On Mon, 14 Feb 2005, Henry Nebrensky wrote:
> The SE dgc-grid-34 at Brunel appears to have had a hardware failure at the
> weekend - it will undergo surgery tomorrow.
The good news is that we've managed to recover the data and move it to the
new server.
The problem was a loose memory module (the manual for the Gigabyte
GA7-VAXP motherboard omits the fact that the correct insertion of a DIMM
into slot 3 requires a smack with a mallet... (slots 1 and 2 are normal)).
The less good news is that, while the old box was down, I've replaced that
service node with an LCFG'd RH7.3 LCG_230 SE as a front-end to a separate
RAID array.
This does actually appear to work as far as the gsiftp server goes, and
the GRIS
ldapsearch -x -H ldap://dgc-grid-34.brunel.ac.uk:2135 -b mds-vo-name=local,o=grid
is running, but there isn't any information provided. /opt/lcg/var is
completely empty, so it looks like something went badly awry during two
separate installs without any obvious error message.
I will try pulling the contents off /opt/lcg/var on the old SE and see
what happens..
The rest of the site is still LCG_220 (with a separate site-config with
different s/w version, install date, etc.) - could there be a weird
conflict somewhere?
Are the LCFG client messages actually logged anywhere?
Thanks
Henry
--
Dr. Henry Nebrensky [log in to unmask]
http://people.brunel.ac.uk/~eesrjjn
"The opossum is a very sophisticated animal.
It doesn't even get up until 5 or 6 p.m."
|