up_the_irons: I'd definitely pay attention to a price table.  (Bigger service providers often have a "we'll fix you up with all the Cloud you can possibly stand!" web page with little or no pricing info, and that's the sort of thing that seems obviously too expensive to pay attention to.)
I can get whole e3 systems for $50 from some of the more budgety providers.  I'd still pay attention at $50 though, just because I know arp to have its act together.  :)
where the Xeon Ds at? :)
What product do folks recommend for auto power cycle on ping failure
First, a warning: Ping failure isn't the most reliable way to detect if a device is down (false positives and false negatives alike)
Second, I used a remote rebooter ages ago that had that very feature. I'm trying to remember its name...
(Damn, I'm really drawing a blank on the name, either vendor or model.) It was a lesser-known name, a small company, some guy's garage, that sort of thing. Sorry
nathani: As an old sysadmin, I'd grab an APC PDU from the scrap shelf and write a contact script for nagios which pokes the PDU. That might not work for everyone though
You could probably do the same with a Smartthings hub and a z-wave outlet, if you wanted to go all IoT on it
Or, you could get kit that doesn't crash and need power cycling ;)
HA! I remembered the name! http://www.synaccess-net.com/np-0808dt/ (not this exact model, but you get the idea)
(They don't seem to list the device I used, an NP-08, but it's basically all the same)
thanks brycec
thanks plett
you could use ipmi
if it's a server that is
its a workstation
dont think it has ipmi
lenovo s30 I think
with linux?
modprobe ipmi_devintf
ipmitool sensor
then you'll know if it has ipmi or not
its running esxi
ahh
some stuff has ipmi without having fancy lights out stuff
the whole box crashes like once a week
hmm
I would think ipmi would go down with it
prob not
it's probably power saving related
is my guess
most workstation/server crashes seem to be :)
there's been a few bugs over time where servers can get unreliable at low load
memory and overclocking are more suspicious on desktops
could it be old capacitors on the motherboard?
i doubt it
what cpu is it?
6 CPUs x Intel(R) Xeon(R) CPU E5-2630L 0 @ 2.00GHz
6 core rather
which version is that?
not sure
not coming up with anything on google suggesting my theory :(
was it always crashing
or did it start crashing?
I think always
only bought it within the last year or so
the place I bought it from said bring it back they will do some diagnostics
Sandy Bridge EP?: http://ark.intel.com/products/64586/Intel-Xeon-Processor-E5-2630L-15M-Cache-2_00-GHz-7_20-GTs-Intel-QPI
not very recent in that case
nathani: Is ESXi purple screening? If the contents of the purple screen don't indicate a driver issue, run memtest on it overnight as it's most likely bad ram
(I'm guessing you're not running ECC ram in a glorified desktop machine)
plett: workstation usually means ECC
And E5-2630L afaik doesn't support non-ECC ram
dont have a monitor hooked up to the workstation
however it crashed after the same amount of time when running both windwows and linux
was linux a hard crash or did it show kernel panic
hard crash
I had run a few iteration of the memtest disc earlier
can run it overnight
maybe a bad mbd
Since when? If anything, workstation=nonECC, server=ECC as a rule of thumb.  15:29:08       @mercutio | plett: workstation usually means ECC
brycec: workstation to me usually means ecc non-registered, single cpu of server style cpu
anyway, memtest should say if he has ecc or not
do you have ecc, nathani ?
I equate workstation with desktop
Same.
oh i don't
i mean you run desktop apps on it, but desktop means something special now
it'd be nice if non-ecc wasn't a thing anymore
yeh
even if motherboards could just do ecc with normal memory
motherboards/cpus/whatever