[00:30] oh_noes: Does sudo apt-key list # show that key as being trusted by apt? Should it be trusted? That doesn't seem to be a key output by gpg --keyserver keyserver.ubuntu.com --search-keys ftpmaster@ubuntu.com [00:30] jmarsden: hey thanks for the reply. No it doesnt list the key [00:30] it's an internal mirror I'm trying to setup, so all my linux servers can sync to it (save bandwidth/downloads) [00:31] it uses to work, and for the most part, an apt-get update works, excepts that line which fails [00:31] So either there is a file in your repo that is corrupted, or else the key is good and you need to grab a copy of the key and add it using apt-key [00:32] I think a file is corrupted, but how do I double check the key isnt the problem? [00:32] i think you can flush your cache repo and start over again [00:32] my apache2 server on the mirror/local repo, isnt running https or anything ... just using apt-mirror and apache2/http [00:33] Can you find that key on a trustworthy keyserver somewhere and obtain a copy and then import it using apt-key add ? If that works, the key was the issue; if not, a corrupted file was probably the issue. [00:34] I'm suspecting the second is closer to reality, since my gpg --keyserver keyserver.ubuntu.com --search-keys ftpmaster@ubuntu.com command does not show such a key. [00:34] what do you mean by keyserver? My ''local mirror repo' is *just* inhouse, internally for my LAN. [00:34] keyserver: Server with lots of GPG keys on it :) Ubuntu packages are signed with keys whereever they are... [00:35] The one I use for Ubuntu stuff is keyserver.ubuntu.com ... see command above [00:37] See https://help.ubuntu.com/community/GnuPrivacyGuardHowto for creating a key and uploading it to a keyserver... not what you want here, but it may help you understand what a keyserver is :) [00:37] jmarsden: ftpmaster sounds like a Debianism... could it be that his mirroring software is trying to use the wrong key because it hasn't been ubuntuized properly? [00:38] twb: No, there are keys that have the ftpmaster@ubuntu.com address, but none have the fingerprint he lists. [00:38] OK. [00:39] Doing gpg --keyserver keyserver.ubuntu.com --search-keys ftpmaster@ubuntu.com shows what the "real" keys are, as far as I know. I'm not a GPG expert though. [00:39] thanks [00:40] *reads quickly up on GPG keys* [00:40] so let me ask this ... if you setup a local repository apt-mirror of "us.archive.ubuntu.com ... then you point a client to this local server [00:40] the client still only needs the original default keys shipped with ubuntu right? [00:41] It doesnt need to import any special keys from the local mirror, because the local mirror is only providing a copy of the software [00:44] Right, you shouldn't need any new keys just because you are making a local mirror. [00:46] jmarsden: thanks ... so back to my original question, the problem can't be key related..? Is that correct? [00:47] I don't know. If you are running Hardy and mirroring hardy software, then I think you are right... but if you are mirroring newer stuff (say for Jaunty) then maybe you need a newer key? [00:48] It might be good just to verify the signatures of all your mirrored packages and see if any fail the check? [00:49] cjwatson: how about you reply to my private message or do you try to force me to take it 1 level further to the irc council? [00:50] privmsg: I do not intend to reply further [00:51] that means i am going to write a blog article about the whole situation [00:51] and i will be happy to mention your person, your name and your role here on irc [00:55] when some people think they can go and infect the ubuntu community with the poisoned spirit of the debian people its about time we do something against it [00:55] jmarsden: yeah thats why I'm a tad confused. Im running hardy, mirror hardy and supplying hardy updates to end users. It's definitely something on my server end. [00:56] jmarsden: right now I'm doing a apt-mirror against a new official mirror, incase the mirror itself is the problem. [00:56] oh_noes: Then my guess is that you have a corrupted file somewhere. Hmmm, OK. Maybe just mirror some small portion of it to start with? [01:02] jmarsden: you mean, re-mirror the hardy-security 'multiverse' URL which seems to be failing? [01:02] How do I do that? [01:03] Edit /etc/apt/mirrors.list to only mirror that one repository? Maybe delete everything in your local copy of that repo and then do that? [01:04] Thanks I'll try your second suggestion [01:05] jmarsden: one interesting thing I found about apt-mirror is that, if you configure it to use two or more official repos to mirror against [01:05] I'm not sure, or it's not documented how to combine them into a single local mirror to share to my clients. [01:06] Maybe my config is wrong, but apache2 is serving /mirror/mirror/isp1.mirror.com [01:06] but if I add a second official repo, isp2.mirror2.com, it'll just create another dir, and another 30G or so at /mirror/mirror/isp2.mirror2.com *and* my apache is still serving from the first dir [01:08] apt-mirror can't really be expected to know that the two repositories are "the same", can it? So it does what you asked, it mirrors each of them, independently. I think that's how it is supposed to work. The /mirror/mirror/ thing feels odd, but that's probably just the way you chose to configure it. [01:10] so it's not expected behaviour, for a local mirror to use 2 or more official repos to sync from? Into a single mirror for a local LAN? [01:10] I don't think so; pick the closest or most reliable one. I'd have to read the apt-mirror docs and/or code to be sure, but I don't think that is common usage. [01:11] I have an intrepid mirror here using apt-mirror, but I've never tried to have it pull from two different remote sources into one "consolidated" local copy. [01:12] yeh I didnt think so either. [01:12] which begs the question ... how does one 'guarantee' the status of the local mirror he/she is syncing from? [01:12] What is a reliable local mirror is no longer any of the above [01:13] You can't. You can guarantee that all signed packages with valid md5sums and valid signatures are "official" and un-tamptered-with, but not how up to date they are. [01:13] ie. my case. how does one confirm his mirror matches that of an official ubuntu mirror without downloading the entire 30G+ again [01:13] jmarsden: yep that makes sense, and when you say it in that context it shows the importance of GPG signatues and keys [01:13] download just the Release file and compare those? [01:13] /dists/*/Release.gpg is signed, /dists/*/Release has checksums of Packages and Sources files, Packages and Sources files have checksums of pretty much everything else [01:14] so you can trace everything from that [01:20] Ok, I'm think I'm on the right track now [01:21] so if a client doing apt-get update, gets this error: com/ubuntu/dists/hardy-updates/main/binary-i386/Packages.gz Hash Sum mismatch [01:22] I'm going to compare my Release file, agaisnt the actual Packages.gz file with md5 [01:26] right. sometimes this kind of thing can happen if you sync in the middle of an upstream mirror push (you have to get fairly unlucky, but it can happen, especially when transparent proxies are involved) [01:27] turn off web proxies as far as possible (or get apt-mirror to send Pragma: no-cache etc., if that's possible) and force a resync [01:30] cjwatson: thanks, by force a resync do you mean just run apt-mirror again? [01:30] I've done that, it downloaded the new index files, then processed them, and said everything is fine. 0 to download, and 0 files to delete [01:30] I'm afraid I'm not familiar with apt-mirror (I use debmirror myself) so I was just speaking in general terms [01:30] (by delete, I mean 0 files that can be cleaned up by clean.sh) [01:56] Is anyone aware of an appliance or .iso that boots a PXE server, so I can boot a workstation and run the installer? I really don't want to install all manner of extra daemons on my workstation, nor do I want to build a server, install the pxe bits, and then finally after several hours actually install the workstation in question. [01:57] nope. [01:58] the workstation you have not accept PXE or CDRom? [01:59] Only accepts PXE, CDRom isn't playing, doesn't boot from USB, very borked. [01:59] ugg [01:59] The scary thing is that I built such a beast last year before I blew away my VMware crap - now I wish I hadn't :) [01:59] heh [02:00] pity you didn't save the VM on a DVD or CD [02:00] 't was nice too, JEOS server, all pretty, small etc. [02:00] I'm hunting through the archives, and it may still live on my backup media, but a download would take less time :| [02:36] Anyone using Simplifymedia server to share music to iphone? [02:37] http://www.simplifymedia.com/linux-server.html [02:39] I have an init.d script that simply just calls the startserver.sh, however for some reason the simplifymedia server will not start properly unless it starts under a user account. How do I execute a command via init.d using another user? [02:39] is it possible to create an OVF with vmbuilder? [03:00] hi [03:00] how is everyone today [03:02] ok [03:03] wish i could figure out simplifymedia [03:03] you? [03:03] foomanshoe: wish i could get inspircd to work [03:04] foomanshoe: everyone who is trying to help me is blaming the ubuntu packagers [03:04] saying it is buggy [03:04] lol, maybe it is buggy [03:04] foomanshoe: idk maybe i need to shoot myserlf [03:05] servers make me feel that way too [03:10] foomanshoe: this is a happy birthday present from hell [03:43] help /connect [03:45] meoblast001: inspircd "works" out of the .deb for me here (Ubuntu 8.10 64bit) -- as in I can edit /etc/default/inspircd and then start it and connect to it. "Works" for a minimal definition of works... [03:46] jmarsden: does it allow you to connect more than 1 user from 1 ip? [03:46] jmarsden: i have reasons i need to do that [03:46] bots and such [03:47] I've not tried that. You said "wish i could get inspircd to work"... I tested it, it works... I only have it on a local desktop machine here... [03:47] let me check multiple local connections to it... [03:47] jmarsden: ok thanks [03:50] Yes, looks fine here... what symptoms are you seeing? I just connected as jm1 and jm2 and they can each see the otehr one in a #junk channel... [03:51] jmarsden: the bot won't connect at all... and i got 2 clients on my laptop to connect but the third was refused a connection [03:52] Any errors in /var/log/inspircd.log ? When the bot "won't connect" what is the actual error message you get back? [03:53] jmarsden: a more complex (time and date and such) version of "Connection Refused" [03:54] What does ctrl-T do in the command line? [03:56] tsrk: swaps last two typed cahracters, usually used after you typo something. [03:56] meoblast001: I think you just need to configure inspircd to allow more users in the relevant class? [03:57] yeah i did that [03:57] uped it to 15 [03:57] does nothing [03:57] I typed "cp ..." and that command was running while I accidently hit ctrl-t [03:57] did it do anything? [03:58] meoblast001: you restarted it after the change, right? Hmmm. well, I have to go AFK for a while, may be back later. [03:58] tsrk: Almost certainly not. [03:58] i'll run a full system restart just for s&g's [03:58] jmarsden, almost certainly? any way to find out? i can redo the copy if necessary, but it's really important files so i'm worried it did something [03:58] is it worth redoing the copy or should i be ok? [04:07] tsrk: You'll be fine. [04:09] jmarsden, ok, good :) [04:24] jmarsden: my server is now unconnectable :P [04:24] :( [04:24] jmarsden: would you like to help me? [04:25] I can try... so if you do /etc/init.d/inspircd start # What happens? Does it start the ircd ? [04:25] jmarsden: yes... it is started [04:25] OK. If you do netstat -ntlp | grep 6667 # do you see it sitting there listening on port 6667 ? [04:26] let me check [04:26] jmarsden: although i'm connected to the server [04:27] ? You just said it was unconnectable... ? [04:27] jmarsden: that's the catch [04:27] jmarsden: i have a bot and myself connected to it right now [04:27] jmarsden: if anything else tries to connect on irc.mysticgalaxies.com, it won't permit it [04:29] Are you seeing anything in its logs when that happens? In /var/log/inspircd.log maybe? [04:30] Slly questions, whats the difference between "deb http://mirror.internode.on.net/pub/ubuntu/ubuntu hardy main main/debian-installer restricted universe multiverse" and "deb http://mirror.internode.on.net/pub/ubuntu/ubuntu hardy main restricted universe multiverse" [04:30] jmarsden: i'm not in debug mode.. should i go into it [04:30] ie. what does main/debian-installer change in regards tp apt-mirror [04:32] meoblast001: I'm not an inspircd expert, but you can try it and see if it gets you better info. I think we;'ve already proved that inspircd does actually work, it just seems to need some configuration to work the way you want it to... [04:33] oh_noes: why would you include main/debian-installer as well as main in the same list of parts of the repository?? One is (I think!) a subset of the other) [04:34] jmarsden: I didnt, Im just cleaning up some sripts I inherited from an old sys admin :s [04:34] I'm a Solaris guy, so learning ubuntu-server from scratch [04:34] but thanks, I'll clean it up to just include main [04:35] OK. The definitions for deb lines are in man sources.lits I think, if you want to verify. [04:35] thx, will do [04:35] That should be man sources.list [04:38] meoblast001: Did you edit the globalmax and localmax parameters to the connect command in /etc/inspircd.conf ? The defaults are 3 which seems rather low... [04:38] jmarsden: on mine they are 15 [04:38] OK. [04:39] And can you connect to the server with say 5 local irssi sessions: irssi -c localhost -n somenick ? [04:40] jmarsden: one second... i'm rebooting hte machine so the boot order will turn everything back on correctly [04:41] meoblast001: Why?? Linux is not windows... reboots are very very rarely needed... [04:41] jmarsden: it's a faster, more error proof way to get the desireable results in this situation [04:42] No, it is desirable to know what you are doing and avoid reboots... anyway... try it when you get the server back up :) [04:42] its back up [04:42] boots very fast cuz i have ubuntu [04:42] jmarsden: since the bot connected at startup, Xchat can't connect [04:42] i'll try irssi [04:43] Xchat? On a server?? Ubuntu server has no GUI in which to run xchat... [04:43] oh no [04:43] on my laptop [04:44] jmarsden: on my loptop i get connection refused [04:45] Test from the machine on which the ircd is running, the server ... get yourself some shells there and run irssi -c localhost -n somenick in each one... does it work? [04:46] ok found the problem [04:46] i can connect as many as i want on 192.168.1.100 [04:46] but irc.mysticgalaxies.com fails [04:46] OK. So is there a router between the server and the public Internet? [04:47] jmarsden: yes.... and it is properly setup to forward [04:47] or the bot wouldnt even get on [04:48] And the router is capable of handling multiple simultaneous TCP connects to one port... you *know* this for sure? [04:48] Sounds to me like you probably have a cheap home router and are blaming inspircd for its deficiencies...? [04:49] jmarsden: it used to at least [04:49] jmarsden: when i reinstalled my server after it's crash, it didnt buy me a new router as far as i know [04:49] Can you get a second NIC into the server and use it to do routing/NAT for you and thereby bypass the hardware router? [04:50] jmarsden: no [04:51] OK, can you get a spare router to test with? Or temporarily hook the server directly to your DSL modem and configure its one and only NIC appropriately?? [04:51] I don't think this issue is all that inspircd related at this point -- do you? [04:51] jmarsden: no [04:52] OK. So that's progress in diagnosis, even if we have no quick easy fix... maybe you can grab a second NIC for the server tomorrow or something. [04:52] jmarsden: what do you mean by second NIC? [04:53] a second hardware network adapter? [04:53] A second NIC, Network Interface Card. Yes. [04:53] jmarsden: i know where i can get a dial up one.... but i have no dial up subscription :P [04:53] So you can route using Ubuntu instead of a SOHO router box. [04:53] ahh ic [04:54] jmarsden: i doubt it is the router [04:54] jmarsden: it worked pre-crash [04:54] my system went under a massive crash a few days ago [04:54] a lot of data was lost due to me not backing up everything i need [04:54] OK... then restore the inspircd config file from your backups and you should be all set. [04:54] Ah... [04:54] jmarsden: i prioritized a new backup system [04:54] which i am about to update right now [04:56] You can use free online disk storage for emergency backups, BTW... adrive or box.net or whatever... [04:56] i made my own backup system [04:56] jmarsden: it's pretty unique i think [04:57] permits a maximum of 1 week of data loss in the event of super massive failure [04:57] 1/2 a day in the event of normal to large failure [04:57] it sends everything to a flash drive [04:57] and i move it to my other systems once a week [04:57] If your house burns down, how much will you lose?? Anyway... so last time you edited /etc/inspircd.conf, what else did you edit? [04:58] jmarsden: if my house burns down... everything i REALLY need is on launchpad [04:58] Except backups of /etc/inspircd.conf it would seem? :) [05:00] jmarsden: well..... i wouldnt have a computer if the house burnt down [05:00] You could get a virtual server elsewhere ... it's not a good backup scheme if it relies on a single physical location for all copies of data... anyway... [05:01] do you have any idea what else you changed in /etc/inspircd/inspircd.conf ? [05:01] jmarsden: my familyl is sort of poor... we have no other option [05:02] jmarsden: i added the cgiirc plugin.. that's it [05:04] Insurance money would get you the $15/month or whatever for a virtual server :) But anyway... something does not add up... same hardware, same OS, used to work, does not work now... *something* is different. Either the issue is the router or its configuration, or you have something strange going on in the Ubuntu box causing it to reject connections via the router... which seems rather unlikely... [05:05] jmarsden: not exactly [05:05] i used to use Ubuntu 8.04 Desktop (that i modded down to act like the server version) [05:05] now i have 8.10 Server [05:07] Shouldn't be an issue... you could test using netcat as a test server on some other port I suppose... nc -4l 192.168.1.100 6666 and set the router to forward that port too... then we can test connections from the outside to port 6666 ... [05:11] Or you could stop the inspircd and do nc -4l 192.168.1.100 6667 so the test is on the same port... [05:11] jmarsden: hmmm... i think i already tried that [05:12] jmarsden: but i think i'm gonna take a break for the night.. i've been doing a lot on this [05:12] thanks for the help though [05:12] OK, no problem. [07:16] New bug: #333711 in dhcp3 (universe) "AppArmor profile for sbin.dhclient3 should handle connman" [Undecided,New] https://launchpad.net/bugs/333711 [08:21] How do I add a python script to startup menu? [08:22] ? [08:29] wut is startup menu ? [08:34] if you mean to say "What is the Startup Menu" [08:35] Then the startup menu is a list of applications that load and launch on startup. [10:42] um, trying to find a simple daemon to use to monitor and output network activity to a log file... at a specified time interval (I'm not 100% sure if I want it total per minute etc).... ill end up reading it into graphs &| a database. [10:46] PC_Nerd: have a look at cacti or maybe munin [10:47] ty [10:47] munin is nice but not very precise for network - just one point every five minutes, you may want something better than that [10:47] depends on your needs though :) [10:49] PC_Nerd: vnstat does it [10:51] yup, I'm mainly wanting to integrate this data into a custom control panel im writing - so ultimately i dont really want somethign like munin that uses a vhost on apache anyway. [10:51] mrtg / rrdtool [10:56] PC_Nerd: i recommend reading about /proc/ , there is tons of info from system if you know what to search [10:57] ah - thanks will do [10:58] PC_Nerd: and /sys/ has info too, not sure where all stuff you need is in [11:00] ok ty [11:00] on the documentation page for vnstat (on ubuntu.com) theres somethign about another program - which taks data from iptables..... [11:04] um..... is ipac-ng still available for ubuntu 8.04 ? - its last listing in the repositories was for gutsy...? [11:07] no, it was removed in Debian (at the maintainer's request; it wasn't being effectively maintained any more and had release-critical bugs), and we removed it for 8.04 [11:10] ok - so is there an equivilent replacement? it seemed like the perfect tool for me [11:11] that I don't know, I'm afraid [11:15] hi [11:15] quick, simple question [11:15] moin [11:15] I need to install Ubuntu 8.04 LST server on a DELL PowerEdge R200 [11:16] I have selected the download for 64bit version, but it downloads the amd64 CD image [11:16] Will this cause an issue on Xeon Processors ? [11:21] anyone ? [11:21] Abracadabra: im not sure.. [11:22] Abracadabra: Should be right from memory. You will get a failure message booting the kernel if it wrong though. [11:23] Abracadabra: try.. you dont loose anything, but i think that there is only i386 and amd64 (32bit for x86 and 64bit for x86_64 [11:25] Abracadabra: I'm running all my dell servers with amd64, but mine are all AMD Opterons. [11:26] Abracadabra: if xeaon are x86 compatible then amd64 is right choice [11:32] mine are Intel Xeon [11:32] Abracadabra: How old is the machine? [11:33] I will try it anyway, I simply found it odd that it doesn't ask what processor type the download is for, and points to amd64 [11:33] honestly, I do not know [11:33] it is a test machine I was given [11:33] I have not yet gone inside it [11:33] <_coredump_> Abracadabra, amd64 is for both intel and amd 64bit, except ia64 thats for itanium. [11:33] it is a quad-core though [11:34] ok [11:34] thanks guys [12:04] mark > around? [12:04] yes [12:05] I read wikimedia was deploying the new sun sans - wanted to know how well they were doing [12:05] any feedback to provide? :) [12:05] sun sans? [12:05] we just have a few thumpers/thors [12:05] nothing fancy [12:06] http://blogs.sun.com/jonathan/entry/open_storage_wins_at_wikipedia [12:07] bla bla ;) [12:08] yeah? so you're not running any new SUN nas for databases and media? [12:08] we have 2 thumpers and 3 thors [12:08] they serve images over http/nfs, and (soon) article text storage in mysql [12:09] sorry, what's a thumper? :/ [12:09] a Sun X4500 [12:09] Thor is the newer version [12:09] x4500? [12:09] k [12:09] a 4u box with 48 disks [12:10] http://www.sun.com/storage/disk_systems/unified_storage/7210/ ? [12:11] http://www.sun.com/servers/x64/x4500/ at a guess [12:11] Deeps > I was asking for thor :) [12:11] mark > did you make your choice on the deal with sun or on performance? [12:12] both :) [12:12] and - do they work well? did you get to compare with netapps? [12:12] we did not, and yeah they work well [12:12] we get nice performance out of them [12:13] Abracadabra: amd64 is for all x86_64 platforms. [12:13] Abracadabra: ...so both Intel and AMD. [12:15] mark > so they are fine as sql servers as well... does the sysadmins of wikipedia have a blog? I am quite interested to read about technology deployment there [12:16] yann2: we have some personal blogs which are aggregated on http://en.planet.wikimedia.org [12:16] (amongst other crap ;) [12:16] but I don't think we've written about that at all === JanC_ is now known as JanC [13:07] mark > if at some point you're bored ;) we (other sysadmins) are more interested by on-site performance feedback than by the prose of sun's ceo :) [13:08] heh I know [13:08] domas writes a lot about that on his blog, http://dammit.lt [13:09] and some of our stuff is on highscalability.com [13:09] but we're rarely bored really, there's so much to do :) [13:10] HI! [13:12] very interesting technical blog, thx for the link [13:13] ditto [13:14] you are welcome! [13:18] I got a question for you domas.. you write in your blog that zfs isn't ready to host innodb [13:18] but a few lines earlier mark says you planned to use 7210 for the article database [13:18] not a 7210, a x4500 [13:18] although the 7210 is relying on zfs isnt it? [13:18] (which I think is the same hw) [13:19] but there only storage space, not performance is important [13:19] well, for text archival we don't really _need_ innodb [13:19] and it's myisam, yes [13:19] we could just store in filesystem, if we wanted [13:19] we don't even need SQL for article store :) [13:19] it's just convenient [13:19] replication and such [13:21] yann2: it's storage of old revisions of (e.g.) wikipedia articles [13:21] ah, ok [13:21] those are almost never accessed anymore [13:22] so they're compressed in a smart way, and then stored for sporadic retrieval [13:26] mark: all revisions :) [13:26] just it doesn't matter much, as all new ones are cached at multiple layers [13:26] I know [14:04] Quick Xen backup related question. I'm using .img files (a disk.img and swap.img per VM), I was wondering if copying the whole image file is the best route, (a 20gb virtual disk = 20gb backup file, regardless of space used). [14:04] *I also plan on using rdiff or rsync on the file system for more frequent backups of critical files and directories [14:24] /quit === Koon is now known as ttx === ttx is now known as ttx_ === ttx_ is now known as ttx [15:33] Greetings and Salutations [16:14] ivoks: I looked back in debian/changelog for Spamassassin and you had the very first Ubuntu entry way back in Dapper. I was just able to file for a sync. [16:14] ScottK: hehe but amavis and clamav were bigger problems to integrate, so... kudos [16:18] hi all I am trying to get sound on one of my servers is there a howto on this? [16:19] I have installed alsa,alsa-tools and alsa-utils [16:19] also turned up the vol but still no sound any help would be great [16:20] Could I have some help on this one? http://pastebin.ca/1345941 [16:21] can apache read openssl/dh.h [16:21] zim: got drivers for the soundcard? [16:22] larswey: alsamixer sees the card [16:22] i dont know if apache can, i dont think that its what im trying to do [16:22] ok [16:22] larswey: libssl-dev may be what you need [16:22] as long as there is a speaker connected it should give sound [16:22] thanks Deeps testing [16:23] xml.h:9:19: error: expat.h: No such file or directory [16:23] next error [16:24] larswey: http://packages.ubuntu.com/search?searchon=contents&keywords=dh.h&mode=exactfilename&suite=hardy&arch=any confirms libssl-dev [16:24] search contents of packages on packages.ubuntu.com to find what packages contain your missing files [16:24] Okey, i'll try [16:27] zim: some soundcards have multiple mute settings, etc... make sure you're not muted in someplace sneaky (like headphone jack instead of speakers or similar) [16:52] yann2: ping [16:52] yann2: regarding your kvm networking issues [16:52] hello kirkland [16:53] yann2: i have one possible fix for your networking problems [16:53] yann2: i have a test package for Hardy, and Jaunty [16:53] yann2: where can you easily reproduce this problem? [16:53] yann2: and test my test package? [16:53] I was just thinking abuot moving my dev-server-host to jaunty :) [16:54] yann2: what is it now? [16:54] hardy [16:54] does your patch fix virtio or the default one? [16:55] yann2: it would affect both, i think [16:55] I would be *very* interested if it fixed virtio - I am hoping of getting faster network performance by adding more cpus :) [16:55] didnt manage to go over 150mbps at the moment [16:55] but yeah, where is your package? [16:55] yann2: https://launchpad.net/~kirkland/+archive/ppa [16:55] thx will give it a try [16:55] yann2: kvm - 1:62+dfsg-0ubuntu9~ppa1 [16:55] is it a hardy+patches or a full new version of kvm? [16:55] k [16:57] yann2: it's the hardy package, minus one patch [16:57] yann2: there's one networking patch in hardy's kvm that is quite clearly wrong [16:58] ubuntu patch? [16:58] yann2: debian [16:58] ok [16:58] will test that now :) [16:58] does it shutdown all the vms if I upgrade kvm? [16:58] No. [16:58] yann2: it doesn't [16:58] ok [16:58] yann2: however, you won't be testing the new one until you launch a new vm [16:59] or restart one :) [16:59] yann2: or power down/up the one you want to test [16:59] right [16:59] yann2: i think you can't just reboot the vm [16:59] yann2: i think you'll need to shut it down, and launch it again [17:00] assuming the guest supports rebooting ;) [17:00] if you got windows xp/2000 guests to reboot that'd be interesting ;) [17:00] * kirkland has absolutely no interest in xp/2000 :-) [17:01] bbl [17:01] deb http://launchpad.net/~kirkland/+archive/ppa hardy-updates main ? [17:02] deb http://ppa.launchpad.net/kirkland/ppa/ubuntu hardy main [17:07] kvm_62+dfsg-0ubuntu7_i386.deb is the one i am using atm [17:07] W: Failed to fetch https://launchpad.net/~kirkland/+archive/ppa/ubuntu/dists/hardy/main/binary-i386/Packages.gz The requested URL returned error: 40 [17:07] got this from your ppa [17:09] yann2: wget https://launchpad.net/%7Ekirkland/+archive/ppa/+files/kvm_62+dfsg-0ubuntu9~ppa1_i386.deb [17:09] yann2: then dpkg -i that [17:11] riight [17:16] wget http://isos.thehumanjourney.net/solaris/sol-10-u4-ga-sparc-dvd.iso -O /dev/null [17:17] 200mbits/sec on average [17:17] better than before (the iso is also on a iscsi share given by the other vm, that may be limiting too) [17:18] aaaaand it doesnt break, even with several cpus - that's the improvement I think [17:18] Resolving isos.thehumanjourney.net... failed: Name or service not known. [17:18] kirkland > it may just be a small patch but it's actually *hugely* important [17:18] Deeps > only local sorry ;) [17:19] yann2: are you confirming this kvm package is better? [17:19] kirkland > I'll do some more tests but if it works flawlessly I'm a happy man. Always hard to spread kvm love when the network is broken :) [17:19] yann2: \o/ [17:19] kirkland > still testing.... [17:19] yann2: please please please update the relevant bugs with your results [17:19] will do [17:19] yann2: irc conversation is good fun, but it's gotta be in the bug or else my buffer overflows :-) [17:19] np [17:22] bguzzardi: ty for the sound tip [17:22] it worked [17:27] kirkland > how safe is it to put into prod? [17:28] yann2: there is literally one line of C difference between this kvm package and the one in Hardy [17:28] - ifr.ifr_flags = IFF_TAP | IFF_NO_PI; [17:28] + ifr.ifr_flags = IFF_TAP | IFF_NO_PI | IFF_ONE_QUEUE; [17:28] yann2: that's the debian patch that was reverted [17:28] quite safe then? :) [17:29] yann2: this flag, IFF_ONE_QUEUE, we're quite confident is *wrong* [17:29] yann2: there's a chance that your execution path won't hit this, in which case we haven't solved your problem [17:29] yann2: and you're in no worse situation [17:29] yann2: on the other hand, if your execution path does hit this code, it's definitely doing the wrong thing [17:30] yann2: in which case this does solve a real problem (and perhaps *your* problem) [17:30] kirkland > downloading a large file with virtio and several cpus usually killed a vm after a GB or so [17:30] very randomly [17:30] not sure if it hits smp only [17:31] yann2: and that has succeeded on the ~ppa1 kvm? [17:31] just doing a loop to download 40GB [17:31] with 3 cpus [17:31] if it holds... ;) [17:31] yann2: k [17:31] yann2: i'll standby [17:32] updated bug report and support ticket btw [17:32] yann2: if this succeeds, and you think it solves your problem, i'll start filing the SRU request to get this out to Hardy [17:32] yann2: which one? [17:32] launchpad #332156 and support #3408 [17:32] Launchpad bug 332156 in kvm "Network issues with hardy guests/hosts, using default nic and virtio" [Undecided,New] https://launchpad.net/bugs/332156 [17:35] I'll do more testing tomorrow with my bacula vm which was the "bad" one [17:35] but so far it looks very good :) [18:07] I just ssh's into my server and initiated a xinit -- :1 from TTY3.. then I did "startxfce4" to start the gui.. when it started it put the session on F10 and moved my current local gui session from F7 to F9.. Why did that happen? === domas is now known as domas|out === domas|out is now known as domas [19:15] New bug: #333977 in openldap (main) "package slapd 2.4.11-0ubuntu6.1 failed to install/upgrade: subprocess post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/333977 [19:47] hey all [19:47] i am trying to open graphical programs on my LCDTV plugged into my server ed. [19:47] where doo i start [19:47] ? [19:50] no Xorg on server [19:53] can i not add it? or am i just making it into a desktop install then? [19:53] right, then it's a desktop [19:54] and if you need desktop application, then you are asking on the wrong place :/ [19:54] sudo apt-get install ubuntu-desktop [19:54] will get you ubuntu desktop, and you should try getting help at #ubuntu channel [20:06] i'm testing jaunty server in virtualbox. i cannot connect to the internet. i've been updating it via cdrom (20090223). i've been looking for the solution via google, ubuntuforums & launchpad. no luck so far. [20:07] IcemanV9: how do you test if connected to the internet ? [20:07] mathiaz: eta for samba upload? :) [20:08] aptitude update, ping -c3 [20:08] IcemanV9: if I remember correctly, ping doesn't work if the guest network is set up with NAT [20:09] cemc: /etc/resolv.conf is there [20:10] do you get an IP address (ifconfig eth0) ? default gateway? what's the host OS? firewall problem maybe? [20:12] * ScottK waves to cemc. [20:12] * cemc waves back to ScottK. [20:12] ;) [20:15] cemc: it has finally worked. i just used full-upgrade command, then reboot. it seems the new kernel (2.6.28-8) did the trick. thanks for your help, though. [20:15] not sure how that resolved it, but you're welcome :) [20:17] cemc: before the new kernel, i checked /etc/resolv.conf, ifconfig eth0, ufw (to be opened), different network card via virtualbox. no connection to the internet. [20:18] cemc: i've used -4, -6, -7 kernel. it seems -8 kernel fixed the internet connection. anyway, i can test it now. [20:25] IcemanV9: http://www.virtualbox.org/ticket/181 you want to look at this too if you're using NAT [20:28] cemc: old bug ... so they won't fix it. alright. thanks. [20:39] Adri2000: still on my todo list - I should get to the SRU soon [20:40] good afternoon eveyrone [22:28] anyone here have inspircd? i'd hate to explain my problem if no one here does.... not too many people have it [22:55] New bug: #310913 in net-snmp (main) "net-snmp process hang on hardy" [Undecided,Incomplete] https://launchpad.net/bugs/310913 [23:20] if I have a new ext3 filesystem created for some random data storage is there any reason to keep the reserved space (tune2fs -m) [23:26] jetsaredim: without reserved space "out of diskspace" will be way more random :) [23:26] jetsaredim: reserved diskspace is just to allow root processes to carry on writing [23:44] jetsaredim: you can make it much smaller than the default if you want, but some reserved space is always advisable AFAIK === WaVeR` is now known as WaVeR