***: bob___ is now known as bob^^ mattx86: LT, btw, I found another provider, one in the netherlands - liteserver.nl LT: you're really searching hard aren't you? I signed up for portfast a few days back... seems ok so far mattx86: if you check WHT, they have a 128MB VPS for 6 EUR (not sure if they still accept it)
hehe yeah :)
good to hear LT: the fact they are also a registrar that does v6 glue was a welcome plus mattx86: I didn't know that.. that's pretty cool
they're a reseller tho right?
not that it matters really LT: not for .uk as far as I can tell, but I think they resell tucows for some other tlds ***: schmir has joined #arpnetworks mattx86: ah
fwiw, so far so good with goscomb here
I'm considering switching to liteserver tho because 1) they're offering 750GB bandwidth (not sure that I'll need it, but just in case) and 2) their routes are amazing from what I've seen here (and about 20ms less on pings than to goscomb)
for the same price or cheaper too
with exchange rates
not sure what their support is like though
well, coupon(s), exchange rates, and plans LT: can't hurt to try it for a month I guess mattx86: yeah, that's what I'm considering doing
I'll mull it over LT: my only worry would be they're xen rather than kvm.. had odd experiences with xen in hvm mode before mattx86: that's actually what goscomb uses (well, Citrix XenServer, the commercial version IINM)
but yeah, I'm using goscomb in HVM mode
no problems (yet) ;)
its not clear tho, whether you can install just whatever you like at liteserver
kinda sounds like a no ("Other OSes are available on request")
in other words, they're probably running in PVM mode.. but I'll check later I suppose LT: I suspect what it means is we have these OSes as PVM but if you want something else you can have HVM and install yourself mattx86: true, it could mean that
that's what goscomb does LT: yisp do that as well mattx86: have you used yisp btw?
their web design kinda scared me off for one thing
I mean, it's not bad.. it, I guess, lacks content more than anything ***: killring has quit IRC (Ping timeout: 276 seconds) LT: yeah... got it half price on wht some time back, their connectivity is ok, but having some issues with the system hanging
seems like sometimes the disks just drop out... think it's some xen hvm issue and as most of their customers are pvm they haven't really put that much time into working it out ***: sentabi has quit IRC (Ping timeout: 264 seconds) mattx86: ah
glad I stayed away then
btw, are you using linux too, or say freebsd? LT: linux... learning freebsd is on my list of things to do on a rainy day mattx86: gotcha ***: sentabi has joined #arpnetworks
ncullen is now known as Guest13192
ncullen has joined #arpnetworks
ncullen has quit IRC (Quit: leaving)
schmir has quit IRC (Remote host closed the connection)
Guest13192 has quit IRC (Quit: Lost terminal)
nakano is now known as nakano_
sentabi has quit IRC (Changing host)
sentabi has joined #arpnetworks
tinono has joined #arpnetworks heda: how does arp share out cpu resources? does it guarantee a certain number of cycles with bursting when the rest of the cpu is idle or what? ***: nakano_ is now known as nakano toddf: heda: not over-allocating means dividing the cpu power in a given system amongst kvm systems so that each kvm system can have its allotted cpu slices fully w/out degredation
aka if there is 8 3ghz cpus in a system, there could be 24 1ghz kvm instances ..
but that is just an example, I am honestly not sure how arpnetworks works that out but they are quite clear about not over-allocating memory and cpu
it is truly amazing but my physical hardware laptop is slower than my kvm instance at arpnetworks.. md5 speedtest results:
my laptop: Speed = 76614086.419157 bytes/second
arp kvm: Speed = 430477830.391735
there have been some benchmarks (unofficial of course) that show arp is much faster in general than other vps providers ***: nakano is now known as nakano_
ziyourenxiang has joined #arpnetworks heda: toddf: thats cool, i know some providers equal split the cpu resources as a 'dedicated' cycles but then if there's a load from one vps and other vps's aren't using their cycles you can burst until the other vps needs them back
toddf: i'd image also that the md5 test results are down to raid 10 throughput rather than anything else ***: nakano_ is now known as nakano
nakano is now known as nakano_
nakano_ is now known as nakano
Hien has joined #arpnetworks Hien: still Sold out ? ***: nakano is now known as nakano_
nakano_ is now known as nakano
Hien has quit IRC (Quit: Page closed)
heda has quit IRC (Quit: heda)
LT has quit IRC (Quit: Leaving) up_the_irons: removing sold out sign now... tooth: o/ heavysixer: up_the_irons: you'll always be sold out to me up_the_irons: heavysixer: whut? ;) heavysixer: up_the_irons: you know like "you sold out" up_the_irons: heavysixer: haha ***: ziyourenxiang has quit IRC (Quit: ziyourenxiang) heavysixer: http://en.wikipedia.org/wiki/Wikipedia
haha up_the_irons: lol ***: nakano is now known as nakano_
wallshot has joined #arpnetworks
killring has joined #arpnetworks
schmir has joined #arpnetworks
heda has joined #arpnetworks
heda has quit IRC (Quit: heda)
schmir has quit IRC (Remote host closed the connection)
tinono has quit IRC (Quit: leaving)
Hien has joined #arpnetworks
schmir has joined #arpnetworks
nynex_ has joined #arpnetworks
Hien has quit IRC (Quit: Page closed)
nynex_ has left
heda has joined #arpnetworks
tinono has joined #arpnetworks
heda has quit IRC (Quit: heda)
schmir has quit IRC (Ping timeout: 276 seconds)
schmir has joined #arpnetworks
iaaron has joined #arpnetworks
iaaron has quit IRC ()
tinono is now known as _tinono_
_tinono_ is now known as tinono
nakano_ is now known as nakano
tinono has quit IRC (Quit: leaving)
mattx86 has quit IRC (Read error: Operation timed out)
mattx86 has joined #arpnetworks
tinono has joined #arpnetworks
schmir has quit IRC (Remote host closed the connection)
mattx86 has quit IRC (Read error: Operation timed out)
mattx86 has joined #arpnetworks
wallshot has quit IRC (Quit: Leaving.) mattx86: hm.. anyone know if nagios would allow you to run a traceroute if latency/packet loss flaps to and from certain thresholds during a specified period of time, and e-mail you the traceroute(s)?
I'm not running nagios (or anything at the moment), but I've been considering writing my own network monitor app
specified period of time = say, within a 20 second window Lefty: you can write arbitrary nagios plugins in whatever language, it's just got a specified output format
so the email would have to be out-of-band, as nagios just gets pass/warn/fail I think
wouldn't be that difficult though mattx86: sounds doable RandalSchwartz: you can give exteneded output in the status
arbitrary text
you'll have to drill down in the interface though
or wait... you can have it be part of the email
might be by default
... http://nagios.sourceforge.net/docs/3_0/pluginapi.html
see the "long text line 1"
$LONGSERVICEOUTPU$ macro for the email
so yeah, you can do it... just gotta write your plugin to do what you want mattx86: I think I'm still leaning towards implementing it for my own network monitor app
I've got a lot of great ideas for one RandalSchwartz: you and 50 others :) tooth: Only 50? mattx86: RandalSchwartz: what was the idea you had the other day, if you don't mind telling me? RandalSchwartz: where 50 means "uncountable" :)
which idea? tooth: Ah, I sit corrected. mattx86: something about when to send notifications, like queue them until the 9-5 period of the day, I believe Lefty: yeah, researching network monitors always felt like a million monkeys on typewriters to me RandalSchwartz: ahh. lemme think mattx86: there's another idea - allow the user to specify the timezone for notifications too RandalSchwartz: that's already there mattx86: e.g., for international cooperation tooth: Just don't bring the pager home. :-) mattx86: lol yeah RandalSchwartz: mv pager /dev/toilet tooth: Mine just stays at work. :D mattx86: the flapping on the wireless network forced my last boss to stop using his pager until I adjusted the thresholds to something that 'worked' ;) tooth: flapping on wireless networks was worthy of a page? RandalSchwartz: nagios at least has that figured out
notify at first breakage mattx86: it was important to me anyways. we'd normally get 1 to 10ms on a good link RandalSchwartz: then don't notify "up again" until up again in a long enough ratio mattx86: maybe 25ms Lefty: I liked the idea of having a "flapping" status, too
we currently use ipMonitor 9, which is way too basic for what we do
I hate it
every other NMS sucks worse though mattx86: btw, does nagios properly do notifications in a network hierarchy? RandalSchwartz: yes
you can tell it "this hides that" mattx86: last time I tried nagios, despite specifying parent nodes, it would notifying me of everything below the parent that was down RandalSchwartz: so it will only complain about the nearest border Lefty: nagios is a bit of a whore to configure, true mattx86: s/notifying/notify/
this was actually nagios 2.0b3 or something :P bob^^: i'm late to this discussion, but if you like nagios, take a look at Opsview
it's nagios (but with improvements) and a *very* useful/usable/stable frontend
also makes scaling nagios across multiple servers very simple indeed mattx86: interesting.. is the community version just as good? bob^^: yup
we use the community version in work at the moment (i work for an ISP)
one master server, three slaves which run checks and a dedicated database mattx86: nice
I'll take a better look at it later
I'll bbl bob^^: there's #opsview if you need info :)
or just give me a shout mattx86: cool, thanks :) up_the_irons: bob^^: will it run on openbsd? :)
i've been meaning to migrate my old nagios setup to my new openbsd atom box jdoe: if nagios works but you hate how it looks, why not just look at the alternate frontends?
I mean I hate how nagios looks, but it's not like I have to look at it on a regular basis or anything. tooth: if you do, you're doing something wrong jdoe: or other people are ;)
I dunno, we have the dashboard thing up on a monitor here, but that's more just a convenient status board. ***: bob___ has joined #arpnetworks
bob^^ has quit IRC (Read error: Operation timed out)
bob___ has quit IRC (Read error: Connection reset by peer)
mattx86 has quit IRC (Read error: Operation timed out)
mattx86 has joined #arpnetworks
bob^^ has joined #arpnetworks
fink has joined #arpnetworks
fink has quit IRC (Remote host closed the connection)
fink has joined #arpnetworks RandalSchwartz: "will it run on openbsd" is related to "will it blend?" :)
up_the_irons - how's your health? tinono: things tend to blend way easier than they tend to run on OpenBSD ~_~
*comming for an obsd guy* :-p
from* ***: tinono has quit IRC (Ping timeout: 255 seconds)
bob^^ has quit IRC (Read error: Connection reset by peer)
tinono has joined #arpnetworks
mattx86 has quit IRC (Read error: Operation timed out)
bob^^ has joined #arpnetworks
fink has quit IRC (Read error: Connection reset by peer)
fink has joined #arpnetworks
mattx86 has joined #arpnetworks
fink has quit IRC (Quit: fink)
killring has quit IRC (Quit: leaving) tooth: Even bricks, for instance ***: nakano is now known as nakano_
tinono has quit IRC (Changing host)
tinono has joined #arpnetworks