mdadm: recovering from an aborted reshape op - to ensure at least one email gets through!

mdadm: recovering from an aborted reshape op - to ensure at least one email gets through!

[Gavin Flower]

Hi Neil,

I'll send this now.

My previous 2 attempts failed.

I had the first bounce due to HTML, but no errors reported for my second attempt.

I replied to you blog post.


Thanks,
Gavin
--

All Adults share the Responsibility

to help Raise Today's Children,

for they are Tomorrow's Society!


      

Re: mdadm: recovering from an aborted reshape op - to ensure at least one email gets through!

[NeilBrown]

On Sun, 13 Feb 2011 15:10:36 -0800 (PST) Gavin Flower <gavinflower@yahoo.com>
wrote:

> Hi Neil,
> 
> I'll send this now.
> 
> My previous 2 attempts failed.
> 
> I had the first bounce due to HTML, but no errors reported for my second attempt.
> 
> I replied to you blog post.
> 

Yes - vger.kernel.org doesn't like html.
Normal attachments should work - and you should get a bounce-message
if there is something odd.
But this and the following messages did get through.

For those on linux-raid: the background is that a reshape for a RAID6
to change chunksize was interrupted by the admin killing the mdadm
which monitors the reshape.  Then after a reboot the array isn't working 
properly.

When I asked for "mdadm -Evs" I should have asked for an extra 'v' - I keep
forgetting that.
However for now:  what would help is:

 mdadm -S /dev/md1
 mdadm -E /dev/md[abcde]2
 mdadm -Av /dev/md1 /dev/sd[abcde]2
 dmesg | tail -50

capture all of the output and post that.

Thanks,
NeilBrown

Re: mdadm: recovering from an aborted reshape op - to ensure at least one email gets through!

[NeilBrown]

On Sun, 13 Feb 2011 16:36:15 -0800 (PST) Gavin Flower <gavinflower@yahoo.com>
wrote:

> Hi Neil,
> 
> You last command seems to be doing something useful, after I had given it the original backup file.
> 
> I will update you when it finishes...
> 
> Will you still want the diagnostic output?  I am happy to send it, if you think it will be useful.

No, not necessary.

I guess whatever boot script tried to assemble the device didn't know
to tell it about a backup file and so the assembly didn't complete,
leaving you to pick up the pieces, but with no clear message explaining
what the pieces were...


I hope that in a few releases we can largely do away with the backup
file so I probably won't put much effort into trying to make this all
run more seamlessly.


> 
> If you want me to run other diagnostics, please let me know.
> 
> You deserve a huge bonus on top of the exorbitant (of zero dollars and zero cents) fee I am paying you! -)

:-)


Thanks,
NeilBrown


> 
> 
> Thanks,
> Gavin
> --
> All Adults share the Responsibility
> to help Raise Today's Children,
> for they are Tomorrow's Society!
> 
> 
> --- On Mon, 14/2/11, NeilBrown <neilb@suse.de> wrote:
> 
> > From: NeilBrown <neilb@suse.de>
> > Subject: Re: mdadm: recovering from an aborted reshape op - to ensure at least one email gets through!
> > To: "Gavin Flower" <gavinflower@yahoo.com>
> > Cc: linux-raid@vger.kernel.org
> > Date: Monday, 14 February, 2011, 13:00
> > On Sun, 13 Feb 2011 15:10:36 -0800
> > (PST) Gavin Flower <gavinflower@yahoo.com>
> > wrote:
> > 
> > > Hi Neil,
> > > 
> > > I'll send this now.
> > > 
> > > My previous 2 attempts failed.
> > > 
> > > I had the first bounce due to HTML, but no errors
> > reported for my second attempt.
> > > 
> > > I replied to you blog post.
> > > 
> > 
> > Yes - vger.kernel.org doesn't like html.
> > Normal attachments should work - and you should get a
> > bounce-message
> > if there is something odd.
> > But this and the following messages did get through.
> > 
> > For those on linux-raid: the background is that a reshape
> > for a RAID6
> > to change chunksize was interrupted by the admin killing
> > the mdadm
> > which monitors the reshape.  Then after a reboot the
> > array isn't working 
> > properly.
> > 
> > When I asked for "mdadm -Evs" I should have asked for an
> > extra 'v' - I keep
> > forgetting that.
> > However for now:  what would help is:
> > 
> >  mdadm -S /dev/md1
> >  mdadm -E /dev/md[abcde]2
> >  mdadm -Av /dev/md1 /dev/sd[abcde]2
> >  dmesg | tail -50
> > 
> > capture all of the output and post that.
> > 
> > Thanks,
> > NeilBrown
> > 
> > 
> 
> 
>       

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

Re: mdadm: recovering from an aborted reshape op - to ensure at least one email gets through!

[Gavin Flower]

Hi Neil,

It completed without apparent error (I did see errors in my very quick look!) - so I rebooted.

Unfortunately, it stopped in the same place.  It ran for over an hour before I went out, it had stopped by the time I got back a couple of hours late.

I should have captured dmesg before I rebooted, but I didn't (I blame walking for almost an hour in the hot Sun...).

I will follow this with email with hopefully useful diagnostics...


Cheers,
Gavin
--
All Adults share the Responsibility
to help Raise Today's Children,
for they are Tomorrow's Society!


--- On Mon, 14/2/11, NeilBrown <neilb@suse.de> wrote:

> From: NeilBrown <neilb@suse.de>
> Subject: Re: mdadm: recovering from an aborted reshape op - to ensure at least one email gets through!
> To: "Gavin Flower" <gavinflower@yahoo.com>
> Cc: linux-raid@vger.kernel.org
> Date: Monday, 14 February, 2011, 15:07
> On Sun, 13 Feb 2011 16:36:15 -0800
> (PST) Gavin Flower <gavinflower@yahoo.com>
> wrote:
> 
> > Hi Neil,
> > 
> > You last command seems to be doing something useful,
> after I had given it the original backup file.
> > 
> > I will update you when it finishes...
> > 
> > Will you still want the diagnostic output?  I am
> happy to send it, if you think it will be useful.
> 
> No, not necessary.
> 
> I guess whatever boot script tried to assemble the device
> didn't know
> to tell it about a backup file and so the assembly didn't
> complete,
> leaving you to pick up the pieces, but with no clear
> message explaining
> what the pieces were...
> 
> 
> I hope that in a few releases we can largely do away with
> the backup
> file so I probably won't put much effort into trying to
> make this all
> run more seamlessly.
> 
> 
> > 
> > If you want me to run other diagnostics, please let me
> know.
> > 
> > You deserve a huge bonus on top of the exorbitant (of
> zero dollars and zero cents) fee I am paying you! -)
> 
> :-)
> 
> 
> Thanks,
> NeilBrown
> 
> 
> > 
> > 
> > Thanks,
> > Gavin
> > --
> > All Adults share the Responsibility
> > to help Raise Today's Children,
> > for they are Tomorrow's Society!
> > 
> > 
> > --- On Mon, 14/2/11, NeilBrown <neilb@suse.de>
> wrote:
> > 
> > > From: NeilBrown <neilb@suse.de>
> > > Subject: Re: mdadm: recovering from an aborted
> reshape op - to ensure at least one email gets through!
> > > To: "Gavin Flower" <gavinflower@yahoo.com>
> > > Cc: linux-raid@vger.kernel.org
> > > Date: Monday, 14 February, 2011, 13:00
> > > On Sun, 13 Feb 2011 15:10:36 -0800
> > > (PST) Gavin Flower <gavinflower@yahoo.com>
> > > wrote:
> > > 
> > > > Hi Neil,
> > > > 
> > > > I'll send this now.
> > > > 
> > > > My previous 2 attempts failed.
> > > > 
> > > > I had the first bounce due to HTML, but no
> errors
> > > reported for my second attempt.
> > > > 
> > > > I replied to you blog post.
> > > > 
> > > 
> > > Yes - vger.kernel.org doesn't like html.
> > > Normal attachments should work - and you should
> get a
> > > bounce-message
> > > if there is something odd.
> > > But this and the following messages did get
> through.
> > > 
> > > For those on linux-raid: the background is that a
> reshape
> > > for a RAID6
> > > to change chunksize was interrupted by the admin
> killing
> > > the mdadm
> > > which monitors the reshape.  Then after a reboot
> the
> > > array isn't working 
> > > properly.
> > > 
> > > When I asked for "mdadm -Evs" I should have asked
> for an
> > > extra 'v' - I keep
> > > forgetting that.
> > > However for now:  what would help is:
> > > 
> > >  mdadm -S /dev/md1
> > >  mdadm -E /dev/md[abcde]2
> > >  mdadm -Av /dev/md1 /dev/sd[abcde]2
> > >  dmesg | tail -50
> > > 
> > > capture all of the output and post that.
> > > 
> > > Thanks,
> > > NeilBrown
> > > 
> > > 
> > 
> > 
> >       
> 
> 


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