=== funkyHat_ is now known as funkyHat === Myrtti_ is now known as Myrtti === Myrtti is now known as Guest71494 [01:16] [Jono Bacon] Ubuntu Accomplishments: Translations Needed For 0.2 - http://www.jonobacon.org/2012/05/25/ubuntu-accomplishments-translations-needed-for-0-2/ [04:43] hey webpigeo1 [05:40] morning everyone [06:19] morning [06:38] morning all [06:38] morning, no sleep for me :D === karl is now known as Guest78962 [06:54] Good morning [06:59] anyone care to tell me how I might make this do what I want it to do? gksudo "echo 0 | tee /proc/sys/kernel/yama/ptrace_scope" [07:06] Morning all [07:07] morning [07:07] here's a fun question for any bash people, if I run this from a terminal it works, if I run it from a desktop launcher it doesn't. http://pastebin.com/FGSHDa0U [07:08] any ideas how to make it work? :) [07:08] yes [07:08] ali1234: care to elaborate on the yes? :P [07:08] no [07:09] :< [07:09] firstly, echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope [07:13] ali1234: what about it? [07:13] secondly, echo 0 | gksudo tee /proc/sys/kernel/yama/ptrace_scope [07:13] * Azelphur slaps ali1234 around a bit with a large trout [07:16] worked around it :D [07:27] write the whole script as if it was root and then put "gksudo script" in the .desktop [07:28] echo | gksudo tee doesn't work because gksudo does not pass through stdio like sudo does [07:29] gksudo /bin/bash -c "echo > whatever" also does not work because gksudo interprets the -c as an argument to itself rather than bash [07:29] there doesn't seem to be any way to escape it either [07:29] you might think gksudo "/bin/bash -c \"echo > Whatever\"" would work, but it doesn't [07:30] also gksudo /bin/bash -- -c "echo > whatever" does not work because the -- gets fed to bash instead of gksudo and then bash ignores -c [07:42] morning all === myrtti is now known as Myrtti [08:27] Could anyone tell me what the equivalent of the following in the old /etc/inittab is for latest ubuntu systems [08:28] h1:35:respawn:/etc/init.d/init.ohasd run >/dev/null 2>&1 "XChat-GNOME" cant save my default. Auto start with "#ubuntu" but i want "#ubuntu-uk" as well.... "auto-join on connect" No result. [08:34] Another problem, Pidgin and Empathy can't us video/audio any more. All MSN stuff suck (sinds it was born), and gmails wont word... === schwuk_away is now known as schwuk === soreau_ is now known as soreau [08:42] But the main problem is the not working item of Bluefish. Sinds Ubuntu 12.04, the option "Finds and Replace" doest work lice before. Just a PHP text, 300 names from $001_ to $002_ ... [08:44] Hello. [08:45] Hi [08:45] Hello Wobbo [08:47] good morning everyone. [08:47] The Bluefish is solved. I am so crumby, no coffee today..! [08:49] Good morning ass-well. [08:57] Morning peeps :) [08:58] Hello bigcalm :) [09:08] howdy all [09:08] anyone know of an open source project to provide a 'cookie' toolbar ? [09:08] to become compliant with the new EU cookie law? === popey_ is now known as popey [09:13] What's the EU cookie law? === dwatkins_ is now known as dwatkins [09:14] vedreamer: all sites that put a cookie on your pc, must provide you a way of accepting it or rejecting it [09:14] unless it's for shopping carts etc [09:15] don't most browsers have an option to get permission each time a cookie is presented? I just assumed this is off by default in everything but lynx. [09:17] dwatkins: yes, but legislators are dumb (and "ask for every cookie" is annoying) [09:17] so now there's an idiotic "cookie law" all over Europe [09:17] MartijnVdS: yup :) [09:18] Because cookies are obviously EVILS [09:18] it could be worse, they could be trying to curb our freedom online.... oh wait [09:18] MartijnVdS: DOWN TO ALL COOKIES [09:18] not even sure how they are going to enforce this [09:19] fines? [09:20] policing how many millions of website? [09:20] so they expect, any site, however big or small, to comply with this? [09:21] brobostigon: yes unless it's cookies related to the navigation of a site ie shopping carts [09:21] MooDoo: ok, so i will have to change drupal, on my server, which has my mum blog on it, simply for the 3/4 people who visit it a month. [09:22] brobostigon: that's what the law says....sucks doesn't it [09:22] pointless more like. [09:22] i know i know [09:24] Just block all IP ranges that belong to the police and you're good :P [09:24] maybe simply better, to scrap cookies entirly, and have everyone remember there login details, and such things. [09:24] MartijnVdS: have a list? [09:25] brobostigon: ah it's all to do with tracking cookies etc [09:26] MooDoo: cant i just simply put a notice at the top of the page, recommending people review the cookie settings in their browser, and leave it there? [09:26] brobostigon: not sure to be honest, afaik people need to be able to allow/reject the cookie [09:27] MooDoo: that was possible already by either blacklisting or whiotelisting cookies from said site, inside browser settings. [09:28] yes but non tech savvy people are not going to want to know or learn how to edit settings in their browser [09:28] hence me suggesting, people who visit, read about doing it. [09:29] brobostigon: nah they won't, people would leave the site rather than read anything [09:30] MooDoo: i wonder then, there is no real solution. [09:30] brobostigon: well for wordpress sites, there is a plugin, visit the site, click yes or no [09:31] MooDoo: so ie, that gives you the choice, as to if the site, puts cookies on your machine or not? [09:31] yes [09:31] link to plugin please? [09:31] brobostigon: I've just set-up a cookie notification work flow for a client: http://www.totalderivatives.com/ (their visual requirements, so please don't blame me) [09:32] ali1234: just search for eu cookie in the add plugin part of wordpress [09:32] totally derivative? [09:32] what's wrong with that visual style? it look sgood [09:32] bigcalm: yes, i was thinking of something similer along those lines. [09:33] the hashsing on the background is making my eyes go odd [09:33] /hashsing/hashing/ [09:33] ali1234: more for the information page /cookies.php [09:33] in fact.. I can't stand to read it :/ .. I see what you mean bigcalm [09:34] My client got me to set it up 2 nights ago :S [09:34] This seems like an effective solution, without causing so much disruption, but whilst still being pretty unavoidable: http://www.civicuk.com/cookie-law/index [09:35] I think they were panicking a little. Hopefully they will improve the layout of the text at some point [09:35] just getting rid of the bars in be background would be nice.. [09:36] Those are the default that comes with jquery-ui :) [09:36] they should hire cookie monster to explain cookies [09:36] That would be awesome [09:36] :) [09:38] Worked myself silly again last night. Lost the will to work this morning [09:38] Ug [09:39] bigcalm: go for a run or something to clear your head? [09:39] I have to do battle with Apple Push Notification now [09:40] bigcalm: a madcap dash round the block then? Always helps for me [09:40] This isn't the weather to be running in :) [09:41] Maybe I shouldn't have had pitta bread and dips for breakfast [09:41] Middle class leftovers yay [09:43] What's tricky is that I don't have an iOS device. Having to rely upon my boss for sending test messages, and he's down in London today. Will be amusing :) [09:45] I was wondering why big sites like facebook etc didn't do this then I noted the 12 months to comply bit. [09:46] ok, i now have that cookie control thing, inside drupal, and seems to be working. === karl is now known as Guest90588 [10:00] morning all and Happy Friday [10:00] happy friday to you too davmor2 [10:01] bigcalm: have a happy holiday! [10:01] morning davmor2 [10:01] AlanBell: you too dude [10:01] MooDoo: me owld mucka [10:03] compiz decided it now needs 40% cpu to do anything [10:03] grey screens everywhere [10:06] davmor2: how's it going? [10:07] MooDoo: sound as a pound there? [10:07] davmor2: thank you :) [10:08] How come just one machine in my house will redirect me to an unknown site when I enter a client's URL? Doesn't happen on any other machine. [10:09] On this one machine, all web browsers do the same thing, including w3m [10:09] * bigcalm pokes /etc/hosts [10:09] nothingspecial: https://plus.google.com/u/1/100694334141523232451/posts/e5NU3idyJzA [10:09] Aha, was hosts fault, phew :) [10:12] davmor2: yes it's ok, sun is shining, it's nearly the weekend, i'm on hols end of next week. so not bad [10:17] MooDoo: nice [10:18] davmor2: yes looking forward to it, need to refresh a little [10:19] http://blog.taylorworld.me.uk/ anyone recognise the drupal error? [10:20] there [10:20] at the top of the page. [10:20] have managed to put lubuntu on an old G4 Imac (17 inch) [10:20] and it works pretty well [10:20] MooDoo: I know that feeling [10:28] is it beer o'clock yes [10:29] s/yes/yet/g [10:31] Morning all o/ [10:32] Any KVM users about? [10:34] Is anyone about? :) [10:35] only us chickens [10:35] RAWR [10:38] no one here [10:40] Moooo [10:40] selinuxium: !ask ;) [10:40] !ask | selinuxium [10:40] selinuxium: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience [10:40] Hey AlanBell [10:41] Hi MooDoo [10:41] selinuxium: hullo [10:41] Hi MartijnVdS, I know about the whole ask thing :) The question I have is around KVM not a specific KVM question. [10:42] selinuxium: go one then ask away [10:44] But to chuck it out there. Who uses it? How much fun are they having with it? I have been using VMware but now have 2 spare servers to play with... OpenStack vs eucalyptus vs anything else? Virt-manager vs convert vs convert2... [10:45] selinuxium: i like vmware, been using esxi for months [10:45] selinuxium: what about xen? [10:45] Wasn't going to ask such a grand question. it is more of a discussion... I used to take these things offline but there doesn't seem to be much traffic in here today so... :) [10:46] MooDoo, never used xen... That would make the above question even more extrapolated [10:46] lol [10:49] hey MooDoo :) [10:49] Happy Friday [10:49] s-fox: hola :) same to you :D [10:51] s-fox: how are you this fine day? [10:54] happy friday s-fox [10:56] \o/ \o/ \o/ \o/ [10:57] 2x1080p running off a laptop \o/ [10:57] popey: you look happy :) [10:57] ah nuff said [10:57] how are they plugged in then popey? [11:03] AlanBell: for a second I thought that said how are they plugged in to popey? I thought you were trying to make out that popey was a teletubbie :D [11:04] davmor2: eho! [11:05] davmor2: one of my client contacts is Lala [11:06] sounds a bit dipsy to me [11:06] bigcalm: rather lala than gaga [11:07] managed to get Lubuntu working on my G4 imac [11:07] runs pretty smooth for an 800 mghz machine [11:09] :) [11:16] AlanBell: I bought an 'ultrabay' which the laptop docks into. it has 1xVGA and 1xDisplayPort [11:22] Isn't there a docking bay that can take a pci-e graphics card and thus give you 2 x dvi ports? [11:47] Argh, great customer service from software supplier on their sales team telephone line "We're away from the office on Thursday 25th May, please leave a message and we'll get back to you"....... Is it me or is today Friday [11:48] DJones: it might be [11:48] I guess I didn't consider they could be in another timezone [11:49] Stockton on tees must be a differnt time zone to the rest of the UK [11:53] I heard you need your passport to get out of there.. [11:54] anyone know what handles 'plugging in your phone and detection' stuff? [11:54] got a bug I want to report [11:55] plugged in my iPad and a message poped up saying 'device is locked, un-lock it first [cancel|try again]'... cancel does nothing, just keeps popping up the error (clicked 10+ times), try again made the error go away === schwuk is now known as schwuk_away [12:59] Well that was fun, kids sports day [12:59] Lots o flovely mummys dressed quite nicely for this heat [13:00] flovely. interesting word. [13:00] Sun stroke causing typos ! [13:00] :) [13:01] :) [13:05] Yummy mummies? [13:06] Oh yes! [13:06] ahem [13:07] Yummy popies? [13:07] \o/ late lunch [13:10] * AlanBell prefers long lunches to late lunches [13:12] oo, bbc says nice weather tomorrow and sunday too [13:12] * popey pops out to get MEAT [13:14] popey: and monday :) [13:19] * daubers has a weekend painting [13:35] aloha [13:36] hello lauraczajkowski [13:37] lauraczajkowski: prod [13:38] I'd normally prod czajkowski but being as this new lauraczajkowski is about I'll prod them instead [13:38] joy [13:39] lauraczajkowski: you know it wouldn't be a normal day if I didn't :P [13:39] at a conference where ssh is blocked :/ [13:39] lauraczajkowski: PROD [13:40] davmor2: tag team on this new lauraczajkowski ? [13:40] lauraczajkowski: well what geek is going to use ssh [13:40] MooDoo: Yeap [15:11] Not sure how many of you use nagios, but just having a tinker with hpasm [15:11] Got to say very good, wish I'd used it earlier === schwuk_away is now known as schwuk [15:33] Anyone fancy giving suggestions for a server shutting down [15:33] HP DL380 G7 [15:33] Shuts down at random times each day [15:34] bad caps [15:34] Overheating? [15:34] Different OS's ( So points to hardware I guess ) but can't find any faults anywhere at the mo [15:34] what does syslog and dmesg say? [15:34] Temps all seem fine [15:34] bad caps man [15:34] That's the problem, they stop outputting [15:34] caps ? [15:34] capacitors [15:34] Capacitors ? [15:34] yes capacitors [15:35] Any way without specifically testing them of saying yay/nay to that ? [15:35] is it happening more often with time? [15:35] Temps looks fine btw, just using this hpasm nagios checker [15:35] Nope, once a day at a random time [15:35] Could be 10pm, 1am, 6am [15:35] how long has it been doing it? [15:36] last month or so [15:36] has it ever done it before? [15:36] If it was a similar time every day, I'd point to a cleaner doing the hoovering [15:36] Nope, new OS is the only extra thing that's been changed recently [15:36] heh yeah me as well [15:37] does it reboot clean? [15:37] I think I've checked most of the standard stuff, was just hoping someone might have some other ideas [15:37] yep [15:37] Not reboot sorry [15:37] powers off [15:37] System halts [15:37] Does it have a UPS fitted [15:38] the G7 is xeon 5600 era. it's not capacitor related. [15:38] There is a ups in the server room it's plugged into yeah [15:38] ran memtest/hdparm [15:38] if it has electrolytic capacitors they can fail [15:38] Only thing I have had come up with the hpasm commands is firmware upgrade needed on the controller [15:39] But it was updated yesterday with newest from HP [15:39] is it *shutting down* or is it powering off unexpectedly? [15:39] iml CAUTION: POST Messages - POST Error: 1770-Firmware Upgrade Required. [15:39] Powering off, logs just stop no entrys for shutdown [15:39] Which points me again to hardware ? [15:40] run if off another power supply? [15:40] yeah tried that, duals - tried it singly on each [15:40] Will see if the nagios checker gives us some info over the weekend I guess [15:41] Or could try leave it running with a live CD for a day or two [15:41] See if that also fails [15:41] you said it was on a UPS [15:41] are *both* power supplies on the same UPS? [15:41] yep [15:42] I don't know for defo but I would expect so, only a very small server room [15:42] move one onto regular power [15:42] Thinking bad feed from the ups ? [15:44] yes. if the UPS is futzed, then it's a single point of failure [15:44] i'd never run two feeds from the same UPS into one server. kinda defeats the porpoise [15:45] A few other servers running from same UPS though, so should affect them [15:46] Really, I used to run all my servers through one UPS, as did a lot of the people I used to go and see [15:46] Couldn't afford/fit two in my last place [15:46] I see your point though, but for me it was there in case of power failure, hadn't thought about if it had caused the issue [15:48] if the UPS fails - and they do fail - then your server is down. better to have one feed with no UPS than both on the same UPS [15:48] and your UPS will drain much faster in the event of power failure if it's feeding redundant power supplies [15:49] I defo see your point [15:49] I hadn't done that on my 60+ servers at my last job :D [15:49] oops [15:51] It was my first time designing a Server room [15:52] when you have 400 servers, you get selective over what to UPS [15:54] So many people plug both/all redundant power supplies into the same UPS [15:56] yeah we got to that point directhex [15:56] BigRedS: Lot's of sites I went to did yes [15:56] Some quite large [15:56] diplo, it's not until you have a complete power failure that you learn why you suck :p [15:57] We had lots of power failures [15:57] Just never a UPS one [15:57] diplo, e.g. "oh bumfarts, the server is covered, but i can't power it off because the kvm and switch are not" [15:57] The genny I bought in could pretty much run the building let a lone the server room [15:58] UPS was only ever used for 30 secs max  [15:58] yeah, at oxford we... used a lot of power. about 1300 units a day iirc for our server room [16:00] All I can remember ( not been there for 2 years or so ) is that we ran about 26/27 amps a phase [16:01] Which was getting close to the limit [16:01] ... yeah. we were specced for two 32A plugs per rack [16:02] yeah same :) [16:02] Right, swimming lessons... catch you guys tonight [16:03] Thanks for all your input! [16:03] My kids lessons btw :) [16:03] 1MW feed, 60% allocated to room power (40% for cooling) makes... 2500A! [16:03] by my sums === Lcawte|Away is now known as Lcawte === diddledan is now known as zz_diddledan [16:20] has it gone all quiet? [16:30] MooDoo: Seems so [16:31] i've got a question [16:32] it may well be the most intellectually challenging question of the 21st century [16:33] 42 [16:33] i asked google, and google doesn't know [16:33] here is my question: [16:34] how does slash play guitar and smoke cigarettes at the same time without setting his hair on fire? [16:36] fire-retardant hairspray [16:36] yup [16:43] wow, i actually found the answer. apparently the sweat puts it out. nice. [17:08] barbecue lit \o/ [17:09] * brobostigon smells a barbecue, from his eeepc, in his back garden, but popey is too far away. === schwuk is now known as schwuk_away [17:44] * kvarley is proud of his little venture today - http://kvarley.co.uk/tmp/RaspberryPi.CardCase.jpg [18:01] evening [18:02] \o9 [18:20] My friend is having difficulty installing proprietary drivers on Ubuntu 12.04. Is there anything obvious in his log? http://paste.ubuntu.com/1006823/ [18:21] "fglrx" [18:26] hmm [18:26] https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/949641 still not fixed, so it's probably that [18:26] Ubuntu bug 949641 in fglrx-installer (Ubuntu Precise) "Installing both fglrx and fglrx-updates results in: error exit status 1 -"/etc/init.d/atieventsd exists during rc.d purge"" [Critical,Triaged] [18:26] "sudo apt-get remove --purge apache2" has left things in an impressively ruined state... [18:35] haaaai [19:09] evening ladies! [20:33] anyone know the current best practice for having iptables rules come back on a reboot? [20:36] sure [20:36] that's one of those things that nobody agrees upon [20:36] I like using shorewall for that [20:36] no, i don't know best practice, i just add a pre up in interfaces file :) [20:37] works [20:37] perfectly fine [20:41] Does anyone know of any software that allows you to use your PC as a bluetooth headset to answer your phone? [20:41] so you don't have to keep switching headsets every time you get a call [20:41] BigRedS: yeah, I am reading lots of conflicting advice, most of it presuming a centos system [20:43] AlanBell: iptables-save [20:43] It's what I use anyhoo [20:43] and then iptables-restore in in if-up script? [20:43] I had a good example for my vps, two secs will make sure [20:44] AlanBell: correct [20:44] you can also iptables-save on post-down [20:44] AlanBell: yeah, iptables-save/restore is a pretty common way of doing it [20:45] nobody's going to be surprised to find it [20:45] so I have the output of iptables-save in /etc/iptables.rules [20:45] yeah that's how I had mine [20:47] Or could try iptables-persistent.. on my list to try [20:48] oooh that would appear to be a better answer [20:49] Not had time to try it yet [20:49] I just installed it and it offered to save the current rules [20:49] time = I am a lazy mofo :) [20:50] yeah, that looks like it should work [20:50] lets reboot the remote server with no emergency console access and see if it comes back [20:51] lol [20:54] down it goes [20:54] why don't people make boot time a priority for server hardware [20:55] they sit around contemplating their raid hardware and beeping randomly and flashing lights for ages [20:55] It's all the checks isn't it [20:55] Post checks, need to update the bios to something new / better, not uefi [20:55] I don't want it to check, I want it to come back up [20:55] And do something for the hardware cards as well [20:56] I used to reboot file server at my old place of work, used to take 6-8 mins from Clicked Restart to fully backup [20:56] if a server needs a reboot I want it to come back up before the phone rings [20:56] Used to scare the bejeebers outta me [20:56] AlanBell: then you're doing it wrong :) [20:56] I feel the same way [20:57] I really dislike all this recent effort towards speedy booting which can only be at the expense of other priorities [20:57] well I don't tend to reboot servers, but when I do, I don't want to wait [20:57] for booting up my laptop, I am fine with waiting, I can walk to the kettle and relax as it warms up [20:58] They do tend to have fscks in the middle which would wreck that couple-of-second optimisation [20:58] servers are the things where the boot time should be engineered faster [20:58] yay, it came back [20:58] well, they're perhaps the last things where anybody cares now, yeah. I don't still don't think it should be *that* far up the priority list [20:59] er, I *do* still don't think :) [20:59] When I'm rebooting servers, normally there's some work to be done on it which takes longer than the boot up by enough that there's more to save by me getting an electric screwdriver than faffing too hard with the boot process [21:01] normally when I am rebooting servers it is remotely as well, so I don't have any visibility of the reassuring flashing lights and beeps [21:01] just 3 minutes of wondering if I mucked up the firewall rules [21:02] and then happily I find that I didn't \o/ [21:02] need to get a DRAC / iLO man :-/ [21:02] haha, yeah [21:02] I'd hate to wait on a boot like that [21:02] then you don't have to hold onto your butt when you reboot [21:03] mattt: yeah, normally we have that on our servers, this is for a customer who wanted a different datacentre [21:03] AlanBell: shoulda gone w/ hostchopper.com [21:03] ours are in hetzner in germany, but this customer wanted it physically in the UK [21:03] do boxes at hetzner have console ? [21:04] good question, they have a console that can reboot the server, and do a rescue pxe boot [21:04] ah [21:05] yeah, no serial console as such [21:05] i've had a box w/ hetzner for ages, need to cancel it [21:06] the one I was setting up today for a customer is at https://www.openmindhosting.co.uk/ [21:07] dual hex core xeon thing, with raid0 SSD [21:08] fancy :) [21:08] what do your customers use these for? [21:08] you run that odd erp stuff right? [21:08] this one is vtiger [21:08] 8 virtual machines running vtiger [21:09] why don't they just put it in the cloud ? [21:09] Still not tried Vtiger [21:10] But I have enough things to play with [21:10] mattt: how is this not in the cloud? [21:10] AlanBell: oh, thought you bought metal and ran your own virtual machines :) [21:11] yeah, which is a cloud :) [21:11] uh [21:11] unless you are defining cloud as "paying amazon" [21:11] no :) [21:11] no, and no [21:12] a cloud is basically a botnet in a cabinet [21:12] if you're paying a monthly fee, and you can't spin up / down at will, then not cloud [21:12] we can spin up and down at will [21:13] the whole point of "cloud" as a term is that nobody knows quite what it means [21:13] I was about to type the same thing BigRedS [21:13] just paying a fixed montly fee that is quite a bit cheaper than running 8 VMs with amazon [21:13] AlanBell: to the max that the hardware supports [21:13] people know what cloud means [21:13] sure, then we buy another lump of hardware [21:13] what are you lot on about [21:13] mattt: yes, and they don't all agree [21:13] right, and wait hours for it to get provisioned :P [21:14] err, ok [21:14] that's not cloud man [21:14] seriously [21:14] for a start, 'cloud' happens at all sorts of levels. I've seen Gmail described as 'the cloud' [21:14] but spinning up another system involves purchasing an office, a warehouse, ~5 vans and recruiting 15 people [21:14] waiting a day for a server isn't really a factor [21:14] AlanBell: but i'm with you on the cost [21:15] a lot of time dedicated hardware at a monthly cost will be far cheaper [21:15] which is why i have my hetzner box :) [21:15] i can slice it up as i see fit, much cheaper than buying individual instances w/ whoever [21:15] amazon makes masses of sense if most of the stuff you occasionally want is turned off most of the time [21:16] if you want stuff that is turned on all the time it doesn't make sense [21:16] yep [21:16] fully agree [21:16] AlanBell: sorry, didn't mean to bash your opinion of cloud [21:16] yeah, it's designed for people who have the odd, predictable spike [21:16] it's engineered for something quite different... [21:17] and if you want a fixed monthly bill from a regular company that will send you nice invoices and give you 30 days to pay by BACS etc. [21:17] mattt: thats fine, it is always good to have traditional thinking challenged :) [21:18] and also, by purchasing metal we can tune it differently, like our small, but fast SSD array [21:19] AlanBell: what are you using for virtualization? kvm ? [21:19] yeah, kvm [21:19] I wanted to do eucalyptus or openstack, but you need too many boxes to run the virtualisation before you run any actual stuff [21:20] i'm guessing you could run everything on a single machine w/ openstack [21:20] there's quite a lot of services tho, probably not an ideal setup [21:21] I've heard 5-6 machines is considered the bare minimum for openstack. and that's for a 'proof of concept' rather than an actual deployment [21:21] shauno: i know for a fact that's not true [21:22] shauno: yeah, I think the documentation is 6 [21:22] i've seen everything running on one, but then VMs on a separate machine [21:22] well, that's what they're telling people when they present it [21:22] I am sure you could do it on less, but I gave up on the idea at that point [21:23] shauno: where did you see it presented? [21:23] so I could buy €50*6 per month to get a system that I can add managed nodes to, or just install KVM and get on with life [21:23] (openstack that is) [21:24] AlanBell: yeah, i'm not sure how beneficial openstack is for 8 VMs anyway [21:24] this isn't for our main infrastructure, this is one particular customer [21:25] we have 4 boxes in hetzner now running an assortment of VMs [21:25] we don't sell hosting as a standalone product, but we do provide hosting for our consulting customers [21:26] ah right [21:26] then openstack to manage that all may work [21:26] one day we might re-visit that and use openstack [21:27] are there more people in here interested in openstack? [21:27] I keep meaning to be interested in it :) [21:27] i never see or hear about london openstack meetups [21:27] but right now KVM works great, virt-manager allows me to see all the VMs across all the hosts, I can't do some crazy dynamic stuff that openstack would allow and I don't really need to do [21:27] (or uk ones for htat matter) [21:27] BigRedS: :P [21:28] I could in theory do a live migration between hosts but I don't think the network configuration would be seamless [21:28] mattt: it's something that's on our radar at work, but we're not interested enough to have got round to using it anywhere, and we're not using it enough for anyuone to be forcibly interested in it... [21:28] really KVM does everything I need, openstack would only be because I want to play with the fasionable toys [21:30] BigRedS: i tested it personally about a year ago in the public cloud [21:31] was kinda neat booting VMs on VMs [21:32] I keep meaning to look more into all these new toys, but they just make me feel like a grumpy old man atm [21:32] AlanBell: do all your hosting clients have VMs, or do any of them have shared hosting? [21:32] shauno: yeah, so much to look at these days, not enough time [21:32] all our clients have VMs, but one of our clients uses his VM for shared hosting [21:32] graphic designer, small sites [21:33] I dislike shared hosting [21:33] for small sites it's fine [21:33] I don't want one customer to be able to cause a problem that affects another customer [21:33] AlanBell: does KVM have IO throttling ? [21:33] I want to sort out disk IO quotas at some point [21:34] no, it doesn't and that is a problem [21:34] we have had VMs OOMing and killing the disk and taking out other VMs [21:34] yeah, that's an issue w/ xen and xenserver i believe [21:35] i know xenserver has some sort of io throttling, not sure if it works properly tho [21:43] * AlanBell kicks of a 10GB sftp between Germany and London [21:44] it keeps getting faster [21:45] how do you back up your VMs? [21:45] 10.9 MB/s [21:45] gosh [21:46] it varies, some of the VMs are actually "warm spares/backups" for on-site physical boxes [21:46] so they have an on-site server, which backs up stuff nightly to the VM, the VM is always running the state of the database at close of the day before [21:47] if the on-site machine fails they can just start using the VM and we will increase the RAM as required [21:47] ah, neat [21:47] others we back up to FTP space on hetzner [21:48] and a few (mostly our internal stuff) with rdiff-backup to a 3TB drive in my house [21:49] I think for this rig of 8 VMs we will back them all up to a single VM in Germany (then back that up to FTP space) [21:50] Quite a setup AlanBell [21:50] well it is pretty small fry really [21:50] But I mean organised [21:51] and we don't manage it as efficiently as a dedicated hosting provider would [21:51] we run a small fry network, but it sucks [21:51] I'm not allowed time to do anything unless it breaks [21:52] AlanBell: the million $ question, how do you monitor it all ? :) [21:52] we would suck on the open market for hosting, we wouldn't be able to compete on price, so we just sell to our consulting customers [21:52] diplo: where do you work? [21:52] #small software company based in Nottingham, me and a colleague work remotely near Bath [21:52] mattt: yeah, that would be another reason why we don't want to compete against the hosting providers with 24/7 staff [21:52] Exactly what we do AlanBell [21:52] The guys in my office know what they know, no more no less [21:53] It's hard work at times, they aren't willing to try things [21:53] we monitor it using the hetzner tools for pings and http fetches, and virt-manager for live VM overview [21:53] we should do better monitoring really, but it is OK right now [21:53] i've been playing w/ nagios these past few months [21:53] a lot of people hate it, i really like it tho [21:53] generally things just work, if they don't then we know about it fairly quick [21:53] Ah that's one thing I have a good setup of [21:53] mattt: you'll learn to despise it :) [21:54] Really, I don't mind it BigRedS [21:54] BigRedS: i started out with zenoss, i learned to despite that very quckly [21:54] oh, one good monitoring trick, put irssi in screen on your most important server :) [21:54] AlanBell: haha! [21:55] so what do you monitor with nagios? [21:56] so far, not as much as i'd like [21:56] since a ton of stuff only runs snmpd, and i despise snmpd [21:56] but it ranges from monitoring network latency/packet loss, SSL certificate expirations, site availability, service availability, rabbitmq queues, etc. etc. [21:56] the openminds server is running snmpd, they monitor the hardware, that isn't our problem [21:57] I monitor mainly http/latency [21:57] But lot's of other stuff on some customer sites [21:57] disk space and IO is what I would like better visibility of [21:57] HBA Cards [21:57] AlanBell: it's great for those :) [21:57] mattt: you can have things running on the hosts themselves send nsca updates to nagios for things like diskspace and loadaverage and the like [21:57] Can do all that, rrd info [21:58] we use mon for that, but it's easy enough to write your own scripts to do it [21:58] That's the part I haven't spent much at yet [21:58] BigRedS: problem is we have a bunch of hosts that are stripped down, and we can't really install anything further ... fortunately they have snmpd running, otherwise we'd be at a loss [21:59] diplo: i like graphite / carbon as an alternative to rrd-based things [21:59] nagios can actually check carbon values too, so you can alert on your metrics [21:59] or performance data [21:59] err, whisper, not carbon [22:00] BigRedS: mon ? let me look [22:01] after UK freeview change their frequency does it mean my crystalpalace dvb-t file no longer works? [22:01] \o/ 10GB transferred [22:02] mattt: ahh, yeah, we've got a load of general-purpose servers so adding crap like that is fine [22:02] AlanBell: yay! [22:03] BigRedS: yeah, i'd just go nagios+nrpe in that situation, until i realize that nsca is better (haven't looked into that at all) [22:04] nsca is for windows ? [22:04] I may be wrong, don't monitor any windows stuff [22:04] fortunately no windows boxes in our env :P [22:05] for windows? No idea, never monitored a windows box [22:05] I monitored about 80 of them at my last job [22:06] but used snmp with Centreon there [22:06] diplo: sorry you had to go through that [22:07] heh, we used to 99% linux when I joined, by the time we left we were 95% windows ( New Finance Director ) [22:08] Who as it turns out was sacked :) [22:08] i was going to say [22:08] But he also implemented SAP, so between the two... [22:08] that seemed like an odd choice, for finance director :) [22:08] We all believe kick backs [22:09] anyway, time for me to crash [22:09] night all [22:10] gn [22:14] mattt: I've no idea which is better, I just got here and we were using nsca so I've stuck with that [22:25] Sod it, im off as well [22:25] gn === Lcawte is now known as Lcawte|Away [23:01] back again [23:01] with a renegade master. [23:02] tidy little discount on nostarch.com -- code geekpride [23:02] christel: hate that song w/ a passion :( [23:02] aww, sorry! [23:03] christel: it's up there with the janga (jenga?) bus [23:03] venga? [23:03] haha [23:04] that's it :P === zz_diddledan is now known as diddledan