On Thu, Jul 05, 2012 at 07:34:17AM -0700, Marc MERLIN wrote:
> On Thu, Jul 05, 2012 at 09:25:44PM +0800, Liu Bo wrote:
> > On 07/04/2012 11:15 PM, Marc MERLIN wrote:
> > I'd expect to get some info from the following one, but I fails.
> >
> > Is it reproducible on your box?
>
> It was reproducible when I reported it, but not anymore now. I couldn't stay
> in a state where I was getting hangs and corruption.
>
> I had problems with btrfs-delalloc- taking most of the CPU and deadlocking
> the machine for minutes or even over half an hour. I'm pretty sure it was
> related to the fact that I only had 10GB free on my partition (including
> lots of snapshots).
> Now that I'm back to >40GB free, I haven't had the problem come back.
>
> How about these I reported:
> [102231.845170] btrfs no csum found for inode 3910562 start 0
> [102231.855415] btrfs_readpage_end_io_hook: 26 callbacks suppressed
> [102231.855418] btrfs csum failed ino 3910562 off 0 csum 3145117582 private 0
> [102231.999555] btrfs no csum found for inode 3910588 start 294912
> [102232.009479] btrfs csum failed ino 3910588 off 294912 csum 697692408 private 0
> [102232.142602] btrfs no csum found for inode 3910588 start 360448
> [102232.142661] btrfs no csum found for inode 3910588 start 393216
> [102232.968939] BTRFS warning (device dm-0): Aborting unused transaction.
> [102234.107914] BTRFS warning (device dm-0): Aborting unused transaction.
> [102235.218139] BTRFS warning (device dm-0): Aborting unused transaction.
>
> I know I got corruption since I saw some of my files being damaged, but
> are they useful in any way, or not so much?
Just to put a lid (maybe) on that thread, I now think that it's very
possible my btrfs hangs when the drive was almost full (10GB remaining
though) were due to my Crucial RealSSD C300, which has just mostly died
yesterday after (I think) the firmware got too many problems trying to remap
blocks (garbage collection) since I wasn't using TRIM.
(yes, I had the latest 007 firmware on the SSD).
I don't know that the SSD caused the corruption I saw in some files and that
went away when I dropped the latest btrfs snapshot, but at this point I'm
pretty sure that the SSD went into emergency garbage collection mode when
its space was running low, and in turn was causing btrfs to hang.
Sorry for the most likely false alarm (except for snapshot corruption which
is likely btrfs related).
I'll post another message about no TRIM with dmcrypt and possible ensuing
SSD problems.
Marc
--
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
Microsoft is to operating systems ....
.... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
