[rescue] Sun T5240 power supply failure?

[Meelis Roos]

>   Of course I can hear the screams of some of the younger admins,

Yess, I can chime in - admin only since 1995 :)

> "Uptime doesn't matter, the only things that matter are availability and
> reboot time!"  Bullshit.

I am afraid of rebooting an old box at $WORK that has had 2000+ days of 
uptime - simply the hardware may have died somewhat and it may not start 
up after a power-off.

Transporting a machine with UPS to save uptime seems like a fun project, 
but not necessary good for the final result (HDD-s and vibration etc).

But uptime itself really does not matter for me - predictability and 
dependability matters.

I want to see my new configuration actually work after a reboot or power 
failure, or the changes are not complete and bite back later in 
unexpected ways.

I want at least known vulnerabilities fixed - software becomes more and 
more complex and security is harder and harder, so postponing patching 
is mostly lying to myself.

If availability matters, I want use a cluster anyway (and if it really 
matters, it has to be geographically distributed). But if I have a 
cluster then reboot time does not matter much - even my old IBM RS/6000 
spending half an hour for POST would do ;)

>  It's no coincidence that the only people who
> say that run Windows. ;)

Nope, running Windows seriously stopped with Windows for Workgroups :)

-- 
Meelis Roos (mroos at linux.ee)