joining from home instead of my vps which is currently down
I'm convinced there's something beyond my software considering both vnc and serial are not connecting
even after a hard poweroff and subsequent boot
gah, network is crawling dog slow again today :(
I can't effectively do anything on my VPS - it's responsive, but only barely.
there's a 10+ second delay for each keystroke over ssh
raptelan_: mtr -4rnc 50 <your-ip> :)
raptelan_: do this from where you are and pastie / pastebin / ...  the results
toddf: replied to your ticket
thanks up_the_irons .. I rarely point fingers, but in this case i'm dubious to believe software that has been running fully stable since february with no changes whatsoever suddenly has problems even booting; will investigate further though
I can confirm that between conserver needing reset (heh, 1st I've ever heard of that one) and me forgetting 'disable mpbios' on an upgrade I did hoping to find a better behavior after the upgrade .. we might have ourselves a reason to not respond well after the initial boots, maybe the console not working and he.net IPv6 issues made me think the system was truly not responding yesterday .. perfect storm as it were .. will see after I get things 
hmm, ok so same behavior as last night, boots up fine, see the login prompt on the serial console, and without further action from me, suddenly I hit enter at the serial console and conserver says the connection to kvm is down
up_the_irons: well of course now it's okay again.  My connectivity from here to everywhere else I could think to try though was good.
running it anyways since there is still some slight delay, but nothing terrible like it was.
up_the_irons: http://fpaste.org/FTqG/
urgle, and now with no action from me *again* I note it is down on conserver and network again
I'm open to suggestions as to what I could be possibly doing wrong; guess I'll ask for the log of when the vps shutdown request was given because I'm not giving it yet it is acting as if someone is telling my kvm 'kill -9 suckah'
ok so has anybody ever experienced anything where the kvm process would exit reguarly?
toddf: i'm in the same situation though.  i have a full server, lots of customers on kvr22, and you are the only one saying you're having problems.  nobody has opened a ticket on kvr22, not one.  if there's a problem, everyone is awfully quiet
toddf: do you remember when we added the 2nd interface to that vm, for the backup server?  maybe we should remove it temporarily and see if the problem goes away
we added that roughly a month ago.
if you wish to convert both nics to intel that would be one thing to consider
it is not responding now
do you see what happens when kvm exits?
i see it shut off now
i only see what is in kern.log and dmesg
(and i looked, nothing interesting)
toddf: ok, switched back to intel NICs and started it
nice thing, i had vio0 and em0 in a trunk(4) so its responding again w/out me doing anything further ;-)
so whatever exec's kvm doesn't log its stdout/stderr ?
(I know some version of virtd or whatnot, but I am not familiar at all with that)
it is seriously from my perspective something like a userland activity that tickles something badly on the kvm userland app which causes it to fatal() or exit() or somesuch
thing is I have no clue what, before I upgraded in the middle of this, i had not changed anything since feb, and it has _never_ done anything like this before
there's nothing in userland doing anything though.  only i have access to the server and i don't tickle VMs ;)
I'm expecting its some obscure thing that we could bash our heads against till time eternal if we chose and maybe then we wouldn't figure it out
I'm not saying userland on the host I'm suggesting something in my guest may have finally tickled kvm just so (theory, not fact, as I didn't change anything, so its really hard to tell)
i can run the kvm process manually (like in a screen session) after it crashes again.  then when it exists, i'll see the output
i'm heading to the data center now... (NTT circuit termination!:)  .. bbl
ack
toddf: did it die again?
splitz
toddf: according to ps, your VM has been running for about 4 hours w/o crashing.  is this better than before?
I should get a t-shirt "I survived the netsplit"
up_the_irons: indeed it is starting to look promising