Networks

Ask something
Question Locked

Disaster recovery testing question!

0 helpful
10 replies

by unimaker2000

Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

We successfully setup a windows server 2003/exchange 2007 server with RAID 1 configuration:
Disk 0 (RAID1) - OS/Exchange install
DISK 1 (RAID1) - Logs
DISK 2 (RAID1) - Database
Before we proceed to the implementation, we want to test our disaster recovery.
We have backup exec 11d with exchange 2007 agent on another server, backup of the exchange 2007 server is successful even to the GRT level.
Now what we are planning to do is to crash probably DISK 0 so that we can test if we can recover it from our backup tape.
Our worry is we want to have an original copy of the disk before we crash it, so that if anything happens we still have the original copy.
Question is, how can we do this? I am able to boot from a BartPE cd with drivers for the server so I can basically image the disks. Shall we go with this approach? Or can we break the mirror on DISK 0 so that the first can be the copy and the second can be crashed for testing and after doing the test, use the first one to build a mirror again. If this is possible, how do we do it?
Sorry if my message is pretty long, but any inputs are very welcome, THANK YOU!

10 replies

by Former member

If you have a second server make sure everything is "mirrored" (copied) so as when you pull out the disk(s) and then reload the server should do a re-write to the disk that you have put in. If you have (on the server) hot swap drives then all should be ok because you should not need to crash it, just take out a drive and replace it with a fresh drive, the new drive should be automatically activated and then your data will be copied over.

Please post back if you have anymore problems or questions.

by unimaker2000

I think this test is only good if only 1 of the 2 physical disk of RAID 1 has failed. But how about if the actual 2 physical disk that make up the RAID 1 fails, how are we going to recover from that?

by cmiller5400

Rebuild

If you want more redundancy look at RAID5 or RAID10.

by tintoman

In your position I would do exactly what you are proposing, image your disk, swap it with the imaged disk to verify it works and everything runs normally, then trash one and try to recover, no drama

by unimaker2000

Ok so after I image the virtual disk, I will restore the image to the same virtual disk to test if the image works. Question! On the restore process, will the mirroring be automatically done at this point? Because if it does then that would be disastrous in case the image doesn't work. The original data on the virtual disk will be permanently lost!
Or after I image, shutdown the server, take out the second physical disk of the array, then boot to cd again, restore the image to the first disk. So in case the image doesn't work, I still have the second disk which have the original data. Will this work though?

by unimaker2000

Didn't get what you mean on the first line. But anyway, in any RAID config there's still a possibility of a major disaster of multiple disk failure at the same time (even if it's somewhat unlikely), where rebuilding won't work anymore.
Now I ask you the question, what would you do if you are in that situation where your backup tape is the only thing that holds your data?

by cmiller5400

(the first line was meant as a joke... well anyway...)

I would keep an image (using Ghost or Acronis) of the machine so that all you need to do in the event of a major disaster is throw the image back down on the machine, and then restore the user data back to the drives. I would probably update this image several times through out the year, especially if you make changes to it frequently.

by unimaker2000

Firstly, thanks guys for all your input!
Since we have a full backup of everything, what I did is shutdown the server and pulled out the secondary disk of virtual disk 0 (RAID 1), powered up the server then reinstalled windows 2003 with sp2 to primary disk. Installed the backup agent and then did a restore from the media tape. The restore went well and restored everything. Test is completed and successful. Shutdown server again, booted to raid bios then failed the primary disk. Shutdown the server again, pulled out the primary disk and placed the secondary disk back, restart and booted to raid bios again and "forced online" the disk. Booted the machine as normal. Shutdown the server again and placed the primary disk back again, boot to raid bios and rebuild! Then boot to normal. Now's everything's in the original state! Whheeeww!!! :)

by Former member

Nice to know that everything worked to your expectations, but i must suggest that you always make a backup of your backup, i have been doing this for eight years and it has not failed me yet. I always test the backup of caurse and then if all is well i back that up (external hdd's).

Please post back if you have anymore problems or questions.

This discussion is closed to new replies.