[Slackbuilds-users] we're back

[slackbuilds at jq.oc9.com]

On Tue, 8 Oct 2013, Rob McGee wrote:

> As many of you already know, we had yet another hardware failure at
> harrier, the SBo mail server on 2013-10-02 at about 20:04 UTC. It
> looked like a SATA controller failure, but instead it seems to have
> been simultaneous failure of BOTH drives in a RAID1 mirror.

Hmmm... You might have won the lottery, statiscally speaking. I think I 
remember a similar issue although and as I kept on putting new drives into 
a linux software RAID1 array, they all kept on failling almost together.

I ain't saying this is what is happening here and I wish it isn't. But 
just keep a close eye on that SATA controller if you see similarities with what 
I experienced.

I am experiencing a problem right now where newer NIC cards don't seem to 
play well with older hardware. I always try to factor in everything so maybe 
we could draw common points with HD controller errors.

Long live your new drives although, 
>
> Go figure. Anyway, between an offsite backup and cautious copying
> from the less dead of the two drives, no data was lost. The dead
> array had all the essential parts of the filesystem" /, /home, /usr
> and /var.

How much space do you need for one snapshot? I got a rsync incremental 
backup scheme going so maybe I could contribute on that level depending on 
space requirements.

>
> Too many eggs in a broken basket (which had not given any signs of
> trouble, strangely enough!)
>
As far as too many eggs in a broken basket I have experienced that 
also. As far as no signs, it might sound like HD controller. Then 
again, I hope I am wrong.

> A problem with getting services restored has been the distance and
> real life obligations: Robby and I were not able to sneakernet a
> replacement drive to the colo facility. He uploaded what he could
> through his slow home uplink, and pulled big stuff from mirror sites.
>

I have a fast link to restore if I can contribute.

Thanks for the great job!
-ls