~# bonnie -u 0 -g 0 /mnt
Error messages:
BEFORE patch:
Without RBD cache:
dom0: no errors
domU: no errors
With RBD cache:
dom0: no errors
domU:
Aug 13 18:18:33 debian-vm-101 kernel: [ 37.960475] lost page write due
to I/O error on xvda2
Aug 13 18:18:33 debian-vm-101 kernel: [ 37.960488] lost page write due
to I/O error on xvda2
Aug 13 18:18:33 debian-vm-101 kernel: [ 37.960501] lost page write due
to I/O error on xvda2
...
Aug 13 18:18:52 debian-vm-101 kernel: [ 56.394645] XFS (xvda2):
xfs_do_force_shutdown(0x2) called from line 1007 of file
/build/linux-s5x2oE/linux-3.2.46/fs/xfs/xfs_log.c. Return address =
0xffffffffa013ced5
Aug 13 18:19:19 debian-vm-101 kernel: [ 83.941539] XFS (xvda2):
xfs_log_force: error 5 returned.
Aug 13 18:19:19 debian-vm-101 kernel: [ 83.941565] XFS (xvda2):
xfs_log_force: error 5 returned.
...
AFTER patch:
Without RBD cache:
dom0: no errors
domU: no errors
With RBD cache:
dom0:
Aug 13 16:40:49 xen-001 kernel: [ 94.954734] tapdisk[3075]: segfault
at 7f749ee86da0 ip 00007f749d060776 sp 00007f748ea7a460 error 7 in
libpthread-2.13.so[7f749d059000+17000]
domU:
Same as before patch.
I would like to add that I have the time to test this, we are happy to
help you in any way possible. However, since I am no C developer, I
won't be able to do much more than testing.
Regards
Frederik
On 13-08-13 11:20, Sylvain Munaut wrote:
Hi,
I hope not. How could I tell? It's not something I've explicitly enabled.
It's disabled by default.
So you'd have to have enabled it either in ceph.conf or directly in
the device path in the xen config. (option is 'rbd cache',
http://ceph.com/docs/next/rbd/rbd-config-ref/ )
Cheers,
Sylvain
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
