up_the_irons: lots of fun, it is a new toy for me!!
does the new beta system support newer libvirt that has the reboot option instead of shutdown/boot only?
That would be nice... feels very clunky not having a "reboot" option
bryc3c: it was confirmed in the recent past that the libvirt in use by current production arpnetworks systems does not have that as an option. not an arpnetworks issue per se, I'm just expecting the new beta, having incorporating newer kvm and such would also have a newer libvirt that supports that
if you had a 'hard reboot' option you could keep your vnc connection as opposed to attempting to time it to still see the f* key sequence to boot off cdrom, for example ;-)
thanks toddf, I wasn't holding it against ARPNetworks
And fwiw I find the "press F12" screen to last too long (about 5-10s)
iirc it was a bios compile time option. not sure if a support@ request to have a faster timeout would result in any action but I'm sure they'll tell you if that is doable.
if the web ui and/or ssh ui were to permit specifying 'boot on cdrom' then the bios timeout would be moot and un-necessary
akin to 1and1's options on physical systems 'reboot into rescue image x' 'reboot into rescue image y' 'reboot' etc
done via tftp of course but anyway
up_the_irons: if i order a Power Up later on, will you grow the existing disk size or can it be a new block device presented to the vps?
ryk: growing existing disk is definately an option... will let up_the_irons answer on the new block device angle ;-)
ryk: RTFM :P http://support.arpnetworks.com/kb/vps/how-do-i-upgrade-or-downgrade-my-vps
(ryk is one of the many people I know from elsewhere. I'm not normally that mean)
yes you are
you are always a big meanie
ryk: there's a difference between being a fair op, and a meanie
 /kick ryk
damnit
ryk: yep, we can do either.  tnx bryc3c  for the link
bryc3c: the timeout was increased since there were so many complaints it was too short
up_the_irons: Yeah I'm not complaining... 'tis just my opinion. But how often am I rebooting my VPS? :P
up_the_irons: by the way, a long time ago we talked about backup
i happened to call rootbsd before i placed this order, because they're pretty local to me
they just have a bug storage box with chrooted homedir's that everyone can ssh to
*big
somehow I recall prices listed with ram/bandwidth/disk increases somewhere. perhaps that was before the website remake. also I understand cpu count is a commodity that can be increased with appropriate funding as well. if one is updating $random_info_page one should add it also. Just my $.02
toddf: when you click learn more on the right side you can see VPS powerups on the left side
whoa... it's true
I totally hadn't noticed that before
daca: there it is indeed ;-)
still it lacks cpu bits. maybe thats a non advertised feature I recall seeing discussed eyons ago...
It's in the FAQ but no pricing
i thought cpu share is assigned proportationally to memory % usage
at least, that is standard practice.
I'm not talking about cpu share, you get 1 dedicated cpu per vps
well not memory usage, but how much RAM is assigned.
I'm talking about multiple cpus per vps
who gets 1 dedicated cpu per vpus?
see "Dedicated CPU resources" on the arpnetworks.com/vps page
i doubt anyone even gets a dedicated core, excpet for maybe the biggest plan
ryk: it's about thread/process concurrency. Same as having a single/dual core CPU in your desktop. With just the single core, we get 1 thread at a time basiucally
if only 8 cores per host
what do you think that means?
I've always been amazed by the cpu my small vps has. 'md5 -t' for example beats any real system I have in my office by quite a bit.
i don't know what it means, but the economics of a box hosting 8 vps's @ $10/mo each with their own dedicated core does not work.
how about mixing small vps's with bigger vps's ?
i would be willing to bet that it means that CPU share is not oversubscribed in terms of GHz
i.e., a small VPS gets 1 core at 1GHz
a 3GHz xeon can server 3 vps's on a single core, in that case
up_the_irons: care to enlighten us on what 'Dedicated CPU resources' means?
toddf: you could just be on a host with friendly neighbors, and are able to consume available cpu that they're not using.
hmm, where'd the 'fbi' bit go that was logging this channel? at one point I could have sworn it existed and was making some parts publically searchable
there is a logs page
lets see if my own logging of this channel produces any jewels based on that phrase
http://irclogger.arpnetworks.com/irclogger_logs/arpnetworks
it hasn't come up since 2010 with that exact phrase, odd
todd: i wonder if you're on a newer cpu than me
ryk: i think it's just that you only get one core
it's still shared
but it means less context switches
which means most consistent performance
err more
it does seem a bit strange not getting like 2 or something on bigger plans though
^
vmware used to say that SMP vms were inefficient
since you had to lock 2+ CPUs instead of just 1.
vmware used to say that 64 bit vms were inefficient
well it depends
virtualisation extensions help to some degree
the newer cpus got faster and faster
err at doing switches
it also copmletely depends on how high your cpu load is
like if you're usually at max 60% cpu
sure, on an idle machine it doesn't matter.
these aren't idle ;)
then everyone having 2 cpu cores wouldn't relaly matter
and would be faster for individual vms
but if you're regularly at 100% load
i imagine a lot of the cpu load comes from qemu
as well
as things like openbsd aren't using virtual drivers
ithink they're 8 core
so if 4 people are using 100% cpu
and have 1 cpu core
then 2 cpu cores are going hard on qemu-dm etc
the 2 cpu cores doing "this and that"
then it'd be full load
but i doubt there's "400% cpu" much
err 4 people using 100% cpu
i wouldn't use 100% cpu for more than like 5 minutes
anyway i haven't had cpu probs
I've heard little about cpu contention here. what I have heard about is disk io contention. if you're hammering the disk consistently for a long period of time not just normal activity but like doing a dd to/from disk constantly, don't expect to be treated kindly by fellow customers on the same host system .. ;-)
if you doing have balls, I'll have them surgically attached, THEN string you up by them.
heh
my disk i/o is reasonably steady slow with no virtio
it hasn't got extremely bad ever that i recall?
jdoe: could you attach an extra set, then string me up by those instead?
I'm... quite attached HAHAHAHAH to my current set.
up_the_irons: http://www.abthorpe.org/2012/10/abthorpe-org-is-back/
:o