What are the odds!

[Lucio]

Arrrghhh! This has to be the day from Hell to be an IT sys admin...

Came in after the long weekend break to discover the power is off, ok, not a massive surprise, it's happened before. Trouble is, it didn't want to come back on. After unplugging every computer in the building, we got power back but the server didn't want to turn on.

Reset the UPS system as it was still alarming, to discover the server STILL didn't want to come back on. Ok, check the UPS and reset it, monitor that's plugged into it comes back up fine but still no server.

Ok, maybe the UPS has been damaged, so I plug a lead directly into the server and still it won't come on. No green LED on the back of the PSU either.


Wouldn't you know it, the PSU has died, despite being on a UPS with power smoothing (a fairly good model I've been told, made by Leibert). On top of that, when we bought the server a few months ago, I wasn't allowed to get a secondary PSU....


What are the odds that a server PSU dies, less than a year after getting it!! Especially given our secondary test server is fine, which wasn't plugged into the UPS.

[peterb]

Slim - but probably betterr than winning the lottery - and someone usually wins that - eventually (just not me!)

Thats the troubl;e with statistics - they count for nothing if you are the statistically unlucky one. Try to resist "Told you so" when you come replace the PSU!

On a more serious note - I don't know if you did a risk assessment when the server was specified, but this is one of those times when it might have been useful to persuade the bean counters. While the liklihood of this happening was low, the impact if it did is high (server unavailable for a working day - cost??) mitigation is a redundant/secondary PSU (cost ???) . Those sorts of arguments and figures can be quite persuasive. (And if they are over-ruled, you can trot them out and suggest that you retro-fit a secondary PSU - as you might want to now)

[switchmode]

Perhaps the PSU did't like the output from the UPS ?

[Lucio]

Perhaps the PSU did't like the output from the UPS ?

I honestly don't know, the UPS has been in place for a year now and apart from reporting that it's in "buck mode" every now and again it's been fine and survived a couple of previous power cuts quite happily.

[Moby-Dick]

Is it a Proliant DL 380G5 ?

[switchmode]

I honestly don't know, the UPS has been in place for a year now and apart from reporting that it's in "buck mode" every now and again it's been fine and survived a couple of previous power cuts quite happily.

It was just a thort, you never know if the output of the UPS is clean, I don't know enough about modern UPS output though.

Maybe the PSU failing caused the power to trip in the first place ?

Hope you get it sorted.

[Lucio]

Is it a Proliant DL 380G5 ?

Close, it's an ML 370G5, the PSU that failed was a 403781-001, had a new part delivered just now and we're back up and running (gotta say I'm glad we aren't relying on HP, they wouldn't be here till about 3pm tomorrow!)

Looking at the server logs, there's no report from the UPS to say it was in battery mode, nor anything to indicate the server was shut down. Best conclusion I can draw is that the PSU failed and THEN took the building with it.

We're gonna get an electrician in to look at things though, because the plug socket for the server has the following

1.5Kw UPS
1.0 Kw UPS for the phones
1 server not currently hooked to the UPS
1 water cooler
2 network switches
1 ADSL modem

All in all, far too much for a standard double socket....

Still, on the plus side, the MD is *now* listening to me!

[Moby-Dick]

wonder if they have the same PSU's as the DL380's we have had something like a 40% failure rate on ours.

[Lucio]

wonder if they have the same PSU's as the DL380's we have had something like a 40% failure rate on ours.

What do you know, it is the same part

HP Parts Store - HP Computer Parts - HP Printer Parts - Compaq Parts

Wonder if we can get a Hexus news story outta this and get a comment from HP. On top of that, their "next day" service is dire, still had no confirmation of when an engineer will be here and despite their parts depot being down the road, they point blank refused to let me go and collect.

After all, these are just plug and play PSU's which took 30 seconds to swap out...

[Moby-Dick]

I doubt HP will comment althought I found they have referred us to a specific team for when we've got failures on those components. Any replacements we've had are fine however.