recovery question?

recovery question?

[Scott Kopel]

I have a raid1 with two disks /dev/hda and /dev/hdc ... three md devices
/dev/md0 made up of hda1 and hdc1
/dev/md1 made up of hda2 and hdc2
/dev/md2 made up of hda3 and hdc3

--I accidently formatted hda1 using #mke2fs -vj /dev/hda1 ... bonehead move 
number 1
--then I booted from a backup drive and resynced the raid
--when I tried to boot from the /dev/md0 I couldn't boot because of 
filesystem errors
--so then I ran e2fsck /dev/md0 and several errors were noted and fixed.. I 
didn't write down the error messages because I was too harried, bonehead 
move number 2, however I think the errors were files without inodes... the 
errors were fixed and a bunch of files and directories were put in 
lost&found ... mostly they were a recently installed phpmyadmin directory 
in /var/www/html/phpmyadmin and the only other file I've found missing so 
far is /usr/bin/info..
--I am curious what happened.. did the format remove the superblock and 
resyncing rewrote the superblock using the inodes for  these files?
any info would be appreciated



Scott Kopel
English Department, FSU
Mail Code 1580
850 644 6177 

Re: recovery question?

[Neil Brown]

On Monday October 13, skopel@english.fsu.edu wrote:
> I have a raid1 with two disks /dev/hda and /dev/hdc ... three md devices
> /dev/md0 made up of hda1 and hdc1
> /dev/md1 made up of hda2 and hdc2
> /dev/md2 made up of hda3 and hdc3
> 
> --I accidently formatted hda1 using #mke2fs -vj /dev/hda1 ... bonehead move 
> number 1
> --then I booted from a backup drive and resynced the raid

Exactly how did you "resync the raid".

doing "mke2fs" would not have been visible to the md driver so it wont
have known that the data was bad.

The thing to do at this point would be

  mdadm /dev/md0 -f /dev/hda1  # fail the drive
  mdadm /dev/md0 -r /dev/hda1  # and remove it
  mdadm /dev/md0 -a /dev/hda1  # and re-add it.

some of the corruption caused by mke2fs might have leaked into hdc1,
so a fsck would have been the next thing to do.



> --when I tried to boot from the /dev/md0 I couldn't boot because of 
> filesystem errors
> --so then I ran e2fsck /dev/md0 and several errors were noted and fixed.. I 
> didn't write down the error messages because I was too harried, bonehead 
> move number 2, however I think the errors were files without inodes... the 
> errors were fixed and a bunch of files and directories were put in 
> lost&found ... mostly they were a recently installed phpmyadmin directory 
> in /var/www/html/phpmyadmin and the only other file I've found missing so 
> far is /usr/bin/info..
> --I am curious what happened.. did the format remove the superblock and 
> resyncing rewrote the superblock using the inodes for  these files?
> any info would be appreciated

format would have zeroed all the inodes as seen on hda1
Any attempt to load an inode from md0 might have found a zeroed one on
hda1.  It if was written back the md0, that would have corrupted hdc1.

NeilBrown

Which drives below to which controller?

[jeff millar]

This isn't strictly a raid question, but it comes up because the machine has
several controllers and drives due to raid...

The machine has three controllers and 6 drives has shown below.  But which
drive letter belongs to which controller?  That information doesn't reside
anywhere in /proc/ide/* that I can see.  Does it exist some where else?

[jeff@maggie ide]$ ls -l /proc/ide
total 0
-r--r--r--    1 root     root            0 Oct 18 10:26 drivers
lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hda -> ide0/hda
lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdc -> ide1/hdc
lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hde -> ide2/hde
lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdg -> ide3/hdg
lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdi -> ide4/hdi
lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdk -> ide5/hdk
-r--r--r--    1 root     root            0 Oct 18 10:26 hpt366
dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide0
dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide1
dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide2
dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide3
dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide4
dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide5
-r--r--r--    1 root     root            0 Oct 18 10:26 pdc202xx
-r--r--r--    1 root     root            0 Oct 18 10:26 via

thanks...jeff

Re: Which drives below to which controller?...clarified

[jeff millar]

The question wasn't very clear.  Most specifically, which drives connect to
hpt366, which to pdc202xx?

> This isn't strictly a raid question, but it comes up because the machine
has
> several controllers and drives due to raid...
>
> The machine has three controllers and 6 drives has shown below.  But which
> drive letter belongs to which controller?  That information doesn't reside
> anywhere in /proc/ide/* that I can see.  Does it exist some where else?
>
> [jeff@maggie ide]$ ls -l /proc/ide
> total 0
> -r--r--r--    1 root     root            0 Oct 18 10:26 drivers
> lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hda -> ide0/hda
> lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdc -> ide1/hdc
> lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hde -> ide2/hde
> lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdg -> ide3/hdg
> lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdi -> ide4/hdi
> lrwxrwxrwx    1 root     root            8 Oct 18 10:26 hdk -> ide5/hdk
> -r--r--r--    1 root     root            0 Oct 18 10:26 hpt366
> dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide0
> dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide1
> dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide2
> dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide3
> dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide4
> dr-xr-xr-x    3 root     root            0 Oct 18 10:26 ide5
> -r--r--r--    1 root     root            0 Oct 18 10:26 pdc202xx
> -r--r--r--    1 root     root            0 Oct 18 10:26 via
>
> thanks...jeff
>
> -
> 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
>