jsbowden: (Default)
jsbowden ([personal profile] jsbowden) wrote2008-07-23 12:46 pm
Entry tags:

Grrr...

Can I start killing users?

Monday, I replace all the tapes in the library, noting that for some reason, all the blank tapes I put in it last Wednesday filled up before full backups ran Friday night (I was out Thurs. and Fri. of last week, but being a good systems guy, I made sure things would continue while I was gone). Two days later? All my tapes are full again. So I start looking around, and what's this? Someone has taken the NetApp from 45% full to 80% full. It seems that one of our engineers decided to automate backing up the output of his MatLab runs. All 400GB of them, all reproducible, every night. Despite the fact that they don't change, but hey, NetBackup doesn't know this, it just sees a brand new 400GB .bkp file every night. Of course, all this data is raw numbers. It's ASCII. This...person...is dumping 400GB of unchanging ASCII into an MS backup file every day. Uncompressed. Filling up my file servers and fucking up my backups in the process.

I swear, I'm going to start throwing people off the fucking balcony. Their brains won't noticeably work any less after the 10 story drop.

[identity profile] theweaselking.livejournal.com 2008-07-23 07:13 pm (UTC)(link)
That's the point where you make a quick business case: The cost of the extra tapes and your time and all the other hassles, as well as the potential for this to cause problems with the REAL backups in case of emergency (large), versus the cost of a 500GB external HDD for this specific guy (~$100).

This is a no-brainer kind of problem. Which is naturally why it's being caused by a user, and why my solution will not be adopted.