Re: Strange Software RAID Problem

["Taylor, ForrestX" <forrestx taylor intel com>]

On Tue, 2003-11-25 at 14:09, list maaghosting net wrote:
> Hi All,
> Here is the problem that I am currently experiencing on a production server.  I
> have a Software RAID1 array that consists of two 36GB SCSI 160 drives.  I cat'd
> out /proc/mdstat and here is what was returned:
> 
>      Personalities : [raid1] 
>      read_ahead 1024 sectors
>      md0 : active raid1 sdb1[1]
>         32507392 blocks [2/1] [_U]
>       
>      md1 : active raid1 sdb2[1]
>         2562240 blocks [2/1] [_U]
>       
>      unused devices: <none>
> 
> The machine is still running, however it looks as if one of the drives has
> failed.  The machine has not been running very long, and I am not 100% sure if
> the RAID array was ever running on both disks.  Is there anyway to tell if a
> drive has actually failed, or if the array has not been initialized.
> 
> The array was constructed during the setup procedue, and the machine boots just
> fine.  Current uptime as of right now is around 29days.  There are no messages
> in /var/log/messages or /var/log/secure about a drive being down, but from the
> Software RAID howto it looks as if the array is not working correctly.
> 
> Any diagnostic steps that can be done without taking the machine down, as it is
> currently a production level machine?

First of all, make sure that you have a useable full backup.

It looks like sda is probably down.  What does your /etc/raidtab look
like?  Did you check /var/log/messages.[1-4]?  You may be able to find
when it went down there.  Are the lights blinking on sda?

Forrest
--