ARCHIVES-L Archives

Archiver > ARCHIVES > 1998-06 > 0896735138


From: Brian Leverich <>
Subject: [ARCHIVES-L] HARDWARE FAILURE (more ... )
Date: Mon, 01 Jun 1998 14:05:38 -0700


> One of the big SCSI hard drives in RootsWeb's main Web server is
> failing.
>
> I'm running backups now, and I hope I can get a current tape made
> before the drive fails entirely.

Done. We're safe. (:

> I have a spare drive ready on the shelf, and if all goes well I'll
> be installing it this afternoon. There will probably be some
> annoying side-effects associated with the move to the new drive, but
> I'm hoping we'll be back to fully normal operations by tomorrow
> night.
>
> We may be up and down quite a bit for the next 36 hours as I work
> through the problems.

We've changed strategy a bit. I'm bringing the current server back
online, failing disk and all, in a few minutes.

I am disabling server-side includes and cgi-bin execution. These
guys don't much affect disk operations directly, but they double or
more the overall load on the server and the disk problems do seem to
be correlated with load.

I am building an entirely new box now. I'm going to install the
operating system on new drives and then restore the failing drive on
a new drive.

Tomorrow morning I will down the old server and move the still
functional user disks to the new server. The server will probably
appear to "yo-yo" up and down as we transition from the old to the
new box.

IMPORTANT NOTES:

(1) If you upload files between now and tomorrow, they may not make
the transition to the new server. I would recommend folks delay
maintenance of their sites until tomorrow afternoon. (That will
also reduce load on the server, which will reduce the probability
that we crash again.)

(2) If the old box crashes again this afternoon or evening, we may
leave it down. Trying to rescue a dying box seems like a bad plan
right now, because it would probably delay us getting the new box
online.

Sorry about this -- we haven't had this much trouble with a server
in almost a year. We're doing everything we can. -B

--
Dr. Brian Leverich Co-moderator, soc.genealogy.methods/GENMTD-L
RootsWeb Genealogical Data Cooperative http://www.rootsweb.com/
P.O. Box 6798, Frazier Park, CA 93222-6798

This thread: