HE.net, Fremont-2 outage?

The mildly conservative estimate is 21 hours minus the guaranteed
turnaround time for your fuel vendor to show up, minus some more
fudge factor to allow for someone to actually hook up and actually
refuel, etc.

The paranoid conservative estimate is more complex; you have to assume
you call the primary vendor, they don't show, and then you have to
call your backup(s). If you have a three hour guarantee in the contract,
you have to remember that this can still represent some scrambling by
your vendor, and if you're lights out, it's quite possible that others
are as well, and hospitals and city hall might rate as more urgent.
It's also possible that the truck'll have a flat, mechanical problems,
or try to rush through the railroad crossing about to be rendered
unpassable by a slow-moving freight train. It'll probably take you
an additional hour to panic and call your backup supplier; now you are
a bunch of hours shorter on capacity than you thought.

Of course, a lot of this is simply how you look at the problem. If
we're talking runtime-until-dry, yeah, 21 hours. If we're talking a
practical number of how long can you go until it's proper for some
panic to set in and calls to get made, it's more like half that. :wink:

With power:

N+1 is usually better than N
Best to assume full load when doing math
Things will go wrong, predict common failures
The best plans are still prone to failure
Safety margins can save your rear
etc

... JG

Joe Greco wrote:

With power:

N+1 is usually better than N
Best to assume full load when doing math
Things will go wrong, predict common failures
The best plans are still prone to failure
Safety margins can save your rear
etc

I find that electrical panelboards, busways, transfer switches, etc. are
often put in the category of things that don't need maintenance or
routine inspections. Big deal if you can start your fancy generator once
a month (I prefer on-load weekly) but the in between stuff is in
disrepair or full of mice. Even a simple dusty transfer switch could arc
weld itself to once side of the contacts.

~Seth

And uncommon ones. :slight_smile:

So as part of a major compute-cluster install, we upgraded our UPS and diesel
generator one weekend, and breathed a collective sigh of relief that we were
now safe from power outages and mostly dodged a bullet. We *did* have some
scary moments when we discovered that (a) of the 400 or so disks on our Sun
E10K, about 10 didn't spin up again and (b) several of the boot disks on said
box weren't mirrored. Fortunately, none of the 10 fails were on a non-mirrored
disk. By Tuesday, all the non-mirrored boot disks were in fact mirrored.

That Friday, a bozo contractor relocating a doorway managed to set off the
Halon. Only lost two disks on the E10K. Guess which two? :wink:

And a month later, we discovered that the nice shiny new automatic cutover
switch was wired in backwards, necessitating another power outage to re-wire it
correctly.

So much for safe from power outages... :slight_smile: