This is the story of the most spectacular data recovery operation I’ve been involved with at Data West. It happened in 1978. The details are fuzzy, and I may not understand what actually happened, but somehow the drive device had slowly and routinely overheated for a while and the drive heads had written to a slightly different place than they should have. The device finally failed, and after it was repaired, I discovered that none of the backup platters could be read, since they had been written to incorrectly. The client had lost all their data. It still boggles my mind that we figured this one out. The hardware repair person and I noticed that we could put a platter on the drive and it could be read for just a few seconds before reading failed. We tried powering down and cooling off the platter and then starting up the drive and reading it correctly again, but it didn’t work long enough to help. I figured out that I could copy a few kilobytes at a time from the platter, but it would have taken weeks to do that, with no guarantee that the technique would work consistently enough. So we got the bright idea of putting the platter in the client’s refrigerator-freezer for a few minutes and we could speed up the time between insertions of the platter and could lengthen the time before failure of the platter to read. Instead of weeks, we were able to restore all the backup platters over several hours, and we literally saved the day. Lessons Learned:
- Hardware is WAY more frustrating to deal with than software.
- Always keep your cool when you’re working with a client’s data!
Thankfully, those early disk drives were replaced with much more reliable ones in the 1980’s and it became much easier to work with backups. This concludes the c’mon back chronicles. If you liked this type of information, let us know and we’ll continue the series.