Wednesday 4 November 2009

Preparing for the Distinct Possibility of a Server Disaster

Besides being an awesome method for migrating SBS to SBS on the same or different hardware, Jeff Middleton’s Swing method brings about a whole new way of viewing disasters and disaster recovery situations.

Because of our previous experiences using the Swing method in past recoveries, we are in the process of preparing our OS backup-less server for the possibility of an all-out failure.

As mentioned in our earlier post this morning on the missing backup and the previous post to that we had a hardware hiccup on one of our client’s SBS servers. The hiccup is enough to cause ShadowProtect to not be able to take an image of the OS partition.

So far, despite the RAID controller throwing the fatal error, no NTFS 55 errors have shown up in SBS’s Event Logs. Hopefully this means that the bad section on the drive is in an empty area. We will be flipping out the array members this afternoon and rechecking the consistency to see if the error has disappeared.

Array member number one’s status:


Array member number two’s status:


Note the distinct lack of error counts for both drive’s media as well as a zero for the Predictive Fail Count on both drives. This is indeed a real puzzle.

So, here are a few of the things we are doing to prepare for the possibility of the main SBS crashing:

  • We have moved the live Exchange private and public stores to the data partition that is being backed up.
  • We have installed Win2k3 using the SBS media into Hyper-V on the second server and using the Swing method have introduced it as a DC on the network.
  • We have the previous backup drive with ShadowProtect Images that can be used to restore SBS into Hyper-V thus providing us with a frame of reference for the setup.
    • It is a little too far out of date to use as a recovery source though . . . maybe.

The second bullet will thus eliminate a step if we need to do a fresh build of the SBS box and re-introduce it back into the network using the Swing method.

We will be that much further ahead for recovery if this happens between the temporary DC and having the existing server recovered on our own lab Hyper-V box for reference.

In this particular case, our client’s second server also has a mirror of their own client data on it. We will be able to mount the ShadowProtect image from the good SBS backup and copy the changed files over using BeyondCompare.

While we are not in an ideal state with this particular server situation, we are in a position to expedite a recovery in short order thus leaving our client with very little down time.

ExchangeDefender will be very helpful in this situation too. If things do go sideways and the stores take a hit, we will have access to our client’s LiveArchive so that they can keep active with their e-mail.

More to come . . . hopefully a happy resolution with the new replacement drives cleaning things up!

Philip Elder
Microsoft Small Business Specialists
Co-Author: SBS 2008 Blueprint Book

*Our original iMac was stolen (previous blog post). We now have a new MacBook Pro courtesy of Vlad Mazek, owner of OWN.

Windows Live Writer

No comments: