The Design of Software (CLOSED)

A public forum for discussing the design of software, from the user interface to the code architecture. Now closed.

The "Design of Software" discussion group has been merged with the main Joel on Software discussion group.

The archives will remain online indefinitely.

Design databases for lifelong storage.

I look at a lot of these projects on the web now: Flickr, Bloglines, Blogspot, Sourceforge, Technorati, eBay, etc, etc.

There hasn't been a major catastrophe where all data has been lost?  I can only imagine if a large set of data is just lost, what the liability would be. 

But, how do you think these databases are designed?  And, do they have policies to enforce that the data should never be deleted or purged.

Even a day glitch might cause panic.

I know we have backup systems, how often do you test the backups and restorations?

Berlin Brown Send private email
Friday, July 29, 2005
Most sensible firms do test their backup recovery plans, although some don't.

There are lots of stories about the ones who don't. I remember a global telecoms company who backed up dozens of SQL Server databases each night. Of course, when they came to restore one, they discovered it didn't restore, and nor did any of the others.

That was a day I was glad I was an external consultant/supplier and not part of the DBA team ;-)
Activate the Xeon Diode
Friday, July 29, 2005
Most database SysAdmin's I've seen are WAY more serious about backing up the database than other computer users.  Daily, weekly, and monthly backups of the database are common.

Note customers usually don't balk at $3,000 tape backup systems, when they are to back-up that priceless data.  This includes ways of verifying the backup is good, of course.
Friday, July 29, 2005
The last shop I worked in hadn't validated their backup data/methods in 3 years.

They had no way of knowing if a crash was recoverable...
KC Send private email
Friday, July 29, 2005
One of the guys I used to work with handled this by asking the SA to restore a random file from a random date at regular intervals.

Guess what -- the first couple of times it didn't work.  But, as the SA got the idea that the backups were going to be checked, that all changed.

Not rigourous, but it's a start.
BillT Send private email
Friday, July 29, 2005
Sshhhhhh!  This is a valid avoidance technique for the government departments to cover themselves!

It was all backed up, but we never checked and no I can't get that incriminating evidence for you it's all corrupted on the backup.

Tuesday, August 02, 2005
Best/Worst? story that I have is from a few years ago. A category 4 hurricane was projected to hit company's headquarters. Concerned with the storm the company choose to send the disaster recovery team away. Once we started the recovery process we knew we had trouble. The tapes for our backend database that served as the core of the multi billion dollar company were complete garbage. The NT/2k network had good tapes... but they had the wrong drive to read those tapes. That was just the tape issues.

Once everyone was back at headquarters, the debrief session had a list issues that numbered in the hundreds. Included on the list included remember to book hotel rooms, don't use a recovery center on the same coast and consider practice drills before another event occurs.
SB Send private email
Thursday, August 04, 2005

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
Powered by FogBugz