Journal Chacham's Journal: Today's First Ramble 2
Well, I got hired to work about 15 days on Oracle. So, working on the test db, and I move the rollback segments to the new drive. As I drop a large column from a table, the RAID controller goes bezerk (mailbox blocked) and the DB is basically trashed. Yeah, I could have recovered, but reappllying what I just did would haven taken way too long. Besides, recovering from a bad rollback segment is hard. The database reported a bad block. So, we decided to get the DB from backup.
Lo and behold, the backup was deleted because a new backup (of the live DB) was being downloaded. "Hey, we don't need this anymore!", sayeth the SA. I'd call him an idiot, but the SA will probably read this
Trashed most of yesterday's work, and today I am waiting for the new backup to be rolled out. A new driver is being compiled. Apparently, the error is fixed in a new RC of the driver.
So, the kernel is being recompiled with the new driver. Hopefully I'll be working soon.
I'd call him an idiot... (Score:2)
No, go ahead and call him an idiot. What kind of dumbass sysadmin uses the ONLY CURRENT backup tape, to do a backup?!?!?! We have 1 weeks worth of tape rotations. Last night's is kept off site, the rest in a fire proof vault. So you can do a backup on Monday, trash the db on Tuesday, and not realize it until the following Monday, and you still have a good backup. (and 4 bad ones, but the good one is what matters.)
Re:I'd call him an idiot... (Score:1)
The backup I was talking about was on the local test machine to be ungzipped into place. That backup was whacked before a fresh one came in.