Major technological problems

We had a thunderstorm this evening. A real corker. The lights flickered and (well you just know what’s coming don’t you?) and the computers blipped for a few seconds.

But in that few seconds, my server went down. It rebooted, but lost one of the three HDDs in there. And it just happened to be the one that was home to all the websites we host, the email for those websites, the MySQL databases…… You’re getting the picture aren’t you?

Other half is busy trying desperately to rescue the situation but as I write now (and I am sure he will update later), my weblog and the Gallery are lost, I have no email and it all seems pretty hopeless.

As a result of this evening’s debacle, we agreed that we would host everything somewhere else commercially rather than our living room from now on. Other half is sorting this out too (he’s a busy little bee right now). I don’t want a repeat episode again, whether it be an act of God or man.

One thought on “Major technological problems”

  1. That sucks – been there, done that and my best wishes to you both (and your hardware !). I would say though that going for commercial hosting isn’t the only answer: given my village and it’s erractic power we have multiple UPSes and they have stopped all of the hardware issues I was getting (the short blips are worse than actual cuts). CPC do some no-name 500VA units for 29+VAT, and you can get a very nice Belkin 1200VA for just over 100 UKP that has both serial and USB monitoring (it even activates the Mac OX UPS tab in the Power Manager :) )

    The other excellent item is a more recent purchase: Western Digital SATA Raid Edition 250GB drives. They are rated for 3 years of 24×7 operation, rather than the normal desktop hard drives which are only supposed to be on for 8hr/day: leaving them on in a server is not in their design and they will fail faster than expected. I got two of them (again, around 100 UKP each) and have put them in a RAID 1 mirror, so hopefully everything is as good as I can hope it to be at home.

    Backups are another issue that aren’t sorted yet here… At the moment the servers are in transition to the new RAID unit, and so there’s a ‘mostly good’ rsync copy from the main one and I might just keep the old system around for a once-per-week dump over the network. Email is pulled via IMAP, with each client caching the data so it’s possible to read old data when the server isn’t around (and scrape the data out to feed back to the server: don’t bring up an IMAP server with empty folders or it’ll wipe the client cache !)

    Truly, this isn’t supposed to be a gloat, but what I’ve done in response to seeing other people loose their data recently in the hopes that I can get things moved over to the new system before it happens to me. Do remember that even hosting has it’s data downside as the recent IndyMedia server raids show: if you’re on a shared server that gets grabbed (or if the warrant is so vague that the entire rack gets taken) then you will have to wait weeks or longer to get it all back again.


Comments are closed.