Re: 2.5.66: slow to friggin slow journal recover

Andrew Morton (akpm@digeo.com)
Tue, 1 Apr 2003 02:28:44 -0800


CaT <cat@zip.com.au> wrote:
>
> The journal recovery rangers from slow to really friggin slow under
> 2.5.66 with definate pauses in disk io stretching for 10s of seconds.
> This does not happen with 2.5.63 and if I hit ^c on fsck and let the
> kernel handle the journal recover for all partitions on mountime
> the recovery under 2.5.66 is either so fast that you don't notice
> it or just a buttload faster. Very objective measurements of time but
> the slowness of a journal recover as done by fsck is so noticible it's
> not funny.
>
> At the time of fsck journal recover or moiunt-time kernel journal
> recover DMA is turned on the drive.
>
> e2fsprogs 1.27 is in use. (1.27-2 from debian woody)
>

e2fsck 1.32 seems to work fine here.

Try arranging for a partition to _not_ be mounted at boot (disable it in
/etc/fstab). Then do a `reboot -f' and when you get a login prompt, run
e2fsck against that partition.

If the journal recovery is still slow then try capturing the state when it is
stuck with sysrq-T.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/