Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

We have a production server (RHEL 5.3 64 bit) where all our domains are hosted remotely. We want to have a standard backup plan for our system so that on any physical damage the system can be quickly reconstructed from a restore point present in another remote system.

Any suggestions are greatly appreciated.

share|improve this question
    
What are your options? Do you have access to a backup device (eg tape library) or funding to get one, and if the latter, what sort of budget do you have for this? Does the colo facility provide any kind of managed or unmanaged backup solution? If none of the above, can you put some kind of external disc pack on the box? If you have no hardware, no budget, and no external provider, you can't do it, end of story; so we need to know what the options are. Also, how much data are we talking about, and how quickly is "quickly reconstructed"? –  MadHatter Dec 20 '10 at 10:21
    
Yes, we have access to a backup device i.e tape, and 2 weeks backup of the database is kept. –  Supratik Dec 20 '10 at 14:19
    
Yes, we have access to a backup device i.e tape, and 2 weeks backup of the database is kept. We also have RAID configured in the Linux box. Our concern is if at the worst condition if entire system collapse and we do not have any other options other than reconstructing the server from Scratch. What would be the best way to restore all configurations from another remote server in less time. –  Supratik Dec 20 '10 at 14:27
add comment

4 Answers

up vote 1 down vote accepted

There are a lot of good questions in the "other" section. All of this boils down to: Depends on what you're doing and how fast you need to recover. Instantly? Very expensive. Can you be down a week? Probably pretty cheap. How often do you get physically next to the server?

But here's what you're probably going to want to back up, at a minimum:

/etc /home your webroot any other application files any database files

For most of this, you can either use rsync or tar. Rsync will require a server "live" somewhere that you can host the data on. Tar will make an archive file (Which you can bzip) and then you'll have to extract it on another server.

You'll definitely need something with enough disk space and cpu and whatnot to run your failover server. It'll need to have RHEL5.3 (keep the same patches on it). You probably won't have to license the 2nd server unless you're running it all the time.

For the database files, you'll probably want to start with mysqldump or pg_dump. If you have Oracle, start with rman. These are the most inefficient, but the fastest & foolproof-ish way to get backups of these data.

Hope this helps as a place to start.

share|improve this answer
    
The rsync options sounds good for us. Is it possible to have a backup of two remote production servers in a single Linux box through rsync. –  Supratik Dec 20 '10 at 14:42
    
Hi Supratik, Yes, it should be possible to do that-- what you'll want to do is make sure you have enough disk space (for versioning) and have the apache configs backed up - and tested that you can cut them over as you need. And of course that the server can handle the load... –  Glen M Dec 20 '10 at 16:38
add comment

How much do you want to spend on this? There are commercial products out there to do this. My favorite is R1Soft. A free version would be to dump the databases to a file then have space elsewhere where you rsync the data to. You can then expand that to keep multiple versions of files.

share|improve this answer
add comment

Given that you have a tape drive, I'd look at bacula. In my experience it just gets on with putting your data safely on tape, and the bare-metal restores are flat-out elegant: you make a USB key customised to your hardware and bacula config, that can be booted on a cold-spare box and will restore it automagically from the most recent backups.

Don't forget that you'll need to dump the databases to disc, as has been said by others, in order for bacula to correctly save them to tape. Restoring the DBs after a cold-metal restore will also be your responsibility.

Whatever you do, for the love of God, test it, and frequently. There's nothing quite so heart-stopping as the discovery that the backup schedule you thought was going to save your arse has been writing zeroes to tape for the last eight months.

share|improve this answer
add comment

i have five servers, two database severs, two web servers, one load balancer and one backup drive. so when one of the web server fails, the other can carry on like nothing has gone wrong. you need to look into load balancers, database replication and software raid.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.