[wp-hackers] [Pretty OT]May I be the first to console Owen...

[David Chait]

This should be a warning to anyone using an unmanaged server of any sort... 
Back up off-server yourself, whether paying the host to do it, or doing it 
yourself.  Don't 'trust' in any 'snapshot' backups in the hosting package 
unless they are expressly defined as off-server or off-site.

I think Matt at unixshell/tektonic has unfortunately tried to offer amazing 
cut-rate VPS service, not factoring in support overhead -- i.e., planning 
and consideration of an 'unmanaged' bullet-proof backup system (i.e., you 
NEVER back up data to the same drive(s) as the original source data).  He's 
continued to get higher demand for his cool services, but this is one of the 
'costs' of a small/one-man show (I don't know for a fact he's one-man, but I 
believe he started that way -- and most cut-rate companies are basically 
remote sysadmins!)

I don't necessarily blame unixshell.  I mean, failed hardware is failed 
hardware -- at that point, I'd start making RAID hardware manufacturers hold 
some kind of 'insurance' policy on card failures that destroy data. ;)  And, 
unixshell/Matt is VERY upfront that it's an extremely-unmanaged system. ;)

OTOH, a host is a host.  You can say you don't charge enough to really 
support folks, and that's how you keep things so cheap, but the moment 
something like this happens EVERYONE on the system realizes that a few bucks 
apiece, per month, to cover a teeny bit of 'support overhead' for disaster 
planning, would be well worth it.  If the plan promises backups, they 
shouldn't be on the same box -- hardware failures happen, don't rely on RAID 
to save your ass. ;) ;)

To Owen: if you need any help from the community, please ask!

-d

----- Original Message ----- 
From: "Owen Winkler" <ringmaster at midnightcircus.com>
| Yes, the server is pretty dead.  I guess the RAID card fried a bunch of
| the drives that were connected to it.  I don't blame Unixshell for my
| inability to restore backups I don't have, and I should have known
| better than to trust only their snapshot system.  Getting the new VPS
| running is another question, however.