... well that was odd. network problems/ ? fink__ is now known as fink_ :) nope weird, my vps just went unresponsive for a minute or three, both from here and from the uk. er... "here" = bc only thing I was doing was untarring perl. things look fine over here yeah I've got nothing useful sorry, didn't think to traceroute. a question for any openbsd users... is there any tuning to be done? I'm relatively new both to obsd and kvm (so I don't pretend to know which I should be looking at) but I'm unimpressed with the performance. It's been compiling/testing perl for about 45 minutes now. LOL hmmm... looks like i will be needing to upgrade my storage soon hehe its .50 a gb a month right? I'm not new to vpses, I know disk performance can suck, but whatever's going on here seems a little above-and-beyond, you know? jdoe: enable soft updates: http://www.openbsd.org/faq/faq14.html#SoftUpdates can I remount with that, or will I need to reboot? jdoe: several tuning things you might want depending on what you're running, no special tuning specifically for kvm beyond 'disable mpbios' jdoe: toddf is the man to ask about tuning :) toddf: 99% of the time just irssi. I wasn't aware there was even an issue (things seemed pokey, but I chalked it up to latency/virtualization) until I tried compiling this. its the soft updates ya need em with obsd or else everything takes years :) some would say your data is less safe, I'm not convinced, but be aware of what you are doing when you go beyond defaults jdoe: you can also add "noatime" to fstab I don't think this is an io issue. softdeps is what I enable by default also, fwiw vmstat shows zero io, ~30-50% cpu, and what looks like a monumental number of interrupts (but might not be) just if you have issues try disabling it as a first stab toddf: not super worried about them, fbsd has used softupdates as the default for quite a while. I dunno, it's showing ~130-140k interrupts/s (presumably because perl is testing Benchmark right now, timer interrupts?) ... I don't know what's normal, but that sure sounds high. building in kvm does take more time than outside yeah I know. ... 55 minutes now though, that's excessive. (I was wrong, fyi I meant syscalls not interrupts...) syscalls and perl regress definately make sense then syscalls are things like 'open' 'close' 'read' 'write' etc yeah I was just reading the vmstat man page wrong... yeah this is insane, I went to lunch and it's still going. a customer is forcing me to not use any dynamic SQL, only stored procedures for this huge complex report i think im going to jump off a bridge I've only half-done it myself, but I don't think it's difficult... well, it's the iPhone side of things that I'm not certain of. I mean, setting up radius, and configuring a switch or AP to use it is pretty trivial. I would imagine you just give the iphone the appropriate credentials. and that's the end of it. though I never got that far (and wasn't going to use it with an iphone anyway) p p p: not found lol RandalSchwartz: you got on VUC a lot? What sorta uptime do the VPS' typically get? i wonder who has the highest uptime here G, uptime is a poor indicator of quality. in fact, its a poor indicator of anything meaningful. what do you *really* want to know? my first server: $ w 19:43:12 up 463 days, 15:59, 3 users, load average: 1.17, 1.04, 1.11 cd $home jpalmer: people that manage to keep their servers up for ages tend to be a good source for stability indicators, I don't normally ask about uptime, and I agree, but yeah up_the_irons: thanks for that G, my VPS has been rock solid, aside from a bad piece of software I installed (memory leak, locked it up a couple times before I found the issue) I like the fact that it's KVM too (but you can't blame the provider when installing poor software.. hehe) jpalmer: I normally measure the decency of machine stability by time since last unexpected shutdown :) G: I restart for freebsd updates sure, but machine stability and provider stability aren't quite the same thing. the provider can be SOLID as hell.. but if your particular setup/configuration is running you out of resources.. it can be pretty unstable. it's important to not confuse the two. up_the_irons: "cd $home"? up_the_irons: "cd" ;) though if that's one of those muscle memory things, I *totally* understand. I still catch myself doing cat | grep occasionally... jdoe: oh that one I do a lot :P in fact once or twice I've down cat | less :P ouch The Unix Programming Environment actually mentions that " [task2] is more natural than [task1] <- [task2] er [task2] <- [task1] haha up_the_irons: you wouldn't happen to be blocking port 514 on our ubuntu vm would you? :P can't figure out why my remote logging no longer works