Sunday, May 30, 2010

Dual Drive Failure

I've always know that a solid backup strategy is important. In my roles as MIS Manager or LAN admin, I always knew that no storage model is without flaw. I've seen issues even with fully redundant SAN solutions where critical issues still impact the sanctity of storage. 

Last week, the issue hit closer to home. I was away on a trips and I returned to an array failure. At home, I like to run a RAID 1 array to protect my music, documents and video files. I like the security of having one drive be an exact copy of the other. That way if a drive fails, you still have your data. The only point at which this falls down, is if both drives fail at once.

About a month ago, one of the 1TB drives in the array failed. I contacted Western Digital and they quickly verified my drive's warranty and I was able to select advanced replacement shipping - all on their web site. It was great. I put the new drive in my computer and then the array rebuilt in about a day. After that, I was all set. However, after running the new drive for about 2 weeks, I came home from a trip and found that not only had the replacement drive failed, but so had the original drive.  I guess Google was onto something with their drive failure study. Clearly, this left me with a total data loss situation.

Luckily, my Windows Home Server backs up my computer every day. In addition, I have a subscription to Mozy which keeps my backups off-site.  The negative with the Windows Home Server is that it seems to only allow a restore to the same drive letter. Since that array had a total failure, that wasn't an option. Mozy, however, allows you to choose your destination. So, I began a restore from Mozy to one of my other 1TB drives. That was a week ago, and it's still running. I believe the restore I've requested is almost 200GB so I fully expect it to take awhile.

Why I am wasting so much time downloading this when the data is on my home server? I have little faith in redundancy right now, so even though the home server is using data protection, I'm feeling better getting a second copy. In the meantime, I've sent both drives off to Western Digital. I hope they get back quickly - as Advanced Replacement was inexplicably not an option when you have a problem with two drives.

6 comments:

  1. I had my hard drive fail in my iMac when I tried to boot it up after returning from a long weekend vacation. It had shut down normally before leaving, but upon return, I got the clicks of death. Obviously, my drive had suffered a mechanical failure.

    I took the iMac to the Apple store and had one of their techs look at it, though I already knew the drive had failed. I could have had it replaced free of charge, since I have Apple Care, however, they required that the drive be sent back to Apple and ultimately to Western Digital. So theoretically, my data could be lifted from it if the drive were taken apart and the data read off the platters in a clean room. Therefore, I opted to replace the drive myself and keep the "dead" drive. I was uncomfortable with handing over my drive with my data on it to Apple.

    Luckily, I already had numerous backups of my data on two external drives. Only a very small subset of my data was lost. I have always said to myself ever since I completely lost a 10 page paper in the middle of the night back in college, "back up often and in multiple places."

    ReplyDelete
  2. You have a good point and it's a wise recommendation for folks NOT to return a hard drive with personal information on it to a vendor. However, in my case, since these drives are members of a mirrored (RAID) volume, you usually would need to know the type of RAID controller I used in order to reconstruct the data. As a result, I felt better about the exchange - both times. :-)

    ReplyDelete
  3. Drive failures are a terrible reality. I had one fail about two years ago. Thankfully I was able to retrieve most of my data with the assistance of Drivesavers. Very expensive but hey I did get most of my photos of the children back. Do you like Mozy. I have 2 redundant copies of everything but I was considering using Mozy as well.

    ReplyDelete
  4. Replacements have arrived! Thank all that is holy! I'm formatting now and hope to be able to begin a restore from Windows Home Server soon.

    As for Mozy - I do like it. The interface is easy and it is great for restoring items here and there. The software is pretty configurable in terms of what to backup and when - as well as how much bandwidth to take.

    If you want to give it a try, they have a free version with 2GB of backup/month.

    ReplyDelete
  5. I have never heard of Mozy. I think I'll give the free version a try to see how it works. I am currently consulting to a law firm that has been using Carbonite and it supposedly works well, though it takes some time to backup a large amount data. I'm not sure if it is an issue with their broadband connection bandwidth, the route that data stream takes to get to Carbonite's data center(s) or some issue with their PCs and/or LAN. Maybe Carbonite's network tends to get congested.

    Have you heard anything about Carbonite?

    ReplyDelete
  6. Mozy and Carbonite are based on the same business model. Both are limited by the available bandwidth at the customer site. For example, the initial backup for Mozy took multiple weeks for me as I was backing up about 120GB at the time. Now, I'm backing up almost 300GB. It would probably take a little over a month to get the initial backup done. After that, Mozy/Carbonite should only backup what is new or changed each day as that is a much smaller amount of data.

    Restores do go faster as most people's Internet connection are faster to you than away from you. But, again, that's all limited to your data connection.

    ReplyDelete