Broken Mirror Repair

Broken Mirror Repair

Post by AKA » Sat, 29 Jun 2002 15:40:41



How do I correct this problem without data loss?  /dev/md5 holds my /usr
directory.
'lsraid -A -a /dev/md5' displays the following;

[dev   9,   5] /dev/md5         5FE1C1BB.D7FDFF17.8BC31E68.AEBEB5F7 online
[dev   8,   2] /dev/sda2        5FE1C1BB.D7FDFF17.8BC31E68.AEBEB5F7 good
[dev   ?,   ?] (unknown)        00000000.00000000.00000000.00000000 missing

The entry in /etc/raidtab shows this;

raiddev             /dev/md5
raid-level                  1
nr-raid-disks               2
chunk-size                  64k
persistent-superblock       1
nr-spare-disks              0
    device          /dev/sda2
    raid-disk     0
    device          /dev/sdb2
    raid-disk     1

Thanks,
Russell

 
 
 

Broken Mirror Repair

Post by AKA » Sat, 29 Jun 2002 15:46:42


Also the boot time message is as follows;

md: considering sdb2 ...
md:  adding sdb2 ...
md:  adding sda2 ...
md: created md5
md: bind<sda2,1>
md: bind<sdb2,2>
md: running: <sdb2><sda2>
md: sdb2's event counter: 00000005
md: sda2's event counter: 0000002b
md: superblock update time inconsistency -- using the most recent one
md: freshest: sda2
md: kicking non-fresh sdb2 from array!
md: unbind<sdb2,1>

Russell


Quote:> How do I correct this problem without data loss?  /dev/md5 holds my /usr
> directory.
> 'lsraid -A -a /dev/md5' displays the following;

> [dev   9,   5] /dev/md5         5FE1C1BB.D7FDFF17.8BC31E68.AEBEB5F7 online
> [dev   8,   2] /dev/sda2        5FE1C1BB.D7FDFF17.8BC31E68.AEBEB5F7 good
> [dev   ?,   ?] (unknown)        00000000.00000000.00000000.00000000
missing

> The entry in /etc/raidtab shows this;

> raiddev             /dev/md5
> raid-level                  1
> nr-raid-disks               2
> chunk-size                  64k
> persistent-superblock       1
> nr-spare-disks              0
>     device          /dev/sda2
>     raid-disk     0
>     device          /dev/sdb2
>     raid-disk     1

> Thanks,
> Russell


 
 
 

1. mirror stripe repair

Look into the output of metastat.

Mostly, you will find that all defect stripes have a diskpartition on
one physicla disk.
Also check /var/adm/messages for SCSI ERROR...

if one of the mirror-halfs is complete intact, then metadetach the
stripeset with problems,
metaclear the defect stripset,
do this on all mirrors with problems.

replace the defect scsi-disk
"format" the new disk with the same partiton-table as the defect drive

metainit the formerly defect stripes,
metattach the mirrors.

Before you do so,
save a printout of :

prtvtoc /dev/rdsk/c?d?t?s2
metastat -p
metadb

Check the disksuite-documentation,
it contains Help for most critical situations.
Most critical situations can be planned in advance in using this
documentation and planning _very_ carefully
the configuration to use.
(dont try to mirror raid5 Volumes).

Juergen.
--

2. serious bug in SunOS rpc.lockd with hostnames >= 20 chars

3. Help repairing Redhat 7.1 -- How can I repair the superblock?

4. /etc/suauth

5. How to repair a broken ext2 fs ?

6. installation problem

7. How to repair broken gdbm file?

8. How to know email of a http client ?

9. Problem with "BROKEN" fileset - how to repair?

10. Smoke and Mirrors: mirrored concat / concat mirrors

11. mirror, unmirror,backup mirror re-mirror howto

12. Breaking a mirrored volume group

13. Breaking a mirrored disk on Solaris 2.6