Hooray for RAID!
If you read yesterday's post, you'll know that:
1) I had been trying to migrate my server to a new box before hardware failure ensued, and
2) I didn't quite succeed.
Well, I was able to boot up the old machine using a Knoppix CD, so that I could get at the hard drives to move data across to the new server. The root (/) partition seemed to be randomly hosed, with data loss only at certain critical places (such as my /etc/aliases file, with over 500 aliases that would be tough to recreate). I had been able to set up the basic structure, though, without some of the flesh:
1) Email was working again, but no aliases, no archive files, etc. New mail was rolling in without a problem
2) Web server was working, old photos are there, but the "glue" (Gallery2 databases) weren't set up correctly.
3) Spam filtering was working, basically. Fine-tuning from Bayes database, blocking lists, firewall rules, and the like were gone.
All in all, survivable but would require some legwork. Looking at the old server, though, the second partition in each RAID group (or whatever the technical terms are) was OK - I was able to get the backup copies of these files as planned originally (all except Gallery2, which needs some sort of arcane alchemy I've not yet figured out).
So... The plan for the future:
Take out the old drives, mark carefully, get ahold of an external USB enclosure for the drives. Put new 500GB drives in the machine, partition and install CentOS 5.2 (just as the new working server has). Bring working configuration to this machine, synch up the data. Power off.
Synch up every week or so, just to make sure, and keep unplugged in between. If there's ever another problem, I won't lose more than a week's data. This time, I ended up not losing any data, as far as I can tell.
Sometimes, things *do* work out.
1) I had been trying to migrate my server to a new box before hardware failure ensued, and
2) I didn't quite succeed.
Well, I was able to boot up the old machine using a Knoppix CD, so that I could get at the hard drives to move data across to the new server. The root (/) partition seemed to be randomly hosed, with data loss only at certain critical places (such as my /etc/aliases file, with over 500 aliases that would be tough to recreate). I had been able to set up the basic structure, though, without some of the flesh:
1) Email was working again, but no aliases, no archive files, etc. New mail was rolling in without a problem
2) Web server was working, old photos are there, but the "glue" (Gallery2 databases) weren't set up correctly.
3) Spam filtering was working, basically. Fine-tuning from Bayes database, blocking lists, firewall rules, and the like were gone.
All in all, survivable but would require some legwork. Looking at the old server, though, the second partition in each RAID group (or whatever the technical terms are) was OK - I was able to get the backup copies of these files as planned originally (all except Gallery2, which needs some sort of arcane alchemy I've not yet figured out).
So... The plan for the future:
Take out the old drives, mark carefully, get ahold of an external USB enclosure for the drives. Put new 500GB drives in the machine, partition and install CentOS 5.2 (just as the new working server has). Bring working configuration to this machine, synch up the data. Power off.
Synch up every week or so, just to make sure, and keep unplugged in between. If there's ever another problem, I won't lose more than a week's data. This time, I ended up not losing any data, as far as I can tell.
Sometimes, things *do* work out.