[dm-devel] dm-raid1: keep writing after leg failure
Lidong Zhong
lzhong at suse.com
Mon May 18 03:49:26 UTC 2015
>>> On 5/16/2015 at 12:11 AM, in message <20150515161108.GA28234 at redhat.com>, Mike
Snitzer <snitzer at redhat.com> wrote:
> On Wed, May 13 2015 at 2:04am -0400,
> Lidong Zhong <lzhong at suse.com> wrote:
>
> > Currently if there is a leg failure, the bio will be put into the hold
> > list until userspace replace/remove the leg. Here we are trying to make
> > dm-raid1 ignore the failure and keep the following bios going on.
> > This is because there maybe a temporary path failure in clvmd
> > which leads to cluster raid1 remove/replace the fake device failure. And
> > it takes a long time to do the full sync if we readd the device back.
> >
> > This patch merges the former five small patches to implement this
> > feature. The userspace should pass "keep_log" into kernel to make this
> feature to
> > take effect when creating dm-raid1, same as "handle_errors".
> >
> > Signed-off-by: Lidong Zhong <lzhong at suse.com>
> > Tested-by: Liuhua Wang <lwang at suse.com>
>
> I've staged a revised patch for 4.2 here:
> https://git.kernel.org/cgit/linux/kernel/git/device-mapper/linux-dm.git/commit
> /?h=for-next&id=164a064dd85801986f3190e6edd119b6fc9a05c5
>
> The big things I clarified/cleaned are:
> 1) keep_log depends on handle_errors feature
> 2) feature reporting via 'dmsetup table' needed to be decoupled
> otherwise any future feature would've forced this same cleanup.
> 3) Cleaned up header and in-code comments.
> 4) bumped the target version
>
> Please let me know if you see nay issues with this version.
>
Hi Mike,
All look good to me. Thanks.
Regards,
Lidong
> Thanks,
> Mike
>
>
More information about the dm-devel
mailing list