We are having 40 factories. Each having suse linux server. We want to set up a disaster recovery solution.


We are having 2 disaster recovery server. Each and every day, we want to take back up of 40 servers data in these 2 disaster recovery server.


Can anyone tell the general idea how to do in shell script very secured way?

A google search gave me all you would want to know, among which is:
http://linux.about.com/od/softbackup..._Solutions.htm

I've heard a lot of good about Amanda and Mondo, although I use none of the existing tools (I created my own).

Yves.

Can you give a brief introduction about your own Disaster Recovery Plan?

i have heard (though have no experience of my own) that bacula is good

Not much actually: a single server to backup.
First there's a backup on tape (a simple tar).
And there's a script I wrote for incrementally backing up using rsync, on a mouted backup partition.

Yves.

I think your question is probably too general and didn't provide much in the way of specific issues of concern.

40 factories, each with a Suse server.

What do these servers do? What is the critical data that differs among them? And how is it stored? If they are simply file servers, that is one kind of solution. If they are mySQL databases, that is quite another. Also, if they are all identical installs with only the data differing, that makes things easier. If they are each in some way different and have different populations of users, then that requires a more complete backup of each.

Then there are the questions about the distances between the factories and the type of connections available to transfer data over. Are those connections on the public net? Are they reliable?

And your disaster recovery servers. What do they have that makes them DR Servers? Lots of disk space? Tape drives? Tape Libraries? How does their capacity compare to the requirements of each of the factory servers?

So, you see, you've asked a question simply, but it is not a simple question. If you refine your question and provide more detail, then you will answer some of your own questions and also come up with questions that will be easier for others to help you with.

Reading between the lines, I rather suspect that you need a complete disaster recovery plan, including mirrored servers at offsite locations, redundancy in production servers, immediate fallover to offsite backup servers if main production servers go down, mirroring, and all that jazz.

Such a plan, to be comprehensive, needs to begin with an audit of your 40 factories, and an identification of the functionality and criticality of your servers. This takes time, and isn't going to happen in a message board context.

That you even ask the question here suggests that you don't know how to go about it. You need to hire a consultant to prepare a plan for you. This will cost you substantially, but not nearly as much as it would cost you to have your production facilities go down for any period of time.

As far as that goes, I'll do it for you, for US$100/hr plus travel, but if you are not in North America you can certainly find someone closer to home to do it for you.

Which all goes back to the question: What are the servers being used for? What happens to the factory when a server goes down? If the server is involved in the operation of the factory in a way that causes the factory to go down when the server goes down, then, yes, you need serious help. If the function of the server is more incidental to the operation of the factory, then perhaps a simpler solution will work. Answers to such questions will determine your course (and cost).

Hopefully it's not that one big company that came to my mind, that provided you with your servers..
EDIT: not referring to MS here, of course.

Well, I assume your 40 servers are already doing backups of their own. If not, your alarms should ring aloud. If they are doing their backups well, and they've being tested once in a while that they actually work (that's one important part often missed when dealing with backups; people do make them, but forget to test if they even work afterwards..), you could also consider backing up the backups in one place. Though in my mind, if the factories are separate, it's better to have them store their own backups..this way you don't lose all backups if it happens. Also if the backups are a little closer to the factories (rather than 5000km away from each of them), it's slightly easier to manage them.

I agree that in this scale (40 factories isn't home games anymore) you shouldn't be just doing the solution yourself. Of course it works if you know how to make it work, take the time and hope that nothing happens before you're ready, but hiring one (or preferrably more than one in this case) person whose profession lies in backup field, saves you money and headache. Also I'm sure your insurance company won't pay a cent if you happen to end up in a catastrophe and they find out you didn't have a proper backup plan.

So..inform us, if you like to continue thinking this here, what the servers are doing (or how they store their data, how quickly it needs to be put back after a disaster in order to continue working = how crucial the servers are) etc., read above for more. Maybe after that we can help. If you have, like you should, a proper backup plan for each of those servers alone, it's not as difficult (maybe) than if you don't have any backup plans at all..

But really, having 40 factories means you're not playing with coins anymore, so you should be fine with hiring some more staff to do this work properly. Either the servers are not important at all and it's ok to you to have them destroyed today without any losses, or they are important and you're ready to do something about it - like pay for getting them secured.