On Mon, 12 Dec 2011 08:42:49 +0200 Anssi Hannula <anssi.hannula@xxxxxx> wrote:
> On Mon, Dec 12, 2011 at 8:24 AM, NeilBrown <neilb@xxxxxxx> wrote:
> > On Mon, 12 Dec 2011 08:02:33 +0200 Anssi Hannula <anssi.hannula@xxxxxx> wrote:
> >
> >> On Mon, Dec 12, 2011 at 7:42 AM, NeilBrown <neilb@xxxxxxx> wrote:
> >> > On Mon, 12 Dec 2011 07:22:17 +0200 Anssi Hannula <anssi.hannula@xxxxxx> wrote:
> >> >
> >> >> On Mon, Dec 12, 2011 at 5:01 AM, NeilBrown <neilb@xxxxxxx> wrote:
> >> >> > On Sun, 11 Dec 2011 09:03:14 +0200 Anssi Hannula <anssi.hannula@xxxxxx> wrote:
> >> >> >
> >> >> >> Hi!
> >> >> >>
> >> >> >> After I rebooted during a raid6 rebuild, the rebuild didn't start again.
> >> >> >> Instead, there is a flood of "RAID conf printout"s that seemingly happen
> >> >> >> on array activity.
> >> >> >>
> >> >> >> All the devices show up properly in --detail and two devices are marked
> >> >> >> as "spare rebuilding", and I can access the contents of the array just
> >> >> >> fine, but the rebuild doesn't actually start. Is this a bug or am I
> >> >> >> missing something? :)
> >> >> >>
> >> >> >> I was initially on 2.6.38.8, but also tried 3.1.4 which seems to have
> >> >> >> the same issue. mdadm is 3.1.5.
> >> >> >>
> >> >> >> I'm not using start_ro and writing to the array doesn't trigger a
> >> >> >> rebuild either.
> >> >> >>
> >> >> >> Attached are --examine outputs before assembly, kernel log output on
> >> >> >> assembly, /proc/mdstat and --detail after assembly (on 3.1.4).
> >> >> >>
> >> >> >
> >> >> > Thank you for the very detailed problem report.
> >> >>
> >> >> Thanks for the quick response :)
> >> >>
> >> >> > Unfortunately it is a complete mystery to me what is happening.
> >> >> >
> >> >> > The repeated "RAID conf printout" messages are almost certainly coming from
> >> >> > the end of raid5_remove_disk.
> >> >> > It is being called from remove_and_add_spares for each of the two devices
> >> >> > that are being rebuilt. raid5_remove_disk declines to remove them because it
> >> >> > can keep rebuilding them.
> >> >> >
> >> >> > remove_and_add_spares then counts them and notes there are 2.
> >> >> > md_check_recovery notes that this is > 0, so it should create a thread to run
> >> >> > md_do_sync.
> >> >> >
> >> >> > md_do_sync should then print out a message like
> >> >> > md: recovery of RAID array md0
> >> >> >
> >> >> > but it doesn't. So something went wrong.
> >> >> > There are three reasons that md_do_sync might not print a message:
> >> >> >
> >> >> > 1/ MD_RECOVERY_DONE is set. As only md_do_sync ever sets it, that is
> >> >> > unlikely, and in any case md_check_recovery clears it.
> >> >> > 2/ mddev->ro != 0. It is only ever set to 0, 1, or 2. If it is 1 or 2
> >> >> > then we would be able to see that in /proc/mdstat as a "(readonly)"
> >> >> > status. But we don't.
> >> >> > 3/ MD_RECOVERY_INTR is set. Again, md_check_recovery clears this. It does
> >> >> > get set if kthread_should_stop() returns 'true', but that should only
> >> >> > happen if kthread_stop() was called. That is only called by
> >> >> > md_unregister_thread and I cannot see any way that could be call.
> >> >> >
> >> >> > So. No idea.
> >> >> >
> >> >> > Are you compiling these kernels yourself?
> >> >>
> >> >> Nope (used Mageia kernels), but I did now (3.1.5).
> >> >>
> >> >> > If so, could you:
> >> >> > - put a printk in the top of md_do_sync to report the values of
> >> >> > mddev->recovery and mddev->ro
> >> >> > - print a message whenever md_unregister_thread is called
> >> >> > - in md_check_recovery, in the
> >> >> > if (mddev->ro) {
> >> >> > /* Only thing we do on a ro array is remove
> >> >> > * failed devices.
> >> >> > */
> >> >> > mdk_rdev_t *rdev;
> >> >> >
> >> >> > in statement, print the value of mddev->ro.
> >> >> >
> >> >> > Then see which of those printk's fire, and what they tell us.
> >> >>
> >> >> Only the last one does, and mddev->ro == 0.
> >> >>
> >> >> For reference, attached is the used patch and resulting log output.
> >> >>
> >> >
> >> > Thanks.
> >> >
> >> > So it isn't running md_do_sync at all. Odd.
> >> >
> >> > Could please add:
> >> > - call "WARN_ON(1);" in print_raid5_conf() so we get a stack trace and can
> >> > see who is calling it.
> >> > - print the value that remove_and_add_spares is going to return.
> >>
> >> Attached. As you can see, remove_and_add_spare returns 0.
> >>
> >> --
> >> Anssi Hannula
> >
> >
> > Please add:
> >
> > diff --git a/drivers/md/md.c b/drivers/md/md.c
> > index 5c95ccb..fa56ac5 100644
> > --- a/drivers/md/md.c
> > +++ b/drivers/md/md.c
> > @@ -7328,8 +7328,10 @@ static int remove_and_add_spares(mddev_t *mddev)
> > }
> > }
> >
> > + printk("degraded=%d\n", mddev->degraded);
> > if (mddev->degraded) {
> > list_for_each_entry(rdev, &mddev->disks, same_set) {
> > + printk("raid_disk=%d flags=%x\n", rdev->raid_disk, rdev->flags);
> > if (rdev->raid_disk >= 0 &&
> > !test_bit(In_sync, &rdev->flags) &&
> > !test_bit(Faulty, &rdev->flags))
> >
> >
> > 'degraded' must be 2 as dmesg contains
> >
> > [ 45.544806] md/raid:md0: raid level 6 active with 8 out of 10 devices, algorithm 2
> >
> > and 'degraded' is exactly the difference between '8' and '10' there.
> >
> > raid disks 3 and 7 must have In_sync and Faulty clear as both of them just
> > show "spare rebuilding" in the 'detail' output.
> >
> > so remove_and_add_spares "must" return 2.
> >
> > Hopefully the above patch will help me understand which of those is wrong.
>
> The output is:
> [ 47.389379] md0: degraded=2
> [ 47.389380] md0: raid_disk=0 flags=4
> [ 47.389381] md0: raid_disk=-1 flags=0
>
> Full assemble log attached.
>
> --
> Anssi Hannula
Bingo.
This will fix it. We don't really need that 'break' there, and it is a
problem.
Thanks.
NeilBrown
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 5c95ccb..440d964 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -7344,8 +7344,7 @@ static int remove_and_add_spares(mddev_t *mddev)
spares++;
md_new_event(mddev);
set_bit(MD_CHANGE_DEVS, &mddev->flags);
- } else
- break;
+ }
}
}
}
Attachment:
signature.asc
Description: PGP signature
