[00:01] osmosis: sorry was afk, 'update-rc.d' is the command that performs functions similar to chkconfig === jjohansen is now known as jj-afk [00:35] SpamapS: are you the father of php in ubuntu? [00:37] god no [00:37] php is definitely a bastard [00:37] drat [00:37] but occasionally I look afte rthe little twit [00:37] whats up? [00:37] was just wondering who cares about the packages [00:37] I do .. I confess ;) [00:37] well low importance, but bug 744371 [00:37] Launchpad bug 744371 in php5 "PHP5-FPM doesn't always restart on 10.10" [Medium,New] https://launchpad.net/bugs/744371 [00:38] FPM is horribly broken on all versions prior to natty [00:38] it's one which will fall off the radar and suddenly have 3 people complainingloudly about how ubuntu sucks... [00:38] this is in natty too, judging byinit script [00:39] ahhh [00:39] I see now [00:39] Thats an easy fix [00:39] hallyn: as soon as I get my upload rights, I'd be happy to make that my first upload. ;) [00:39] yeah, np for natty. hard to justify an sru unless someone else confirms the problem [00:40] when is that projected? [00:40] i'm a big jealous of the fast-track you've taken :) [00:40] Not sure, whenever the DMB enacts their voting results [00:40] on full server suite? [00:40] I wouldn't mind having a nice well defined set of concerns to focus on like you. ;) [00:41] no they voted core dev for me :) [00:41] well somehow lxc ended up not in my set. which is a real fing bugger [00:41] wow [00:41] big time congrats [00:41] * hallyn starts building up a sponsor queue to ship to SpamapS [00:41] you dont' have much else to do right? [00:41] * SpamapS is quite humbled by the names of the team members [00:41] No nothing really to do at all. :) [00:49] SpamapS: well, i think i'm goign to have to do another request to add lxc. suspect they'll tell me to shove off, but... [00:50] hallyn: definitely not.. I think they'll give it to you in a blink [00:51] hallyn: the pain is that you have to sit through the meeting again.. I think they should have async meetings where votes/questions can be done via email. [00:58] jibel_, yes === dendrobates is now known as dendro-afk === jj-afk is now known as jjohansen [01:20] i need some help, im getting ABYSMAL mdadm resync performance on a brand new raid6, for a 4 x 2T array, it's taking nearly 22 hours, what is the problem? thanks. [01:23] ejv, that's normal [01:24] especially for such large disks [01:25] you can speed up the process a bit by changing /proc/sys/dev/raid/{speed_limit_max,speed_limit_min} [01:25] yea i changed those, and it didn't realy impact performance in a noticeable way [01:25] it's holding a steady 25MB/sec [01:25] yeah, that's about all you're going to get [01:25] what if i changed the stripe_cache_size in /sy/block/ [01:26] sys* [01:26] since raid 6 is double parity [01:26] i just figured since they disks are empty and have no data, initializing a new array should be a quick and simple operation, not 22 hours lol [01:27] the* [01:27] nope [01:27] it has to build the entire array [01:27] if I bought a better CPU, could I perhaps get a boost? [01:27] yes [01:28] another factor is how fast your disk controller is [01:28] I'm guessing all four disks are on one controller [01:28] cpu is a previous gen athlon ii, it's a software raid, via mdadm [01:28] if that controller is only PCI or only PCIe x1, that's going to be a significant limitation [01:29] see how high your iowait stat is, top can tell you [01:29] qman__: http://pastebin.com/raw.php?i=n8md9gB8 [01:29] (sdb-e are the md6 raid6 members) [01:29] not too much [01:30] what did you set the two values to? [01:30] 500000 for max, 50000 for min [01:30] yeah, that's not high enough [01:30] try 2000000 for both [01:30] i thought that's in megabytes [01:30] im only seeing a resync speed of 25mb/sec [01:30] kilobytes IIRC [01:30] hmm [01:31] one moment :) [01:31] will adjust [01:31] setting it too high won't have any ill effects on the system, it won't do more than it can handle [01:32] it's just set that low for in-production operations, so they won't slow you down [01:32] ok changed the values on both [01:32] is this operation swap intensive, should i enable my swap (it's off right now) [01:33] no [01:33] kk [01:33] it's pretty much all CPU and I/O [01:33] also, what type of drives do you have? [01:33] these are low power green drives, segates 5900 rpm [01:33] yeah, that's going to hurt it [01:34] well how does this math work, if each drive via iostat is reporting 25mb/sec, that means my system is doing 100mb/sec of disk i/o, that's not too shabby right? [01:34] yeah [01:34] ok lol [01:34] * ejv needs to feel good about his purchase [01:34] should mutt update itself when a new mail in received without any user input? it appears for me i have to do some sort of input before a new mail shows [01:34] 100MB/s isn't too bad [01:35] for the class of hardware you're using, anyway [01:35] right, older cpu, low power drives [01:35] etc. [01:35] you're probably being limited by the disk controller [01:35] you know what [01:35] i didn't think of that! [01:35] those drives should be capable of 60-80MB/s each, but if they're all on one controller, on one PCIe lane, that's slowing you down [01:35] it's on a pci sata 4 port [01:36] yeah [01:36] SATA 150 for each port [01:36] man i TOTALLY didn't think of that [01:36] the PCI bus is shared with all other PCI devices [01:36] and is only 133MHz [01:36] these are SATA II drives on a SATAI ports [01:36] yea [01:36] wow totally didn't think of that [01:36] so i'm probably pushing the limits of the controller [01:37] all my onboard motherboard SATA slots were occupied, so I had to use the 4port SATA PCI [01:37] i can bet you the onboard SATA ports are MUCH faster [01:37] very likely [01:37] *sigh* [01:38] but 22 hours still isn't bad [01:38] thank you very much, you made my day with that comment [01:38] my last resync took three days [01:38] a light bulb went off in my brain lol [01:38] yea, well here's my situation [01:39] i have an old raid6, 500G members, in operation since 2008, that are occupying those precious onboard controller ports, my plan is to migrate the data, and then completely phase them out [01:39] then i'll move the new 2T 5900rpm discs to the onboard spots and remove the PCI controller until I need it again [01:40] sounds.... sane, right? lol [01:40] yeah [01:40] wonderful :) [01:41] shouldn't have any problems, as long as you actually go offline for the switch [01:41] according to newegg, it looks like 4 port PCI SATA300 cards are really common now [01:41] only need a few minutes, but trying to swap it in live is usually a bad idea [01:41] back in 2008, they were *expensive* [01:41] yup i'll be shutting the server down completely, and very carefully moving everything [01:42] I lost three drives in one day on my last array [01:42] god, what happened [01:42] were they from the same manufacturing batch? [01:42] seagate firmware, I suppose [01:42] yeesh [01:43] sry to hear that [01:43] that was a raid 5 [01:43] so that wasn't able to completely protect [01:43] the replacement is a raid 6 with hot spare [01:43] i wonder if mdadm supports hot spare [01:43] it does [01:43] ok i might consider that then [01:43] that's what I use [01:44] other tip [01:44] heh, raiding with 5900rpm drives are never run [01:44] set up smartctl to email you [01:44] damn sleep mode on them [01:44] you recommend it for such large members, because of the really bad URE statistics these days? [01:44] sleep mode? my drives better not sleep... [01:44] it's built into the firmware [01:44] I don't buy green drives, too much hassle [01:45] and too many horror stories [01:46] but anyway, set up smartctl to email you [01:46] my new array has already had two drives start to fail, I RMA'd one and I'm going to RMA the other next time I get a chance to pull it apart [01:46] well i did my homework the best i could, and didn't ever see any issues with the ST32000542AS drives for use in raid, i'll keep looking [01:47] hmm email... scary [01:47] don't i need an MTA and SMTP and all that gobbly gook then... [01:47] all you need is postfix on the server [01:47] if you have a mail server you can smarthost it, otherwise you can set it up with gmail or something [01:48] hmm ok i'll look into doing that [01:48] it's not too involved and it's worth the advance notice [01:48] yea definitely [01:48] i have a mail server managed by a friend on another remote host, I imagine I can point smartctl to that [01:48] or sure, google like u said [01:48] err gmail* [01:53] thanks for your help again qman__ [01:53] i shall return :D [01:53] 1175 minutes to go [01:54] no problem === dendro-afk is now known as dendrobates === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [05:11] would you recommend raid 10 for a high i/o's server? === MetaJake__ is now known as MetaJake === dendrobates is now known as dendro-afk [06:19] See http://www.debian.org/security/2011/dsa-2208 Ubuntu 10.04 bind9 users might be interested in running the backport I setup at https://launchpad.net/~cp/+archive/net === Guest79834 is now known as Jem === Jem is now known as jemofthewest [06:30] if i want an encrypted raid and lvm what is the correct order for setup? ie make an lvm of 4 drives, then raid then encrypt or what order should it be in === Jem is now known as Guest38942 [07:56] I have a problem [07:56] after sudo poweroff the fans and cpu fan keep spinning [07:57] and I know it's not an hardware problem because I have just shutdown the computer and fans with power button while I was in the BIOS. [08:02] :| [08:09] rcconf: I've had that happen before with some hardware.. seems like some of them just ignore that final instruction to power off. [08:10] SpamapS: how did you solve it? [08:10] hammer [08:11] Daviey: you coming to the openstack design summit ? If not, anyone from Canonical ? [08:12] lifeless: shut up thanks :) [08:22] rcconf: I pushed the power button. :-P [08:22] eh :\ [08:22] that doesnt solve my issue [08:55] rcconf: I noticed that my brand new PSU doesn't immediatley power off after shutdown, it keeps running for about 15 seconds to make sure all the drives safely spin down, are you components on perpetually? [08:55] s/you/your/ [08:56] ejv: yes, i think so [08:56] i have tried to wait like 30 minutes [08:56] yea that's far too long === skaet is now known as skaet_zzz [08:57] at that point i'd toggle the manual switch on the power supply [08:59] ejv: it does not have one [08:59] ejv: computer is old [08:59] :) [09:01] ejv: and [09:01] CPU fan is using 5300RPM [09:01] spinning i mean [09:01] and I cant control the speed [09:01] old BIOS [09:09] dunno if you can complain toooo much if the box is a dinosaur rcconf :p [09:09] ejv: 2000 computer [09:09] You can push it off the roof, tho [09:09] not that old [09:09] twb: if you give me another one [09:09] assuming it's small enough to fit through the door [09:10] a power unit from 2000... that's kinda pushing it [09:10] you could prolly do yourself a favor, electric bill wise, and get a more efficient PSU :P [09:11] it's some good old machine I found [09:11] but is it an excuse for remaining on, of course not :P [09:11] and gave it life [09:11] im God [09:11] :| [09:12] rcconf: a binge-drinking roughneck with a penchant for turning lovers into vases? [09:13] english please [09:14] http://en.wikipedia.org/wiki/Greek_mythology ? [09:15] I dont like false gods [09:15] wont read [09:15] All things are true and false and meaningless. [09:15] http://en.wikiquote.org/wiki/Principia_Discordia [09:16] sudo nano /etc/hosts 127.0.0.1 twb-links [09:17] You should use sudoedit [09:17] ? [09:17] Instead of "sudo nano" [09:17] no. [09:24] ttx, I'm /probably/ not. zul certainly is [09:24] ttx, Clashing with the release week is a mega PITA. [09:35] zul: did you register yet ? [09:37] morning [09:37] hello ttx ! [09:37] huats: yo [09:37] hello === squishy is now known as SquishyNotHere [10:23] it would be awesome to have a php based manager like webmin, but php... I know there is a php extention to the webmin pl files, but this is old === SquishyNotHere is now known as squishy === cs278| is now known as cs278||work === koliz is now known as koltroll [10:54] Mr. Hallyn??? [10:54] Are available? [10:55] ? [10:57] hallyn I've got a question regarding your kvm spice ppa === squishy is now known as SquishyNotHere === SquishyNotHere is now known as squishy [11:15] wout-lnx: It's 5:15 in the morning where he is. It'll probably be a couple of hours :) [11:20] HELP!!! how to upgrade from 10.04 to 10.10 ? [11:24] ttx not yet hoepfullly today [11:24] phoenixsampras: sudo-apt dist-upgrade [11:24] phoenixsampras: sudo apt-get dist-upgrade [11:25] or release.. [11:25] wait [11:25] phoenixsampras: sudo do-release-upgrad [11:25] phoenixsampras: sudo do-release-upgrade [11:26] rcconf: IIRC there are at least 3 ways to power down an "IBM PC compatible" computer, and it sounds like linux guesses wrong about what your hardware supports... ;) [11:27] JanC: motherboard is amd pc chips m810lmr [11:31] phoenixsampras: edit /etc/update-manager/release-upgrades and set "Prompt=normal", then do-release-upgrade [11:36] thank you xampart, you are wise [11:36] phoenixsampras: my command doest work? [11:36] doesnt [11:36] hi... I'm using backupninja for backups. I'm pretty sure when I set it up (2 yrs ago) I created config files in /etc/backup.d/. I edited them a while back only to see that edits have no effect. I just discovered that backupninja apparently uses files from /usr/local/etc/backup.d which appear to be copies of the files I originally created in /etc/backup.d (old versions). Why could that be? Should I simply edit those in /usr/local/backup.d fr [11:36] om now on? [11:51] xampart: how to verify the ubuntu server release version? [11:51] cat /etc/issue [11:52] for starters [12:02] or lsb_release -a [12:12] draven_sol, raid first, then lvm, then encryption [12:16] rcconf, by default, LTS will only upgrade to another LTS, so to go from lucid to maverick you need to change that setting first [12:23] k [12:43] how do I check which repository a package is installed from? === squishy is now known as SquishyNotHere === SquishyNotHere is now known as squishy [12:45] n/m, found it [12:59] muszek: what was the answer> [12:59] ? [13:05] a7ndrew: apt-cache showpkg _packagename_ (doesn't show exactly which version is installed, but lists all installable versions and some data about them, incl. the origin... which is good enough for me) [13:14] Daviey: good morning [13:19] muszek: thanks, interesting to contrast that with what 'aptitude show' reveals. It runs faster too. [13:23] hello, whats the best virtualization software to use? [13:25] bda: kvm of course [13:25] ;) [13:26] how can i log user actions on server? [13:26] pmatulis: i thought ubuntu recommended it for reason :) would it be better than a headless virtualbox, though? [13:30] bda: try both and decide for yourself but i suppose kvm will become more and more integrated and supported in ubuntu than virtualbox [13:30] bda: so you need to think of the future a bit [13:31] wout-lnx: hi [13:32] Hello hallyn [13:32] wout-lnx: fwiw i'm not currently touching myspice ppa, bc dev-zero is going to update his and push to universe [13:33] hallyn: Ok but dev-zero's ppa doesn't contain a spice compatible kvm bin [13:34] pmatulis: right. thanks for your help! [13:35] hallyn: It is my understanding that kvm in ubuntu is not spice enabled because spice libs are not yet packaged. Correct? [13:36] hggdh, hello! [13:36] wout-lnx: yup [13:36] Good news about the patch [13:36] The PPA has been updated btw. [13:37] wout-lnx: yes, dev-zero is going to (AIUI) use my qemu-kvm packaging [13:37] wout-lnx: so he'll basically end up with my qemu (quasi) package and his own updated library packages [13:37] hallyn: Can I help? [13:37] lun0: I guess it depends which actions you want to log. Do you mean their shell commands? If so apparently you can use the 'script' command to do it, I'm not too sure of exactly how its done though, as Ihaven't had a need. [13:37] hallyn: Testing needed? [13:38] wout-lnx: sure. [13:38] wout-lnx: are you on gimp#spice? [13:38] a7ndrew: ok, will check it [13:38] hallyn: I'm not a coder, nor a packager === war10ck is now known as pan1nx [13:38] hallyn: not yet [13:38] wout-lnx: ok. well, i'm starting to think dev-zero may have gotten distracted... (real life does sometimes get in the way if you're doing this for fun :) [13:39] wout-lnx: anyway the stuff i have *should* work. kim0 has tested it [13:39] hallyn: gimp#spice. gimp=gimpnet?? [13:39] wout-lnx: actually maybe he never got audio to quite work [13:39] yeah [13:40] you can ping dev-zero and ask him if he has new paackages he can push soon to his ppa [13:40] ivoks: hey, od you have any currently-unused manzanita chassis which are not behind a firewall? [13:40] feh [13:40] wout-lnx: I got as far as running a vm with spice from cli [13:40] hallyn: I got as far as running a vm with spice from cli [13:41] wout-lnx: ok, [13:41] wout-lnx: i've gotten redhat to boot graphically [13:41] hallyn: running a vm using libvirt, with virt-manager with spice devices [13:41] wout-lnx: but i was doing funky things with the library versions anyway [13:42] wout-lnx: so that's another reason to use dev-zero's packages. If he says he doesn't have time then i can go ahead and update his packages and create new ones in my ppa [13:42] hallyn: funny thing is, I could connect to the cli vm with spice client just fine, I couldn't connect with the libvirt one.... (only got frozen screens and then client would die) [13:42] oh, not sure that's funny. [13:43] how does libvirt do the connection? [13:43] It's probably due to my not properly falling back tothe old protocol [13:43] which, again, is why we should switch back to the simple old libs :) [13:43] ;-) [13:43] Well I'm connected to gimnet... [13:45] wout-lnx: ok, i've pinged dev-zero several times over the last week, so he may just be on vacation. let's see if he was just ignoring me :) [13:50] Hello, is it recommended to apt-get remove consolekit on Ubuntu Server? What is console-kit-daemon used for? === PrestonConnors1 is now known as PrestonConnors [13:53] PrestonConnors: If it was recommended to remove it, it woulnd't be there to begin with. [13:54] soren: Do you know why console-kit-daemon would consume many resources for a server with a lot of ssh connections? console-kit-daemon is consuming about 3.6GB of RAM and about 60% of CPU on these servers. [13:54] Probably becuase it's broken somehow. [13:54] soren: When I kill console-kit-daemon it automatically restarts with normal usage and starts to linearly consume more resources over time. [13:55] PrestonConnors: Which Ubuntu version is this? [13:55] 10.04.2 64bit [13:55] fully updated [13:55] it is across 20 servers, 16 cores 96GB ram each [13:56] looks like bug 232557 [13:56] Launchpad bug 232557 in consolekit "console-kit-daemon leaks memory" [Medium,Confirmed] https://launchpad.net/bugs/232557 [13:56] is there a safe way to disable console kit daemon [13:57] i am killing it every day via crontab but would like a better resolution. some people recommended to apt-get remove it because it is quite buggy. === twister004_ is now known as twister004 [14:23] jamespage: im dropping the cobbler4j stuff, its no longer shipped in 2.1.x [14:24] :-( [14:24] no this is a good thing ;) [14:24] only joking! [14:24] so was i [14:25] what that it was a good thing or that you where dropping it? [14:25] * jamespage is really confused now [14:25] upstream no longer ships it [14:25] TBH not surprised.... [14:25] neither am i [14:27] tbh i dont think anyone installed it either [14:39] Has someone practical experience with Request Tracker? [14:43] zul: don't forget to add python-koan :) [14:44] Daviey: howdy! Hey the eucalyptus powernap config and the separation of the tools have not yet been merged into the branch right? [14:45] RoAkSoAx: frig ill do it after beta [14:48] is there a way to do a non-interactive install of Postfix that automatically chooses "No configuration"? [14:48] zul: ok :) [14:52] Anybody know how I can connect to rdp on a non-standard port using remmina or terminal server client apps? [14:53] Hey Everyone, What is a good virtual server platform to use on ubuntu? [14:54] vmware [14:54] ok would prefer something else because the stupid interface is buggy [14:55] Doonz, I like Virtualbox [14:55] I use that daily, and I like the RDP thing you can use on it's virtualmachines. That's what I'm trying to connect to right now. [14:56] If I can just figure out how to change port on client. ughh [14:56] kool [14:56] wich client [14:57] I've got Remmina and Terminalserver client. But doing ip:port doesn't seem to do the trick [14:57] melter: an easy way to do it is to install postfix interactively the way you want it, then use 'debconf-get-selections | grep postfix > postfix-options.debconf' to save your interactive responses, then do to unattended installs just do a 'debconf-set-selections < postfix-options.debconf' before you install the package [14:58] melter: or something close to that [14:58] kpettit: may sound stupid but have you tried ip [space] port [14:59] semiosis: thanks [14:59] Doonz, no. I'll try that... [14:59] RoAkSoAx, no.. i was going to merge and upload it concurrently... is that ok? [14:59] haha, that was it. good call Doonz [14:59] woot! /me chalks up his deed for the day [15:00] Daviey: yeah that's fine :) I was just wondering [15:00] i forget wich client i was using and thats how it wanted the format to be [15:00] VirtualBox does this thing where you can turn on RDP to any virtualmachine. You can see boot screen and everythign. Very nice, espically with OS's that don't have RDP [15:00] ill have to install that when i get home [15:01] basically i have a quad core with 16gb ram and would like 5-8 linux servers on it to play around with carious things [15:01] The default virtualbox setup is great for desktop vm stuff. But I use the "headless" stuff when I need to run it on a server. [15:01] RoAkSoAx, looking to upload when b1 is due. [15:01] yeah it will be headless so remote management features are key [15:01] VirtualBox should do great for you then. Has really good snapshot and import/export stuff [15:01] does it allow you to mount iso's for install os's? [15:02] yes [15:02] You can do it all CLI or through the GUI [15:02] i like gui.... [15:02] * RoyK wonders how long it'll take for Oracle to ruin vbox as well [15:02] I normally use the GUI myself. But when I install it remotly I do the CLI stuff. [15:03] It actually did better than vmware server. I had a couple of vmware images that kept crashing vmware, but worked with virtualbox for whatever reason. [15:03] kool im a linux newb so i like to mess around with vm's so i dont hoze my system [15:03] RoyK, No kidding. Hope they don't mess with it to much like with OpenOffice or MySQL [15:04] Doonz, smart. I get too anxious sometimes trying out crap and mess up my machines frequently. [15:04] I'm doing the alpha of 11.04 right now on my desktop. Bad idea, I should have just done it in a vm [15:04] yeah and when i break things that make it so my wife can do what she wants i get to sleep on the couch [15:04] haha, same goes here. [15:05] kpettit: I was an OpenSolaris user, but that abruptly stopped after Oracle pulled the plug - using OpenIndiana for storage now... [15:06] I liked alot of the ideas OpenSolaris came up with. Never had a chance to play with it much though [15:06] qman__, thanks for providing the answer [15:06] kpettit: we have ~350TB on OI atm [15:06] works well [15:06] that sounds sexy [15:06] lol [15:07] ive only got like 12tb [15:07] man. that's a monster. [15:07] :( [15:07] I've got TB envy now. I was happy to have 8 [15:07] RoyK: Did you ever manage to find why firefox wouldn't run but xterm would? Just curious [15:07] Hostname: ss1 - OS: Linux 2.6.31-23-server/x86_64 - CPU: 4 x Intel(R) Xeon(R) (1600.000 MHz) - Processes: 162 - Uptime: 3d 18h 56m - Load Average: 0.05 - Memory Usage: 537.29MB/5976.13MB (8.99%) - Disk Usage: 4896.45GB/9745.55GB (50.24%) [15:07] I only have 6TB :( [15:07] thats my fileserver at home [15:07] I'm using one of those Drobo's. I really like it [15:07] We were moving to more ZFS storage until oracle ruined it. Now it is mostly xfs on rhel-derivatives or ubuntu (depending which subsite is managing it). Alltogether it is a few PB though. [15:09] Has someone practical experience with Request Tracker? /msg me === skaet_zzz is now known as skaet [15:23] Regarding VM, any reason _not_ to use the native KVM tools on Ubuntu ? [15:30] What uses the .conf files in /etc/init/ ? [15:31] queso: upstart [15:31] Pici: How would I restart a service that has a conf file in there? [15:32] queso: service restart [15:33] cloakable: Is the service's name equal to .conf? [15:33] queso: Think so. [15:34]