|Dedicated server for Backup|
| 4:12 pm on Nov 26, 2011 (gmt 0)|
I like to set dedicated server which is reliable. Need multiple daily backups af all data on another independent dedicated server to make sure no data is lost.
Is there any feedback for this?
| 6:48 pm on Nov 26, 2011 (gmt 0)|
You should define exactly what it is you think is going to fail. Power supply? Hard drive? Motherboard?
If you think the whole server is going to disappear somehow, and you want 0 downtime, then you don't want two servers backing up between each other. You want two machines that are continually synced together. And that takes some technical know-how to get up and running (I think it's called a round robin config or something, with 'heartbeat', but I could be wrong).
If you just want backups a few times a day, then if the server dies, realize it will probably NOT be as simple as flipping a switch and having the second machine take over. Nice idea, likely not quite as sane in practice.
Here's the kind of thing you can do:
- two machines running in sync, so if one dies, it's a seamless transition. This takes some work to setup, costs you two servers and costs two colo fees.
- One machine, with another machine doing backup at the same location. This doesn't make a lot of sense to me, you're paying for two servers and two colo fees, for no advantage.
- backup offsite multiple times a day. This is close to what I do.
- keep spare hard drives in the server, and backup multiple times per day to the spare hard drive. All that costs is the cost of a hard drive.
For the last two options, just buy some spare hardware and keep it handy.
Here's what I ended up doing (just ordered new servers this week):
- 1u server, redundant power supply, 16 gigs ram, 2 cpu's.
- 3 hard drives. raid striping (for speed) plus a spare hard drive that's probably going on a shelf.
- spare CPU,spare motherboard, spare power switching device, spare PS.
Then I backup offsite.
Now if something dies, I have backups offsite (at my house actually. I have spare parts for everything likely to go wrong (except memory, but that's readily available, so are hard drives). I drive to the colo facility and swap out parts, I'm back up and running in < an hour from almost any fatality.
Server company tells me that the only things that really fail are the hard drives and power supply. So if nothing else, have backups of those.