Author: James Courtier-Dutton Date: To: Hampshire LUG Discussion List Subject: Re: [Hampshire] SATA drive went read-only
On 19 October 2010 20:39, Chris Dennis <cgdennis@???> wrote: > Hello folks
>
> A client's Debian-based HP Proliant server suddenly decided to put make some
> of its partitions read-only the other day -- see the syslog extract below.
>
> I need to work out whether it's a problem with the SATA disk drive, or the
> controller, or whatever, and I can't make head or tail of the error codes --
> googling hasn't been much help.
>
> And the server is a production machine that the client can't live without,
> so I can't run tests very easily.
>
> Can anyone point me to some information about SATA error codes, so that I
> can begin to make a diagnosis?
>
> My first thought is to replace the hard drive, but I don't know if that
> would solve anything.
>
> cheers
>
> Chris
>
Those are hardware problems.
Linux remounts partitions read only when it gets certain hard errors.
My gut feeling in times like this is to take the server offline and
immediately fail over to the backup server.
Reasoning: Follow disaster recovery procedures.
Try a smartctl -a /dev/devicename
See if there is anything in the log.
This should give you the answer as to whether the problem hardware is
the HD itself, or the controller.
sd 0:0:0:0: [sda] Add. Sense: Recorded entity not found
end_request: I/O error, dev sda, sector 167402719
This implies a HD has failed or is failing.
If you really cannot do without the server, I hope /dev/sda is part of
a RAID array, because you need to replace it NOW.
Do not wait.
