When we got fired up on the new server, we had a software RAID array. We had two harddrives, one mirroring the data on the other; sort of an "always backed up" setup. Unfortunately software RAID setups often have issues and we got kicked hard. The mirror failed and we ended up with some corrupted data. However, we got lucky and found out that only a small number of files got corrupted, and if they were corrupted on one of the disks, the other disk probably had a good copy. So we were able to recover almost all the data. We decided to knock one of the drives down to "emergency use only" since the file system was more corrupted than the other one.
Today, however, the worst happened. It turns out the cause of our software RAID setup working wasn't a software glitch or anything relating to the RAID itself, but was a harddrive failure in the making. It started out as a series of filesystem errors and then things got really flaky. (If any of you use Linux, take a look at the dmesg command and imagine <i>all</i> that output replaced with harddrive errors.) Luckily, we were able to recover data. Unfortunately this time we were unable to get an almost "fresh" copy of it, instead everything has rolled back to April 13th. Unfortunately this is bad news for a board such as this, since there were alot of interesting posts in the past few days.
In the mean time we've moved all of our customer sites to another server on the network until we can get that drive replaced. We will not move back to the regular server until we are assured that it is solid and reliable and we have a good backup solution in place. We plan on having a backup image updated every hour on the hour when we put it in place.
<b>Also</b>, and probably most importantly, due to the problems we've been experiencing lately, ketnar.org is now having a billing freeze. All outstanding balances will be reset to zero owed when billing resumes, and 2 months of service will be credited to all current accounts, in addition to the time that our billing freeze is in effect. We apologize for the problems we've been having lately, and are trying to get things back in order so that we can continue to offer the outstanding service and support our customers deserve.
<b>*</b> For the curious, the drives that failed were both Western Digitals. They're pretty good disks for light or medium work loads, but obviously not suited for a server environment.
Again, we apologize deeply for the mess that's been happening lately. Hopefully the road from here on out will be much smoother.
Sincerely,
Damon Canine
-- ketnar.org Systems Administrator