I work for a large company, who have somewhere in the region of 400 server running RedHat Enterprise Linux. This week has been a nightmare...

Last Friday we had a problem with one of our Linux machines which resulted in us having to recover it from scratch. This process took an entire working week to complete and even now we're not in a position where we have a server that is identical to how it was before the problem occurred.

As such, we need to find a good tool for backup and recovery of OS images. For those of you familiar with HP-UX, we're basically looking for something along the lines of Ignite-UX.

There are a number of essential requirements that need to be met by whatever product we end up going for and they are...

* It has to be able to take a live system image, e.g. we don't want to have to have downtime on the box to create a backup image.

* The image needs to be made across the network, probably to an NFS share on another Linux machine, rather than ISOs or to CD/DVD etc.

* It needs to be able to cope with LVM patitions.

* It needs to be RELIABLE!

I've already done some digging around, and had a play with MondoRescue (www.mondorescue.org), but this seems to be far from stable, as each time I've attempted to use it to create a system backup it bombs out with Segmentation Faults.

Granted, I've not spent much time looking into the problem specifically because I've been working on recovering our downed server.

Can anyone recommend (GPL or proprietary) a product that fits the above criteria as I'm running out of idea to be honest?

Many thanks in advance - ech310n

My coworker got MondoRescue a few months ago and has used it successfully to restore a Fedora Core 6 system multiple times. He also successfully used it to load the image on another server (same class) with a different drive configuration and noted that it discovered the additional drives and used them.

I saw mention of Clonezilla and Sysrescuecd today in another forum.

I've just started working in a company that uses NetWorker (http://www.emc.com/products/detail/s.../networker.htm) which seems to be a good product for backup and recovery. Maybe you should look into it if you haven't already deciced on another product.

I don't think his question was about backup/recovery. It was about making a golden image and then using that as the basis for installing new systems of the same type so they don't have to do all the post install tasks manually.

There are OSS backup solutions if he just wanted a backup solution.

Networker isn't OSS - it is something you pay for and not my favorite of the commercial ones. NetBackup and BackupExec from Veritas (now Symantec) are the things I've seen in most commercial environments and what I would recommend over Networker.

In fact NBU has a Bare Metal Restore option (you pay extra for it of course) that would likely do what the user wanted. However, for what was originally asked I think Mondo was a good answer and Mondo is OSS.

Just a quick update... this is a pretty old thread, but here goes. We've now got a really good automated kickstart process for Linux that works great. In the event of the complete loss of a machine, we just rebuild the OS with kickstart to the relevant OS release (as all are standard in their configuration) then restore the application file systems. Works OK. We are also about to being a trial of NetBackup BMR.

Cheers!