[clue-tech] removing foot with gunfire.

[Jack Parker]

(follow up on losing / disk and rebuilding)

I've got the toothpaste back in the tube - I wound up losing /etc which is a
pain, but I've managed to preserve my data and code.  A useful disaster
recovery experience.   It took me way too long and I lost stuff.  I would
find it useful to see what y'all think in terms of planning for the next
one.

1 - Losing the / disk took a long time to recover from.  How to mitigate?
   - Mirror /
   - Keep a hot swap ready (but then it falls out of date)
   - standardize on never putting anything of value on / - keep it all
elsewhere.  So that a minimum / can be thrown together and then a quick
script run to copy everything from a savepoint back into it's proper
location.

2 - Back up - no brainer.  This is my 4th scare in the last 4-6 months.
Yes, I'm too loose with my procedures and half the time the skin is off the
machine.
I like Brians suggestion for either USB/Firewire disk or even a dummy
machine which serves as file storage to back up to.

3 - Decide to have a 'production' machine which is not touched and have a
second machine to play around with, but then my work is all sandbox work -
hmmm, separate sandbox A (work - somewhat controlled) from sandbox B (I
wonder what this does).

What have I missed?

scares to date:
  - today, lost /
  - 2 weeks ago, left the window open over a cold night, SATA contacts
contracted and the primary disk disappeared.  Again, took me too long to
figure out that the BIOS shifted the boot order because the primary disk was
missing.
  - 2-3 times eth0 has 'dropped' and I've had to play games to get it back.

Regards,
Jack Parker