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.

I'd like to mirror a drive between two windows servers for fail over/disaster recovery. I assume there is a software package for this. What do you use?

share|improve this question

7 Answers 7

I've used two.

DoubleTake is a common one for applications that don't have much disaster recovery capabilities built in.

Veritas Volume Replicator is a really slick enterprise-quality one. I've seen that used with Microsoft Exchange, for example, to replicate data between two different SANs.

If I had to recommend one over the other, it'd be Veritas.

share|improve this answer

There are several options.

DFS-R is built into Windows 2003 R2 and 2008 (and I use it) FRS is built into 2000 and 2003 R1, but I wouldn't recommend it.

These are really aimed at file shares and are file-level mirrors, not block-level.

If you have a disk that you want two servers to see, then you need windows clustering, but you need a hardware base for that - either a self-contained RAID array (e.g. Dell MD3000) or SAN.

If you need to do this block-level, then you're looking at hardware solutions from SAN providers - I've used Dell EqualLogic, but there are dozens in the market, all the way up to the linkes of EMC.

If you're trying to mirror something like an Exchange server for backup, then a backup system that integrates with Windows Shadow Copy would be the right approach. I know that BackupExec (now part of Symantec) have one of these.

share|improve this answer

I implement this by running Windows in a Xen VM under Linux (CentOS) and using DRBD to mirror the VM's disk images. Okay, so it's an extra level to figure out, but it is free and gives you the flexibility to start adding further VMs to your machine. I also use Linux LVM to make easily resizeable disk volumes to use as VM disks. Hardware wise, other than two machines there's not really anything else necessary, although an extra network card per machine dedicated to DRBD is always handy.

share|improve this answer
    
Is there a good writeup on the web for how you set this up? –  Nathan Feger Sep 26 '09 at 23:56
    
I have my own sketchy checklist of notes assembled after searching Google for solutions to problems as I came accross them. I use that to set up a new server - I was using it today, actually. That is for CentOS 5.1 though, which is several revisions behind being current, and very specific to our setup. I'll see if it's worth putting those notes on the web somewhere. –  David Hicks Sep 30 '09 at 23:20

We use MirrorFolder for this. It's file level, but it's affordable ($40) and easy to configure and use. We use it on server doing site to site backups of about 250GB. It manages differentials and gives you logs of progress. Good luck!

share|improve this answer

One thing to consider is what exactly you are protecting against.

If you are protecting against data corruption or accidental file deletion, you do not want a continuous-sync solution, as the corruption/deletions will be replicated to the backup before you can benefit from it.

If on the other hand you are wanting to achieve 100% uptime on your web farm while rebooting one of the servers - synching every hour may not be enough.

share|improve this answer
    
The intended protection would be against a significant hardware failure (like multiple failed power supplies, motherboard blowing a capacitor, multiple drive failing in the RAID, etc.) on a system that absolutely, positively has to be up and running for critical periods of time. –  BIBD May 26 '09 at 15:33

VVR is an option. I've also designed solutions where the critical data sits on a SAN and can simply be remounted on a different server if the server goes down.

As always, it depends on the application and your requirements for RTO/RPO.

share|improve this answer
up vote 0 down vote accepted

I ended up using SteelEye's Data Replication and High Availability suite. They have version fro Windows and Linux (which wasn't a requirement). The Data Replication protion allowed for block level replication (Synchronous or Asynchronous). The High Availability portion allowed services to be switched over from once server to another.

share|improve this answer

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.