Online resize

Online resize

[Wayne Gemmell]

Hi all

I'd like to do an online resize for my RAID5 drives now that I've upgraded 
them to 500GB drives. Now I'm a bit stuck. I assume the steps would be the 
following.

1) Resize partitions
2) Drop EXT3 journal
3) run e2resize
4) Recreate journal.

Are there any raid specific things I have to do to get this to work. Am I on 
the right track?




-- 
Regards
Wayne 

Re: Online resize

[Robin Hill]

[-- Attachment #1: Type: text/plain, Size: 1195 bytes --]

On Tue Mar 11, 2008 at 11:19:55AM +0200, Wayne Gemmell wrote:

> Hi all
> 
> I'd like to do an online resize for my RAID5 drives now that I've upgraded 
> them to 500GB drives. Now I'm a bit stuck. I assume the steps would be the 
> following.
> 
> 1) Resize partitions
> 2) Drop EXT3 journal
> 3) run e2resize
> 4) Recreate journal.
> 
> Are there any raid specific things I have to do to get this to work. Am I on 
> the right track?
> 
Sort of.  The steps you'll need are:

 1) Resize partitions (usually a drop & recreate step - this doesn't
    affect the data though).
 2) Force the OS to be aware of the new partition sizes - either a
    reboot or using partprobe.
 3) Grow the array (mdadm --grow) - check the manual page for details.
 4) Resize the filesystem (resize2fs).

There shouldn't be any need to drop & recreate the journal (I've never
done anyway - I guess it may be needed if you want to resize it).

HTH,
        Robin
-- 
     ___        
    ( ' }     |       Robin Hill        <robin@robinhill.me.uk> |
   / / )      | Little Jim says ....                            |
  // !!       |      "He fallen in de water !!"                 |

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Online resize

[Wayne Gemmell]

Well that didn't go so well

On Tuesday 11 March 2008 13:12:19 Robin Hill wrote:
>  2) Force the OS to be aware of the new partition sizes - either a
>     reboot or using partprobe.
>  3) Grow the array (mdadm --grow) - check the manual page for details.
I ran partprobe and then mdadm --grow --size=max /dev/md4 and nothing 
happened. Then I decided to reboot thinking that partprobe hadn't picked up 
that change and now it says that sda6 doesn't have a superblock. Any ideas 
what went wrong?

>  4) Resize the filesystem (resize2fs).


-- 
Regards
Wayne
--
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: Online resize

[Robin Hill]

[-- Attachment #1: Type: text/plain, Size: 1131 bytes --]

On Tue Mar 11, 2008 at 02:44:17PM +0200, Wayne Gemmell wrote:

> Well that didn't go so well
> 
> On Tuesday 11 March 2008 13:12:19 Robin Hill wrote:
> >  2) Force the OS to be aware of the new partition sizes - either a
> >     reboot or using partprobe.
> >  3) Grow the array (mdadm --grow) - check the manual page for details.
> I ran partprobe and then mdadm --grow --size=max /dev/md4 and nothing 
> happened. Then I decided to reboot thinking that partprobe hadn't picked up 
> that change and now it says that sda6 doesn't have a superblock. Any ideas 
> what went wrong?
> 
What version metadata are you using?  I'm not sure how 1.0 will handle
partition resizes but any other should certainly be okay.  Have you
tried running mdadm -E on each of the component partitions?  I assume
the new partitions are starting from the same position as the old ones?

Cheers,
        Robin
-- 
     ___        
    ( ' }     |       Robin Hill        <robin@robinhill.me.uk> |
   / / )      | Little Jim says ....                            |
  // !!       |      "He fallen in de water !!"                 |

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Online resize

[Wayne Gemmell]

On Tuesday 11 March 2008 16:58:32 Robin Hill wrote:
> What version metadata are you using?  I'm not sure how 1.0 will handle
> partition resizes but any other should certainly be okay.  
I'm not sure, the version in mdadm -E is 00.90.00. Is that what your talking 
about? They were done at the same time.

> Have you tried running mdadm -E on each of the component partitions?
Yes, It just says "no md superblock detected on /dev/sda6". Same with the 
rest.

> I assume the new partitions are starting from the same position as the old 
ones?

Yes, I just dropped the partition and made a new one in place.



-- 
Regards
Wayne 
--
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: Online resize

[Wayne Gemmell]

On Wednesday 12 March 2008 09:27:31 Wayne Gemmell wrote:
> On Tuesday 11 March 2008 16:58:32 Robin Hill wrote:
> > What version metadata are you using?  I'm not sure how 1.0 will handle
> > partition resizes but any other should certainly be okay.  
>
> I'm not sure, the version in mdadm -E is 00.90.00. Is that what your
> talking about? They were done at the same time.
Just to clarify this I ran mdadm -E on a partition in a working array that was 
done at the same time assuming that they are the same. 

-- 
Regards
Wayne 
--
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: Online resize

[Michael Tokarev]

Robin Hill wrote:
> On Tue Mar 11, 2008 at 02:44:17PM +0200, Wayne Gemmell wrote:
> 
>> Well that didn't go so well
>>
>> On Tuesday 11 March 2008 13:12:19 Robin Hill wrote:
>>>  2) Force the OS to be aware of the new partition sizes - either a
>>>     reboot or using partprobe.
>>>  3) Grow the array (mdadm --grow) - check the manual page for details.
>> I ran partprobe and then mdadm --grow --size=max /dev/md4 and nothing 
>> happened. Then I decided to reboot thinking that partprobe hadn't picked up 
>> that change and now it says that sda6 doesn't have a superblock. Any ideas 
>> what went wrong?
>>
> What version metadata are you using?  I'm not sure how 1.0 will handle
> partition resizes but any other should certainly be okay.  Have you
> tried running mdadm -E on each of the component partitions?  I assume
> the new partitions are starting from the same position as the old ones?

Robin, you're completely wrong here.

Only metadata at the START of the partition will work here.  Which is
either 1.1 or 1.2, but NOT 1.0 or 0.90, - which is at the END of the
partition.  When the end of the partition goes away, the superblock
becomes "orphan" somewhere inside the partition, and can't be found
anymore.

What should be done currently is:

a) re-create the array with new partitions, with THE SAME DISK ORDER,
   METADATA VERSION AND CHUNK SIZE as it was before.

b) restore partitions to the EXACT size as they were before, and
   start from scratch

/mjt

Re: Online resize

[Wayne Gemmell]

On Wednesday 12 March 2008 10:04:56 you wrote:
> a) re-create the array with new partitions, with THE SAME DISK ORDER,
>    METADATA VERSION AND CHUNK SIZE as it was before.
Is this possible without loosing data?
>
> b) restore partitions to the EXACT size as they were before, and
>    start from scratch
Thats easy enough. I'll do that now.


-- 
Regards
Wayne 

Re: Online resize

[Robin Hill]

[-- Attachment #1: Type: text/plain, Size: 1221 bytes --]

On Wed Mar 12, 2008 at 11:04:56AM +0300, Michael Tokarev wrote:

> Robin Hill wrote:
>> What version metadata are you using?  I'm not sure how 1.0 will handle
>> partition resizes but any other should certainly be okay.  Have you
>> tried running mdadm -E on each of the component partitions?  I assume
>> the new partitions are starting from the same position as the old ones?
>
> Robin, you're completely wrong here.
>
> Only metadata at the START of the partition will work here.  Which is
> either 1.1 or 1.2, but NOT 1.0 or 0.90, - which is at the END of the
> partition.  When the end of the partition goes away, the superblock
> becomes "orphan" somewhere inside the partition, and can't be found
> anymore.
>
That's why I said I wasn't sure about 1.0 metadata.  I've grown 0.90
arrays in exactly this way (though I may not actually have rebooted) so
I assumed they'd work okay (and the manual doesn't specify where the
metadata is for these).

Cheers,
        Robin
-- 
     ___        
    ( ' }     |       Robin Hill        <robin@robinhill.me.uk> |
   / / )      | Little Jim says ....                            |
  // !!       |      "He fallen in de water !!"                 |

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Online resize

[Michael Tokarev]

Wayne Gemmell wrote:
> On Wednesday 12 March 2008 10:04:56 you wrote:
>> a) re-create the array with new partitions, with THE SAME DISK ORDER,
>>    METADATA VERSION AND CHUNK SIZE as it was before.
> Is this possible without loosing data?

Yes.  But it's unreliable, sorta.  Too easy to screw things up, that is.

>> b) restore partitions to the EXACT size as they were before, and
>>    start from scratch
> Thats easy enough. I'll do that now.

And now, the original question comes up again: what to do, how to
perform the actual resize of the raid array...

To be fair, I don't really know.  I always used my own tiny program
to assist me in a situation like this - to back up and restore
v0.90 linux raid superblocks.  It's available at
http://www.corpit.ru/mjt/mdsuper.c .  Here's how I use it:

First, save the original superblocks in files:

  for f in sd{a,b,c,d,e}1; do
    ./mdsuper read /dev/$f > $f
  done

this instructs mdsuper to read the superblock (v0.90 only!) and
write it to standard out.  This has to be done while the array
(on original partitions) is STOPPED.

Now, resize the partitions, reboot if necessary (for the kernel
to recognize the new sizes).  You can save a dump of your current
partitions using `sfdisk -d' (redirect its output to a file as
with mdsuper) - just in case if something goes wrong.

At this point, your array is unaccessible - mdadm will tell you
that it can't find the superblocks.  But we have the superblocks
saved - so we only need to restore them back:

  for f in sd{a,b,c,d,e}1; do
    ./mdsuper write /dev/$f < $f
  done

(note the redirection is opposite!, and read was changed to write!).

Now, try to assemble the array - it should be here.  We just moved
(actually copied) the superblock from the now-middle-of-partitions
to the end, where it should be, and mdadm is now able to find it.
The array is still of the old size - but now mdadm --grow will
work.

That's basically it.  Just watch the mdsuper invocation.

/mjt

Re: Online resize

[Wayne Gemmell]

On Wednesday 12 March 2008 11:38:25 Michael Tokarev wrote:
> Wayne Gemmell wrote:
> > On Wednesday 12 March 2008 10:04:56 you wrote:
> >> a) re-create the array with new partitions, with THE SAME DISK ORDER,
> >>    METADATA VERSION AND CHUNK SIZE as it was before.
> >
> > Is this possible without loosing data?
>
> Yes.  But it's unreliable, sorta.  Too easy to screw things up, that is.
Thankfully I didn't need to do this. Resizing the partitions fixed that.
And now for the dangerous (looking) part.

> And now, the original question comes up again: what to do, how to
> perform the actual resize of the raid array...
>
> To be fair, I don't really know.  I always used my own tiny program
> to assist me in a situation like this - to back up and restore
> v0.90 linux raid superblocks.  It's available at
> http://www.corpit.ru/mjt/mdsuper.c .  Here's how I use it:
>
> First, save the original superblocks in files:
>
>   for f in sd{a,b,c,d,e}1; do
>     ./mdsuper read /dev/$f > $f
>   done
>
> this instructs mdsuper to read the superblock (v0.90 only!) and
> write it to standard out.  This has to be done while the array
> (on original partitions) is STOPPED.
>
> Now, resize the partitions, reboot if necessary (for the kernel
> to recognize the new sizes).  You can save a dump of your current
> partitions using `sfdisk -d' (redirect its output to a file as
> with mdsuper) - just in case if something goes wrong.
>
> At this point, your array is unaccessible - mdadm will tell you
> that it can't find the superblocks.  But we have the superblocks
> saved - so we only need to restore them back:
>
>   for f in sd{a,b,c,d,e}1; do
>     ./mdsuper write /dev/$f < $f
>   done
>
> (note the redirection is opposite!, and read was changed to write!).
>
> Now, try to assemble the array - it should be here.  We just moved
> (actually copied) the superblock from the now-middle-of-partitions
> to the end, where it should be, and mdadm is now able to find it.
> The array is still of the old size - but now mdadm --grow will
> work.
>
> That's basically it.  Just watch the mdsuper invocation.
>
> /mjt
> --
> 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



-- 
Regards
Wayne Gemmell
Flash Media Group
Tel:	27 (0)12 430 7597
Cell:	27 (0)83 666 3342

Re: Online resize

[Neil Brown]

On Tuesday March 11, wayne@flashmedia.co.za wrote:
> Hi all
> 
> I'd like to do an online resize for my RAID5 drives now that I've upgraded 
> them to 500GB drives. Now I'm a bit stuck. I assume the steps would be the 
> following.

What exactly do you mean by "now that I've upgraded them to 500GB
drives".
Don't tell my you got new 500GB drives, created little partitions on
them to match the old drives, and added those to the array.  That is
the wrong way around.

md is quite happy having using some partitions much larger than
others.
So if you have an array with (say) 4 200GB drives and want to replace
them all with 500GB drives, you should make 500GB partitions on those
drives (assuming that you really want to use partitions) and then
one at a time, fail a working drive and add a new drive, and wait for
recovery to complete.

Once that is done you still have a working array of the same size as
before but all the devices (and all the partitions) are larger.
Then just
  mdadm --grow /dev/mdX --size=max

and the array will resync the new space and be larger for you.  Then
you can resize the filesystem.

NeilBrown

Re: Online resize

[Wayne Gemmell]

On Tuesday 29 April 2008 03:00:13 Neil Brown wrote:
> On Tuesday March 11, wayne@flashmedia.co.za wrote:
> > Hi all
> >
> > I'd like to do an online resize for my RAID5 drives now that I've
> > upgraded them to 500GB drives. Now I'm a bit stuck. I assume the steps
> > would be the following.
>
> What exactly do you mean by "now that I've upgraded them to 500GB
> drives".
> Don't tell my you got new 500GB drives, created little partitions on
> them to match the old drives, and added those to the array.  That is
> the wrong way around.

That sounds about the way I did it. Since the hdd were so much bigger than the 
original drives I managed tho fit all 6 partitions on 2 hdd. Then created a 
new raid and copied everything across. Deleted the 6 partitions and grew the 
raid to fit the next 2 hdd. 

Your way sounds easier....


-- 
Regards
Wayne 
--
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