- This topic is empty.
-
Posts
-
Yes, I know, old topic. I just wanted to add a few things.
I just skimmed over the article, but I did read the section on data loss. I know from experience that data loss is a pretty ugly thing. I had a cheap power supply (my first mistake) fry, and took out 2 hard drives — one had just programs, and the other had all of my work, writings, art, and data on it. And, I did not have a recent backup (my second mistake). The thing is, I needed this data. So, I found a data recovery service. After contacting them (and getting a $300 discount for being a poor college student), parts came to $400, and labor was $700 – a total of $1,100!! I got my drive back with a 99.9+% recovery rate (only 1 file was corrupt). I was happy, but my bank account was not.
I learned 2 good lessons from that: one, my data drive is now kept in an external enclosure, away from the internals of the computer (although I only run quality PC Power and Cooling PS's now); and two, I BACKUP my data, and I do if often (not as often as I should though).
So, Tim made some real good points in that article about saving your data.
Another thing to mention here is verification of your backups. Its one thing to implement a backup strategy. However, it is INCREDIBLY important that you check your backups often to make sure you're getting a good backup.
About a month ago, we had a customer who lost power as a result of Hurricane Ike (if you didnt read about it, the resulting low pressure snuck around to hit most of the mideastern part of the country, knocking out power to many states including Ohio and its neighbors.). They were without power for several days. When the power came back on, they found they couldnt get into the software. After some research, I found that the database server installation had become corrupt. “Thats easy, just reinstall then restore from backups”, I said to myself and to my boss. Well, the database server installation wasnt the only piece that was broke. A bit more research found that the physical data files had torn pages as a result of losing power. The Friday before this customer lost power, I had made a change to the backup plan to resolve some disk space issues. The change I made worked one night, but not the following night. The only good backup we had was a day older than what we would have liked. I spent two days attempting to recover as much data as I could, but the end result was that I was not able to recover everything, and the customer had to re-key some data.
There are two lessons to be learned from this.
1) Make sure that your backups are working as expected. If you need to have an up to date copy, check your backups often and report problems with the backup as soon as you see them, especially with automated backup plans. Chances are, if caught soon enough, a manual backup can be taken and then the problem can be resolved.2) On mission critical systems, always make sure you have a battery backup device connected to your server. Also make sure that it is configured to shut down the system whenever it finds itself using the battery for power, instead of the electrical outlet. This will allow the system to power down normally, preventing any kind of loss of data. Once the system has shut down, UNPLUG it. Leaving the system plugged in can also sometimes result in a surge of power that the system doesnt expect, causing hardware failure.
If there is any one particular thing I hate about working in IT, its data recovery on a failed system. Backups make the job a lot easier. Had the power went out before the backups ran, there wouldnt have been anything I could have done other than what I did. However, had the battery backup been configured to work as suggested above, none of this would have happened.
The story from MI: http://midev.mightorindustries…..#038;id=37.
Have comments or additional suggestions, please post them here.
- You must be logged in to reply to this topic.