=== Lcawte is now known as zz_Lcawte [01:07] apple UUID leak sure is interesting [01:08] er, UDID [01:11] ali1234: just fyi people are lawyering up over pirate, and he has registered businesses at his address :p [01:17] http://gawker.com/ [01:17] lulz [01:17] hahahaha === NimChimpsky is now known as HennaKaur [02:28] hi, I've recently set up my thunderbird with my calendar settings, and I'm looking for a way to run it through a screensaver - either directly or a screensaver that just displays the content behind it in some way. I want to leave my calendar running in the background on one of my machines, but I want to try and avoid image burn etc. [02:31] image burn does not happen any more [02:33] ok, it does, but it's not something you need to worry about [02:35] are you sure - I could be leaving my screen (LCD tv) on for hours at a time on a static calendar image (well static save for updates) [02:35] 24 hours a day? [02:36] possibly not quite 24, as I sleep for some, play video games for others - but you could be talking at least 6 hours a day, often non-stop [02:36] because of the nature of my work, a calendar like that would be really useful [02:37] LCDs don't burn in the same way that CRTs do [02:37] yeah - I've researched enough to see that, just saw that it's still techinically possible to get a similar effect, which is why I was just a bit concerned - I love my tv :D [04:14] morning all [05:28] morning === em is now known as em_ === em_ is now known as em [06:13] mrgh... [06:14] morning I guess [06:19] \o [06:19] you're an hour late :) [06:22] hehe [06:22] christel: he was supposed to be at work at 8, and it's past 9 (for him) now :) [06:22] oopsie! [06:23] * MartijnVdS has fresh tea! [06:23] \o/ [06:25] :D [06:25] i have lukewarm coffee :( [06:25] hehe [06:25] eww [06:31] ikr? [07:02] Morning all [07:04] \o diplo [07:08] diplo: :D [07:09] So an actual on ish topic this morning but probably to early [07:10] It looks like my company are going to brand some 1u servers to sell as a new software/hardware stack [07:11] So all the machines will be exactly the same, basically what I'd like to know is, is it worth trying to optimise Apache/php/mysql to this hardware or will the differences from stock to an optimised install not make a big enough difference [07:11] This could be from 6 - 10 users up to 80-100 [07:13] I'd only worry if you're hitting limits somewhere [07:13] can you scale up with even more hardware? Why not do that.. unless you're google-scale or Amazon-scale, it's probably not worth it [07:13] (unless, as I said, you're hitting a specific bottleneck) [07:13] It's not been tested with lots of users yet, I think it's quite heavy on apache/mysql with a few ( 3? ) users [07:14] time for some profiling :) [07:14] OK, sounds promising.. point me in the direction of what to look at ? Please [07:14] optimise the sql indexes of the app [07:15] Yes, learn how to tune mysql indexes first [07:15] then if that's not enough, tune Apache [07:15] log all slow queries and queries not using indexes then fix them all [07:15] and if that doesn't work, switch to a real RDBMS :P [07:16] ah ok, I've optimised sql indexes before.. will enable that [07:16] Is their a way of logging slow processes with apache like mysql slow queries [07:17] no [07:17] there's http://httpd.apache.org/docs/2.2/misc/perf-tuning.html thoguh [07:18] the reason for slow page loads will be slow queries [07:19] separating static files out to a separate, dedicated server might also help [07:19] You can use nginx or lighttpd on that instead of Apache (even more light-weight) [07:20] yeah, it's an internal app ( our erp system ) much like OpenERP [07:20] but only optimise stuff that is actually slow [07:20] but it's not hugely mysql dependant, we have our own middleware that writes db to a cisam ( legacy ) database and reads as well [07:21] But when I watch top, each time a link is clicked I notice a 3% spike on apache [07:21] Only optimize things that are slow, and only optimize after profiling :) [07:21] and I was thinking 3% x 20 users might start bogging it down [07:21] diplo: can you "script" user activity? Even crudely? [07:22] php-apc is a free performance boost [07:22] On a test setup? [07:22] That will help finding bottlenecks [07:22] You mean like run the stuff they would be doing ? [07:23] yes, or similar things [07:23] Well I saw a firefox pluging which the name has eluded me for now that does the job for you that you can record and then keep running [07:24] selinium [07:24] heh, just found it in my mail, yeah [07:24] kind of good, but does need some manual tweaks to recorded sequences [07:25] and can break on firefox upgrades [07:28] Thanks for input guys, will take a look at stuff today [08:18] Morning earthlings [08:18] hmm, 12.10b1 tomorrow [08:18] morning Sourcererling [08:20] lol === schwuk_away is now known as schwuk [08:21] whoa.. a schwuk :) [08:22] MartijnVdS: :) [08:23] christel: Are you around? Still getting a load of IRC spam :-\ [08:25] n1md4: perhaps ask in #freenode for help there [08:26] czajkowski: sure, only that christel was advising yesterday. I'll ask on freenode. [08:26] n1md4: yup but that is the channel to get more advice [08:29] czajkowski: Appreciate that :) [08:32] Good morning all! :D [08:35] Good morning peeps :) [08:37] I get the feeling that my router is in .be - doubt I'll see it today [08:38] Amusing that the last status on the tracking website is "Parcel is routed" [08:44] Not switched? === zz_Lcawte is now known as Lcawte [08:52] So it'll turn up tomorrow. One of the 2 days a month when I'm not in my home office [08:52] Bah [08:58] aww [09:03] hmm you don't need to sign the CoC to make a PPA any more. when did that change? [09:06] anywhoo, i approve of this [09:08] https://help.launchpad.net/Packaging/PPA is out of date and still says oyu need to sign the CoC [09:12] what's wrong with CoC? [09:12] nothing wrong with it's content [09:12] By signing the CoC, I learnt more about how one uses keys and encryption :) [09:12] i already know that stuff [09:12] I didn't :) [09:13] You tend not to learn stuff until it actually becomes a need [09:13] and i find the idea of signing an agreement that says i will be nice to be rather silly [09:13] bigcalm: that's where you're wrong ;) [09:14] but that is irrelevant [09:14] ali1234, it means if you're not nice, your account can be closed without argument. because you knew you were supposed to be nice [09:14] the point is, i just made a PPA without signing the CoC, so the requirement has been quietly dropped [09:14] MartijnVdS: I don't know how to do a lot of network stuff, but I'll learn if I find I need to [09:14] mungojerry: exactly. and it is extremely vague [09:14] but i like the sentiment [09:14] mungojerry: which means by signing it i am essentially allowing it to be used against me [09:15] * bigcalm kicks nginx into submission [09:15] * daubers hands bigcalm the nginx LART [09:15] what's the most reliable sw for copying isos to usb? the ubuntu one works for ubuntu only, unetbootin seems to not work quite a lot [09:17] bigcalm: Are you nginx savvy ? [09:17] heh, I am [09:17] I installed it just the other day [09:17] well, actually I wouldn't say I'm savvy [09:17] but I got it working for the office here [09:18] I've got it working ok for the most part, want to play with concrete5 CMS [09:18] http://85.119.82.250/concrete/ [09:18] diplo: not in the slightest :D gitlab kind of forced it upon me [09:18] oh right [09:18] Supports concrete5 request URLs [09:18] Is the part I can't get to tick :) [09:18] I've tried lot's of different things but not worked it out yet. [09:19] wait [09:19] wait [09:19] I thought by nginx you meant ngircd [09:19] LD [09:19] stupid names! [09:19] :D [09:19] mungojerry: many ISOs are hybrid these days. you can just dd them [09:19] having said that we do use nginx here.. but I don't know anything about it [09:19] ali1234, dd to the /dev/sdd or sdd1? [09:19] Thought I better start using my VPS I pay money for [09:19] mungojerry: to /dev/sdd [09:20] and should my machine freeze for the entire time? cos it does :( [09:23] i thought there was a kernel fix for the desktop freeze during heavy i/o [09:23] maybe it's just my machine thuogh [09:23] you wish there was a fix for it [09:23] i thought there was lots of news about it [09:23] a year ago [09:24] that was not the IO scheduler. that was the process scheduler [09:24] i assume you are refering to the BFS thing? [09:24] Hows the game coming along ali1234 ? [09:24] con kolivas? [09:24] no, more recent [09:25] my memory's not what it was though [09:25] anyway, my machine's still working atm ..which is unusual [09:25] usually copying a DVD sized iso to the usb kills the machine for 20 mins [09:25] not sure of the name, racing through tubes etc [09:25] diplo: been busy working and playing tekkit recently, so not done wanything with it [09:25] :( [09:25] looks good though [09:26] it has no name yet, and i've been working on it for nearly 10 years so don't hold your breath [09:26] wow [09:26] what's your day job ali1234 [09:26] freelancing [09:26] so when i've got no work, i do my own stuff :) [09:27] nice :) [09:27] sometimes i get to do my own stuff for work purposes :D [09:28] project i've been building: http://drumoff.tv [09:28] just finally got it online yesterday [09:28] totally not interesting unless you play drums [09:29] i know a few drummers [09:29] send them the link please :) [09:30] will do [09:30] i recommend having a youth category too for under 12s [09:30] or similar. there's some wicked young drummers out there [09:30] i'm not sure on the legalities of doing that [09:31] we're not even sure we will get enough entries for 1 category yet [09:31] tell popey , he has lots of friends on the internet :D [09:31] for U12's, believe it's fine here, not-so-fine in the US [09:32] running the competition isn't my department, i'm just the tech support [09:32] you don't even have to show the drummer [09:32] I'd recommend you show jono [09:32] jono likes rawk [09:32] * mungojerry submits a video of air dumming [09:33] s/dum/drum [09:33] technically i think that is allowed lol [09:33] making mouth noises [09:33] rules say you can only use bass hat and snare but it doesnt say you have to use all or any of them [09:35] does this count ;) https://www.youtube.com/watch?v=F7SrSFn7isk&feature=plcp [09:36] it's longer than 30 seconds, so no [09:36] :P [09:36] kit is pretty much right though [09:42] Morning all [09:43] yo [09:44] running hdparm (non cached) on a thin provisioned disk yields 600MB/s , over a 125MB/s link? [09:45] real test (e.g. dd) yields 100MB/s [09:45] what's hdparm doing wrong? [09:48] megabits vs, megabytes? (just a guess) [09:51] nope [09:51] sitting in a cache somewhere? [09:51] dd yields more realistic results. [09:51] hdparm -t uses buffers but not cache [09:52] does that no only work on real disks though? Rather than networked storage? [09:52] i.e. is it just sitting it in RAM? [09:53] it's an iscsi connected disk [09:53] urgh [09:53] running it on thick provisioned disk works [09:53] nuff said [09:53] it's enterprise quality storage [09:53] * daubers has never like iscsi [09:54] i am running against a 10Gbit and 1Gbit iscsi connected disks [09:54] mgdm: glad to see you haven't given up irc :D [09:54] it's a stopgap until FC [09:54] What do you get on the 10Gb link ooi? [09:55] well that's the beginning of the problem [09:55] much less than we shou,ld [09:55] which is? [09:55] can't get IBM to admit the problem [09:55] hdparm giving around 130MB/s and i would expect more than 3x that on a V7000 [09:55] How many drives in a V7000? [09:57] there's 12 600GB SAS 10k atm [09:58] from 16x3TB SATA drives I can get ~800-900MB/s over a single 10GbE link (using NFS) [09:58] Raid 6'd ^ [09:58] we have a serious problem and i think it's IBM fault [09:59] they supply the blades, and the SAN and the customised vmware [09:59] last problem we had was very similar for sas connected storage, and it was with their customised vmware [09:59] dodgy lsi VIB [09:59] ooof... SAN..... I'll step away from that now [10:00] daubers, is that hdparm speed or dd'ing a 1GB file for example [10:00] mungojerry: That's dd'ing a 20GB file (we always test with files bigger than the amount of RAM in the NAS) [10:00] the consulting company didn't even notice the problem and tried handing over to us [10:01] If I do it smaller than the free RAM in the NAS I'll get it running as 1200MB/s (i.e. near line speed) [10:01] 1048576000 bytes (1.0 GB) copied, 15.7809 s, 66.4 MB/s [10:01] lol [10:01] that's read speed [10:02] gonna have to install linux directly onto a blade instead of vmware to find out where the problem is [10:02] the iscsi on the 1gb link is faster than the 10gb v7000 [10:02] i didn't choose this hardware btw [10:04] have a support call open with IVBM atm but they aren't looking holistically [10:04] just at the SAN logs [10:08] daubers, do you dd the file locally or to /dev/null? [10:09] Read tests I'll dd to /dev/null [10:09] local dd will be limited by the local disk [10:09] yes, just checking you wre doing to same test as me [10:09] An interesting one I tend to run as well is do it with lots of files simultaniously [10:10] CIFS tends to work better that way for some reason [10:10] or give better results anyway [10:10] have you used iozone? [10:10] http://www.iozone.org/ [10:11] Yes..... but I'm generally not interested in IOPS for my final use of these systems [10:11] the streaming read tests tend to be more reliable for what I'm doing [10:11] it helps with tuning nfs too though [10:12] Again... not really with what I'm doing :) End use is video editing, so lots of large streaming reads. Most of the IOP based things are based around random read/write cycles [10:12] Which is what most servers will end up doing [10:12] (i.e. databasing or working with smaller files) [10:13] are you also involved in backups? [10:13] for your work systems [10:14] Yes, but the backups are quite slow (and not much changes in the office itself) [10:14] having a running argument as to the expected daily % change on our systems [10:14] heh [10:15] I was arguing the toss with someone the other day who's recorded daily changes was less than a tenth of a percent of the total dataset [10:17] :-o [10:17] i think that's called static :P [10:18] our data includes VMs [10:18] windows has quite high daily change on quiet system [10:18] linux is much lower. [10:18] and is more tunable [10:20] do you backup things outside of the users home directory? [10:21] yes, all data and entire VMs [10:22] just discovered http://sugru.com/ looks like someting every man should own [10:24] i'm increasingly convinced that only backing up home directories, conffiles that have changed and data like mysql dumps and Maildirs is good for most use cases [10:24] ubuntu + overlay of stuff that makes it server X [10:25] unfortunately the enterprise has all sorts of apps that are complex [10:25] virtualisation makes restoration of service a million times easier now though [10:26] and consultations have tendency to only recommend windows, and PHBs have tendency to do what highly paid 3rd parties tend to recommend [10:27] i've started developing in a VM recently, its so much nicer to go, oops. shouldn't of upgraded, everythings broke. restore snapshot! [10:27] gord: christel needs to annoy you [10:28] Morning davmor2 [10:28] then i'm closed for business! no more customers today [10:29] oh i cant remember what i wanted you for [10:29] bigcalm: hey dude, steak minus 8hours and counting [10:29] fantastic [10:29] oh sweets.. davmor2 said you had to tell me about some Amazing Japanese Sweets [10:29] japanese colleague brought amazing sweets. found out they were sweets but dried white bait UGHGH [10:29] davmor2: we shall see. I had steak at the weekend. Don't know if WS's one will live up to the high standard [10:30] christel, http://gordallott.com/+ i share them on my google + thing every two weeks, have a service that sends me japanese candy twice a month :) [10:30] christel: gord likes to eat the most disgusting things. I'll swallow almost anything, but not what he has to offer... [10:30] o_O [10:31] Please don't read bad things into that statement! [10:31] ^-^ [10:31] the sweet potato candy i got last time was delish [10:31] sweet potato candy.. fantastic [10:31] bigcalm: :o [10:31] bigcalm: what were these ones he made you guys try? :) [10:31] ali1234, the dd direct to usb didn't work :( [10:32] christel: That's why I said annoy gord, I'm pretty sure it was sour pear something or other [10:32] christel: a question for gord [10:33] gord: is it like Surprise stuff or do you select preferences or something? :) [10:33] That 'sweet' was one hell of a surprise [10:34] * mungojerry wants sweets now [10:34] it was umboshi candy, umboshi is a pickled plum dish that japanese people love but its an aquired taste that i think is fun to surprise people with [10:34] eric the elephant mmm [10:34] mungojerry: the ISO does have to be specially made to do it. ubuntu ISOs should work [10:34] what ISO are you trying to use? [10:34] I tried some recently that were licorice & chilli. that was .. confusing [10:35] christel, its a surprise, basically the guy running it selects 2-3 candies and stuffs 400 envelopes with them and mails them out to people all over the world. he sends out really nice emails explaining what everything is i think the how it works on http://candyjapan.com explains it better than i ever could [10:35] ali1234, i used an elementaryOS nightly build. however since this is based off ubuntu i'm using the ubuntu startup disk creator now, whic [10:36] h should work i hope [10:36] gord: ooh danke [10:51] elementary is looking really neat atm [11:03] Any body know of a good tut on writing init.d scripts? [11:07] they're not that standardized [11:07] bigcalm: for what system? Don't forget you have upstart and systemd now instead [11:07] Debian 6.0 (squeeze) [11:08] they just respond to !$ and start/stop/restart the process with whatever mechanisms available [11:08] Uh, $1 even [11:08] bigcalm, cp /etc/init.d/skeleton /etc/init.d/myscript [11:09] directhex: ta [11:09] upstart/systemd is much more logical, but upstart has some limitation and systemd isn't really implmented anywhere [11:09] bigcalm, nano /etc/init.d/myscript [11:09] directhex: latter part a bit patronising :) [11:09] "nano -w" ftw [11:10] i've been a professional linux systems manager for 8 years, and i use nano :D [11:10] i once went to a job interview and they asked me "vi or emacs?" [11:10] and i said "nano" [11:10] directhex: that's cause nano rocks [11:11] directhex: sounds like you're introducing yourself at an AA meeting [11:12] directhex: plus it sounds like an episode of mork and mindy when you say it twice :D [11:13] daubers, the 800-900MB/s was against how many disks? [11:13] ali1234, did you get the job? [11:13] mungojerry: 16 [11:13] i grew up on solaris, so vi all the way [11:13] daubers, thanks [11:14] what's the SAN/NAS? [11:15] mungojerry: Something I designed/built :) Was a xeon based nas with an areca card in it [11:16] mungojerry: no [11:19] i grew up on a zx spectrum, so uh... zx basic interperator all the way? [11:22] gord: I bet you are really glad that there is that spectrum emulator then :D [11:22] wouldn't get any work done without it! [11:30] daubers: happy birthday (if facebook isn't lying) [11:31] who would do that? just go on the internet and post lies? [11:43] dutchie: Ta [11:43] directhex: Everyone eva? [11:46] daubers: lies? what are those? [11:47] MartijnVdS: According to various governments, facts. [11:47] ah ok, that's good then [11:49] hi folks [11:50] hey hoover [11:50] happy daubers day \o/ [11:50] is it daubers day? [11:50] * davmor2 congratulates daubers on making it another year [11:51] congratifications! :) [11:51] ta [12:16] * popey waves to hoover [12:34] hey popey [12:34] thanks for your support mate ;-) [12:34] np [12:36] I hope all goes well tonight. I would have expected 12.04 to work well on such a relatively recent machine, but who knows [12:37] one thing I noted is that the gpu fans run at full speed with the latest nvidia binaries, so I hope it's possible to downgrade to an earlier version [12:38] (unless the bug's been fixed in the newest drivers, of course) [12:42] hoover, what version of nvidia driver you on? [12:43] chaps.. [12:43] Quadro NVS 290/PCI/SSE2 if someone has that card... [12:44] 195.xx ATM [12:44] and they have the 173 nvidia binary driver.. [12:44] 12.04.1 lts? ;-) [12:47] hmmm === issyl0 is now known as Guest17314 [12:49] * bigcalm wibbles back into the office [12:50] i seem to have 304.43 nvidia driver version [12:52] yep I ended up with 195.xx by trial and error, seeing which one would not show the 100% fan speed problem [12:52] I think the newest one was 295.xx or thereabouts back then [12:53] hoover, which card you have? [12:53] popey: Is their a way to find out what package a meta? package installs like mail-server^ ? [12:54] i am using a PPA to get the newer one on 12.04 [12:58] hey biggie [12:58] mungojerry: and 8800GT [12:59] Afternoon hoovie [12:59] hoover, 304.43 is supported for 8800gt http://www.nvidia.com/object/linux-display-ia32-304.43-driver.html [13:00] mungojerry: thanks [13:01] as mentioned , i'm getting it from a PPA on 12.04, but it's been stable for me, and newer is generally better for nvidia drivers [13:01] I'd use the latest drivers if the fan problem has been fixed. I guess I'll find out tonight. [13:08] diplo, section: metapackages [13:09] alan@deep-thought:~$ apt-cache show ubuntu-desktop | grep Section [13:09] Section: metapackages [13:09] ta [13:16] apt-get install foo^ installs a task [13:16] tasksel --list-tasks [13:19] I just wanted to see what packages the task installed [13:19] Or whether I should just install manually [13:20] AlanBell: you about at the moment? [13:23] ok folks laters all [13:26] hi [13:29] I could use some help trying to sort out a problem with an external hd, can anyone help? [13:29] Goshawk: we wont know until you tell us what the problem is :) [13:38] ok, long story but basically i installed ubuntu on to an eternal hd by accident and i need the data that is on it. I ran a recovery program and can find all the data (mp3s etc) but i cant do anything with it because its locked, ubuntu is telling me i dont have permission to use it. So the question is how do i get ubuntu off the ex-hd without loosing all of the other stuff? [13:39] run as root [13:39] just to be clear, the recovery program has made it such that you can see all these files in a normal file browser? [13:39] yes, but all the files have a padlock sign on them [13:40] <_serial_> chown it [13:40] <_serial_> them even [13:40] run as root is less change :) [13:41] how do i do that? [13:41] jacobw: less change, more danger :P [13:41] <_serial_> sudo chown -R username:username /path/to/files/* [13:41] ♫ I ain't afraid of no root! [13:42] yes, you're right, it would be safer to chown them to be readable to your user only. [13:42] MartijnVdS: welcome to windows... [13:43] is there a way to just uninstall ubuntu from the hd? [13:44] <_serial_> run gparted and remove the partition, you can do that from live cd/dvd/usb or install it on your system [13:45] <_serial_> remember to backup though [13:47] that will delete everything in the partition, though [13:47] <_serial_> does the task though? [13:49] yes, but i think it would be much better to recover and verify the 'lost' data first [13:49] for sure [13:50] <_serial_> true but i did say backup :) [13:52] Ok, so find the path of one of the files that you can see in the file browser, and test using `chown youruser. /path/to/file` to make it accessible to your user [13:54] there's no way, given a volume with a Ubuntu install and some other stuff on it to simply remove allthe Ubuntu bits and leave the other stuff [13:55] That took forever to actually be sent... [14:01] ok, working on it now, thanks for the help [14:44] o/ jussi [14:47] jussi: that's administrator though right [14:58]