Microsoft says it had Sidekick data backup, but it was hit, too

Wish this article had more details than just "the backups were impacted too".

A few months ago Google had a large multi-hour App Engine outage. Afterwards (within 24 hours?) they posted a detailed technical explanation/postmortem online for anyone interested to read, with a play-by-play that even went minute-by-minute.

Microsoft/Danger owe their customers more of a real explanation.

> Will this prompt Microsoft to review its policy for data backups? "This was an extraordinary situation [...]. We greatly apologize [...]. We have a full team working in parallel with the restoration team to augment the redundancy and stability of this acquired service over the long-term."

That's got to be the best non-answer answer I've read in a long time. And no - redundancy and stability of the service have nothing to do with the backup policy.

Wrong, wrong, wrong. If you have a backup that CAN be impacted by a problem in the main server then you do not have a backup at all, you have, at best, redundancy.

The very reason to have a legitimate backup is that it cannot be damaged by the same mechanisms that can damage your primary data. It should be inert, if possible, or at least to some extent immutable, and it should be offsite.

RAID is not a backup. SAN is not a backup. Mirroring to a redundant storage array is also not a true backup.

Backups are hard to get right, and expensive. But the cost of getting it wrong can be monumental.

In my experience, new "policies" put in place in these situations don't really do anything. They just inject bureaucrats where there were none before, and aggravate the few people remaining who might really help. And you can bet managers will start a witch hunt, firing some mostly-competent guy as a scapegoat (which may "feel good" at the time, but ironically will only slow down their recovery and set them up for their next failure).

The smart thing would be for Microsoft to think of the employees: what management decision is likely to make working in that group least likely to suck over the next few months? If it were me, I would invest in technology and hires/transfers, not in bureaucracy. I would give that group everything it needs to build and rebuild that system twice. After all, doesn't Microsoft have the money for that?

You see teacher, I have two dogs...

While this is very unfortunate for SideKick users I suddenly feel pretty good about my own backup solution. Redundancy, rsync weekly to a live volume, multiple offline backups including one off-site (although it's a few months out of date right now. The important data exists on at least half a dozen sources at any given time. I am of course referring to my personal MP3 collection.

Reminds me of something happend to a friend.

Friend: Sir, my hard disk partition crashed and the project files are all gone.

Lecturer: Don't you have a back up?

Friend: The backup was on the same partition.

I find it almost unbelievable that there were no cyclic backups.

This highlights a very reasonable fear that people and corporations have about the cloud.

It might even kill off some of the smaller cloud-based players...