we're professional idlers
this would be a good channel for idlerpg, though :P
zeshoem: the fact that we can sit here for days on end, idling (many from our vps's)  says a lot about the stability of the service.   otherwise,  we'd be pretty vocal and complaining,  OR.. not idling here at all cuz we've cancelled and moved on :P
heh
jpalmer: lol, good point :)
what's idlerpg?
up_the_irons: it's a game where you level up by idling for long periods of time
LOL
up_the_irons: a rocket propelled grenade that hasn't been loaded into a rocket :)
requires exactly zero input or skill from the user
lteo: hahaha
Lefty: not saying anything at all takes skill for certain people...
jlgaddis: updated your ticket
thank ya sir.  i should be able to get the install done tonight, it's just a matter of firing off a kickstart.
cool
jpalmer: do it
then share it with me...
;)
lol
I'm actually building a kvm and libvirt module now.  I'll be using it to setup some kvm hosts.  but beyond that, I'm sure you have some custom/proprietary scripts and settings that I wouldn't know about ;)
up_the_irons: my goal though,  is to provision a machine from bare metal to a full working KVM host,  100% automated.
and then from there,  using libvirt.. I can automate the provisioning of VM's
up_the_irons: done, thanks again.  ping me after you drop it back down and i'll do another hard shutdown/boot.
jpalmer: nice, that would be a cool setup
jlgaddis: just reply to the ticket so it goes in the queue; will drop it down later tonight
*nod*  already did, just mentioned it here in case you wanted to do it tonight.
s/tonight/now/.
thanks again, much appreciated.
np!
up_the_irons: if interested, I'll share the git repo with you when done.
up_the_irons: I'll be doing it for CentOS.  so you'd have to change it for ubuntu IIRC
jpalmer: right..  definitely would like to take a look!
but I'll make it so that multiple OS support would be trivial
kickstart + puppet?
i'm about to work on the same sort of stuff for this isp
jlgaddis: well, I'd use a tool that dynamically generates the kickstart, and integrates with puppet (foreman)   but basically, yes
jpalmer: u use puppet and foreman together?
yeah
ahh, i just planned on a generic kickstart setting up the stuff that's the same across all hosts, then using puppet to take care of the host-specific stuff
*nod*  lots of people do that works for many.
I just like the 'cradle to grave' aspect of foreman
that, and it's reporting is fantastic.
is it hard to maek a full working kvm host?
mercutio: not hard,   but when you have a lot of them..  like any farm, it's hard to make sure they're all configured identically without automation
yep, and to really be able to manage them at that scale, you *have* to automate
ad it's nice to be able to handle maintenance and upgrades in 1 place.. and have them applied to all relevant hosts
I'm sure up_the_irons has a ton of scripts he uses.
yeh i would figure maintenance would be the hard part
oh yeah, has anyone else noticed vnc issues when using the most recent tightvnc?  using an older version seems to work better.
otherwise was getting weird disconnects something about psuedo encoding.  and if reconnecting a black screen.
git + dsh are my best friends
i'm just getting to the point where i need something like puppet
up_the_irons: when I get this setup,  I'll let you attach a raw host.. and see if it works for ya ;)
jpalmer: tnx!
well i suppose one advantage of having lots of hosts is that they can each have the 496 day issue so it doesn't seem as random
(that effected multiple hosts right?)
mercutio: i wouldn't call that an advantage for the 497 day issue ;)  But after it happen a few times, and I noticed the pattern, it did give away the 497 issue.  i probably would not have figured it out if i just had, say, 2 hosts
a bug that happens only every 497 days, well, that is indeed a hard one to track down...
i'd say it's an advantage in some ways.. versus completely diff bugs with each host.
up: i just googled it - adn apparently linux 2.4 had issues with it
it's something to do with jiffies... with hz and timestamps
like a wraparound
so in theory it should be able to be checked for.
but then, who's going to check for that?
windows 95 used to have a bug with like 47 days of uptime i think
oh windows 95 bug is 49.7
and strangely that wasn't noticed that frequently...
jiffies!
i wonder if that issue has mutated with dynamic ticks.
mercutio: yeah, i already know _why_ it happens; i would just like there to be a firmware fix to correct the problem.  otherwise, i simply put on my schedule, for each host, after integer multiples of 497 days, to disable the watchdog around that period.  a simple workaround, but really something i should not have to worry about in 2012
cd $food
oh it's a watchdog timer doing that.
a fix would be nice yeh.  it would be frustrating having to do heaps of reboots when there's no power failure or any real reason