Change disk problem

Change disk problem

[Jon Lurås]

Hello

I got a big problem with replacing disk on Software-RAID.
Whatever I do it just ends up as a spare disk.
I guess the problem is that the faulty disk was not removed from the RAID.
We just shut down and replace'd the disk.
Any idea of how to solve this?

Thanks in advance 

Jon 

DETAILS:

xxxxx:~# mdadm --detail /dev/md0
/dev/md0:
        Version : 00.90.00
  Creation Time : Sat Jun  7 22:04:01 2008
     Raid Level : raid5
     Array Size : 490223232 (467.51 GiB 501.99 GB)
    Device Size : 245111616 (233.76 GiB 250.99 GB)
   Raid Devices : 3
  Total Devices : 3
Preferred Minor : 0
    Persistence : Superblock is persistent

    Update Time : Sat Jun  7 22:04:01 2008
          State : active, degraded
 Active Devices : 2
Working Devices : 2
 Failed Devices : 1
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 64K

           UUID : 7edd5dfe:14b3ffb2:304b0607:0d821dcd
         Events : 0.2

    Number   Major   Minor   RaidDevice State
       0      56        1        0      active sync   /dev/hdi1
       1      56       65        1      active sync   /dev/hdj1
       2       0        0        2      removed


xxxxx:~# mdadm --version
mdadm - v2.5.6 - 9 November 2006

xxxxx:~# uname -a
Linux xxxxx 2.4.27-3-386 xxxxxxxxxxxxxxxxxxxxxxxxxx

xxxxx:~# cat /etc/debian_version
Debian 4.0

Re: Change disk problem

[David Greaves]

Jon Lurås wrote:
> Hello
> 
> I got a big problem with replacing disk on Software-RAID.
> Whatever I do it just ends up as a spare disk.
> I guess the problem is that the faulty disk was not removed from the RAID.
> We just shut down and replace'd the disk.
> Any idea of how to solve this?

You say 'spare' but I think you mean 'that md/mdadm is ignoring it completely
(on the grounds that it's not showing literally as a 'spare' in the --detail output.

This is expected behaviour - md does not just randomly grab new disks; you have
to use 'mdadm /dev/md0 --add /dev/hdk1' (use the correct component) to give md
permission to use the new disk.

Note that you should have partitioned it too (ie 'just shut down and replace'
becomes, shut down, replace, partition, mdadm --add )

> 
>     Number   Major   Minor   RaidDevice State
>        0      56        1        0      active sync   /dev/hdi1
>        1      56       65        1      active sync   /dev/hdj1
>        2       0        0        2      removed
> 


Look at http://linux-raid.osdl.org/ for more info...

HTH

David

--
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: Change disk problem

[David Greaves]

Jon Lurås wrote:
> Hello 
> 
> Thank you for your input.
> 
> I gave to little information in the original mail. I had tried to 'add'
> the new disk. It shows up like this - se below. The bad thing is that then
> the device '/dev/hdj1' suddenly changes to 'spare'...
> and everything looks very bad :-(

OK, you also wrote (forgot to cc the list)
Jon Lurås wrote:
> Hi again
>
> A little more information:
> The only solution I found to the problem which arised after the 'add' og
> /dev/hdk1 is
> "mdadm --create /dev/md0 --verbose --level=5 --chunk=64 --raid-devices=3
> /dev/hdi1 /dev/hdj1 missing"
>
> The it looks like the original 'mdadm --detail /dev/md0' result.
>
> Jon

So I'm now quite confused.

Can you tell us what you did a step at a time.
This is what should have happened:

1 Original disk failed.
2 md0 went into degraded and mdadm --monitor sent you an email
3 you shutdown and replaced /dev/hdk (?)
4 you partitioned /dev/hdk to make hdk1
5 you used --add to add hdk1 to the array
6 all was well

now you seem to have done several 'interesting' things in addition (like issuing
a --create).
However, if you can go back to the original --detail given, a simple --add
should resolve your problems...

You may want to check your dmesg to make sure you don't have a hardware issue.

> demeter:/etc# mdadm --detail /dev/md0
> /dev/md0:
>         Version : 00.90.00
>   Creation Time : Sat Jun  7 22:04:01 2008
>      Raid Level : raid5
>      Array Size : 490223232 (467.51 GiB 501.99 GB)
>     Device Size : 245111616 (233.76 GiB 250.99 GB)
>    Raid Devices : 3
>   Total Devices : 4
> Preferred Minor : 0
>     Persistence : Superblock is persistent
> 
>     Update Time : Sun Jun  8 12:05:02 2008
>           State : active, degraded
>  Active Devices : 1
> Working Devices : 3
>  Failed Devices : 1
>   Spare Devices : 2
> 
>          Layout : left-symmetric
>      Chunk Size : 64K
> 
>            UUID : 7edd5dfe:14b3ffb2:304b0607:0d821dcd
>          Events : 0.8
> 
>     Number   Major   Minor   RaidDevice State
>        0      56        1        0      active sync   /dev/hdi1
>        1       0        0        1      removed
>        2       0        0        2      removed
> 
>        3      57        1        3      spare   /dev/hdk1
>        4      56       65        4      spare   /dev/hdj1

This doesn't look good.

what does --examine say about each of the components?

David
--
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: Change disk problem

[Jon Lurås]

Hi

I didn't check in dmesg all the time. Now I check again - and your are
correct there is a hardware problem (maybe) with my hdj. This has nothing
to do on the linux-raid list... "hdj dma_intr: status=0x51 { DriveReady
SeekComplete Error }"

Jon

-----Original Message-----
From: David Greaves <david@dgreaves.com>
To: Jon Lurås <jon@irisinfo.net>
Cc: "linux-raid@vger.kernel.org" <linux-raid@vger.kernel.org>
Date: Sun, 08 Jun 2008 12:26:45 +0100
Subject: Re: Change disk problem

> Jon Lurås wrote:
> > Hello 
> > 
> > Thank you for your input.
> > 
> > I gave to little information in the original mail. I had tried to
> 'add'
> > the new disk. It shows up like this - se below. The bad thing is that
> then
> > the device '/dev/hdj1' suddenly changes to 'spare'...
> > and everything looks very bad :-(
> 
> OK, you also wrote (forgot to cc the list)
> Jon Lurås wrote:
> > Hi again
> >
> > A little more information:
> > The only solution I found to the problem which arised after the 'add'
> og
> > /dev/hdk1 is
> > "mdadm --create /dev/md0 --verbose --level=5 --chunk=64
> --raid-devices=3
> > /dev/hdi1 /dev/hdj1 missing"
> >
> > The it looks like the original 'mdadm --detail /dev/md0' result.
> >
> > Jon
> 
> So I'm now quite confused.
> 
> Can you tell us what you did a step at a time.
> This is what should have happened:
> 
> 1 Original disk failed.
> 2 md0 went into degraded and mdadm --monitor sent you an email
> 3 you shutdown and replaced /dev/hdk (?)
> 4 you partitioned /dev/hdk to make hdk1
> 5 you used --add to add hdk1 to the array
> 6 all was well
> 
> now you seem to have done several 'interesting' things in addition
> (like issuing
> a --create).
> However, if you can go back to the original --detail given, a simple
> --add
> should resolve your problems...
> 
> You may want to check your dmesg to make sure you don't have a hardware
> issue.
> 
> > demeter:/etc# mdadm --detail /dev/md0
> > /dev/md0:
> >         Version : 00.90.00
> >   Creation Time : Sat Jun  7 22:04:01 2008
> >      Raid Level : raid5
> >      Array Size : 490223232 (467.51 GiB 501.99 GB)
> >     Device Size : 245111616 (233.76 GiB 250.99 GB)
> >    Raid Devices : 3
> >   Total Devices : 4
> > Preferred Minor : 0
> >     Persistence : Superblock is persistent
> > 
> >     Update Time : Sun Jun  8 12:05:02 2008
> >           State : active, degraded
> >  Active Devices : 1
> > Working Devices : 3
> >  Failed Devices : 1
> >   Spare Devices : 2
> > 
> >          Layout : left-symmetric
> >      Chunk Size : 64K
> > 
> >            UUID : 7edd5dfe:14b3ffb2:304b0607:0d821dcd
> >          Events : 0.8
> > 
> >     Number   Major   Minor   RaidDevice State
> >        0      56        1        0      active sync   /dev/hdi1
> >        1       0        0        1      removed
> >        2       0        0        2      removed
> > 
> >        3      57        1        3      spare   /dev/hdk1
> >        4      56       65        4      spare   /dev/hdj1
> 
> This doesn't look good.
> 
> what does --examine say about each of the components?
> 
> David


--
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: Change disk problem

[Justin Piszcz]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 373 bytes --]



On Sun, 8 Jun 2008, Jon Lurås wrote:

>
> Hi
>
> I didn't check in dmesg all the time. Now I check again - and your are
> correct there is a hardware problem (maybe) with my hdj. This has nothing
> to do on the linux-raid list... "hdj dma_intr: status=0x51 { DriveReady
> SeekComplete Error }"
>
> Jon

What does:
smartctl -a /dev/hdj

Show?

Justin.