-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Until Mark has a chance to build a new mail server setup from scratch on new machines (targetted within the month), we attempted to forestall additional crashes on goeje by swapping its hard drive to another machine of the same physical type.
The hope was that if the intermittent crashes were due to a hardware problem, that another box might not exhibit the same defect.
Goeje and Harris's drives were swapped about three and a half hours ago, pretty much without a hitch.
Unfortunately, the new goeje then crashed within an hour, and would not reboot. POST failures indicated major failures, perhaps of memory. (Harris had not been having any problems that we were aware of, though it had been out of regular rotation and used for some experimental service testing.)
The drive has been switched back to what we hope is its original location by the fine friendly late-night techs, and mail should be back up and streaming through.
We may have to consider a more annoying move of the mail software, configuration, and archives (which is half the work of the full move, and will have to do it all over again for the final setup, hence the reluctance to do it in the first place).
(We are maintaining a periodically updated backup of goeje's data, so we can restore to another machine if the drive also shorts out in case of emergency.)
- -- brion vibber (brion @ pobox.com / brion @ wikimedia.org)
wikitech-l@lists.wikimedia.org