[00:26] is there anything that tells me EXACTLY whats all included as part of the ubuntu 7.10 server iso? [00:26] every single doc i've read seems to point taht lamp stuff isnt and other related stuff [00:26] yet i know for a fact they are [00:26] it seems redundant to talk about "installing apache2" if it's already installed [00:26] VoiDeT: Can't remember when I set that up last, but I guess it's not that different from setting it up to work with a normal iptables installation? [00:27] baffle: even with iptables its giving a port missmatch error. I'm reading up on ubuntuforums alot of people are having issues with nat, producing the same error, but i can't seem to fix it like they can :S [00:28] the ubbuntu server guide seems to be more of a "how to make ubuntu into a server" than anything related to the ubuntu-server iso [00:28] sommer: ^^^^ [00:28] VoiDeT: If I remember correctly, the most important things is to define the range to use for passive ports in proftpd.conf, and also specify your external [00:29] VoiDeT: If I remember correctly, the most important things is to define the range to use for passive ports in proftpd.conf, and also specify your external IP address to some option.. And to forward the range + normal ftp ports in to the server in question. [00:29] baffle: range is defined as 60000:65534, ports 20:21 are forwarded to the box, also are the passive ports, and MasqueradeAddress is set to my public ip [00:30] VoiDeT: Well, then it works. :) [00:30] (Clearly not, but anyway) [00:30] but its not :( [00:30] hehe [00:31] As far as I can remember, that was the only options I set up last time I did this. Well, actually I wasn't using NAT, but except the MasqueradeAddress the setup is identical. [00:31] Have you run tcpdump on the various servers? [00:31] So that you know that the responses are correct etc. [00:31] well i see whats happening [00:31] And are you actually checking from a remote host.. :) [00:32] they connect fine, login all that, but when a command to ls etc. is made, the client is told to use its local ip, and they try use that to execute the command, [00:32] because im seeing the ppls local ips, even though their on the other side of the world [00:33] Uhh, that's strange. [00:41] ScottK: thoughts on 180551? [00:41] I rather suspect this is a simple case of freebsd(?) defaulting to actually setting hostname to the fqdn, not just one component... at least in Wietse's world. [00:46] * lamont smacks bug 200308 into oblivion. iz PEBCAK [00:46] Launchpad bug 200308 in postfix "Upgrade doesnt work if using mysql dict types" [Undecided,Invalid] https://launchpad.net/bugs/200308 [00:47] * lamont always loves it when the submitter finally gives sufficient information to verify that things are (not) working in exactly the manner that they told postfix to [00:49] maybe I should have preinst and start check for stupidity in alias_database maptypes [00:49] Anyone from the server team (or someone else) with insight in the multipath-tools-boot package? [01:05] morning all [01:06] so anyone have an ubuntu-server guide that doesnt suck? [01:08] or perhaps a ubuntu-server document that actually says whats all installed instead of telling you how to install things that already exist on the server? [01:08] its great to say "lamp" but if it doesnt tell you what mod_xxx's exist or what versions of the lamp components are installed in the iso [01:09] man, someday im gonna remember how useless irc is most days.. off to centos! [01:10] someone was here to make friends [01:10] Yeah. He didn'y even have a question. [01:10] haha [01:10] admin overlord [01:10] (Altho he did have a point, that server guide sure is pointless) [01:12] he did have a question, just it wasnt obvious (and i wasnt looking...) [01:12] baffle: how so? [01:13] kgoetz: Well, maybe it isn't pointless for "most people". But nothing I've ever wanted to have answered has been there at least. :) [01:14] baffle: tbh, i've never looked there *before* asking, so i dont really now :$ [01:14] *know [01:15] kgoetz: I would expect an "advanced topics" section to include things like high availability, clustering of services, clustering of filesystems, lvm in depth, iscsi/fc storage information, multipath information etc. [01:15] * kgoetz goes to look at what it actually inclused [01:19] actually, no i wont, because works just deced i have stuff to do ;| [01:22] baffle: do you have the conntrack_ftp module (or whatever) loaded? [01:23] * lamont considers asking "there's a server guide?", then decides not to troll [01:23] hehe. [01:23] will be linked with hardy :) [01:23] I generally just install the server bits, and then add what I want/need as I go. [01:24] unless it's a desktop box, in which case s/server/desktop/ and lather/rinse/repeat [01:24] lamont: I think most people do. [01:25] well, and any of them tend to get 'home-config' from the repo at the house, which violates debian policy left right and sideways to make things what I want them to be on the home machines [01:25] "because once you have tac nukes, everything starts to look like a small city" [01:25] * kgoetz didnt follow that last rant [01:26] tac nukes comment is my version of the hammer & nail cliche [01:26] home-config is the "smack the config around right, and depend on those other packages I want" not-quite-meta package [01:26] which also delivers my version of several other packages config files, etc. [01:26] iz very policy-violating [01:26] ah right [01:31] Jeeves_: you around? wondering if you did hardy or an earlier ubuntu [01:33] lamont: is 180551 and affect of Bug #8980 [01:33] Launchpad bug 8980 in netcfg "hostname -f does not return a proper FQDN" [Medium,Confirmed] https://launchpad.net/bugs/8980 [01:36] hrm... quite possiblt [01:36] possibly, even [01:36] Bonus points for fixing bugs with numbers less than 10,000. [01:39] Fix bug #1. [01:39] Launchpad bug 1 in ubuntu "Microsoft has a majority market share" [Critical,Confirmed] https://launchpad.net/bugs/1 [01:40] * lamont remembers getting the email from bugzilla when that bug was originally filed [01:47] * lamont rapidly approaches time to head back home [01:48] :) [02:01] * lamont -> home [02:49] ah, he left [02:57] * googlah|sleepy will give a try to install Spamassasin along with Postfix [03:01] googlah|sleepy: You'll be happier in the long run if you use amavisd-new to glue them together. [03:02] I'm getting "unable to determine geometry of file/device" from gparted on a Compaq DL360 (orginal pIII) [03:03] darn. does not recognize the array. [03:04] baffle: the server guide currently is geared more to admins new to linux or ubuntu [03:04] baffle: that doesn't mean it has to always be that way, do you have any other topic suggestions? [03:05] yeesh. I get to make a custom CD. [03:11] ScottK2: Don't seem to be any good tutorials out there though. But amavis seem familiar, isn't that which puts a line in each mail, that is has been checked? [03:17] googlah: I missed part of your conversation. What is it that you are trying to accomplish? [03:20] ah, just said that I would give a try to install Spamassasin along with Postfix, but ScottK2 said Amavisd would be better in the long-run [03:21] googlah|sleepy: ok. [03:21] googlah|sleepy: the "Mail Filtering" section covers amavis: http://doc.ubuntu.com/ubuntu/serverguide/C/index.html [03:24] Ah, thank you. great guide btw. [03:26] googlah|sleepy: thanks, if you see any mistakes or issues please let me know... it can still be updated until thursday [03:27] Oh, its yours?! keep up the good work. Is it for the release of Hardy Heron? [03:28] googlah|sleepy: yep, but most of it should work fine on gutsy, I don't think those packages changed much [03:29] Noo, me neither. [03:29] googlah|sleepy: Also for postfix you can get the latest (2.5.1) in gutsy-backports [03:29] ScottK: and 2.5.2 is due to be out sometime soonish (iz RC now) [03:30] sommer: May you not mention the --purge option in the Apt-Get section? [03:32] * lamont sleeps [03:33] googlah|sleepy: that's a good point [03:35] I'll add some information on that [03:37] Yeah, sometimes you want to remove the configuration files as well [03:38] definitely I use it frequently :) [03:44] hey [03:44] http://galatea.neoturbine.net/projects/phpsysinfo/ [03:44] any reason why it doesnt show the mount points there [03:51] print_r (phpinfo()); [03:51] uh [03:51] oops. tried the stable release of phpsysinfo? [03:52] is mine old? [03:52] it's always been like that though [03:53] no, yours is newer, but it's just an RC [03:53] beta [03:53] oh [03:54] i dunno, the stable one looked kinda old [03:54] but i mean it works with the beta one on different systems [03:55] it sure does, but try and see if thats where the problem is [03:55] perhaps i should try to read the code and see where it gets that information from [03:55] i dont think i feel like doing that right now [03:57] heh, no. but whats the time over there right now? it's not 5am as it is here. :p [03:58] lol, no [03:58] midnight [03:59] heh, yeah [07:12] <_ruben> mornin [07:21] kgoetz: Awake? [07:23] <_ruben> Jeeves_: 32-way web server? not bad ;) [07:24] _ruben: The T1000 has a cpu with 8 cores, and four threads per core. [07:24] <_ruben> Jeeves_: ah, nice [07:26] <_ruben> cant wait to start my virtualization project .. will be using a dual quad core with 32gb initialy [07:28] Nice :) [07:33] <_ruben> too bad another collegue gets to play with it first ... test with running office apps in an rdp application server [07:43] <_ruben> damnit .. too many ppl drinking hot choco, them bastards .. machine's out of cacao, same as yesterday morning .. sigh :p [08:10] <_ruben> Jeeves_ (or anyone else with archive access): could you tell me whats the size of the archive for just gutsy i386+amd64? [08:21] _ruben: Can't tell [08:21] All the files for the various releases are in the same directory [08:24] <_ruben> ah crap [08:24] <_ruben> not feeling like syncing releases i wont use anyways [08:26] _ruben: apt-proxy ? [08:26] <_ruben> thats one i havent looked into yet [08:27] <_ruben> apt-proxy doesnt build a 'full' mirror, only packages that have been requested once .. then again .. that might a decent solution afterall === \sh is now known as \sh_away [08:38] moin [08:41] <_ruben> mornin kraut [08:41] mornin _ruben === \sh_away is now known as \sh [08:50] morning [08:56] Jeeves_, around now === joerlend_ is now known as XiXaQ [08:57] Jeeves_, whys your system got 32 cores, ours is only 24 :( [08:57] Kamping_Kaiser: Hi! [08:58] Upgrading to Hardy as we speak [08:58] _ruben, use debmirror and sync with that, it'll tell you how big the release is, then you ^C it and move on :) [08:58] Jeeves_, ah, so didnt install with gutsy/ [08:58] ? [08:58] Kamping_Kaiser: I installed gutsy now [08:58] Dapper didn't work [08:58] Haven't tried the Hardy installer [08:58] So if you try that [08:59] That would be handy [08:59] hardy Doesnt Work For Me (TM) [08:59] Hmm? [08:59] well, the snapshot from 20080225 doesnt [08:59] On the Sun? [08:59] on a SunBlade 150 (workstation) [08:59] Ah ok [08:59] i havent tried the T1000 yet [09:00] * Kamping_Kaiser has been upsettingly busy - your going to finish both systems at the rate your going [09:00] Kamping_Kaiser: I've fix /etc/ethers for you so you can edit that too [09:01] Jeeves_, cheers [09:01] I'm going into a meeting [09:01] can you email me the mac of the system your playing with and i'll leave it alone [09:01] will be back in about an hour [09:01] if you want :) [09:01] np [09:23] <_ruben> Kamping_Kaiser: debmirror seems useful btw, will look into that some more [09:24] _ruben, np. its handy [09:24] <_ruben> Kamping_Kaiser: you're using it yourself as well ? [09:25] sommer: I have loads of topic ideas. :-) [09:26] _ruben, currently, yes i do. planning to switch to the proper archive mirror scripts probably, but debmirror for the moment [09:26] sommer: What about "Secure and distributed AAA in the enterprise"? :) I.e. setup of kerberos kdc + slaves, replicated LDAP masters, libnss_ldap and all the friends. [09:27] <_ruben> Kamping_Kaiser: why if i may ask? ;-) [09:29] _ruben, because debmirror doesnt get "metadata" - eg lots of stuff required by debian-cd [09:30] <_ruben> Kamping_Kaiser: ic [09:31] if all you need is .deb files, debmirror is great. if you need to make cds... debmirror makes your life harder :) [09:32] <_ruben> perhaps i should just go down the full rsync route, and put some harddisks and our internet link to some good use ;-) [09:32] hehe [09:33] * _ruben does have a few spare Mbps (95-percentile wise) [09:34] :O bugger [09:35] * Kamping_Kaiser needs more nybles [09:36] * _ruben bitchslaps some co-workers [09:36] <_ruben> they ordered (paralel) ata disks for our new backup server instead of sata [09:36] * Kamping_Kaiser wonders if _ruben is a sysadmin [09:36] ahahahah [09:37] <_ruben> 8x500GB [09:37] <_ruben> wonder if i can take em home to upgrade my 6x200GB server ;) [09:37] hehehe [09:37] i'm going to need more hdd if i want to play with fossology properly :( i hate buying disc [09:38] <_ruben> fossology? [09:38] <_ruben> disks become cheaper way too fast .. buy today, half-price tomorrow [09:39] http://fossology.org/ [09:39] i sleep with my computers in the same room. if i throw in more discs i need more servers (so another machine) andi have much more noise :( [09:41] _ruben, i want to feed debian+ubuntu+gnewsense into it. so i'll need *terabytes* [09:42] <_ruben> luckily i dont sleep in the same room as my computers no more [09:42] <_ruben> currently i have about 2-3TB of storage at home [09:43] i'd like some of that sort of storage at home :( [09:43] stuff we have at work doesnt count [09:43] <_ruben> got one server with 6x200GB ata disks .. and another with 6x250GB sata disks [09:43] <_ruben> sw raid5 [09:44] <_ruben> but those you dont want to sleep next to ;) [09:44] eneco must be happy with you :) [09:44] <_ruben> henkjan: nah .. but essent might ;) [09:45] <_ruben> tho they actually reduced my monthly fee this year .. then again .. that was based on flawed info (moved there last august, and the first few months we didnt use much power) [09:48] <_ruben> henkjan: and further more, those two arent on 24/7 ;) [09:48] <_ruben> the only 'box' that's on 24/7 currently is an epia 500MHz mobo with a single hdd .. rest of the boxes have WoL [09:51] ah, i'm using an epia too, 533Mhz [09:51] my ubuntu server always frezes network service after 15-20min. When need to connect, I have to "ping" it then it's fine [09:51] anyone has such a problem? [09:54] <_ruben> henkjan: could be 533MHz indeed, got a 800MHz one as well, but that one isnt passively cooled [09:55] <_ruben> 533 it is [10:07] Morning all. I have a quick question regarding my ubuntu server set up. [10:08] !ask [10:08] Please don't ask to ask a question, ask the question (all on ONE line, so others can read and follow it easily). If anyone knows the answer they will most likely answer. :-) [10:09] currently running a caching name server. It appears that I can resolve ip addresses inside and outside of my private network, however, when I try to ping my internal name sever I don't get a response. [10:09] idon'thave any firewall set up at this time [10:10] <_ruben> so you can resolve but not ping? [10:10] <_ruben> and you're sure you're using the same machine to resolve and are trying to ping? [10:10] no can't resolve either, but I can resolve other a recordson the internal network [10:11] for instance i can resolve gateway which is an A record of 192.168.x.x [10:12] I can ping it and I can typein http://gateway ....................and I get the routers web page [10:12] so that appears to be resolving for me but I can't resolve to the main server [10:13] which is my apache, ldap, samba server. [10:13] <_ruben> im guessing that your using your router's dns server which doesnt know about your own ubuntu nameserver yet [10:14] well my client is only pointing to my caching dns server but it also has the an entry for gateway [10:16] my caching name serveralso has forwarders set up for external dns [10:19] <_ruben> cjsstables: are you trying to ping by hostname or ip? [10:19] both, neither work [10:20] oops hold on that.. one second let me verify [10:21] I can ping ip ok but not by host name [10:21] however i can ping my router by host name [10:22] it has an A record in DNS gateway 192.168.1.1 [10:22] <_ruben> which means there's nothing wrong with ping, just the dns records being fubared [10:23] but there is also an A record for the server DC01 192.168.1.2 [10:24] I don't understand how I can resolve one but not the other [10:24] oops. hold on.. I just did a ping on dc01 and it resolved [10:25] it wasn't doing that before [10:26] ok.. here is some moreinfo. if I ping gateway I can resolve. if i ping gateway.cjsoffice.local I can't resolve [10:27] is it because my domain ends in local? [10:31] <_ruben> it depends on both your local machine's resolver settings and the nameserver configuration [10:31] <_ruben> .local itself shouldnt impose any problems [10:34] ok. I haven't done anything with the resolv.conf on the localmachine. could that be the issue? [10:36] any clue is that I get an error message after boot as follows.... Could not look up internet address for client01. This will prevent Xfce from operating correctly [10:36] what would I need to putintomy resolve.conf [10:37] <_ruben> "nameserver ip.address.of.nameserver" ... and optionally "search cjsoffice.local" [10:38] ah..... that makes sense. I'll try that. what service will I need to restart after I make those changes? [10:38] or should I just do a reboot [10:38] <_ruben> nothing .. should work right away [10:38] k. brb [10:40] Kamping_Kaiser: Still awake? [10:40] Jeeves_, still here [10:41] Hardy is broken for sparc [10:41] http://nl.archive.ubuntu.com/ubuntu/pool/universe/d/db4.4/libdb4.4_4.4.20-11_sparc.deb is a symlink to a non-existant package [10:41] oklooking at my resolv.conf I have domain cjsoffice.local then nameserver 192.168.1.2 nameserver "my ispsdnsaddy 1" nameserver "my ispsdnsaddy2" [10:41] <_ruben> i thought sparc support was dropped completely for hardy ? [10:41] Jeeves_, :( i knew it was unsupported, broken as well is upsetting [10:42] Kamping_Kaiser: I'm checking if it might be my mirror [10:42] <_ruben> Jeeves_: i'd blame the mirror in your case ;-) [10:42] hehe [10:42] * Kamping_Kaiser might [re]suggest debian to his work [10:43] Kamping_Kaiser: Sorry, what?! [10:43] _ruben Should I remove the external nameserver entries [10:44] Jeeves_, i might suggest a platform thats going to receive support to my workplace ... not being able to get things fixed on a server is a bit of a ... pain :| [10:44] * Kamping_Kaiser sighs. i'll have to mentally thrash that out later [10:44] <_ruben> cjsstables: if your own nameserver is setup properly and always available, then they shouldnt be needed [10:44] Hmm, it seems that archive.ubuntu.com does have the needed file [10:45] * _ruben suggests Jeeves_ and Kamping_Kaiser to team up to write a proper mirror script ;-) [10:45] hehehe. :P [10:45] ok. be back in a bit... thanks _ruben [10:45] <_ruben> ok .. thats enough makeing fun of innocent ppl for today ... [10:46] <_ruben> tho cant promise anything [10:46] * Jeeves_ is syncing as we speak [10:46] * _ruben probably should go try to get some actual work done for a change [10:47] * Kamping_Kaiser laughs at doing work [10:53] pschulz01, hello [10:53] you just missed us talking about suns :) [10:59] Kamping_Kaiser: The box is upgraded to hardy [10:59] Jeeves_, and running ok? i may have to rip off a gutsy cd at work and try that. [11:00] its a shame to see sparc in ports again :( [11:00] Kamping_Kaiser: I'm rebooting as we speak [11:02] {0} ok boot disk [11:02] Boot device: /pci@7c0/pci@0/pci@8/scsi@2/disk@0,0 File and args: [11:02] Let's see what it does! [11:03] Loaded kernel version 2.6.24 [11:04] :) [11:04] bit-beheer@webserver2:~$ uname -a [11:04] Linux webserver2 2.6.24-12-sparc64-smp #1 SMP Wed Mar 12 23:52:17 UTC 2008 sparc64 GNU/Linux [11:04] Works [11:05] (Fixed our mirror too, btw) [11:05] neat [11:05] so it can be made to run, but the support level may be less then ideal [11:06] Kamping_Kaiser: It seems, yes [11:07] * Kamping_Kaiser is wgetting isos at work to test on the sunblade150 [11:07] Jeeves_, so are you dualbooting that server, or just ubuntu on it? [11:08] Kamping_Kaiser: I've trashed Solaris [11:08] root@webserver2:~# cat /proc/mdstat [11:08] Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] [11:08] md1 : active raid1 sdb4[2] sda4[0] [11:09] Jeeves_, ok. i still want to try a multiboot if the 2nd T1000 is still unused [11:13] Kamping_Kaiser: Feel free, but I want to test the whole bunch by monday [11:13] Jeeves_, right. [11:20] B - Backspace key sends : BS [11:20] terminal emulators rock [11:23] Jeeves_, can you give me permission to use /dev/ttyS0? i need it for minicom. either changing its group, or adding me to dialout would work [11:23] Kamping_Kaiser: Ehm [11:24] or i could juse use ssh i guess [11:24] YOu should ssh to the ilom [11:24] console -f [11:24] ssh webserver1.ilom.ubunturelease.bit.nl [11:24] yeah, ssh works *smacks self* === \sh is now known as \sh_away === \sh_away is now known as \sh [11:42] hi all [11:43] Hi [11:43] kgoetz: pong [11:44] people, I have ubuntu 6.10 in my server in production, but now I need python 2.5 for my app and default is 2.4 on 6.10. There are somes packages in python 2.5, but not all like as python 2.4. So I think what is better choice. Is update to last version of ubuntu 7.10 right ? [11:46] kgoetz: Any idea how to get the ICA protocol working in the gnome 'Terminal Services Client'? [11:46] ICA? [11:47] Kamping_Kaiser: Citrix [11:47] ICA is citrix [11:47] Kamping_Kaiser: The option is there.. but it's greyed out. [11:47] pschulz01, you need to install the citirix client, which gnome TSC wraps around [11:47] pschulz01, with luck they have fixed the client to Not Suck (eg, to allow it to do encryption) [11:47] Ta.. alian it is then [11:48] any idea for my question ? [11:48] alien [11:48] _ruben stillout there? [11:49] Kamping_Kaiser: Is there a package? [11:49] pschulz01, for citrix? i dont remember. i think it was staticly linked tarball stuff (installed it on 6.06 at the hut) [11:51] <_ruben> cjsstables: more or less ;) [11:53] ok. still no good. I can't seem to resolve dc01.cjsoffice.local to an ip, however i can resolve dc01 to an ip [11:54] is possible update ubuntu 6.10->7.10, or for need 6.10->7.04->7.10? [11:54] J_P, latter [11:56] Kamping_Kaiser: thanks [11:56] np [11:56] Lunch! [11:57] enjoy :) [11:57] Kamping_Kaiser: Wants openmotif [11:57] Kamping_Kaiser: sheesh. [11:58] heh [11:58] i hate citrix, and its over that one experiance with their client. its going to take a lot of technical brilliance to be unscarred [12:03] I trying find page on ubuntulinux.com where say relation of number and nickname, but not find, anyone know ? [12:14] Kamping_Kaiser: Doesn't look like the Linux client supports SSL yet. [12:15] pschulz01, i remember trying the native linux and java client, both sucked (wouldnt do crypto properly) [12:16] "Citrix SSL server not accepting connections" [12:16] pschulz01, what level of encryption is the server using? [12:16] if its > 128 (iirc) bit you're stuffed (unless they fixed the client in the last few years) [12:18] * Kamping_Kaiser frantically tries to remember how to send 'stop a' without a 'stop a' key === \sh is now known as \sh_away [12:20] ^break i think === \sh_away is now known as \sh [12:22] nope :/ [12:25] oh well. tomorrows problem :( [12:25] Jeeves_, hope to catch you tomorrow. [12:25] night all [12:26] Kamping_Kaiser: yet about update, I need reboot after each update version or Can I reboot only in last update ? [12:26] J_P, every update [12:27] Kamping_Kaiser: ok [12:37] baffle: Those are good ideas, I had planned on expanding the LDAP, and Samba sections for the next release. [12:38] baffle: usually the first step in the process I've been using is to update or create an article in the wiki first, then place the content into DocBook for the official docs [12:38] baffle: if you can help document any of those areas it would be greatly appreciated, either way thanks for the ideas... good stuff === \sh is now known as \sh_away [13:31] soren: Is there anyone on the serverteam hanging around that can assist in bugfixing multipath-tools-boot? I've discovered what has happend and created an ugly workaround. But I'm not familiar enough with it to implement it in a proper way. [13:34] baffle: "has happened"? [13:37] soren: The initramfs first runs dmsetup (probably the lvm2 scripts) and discovers the logical volumes before the multipath script is run. This means that it can not create the /dev/mapper/mpath0, because there are allready mappings for it. If I stop the script, use "dmsetup remove /dev/mapper/", "multipath", "kpartx /dev/mapper/mpath0", "lvm vgscan" then everything seems to work. :) [13:39] baffle: Hm.. Ok. Let me check something. [13:40] soren: I.e. i'm going into a shell in the scripts/local-top/multipath script, entering the commands "dmsetup remove kvm1--5.v.kq.no-root" "dmsetup remove kvm1--5.v.kq.no-swap_1" "multipath" "kpartx -a /dev/mapper/mpath0" "lvm vgscan" and then exiting. [13:43] baffle: It might be a simple fix.. [13:44] soren: Yes, maybe. I'm also not sure how or when the "local" scripts might be run. [13:46] so I'm setting up a computing cluster (high-performance for solid-state physics). What kind of server management support does Ubuntu have? [13:47] I'm looking at managing about 20-30 computing nodes [13:47] links are great [13:47] baffle: Are you at libery to fiddle a bit with the system? [13:47] soren: Yes, very much so. [13:48] baffle: Cool. Could you try renamed /etc/udev/rules.d/95-multipath.rules to /etc/udev/rules.d/65-multipath.rules ? [13:48] s/renamed/renaming/ [13:48] I'm also interested in how Canonical support would work (and cost) for such a system. [13:48] And the run an update-initramfs -u -k `uname -r` [13:49] Solarion: Hang on. [13:49] soren: thanks [13:49] Solarion: Canonical offers http://www.canonical.com/projects/landscape [13:49] Solarion: You get it for free along with a support contract. [13:50] I'm not sure what Ubuntu itself has in that area. [13:50] soren: Hmm. What level of support contract? [13:50] It's too bad landscape can't be bought to run locally [13:51] soren: Also, how much would support cost for a 20-30 node cluster? [13:51] soren: And rebooting I guess. :) [13:51] since it's kind of one and kind of many boxes [13:51] baffle: Yes :) [13:51] Solarion: You'd have to ask them. [13:52] http://forms.canonical.com/dashboard/survey/response.jsp [13:53] soren: ah, I thought you were a Canonical employee. :) [13:53] Solarion: I am. :) [13:53] soren: do you have any detailed knowledge of how Landscape works? [13:53] Solarion: I don't sell support, though :) [13:53] Solarion: What do you want to know? [13:53] heh. Fair 'nuff. :) [13:54] so Landscape runs on Canonical's servers? [13:54] The server part of it, yes. [13:54] the page is kind of sparse on details [13:54] The agents fetch information from the server every once in a while and act upon the data it gets back. [13:54] soren: How would computing nodes hidden (and not routed to the Intarwebs) be managed in that case? [13:55] Stuff like "Add this user", "install this package", "run this script".. [13:55] Solarion: Dunno. [13:55] computing nodes can talk to the head node and shouldn't talk beyond it. [13:55] so it's kinda like Red Hat's management stuff then? [13:55] I'm not sure how that'd work. The good folks in our support centre would know, though :) [13:55] heh [13:55] Solarion: I have no clue about redhat's management stuff :) [13:56] it's small enough that I could always just do NFS and potentailly run an interior mirror, but the easier and more Canonical-supported, teh better. [13:56] Sure. [13:56] soren: Nopes, same error. As I mentioned, I spawn a shell in local-top/multipath, and I checked, the path could not be created, and /dev/mapper/ existed. I also noticed that it seems to take a little while before the devices from the FC-bus actually gets registred, because the initial run of "multipath" didn't actually create *any* paths, but then some kernel messages popped out and running multipath again created mpath1 (wich doesn' [13:57] baffle: I'm struggling to remember why I moved it to 95.. [13:57] soren: So it is udev that runs dmsetup and creates the devnodes for lvm logical volumes? [13:57] baffle: You shouldn't need to run multipath manually. [13:58] soren: another reason to go Ubuntu is system76. :) [13:58] baffle: udev should notice when new stuff shows up and ask multipath if it wants to use it for anything. [13:58] anyone have experience with them? [13:58] Solarion: From what I've understood from Landscape it is quite simmilar to RHN, yes. [13:59] (We have RHN for some customers) [14:00] baffle: Are you still in the initramfs? [14:00] soren: No, I just booted, but I can boot back in. [14:01] soren: Maybe it should be 64, not 65? I.e. so it is run before dmsetup. [14:01] baffle: Could you please check that the multipath rules files actually has 65- in /etc/udev/rules.d/ ? [14:01] baffle: Heh. You could try. [14:01] soren: I can just unpack the initramfs. [14:02] baffle: who is "we"? [14:02] baffle: Sure. [14:02] Solarion: It's them. [14:02] :p [14:02] Solarion: A company. [14:02] *g* [14:02] * soren goes to fetch some water [14:02] baffle: ok [14:03] soren: Actually.. I can't seem to find *multipath* in etc/udev/rules.d/ from the initramdisk... [14:03] looking forward to the next LTS release btw [14:04] any estimate when gcc 4.3 will be available? [14:04] Yes, with the new LTS release I know a lot of our customers will start using Ubuntu instead of RHEL/SLES/Debian.. [14:04] baffle: Er.. That's... Er... [14:04] Solarion: It is a teleco/isp called Ventelo in Norway. [14:04] soren: Maybe that is the problem. :) [14:05] baffle: ah, sounded like you were a consulting biz [14:06] Solarion: I am. [14:06] Solarion: I work in the part of the company that does hosting of larger (mostly internet-facing) installations. [14:07] Solarion: 4 datacenters, and the largest websites in .no. [14:07] COOL [14:07] dangit [14:08] Solarion: Working with large clusters sounds more fun. :) [14:08] nijaba: sorry; not registered so I can't msg back. I've put in a sales request so we'll see what they send back. [14:08] soren: The comment "15:06 < baffle> Solarion: I work in the part of the company that does hosting of larger (mostly internet-facing) installations. [14:08] 15:07 < baffle> Solarion: 4 datacenters, and the largest websites in .no. [14:08] 15:07 < Solarion> COOL [14:08] Ooops. [14:08] accidental posting of accidental caps. heh. [14:09] soren: The comment "# Only copy across relevant rules" in hooks/udev seems to be to blame? [14:09] baffle: so if I need to set up a website in Europe, I know whom to contact? :) [14:09] Solarion: As long as you want to pay a lot of money, sure. :) [14:09] heh [14:09] baffle: No, I think I am. [14:10] soren: You are? Isn't this package just imported from Debian? :) [14:10] baffle: Nope. [14:11] baffle: So... You have the multipath binaries in the initramfs, right? [14:11] soren: Yeah, they are there. [14:11] baffle: Interesting. [14:11] baffle: Er.. This is hardy, right? [14:12] soren: As I mentioned, I can change things around so that it works. [14:12] soren: Yes, hardy. [14:13] Clean install from [14:13] monday I think. [14:13] From netinstall. [14:13] You have multipath-tools installed, and not multipath-tools-initramfs, correct? [14:14] soren: Yes, multipath-tools and multipath-tools-boot [14:14] "un multipath-tools-initramfs" [14:15] God, I wish I had some hardware to test this.. [14:15] soren: You can use mine? [14:16] soren: It's not like it is in production. [14:16] baffle: Nah, not quite yet. [14:16] baffle: Thanks, though. [14:16] soren: Ah, you mean like in your lab so you can do regression testing etc. :-) [14:17] soren: Don't you have a SAN lying around to test with? Jeez. :) [14:19] Ah... I think I see what's going on. [14:20] soren: But if missing udev rules is to blame, maybe a "cp -p /etc/udev/rules.d/95-multipath.rules ${DESTDIR}/etc/udev/rules.d" should be added to multipaths hook? [14:21] baffle: multipath has no hooks. They were moved to multipath-tools-initramfs. [14:21] Which I somehow missed. [14:21] Oh. [14:22] But /usr/share/initramfs-tools/hooks/multipath exists? [14:23] Er... No? [14:23] Where does that come from? [14:23] dpkg -S /usr/share/initramfs-tools/hooks/multipath [14:23] # dpkg -L multipath-tools-boot | grep hooks/mul [14:23] /usr/share/initramfs-tools/hooks/multipath [14:24] multipath-tools-boot: /usr/share/initramfs-tools/hooks/multipath [14:24] soren: You have -boot installed as well, right? [14:25] Oh, I *thought* I did. [14:25] Solarion: We've been working on a management system for all our customers, actually.. It incorporates package stuff from RHN, but we're trying to make it multiplatform. I.e. win32/win64/rpm-based linux/deb/freebsd/solaris. It's a challenge, but quite fun. [14:26] Lovely. that should probably install the multipath rules as well. [14:26] soren: Where? It's not mentioned in the hook? [14:26] Right. [14:27] soren: It just copies binaries etc. [14:27] Add: [14:27] cp /etc/udev/rules.d/*-multipath.rules $DESTDIR/etc/udev/rules.d [14:27] to the bottom. [14:27] of that hook script. [14:27] Rebuild, reboot, report :) [14:27] Allready on it. [14:27] It's 65 now? [14:27] Or 95? [14:27] No, I renamed back to 95. [14:28] Wonderful. [14:28] Carry on. [14:28] Remote IP console sure rocks. [14:30] I believe you. [14:31] Yay. [14:31] It created mpath0 [14:32] So we're happy? [14:32] Not yet, I'll have to continue booting. And I think the management module of my bladecenter just rebooted.. [14:33] No, it was firefox exploding. [14:33] And that is why it's called an alpha release. :) [14:33] <_ruben> heheh [14:34] No, something else seems to have exploded.... [14:35] soren: No, not there yet. "Iii, panic! No root!" [14:35] :-) [14:36] Ok. [14:37] soren: I had to run "kpartx -a /dev/mapper/mpath0" for it to create the device nodes for the partitions. [14:37] (And then follow up with LVM volumes) [14:37] baffle: Oh. [14:37] baffle: Right, rihgt. [14:37] soren: But except that, it worked. [14:38] baffle: Add: [14:39] cp /etc/udev/rules.d/*-kpartx.rules $DESTDIR/etc/udev/rules.d [14:39] to the hooks script, too. === \sh_away is now known as \sh [14:40] rinse, repeat. [14:46] soren: Doesn't seem like kpartx was run, no root found. [14:49] (95-kpartx.rules exists inside initramfs) [14:57] soren: Running out of ideas? :) [14:57] * soren is scratching his head [14:58] No, I'm not out of ideas.. My usual approach is a lot of poke, check, fiddle, check, ponder, poke, check.. It's not as easy when I need other people for the poking, fiddling, and checking :) [14:59] soren: As I mentioned, you can get access to it if you want. [15:00] It's tempting. [15:00] Hi, I'm looking for a good HOWTO on setting up reverse DNS for a block of 64 addresses. Many docs point to a site called Acme Byte & Wire (http://acmebw.com) which seems to be down/gone. [15:01] dthacker-work: Talk to your ISP. [15:01] baffle: What sort of access can you get me? [15:01] soren: Full KVM I guess. [15:02] baffle: I mean how? A java applet or some such? [15:02] soren: Yes, java applet. [15:03] Won't work :) [15:03] * soren is on amd64 [15:03] No java applet love right now. [15:03] In a few days, sure. Right now? Not so much. [15:05] soren: Oh, doesn't java work with amd64? Well, I can fix that as well. :) [15:06] baffle: So can I. In fact, I already did. It just needs to get uploaded :) [15:06] soren: I meant in another way. [15:06] Do tell. [15:06] soren: rdekstop to another machine, access from there. [15:07] baffle: Let me just think a little while longer before I get into something that involves rdesktop. [15:08] soren: Okay. :) [15:08] Ok, so the multipath rules work fine.. [15:09] New devices are created and they go through udev's processing. [15:09] So far so good. [15:10] I don't see why the kpartx rules shouldn't see it. [15:11] baffle: Could you -- just for fun -- rename the scripts to 65, rebuild, reboot, etc? [15:11] kpartx or multipath? [15:11] Or both? [15:12] baffle: I was thinking both. [15:14] booting [15:16] brb [15:19] back [15:20] soren: No, doesn't seem to be a change. What puzzles me is that I in the local-top/multipath script I added the "sh" line after "multipath -v $VERBOSITY" (wich is set to 5 right now) and when I do an "ls /dev/mapper" only control shows up.. While I'm waiting for that, I get some kernel messages saying that the link on the FC adapters are up, some errors reading paths that are down etc, and then after some seconds I get /dev/mapper/mpath0. [15:20] But no partitions. [15:21] Is the scripts run alphabetically? Wouldn't kpartx udev run before multipath, now that both are at 65? [15:22] (I never really understood udev, to be honest) [15:22] Doesn't matter. [15:22] They're not acting on the same devices anyway. [15:22] True.. multipath is creating mpath0, and I guess then the same scripts are run over that device.. [15:23] baffle: I don't quite understand what the point of the local-top/multipath script is. [15:23] udev should totally handle all of that. [15:24] baffle: What is 'the "sh" line'? [15:24] Oh! That's how you break out of the boot process? [15:25] soren: Yes, is that wrong? I just exit the shell to continue.. [15:25] soren: I think basically what is does is to run "multipath", to detect paths for multipathd. [15:25] It's probably fine.. I just use "break=bottom" on the kernel command line. [15:26] baffle: right, and that shouldn't be necessary. [15:26] soren: It automatically breaks at the bottom as well, when it can't find root. :) [15:26] baffle: I thought you said it paniced? [15:26] soren: No, not panic. [15:26] 15:35:52 < baffle> soren: No, not there yet. "Iii, panic! No root!" [15:26] :) [15:26] soren: Ah, sorry. :) [15:27] soren: It just runs the normal loop w/counter waiting for /dev/mapper/root-logical-volume to be available (not, it's not named that), and then exits to the shell. [15:27] baffle: The mpath device you mentioned earlier is a symlinnk, rigt? [15:28] Er.... [15:28] ...is a symlink, right? [15:28] soren: No.. [15:28] Hm. [15:29] Ah, no. I'm living in the past. Never mind. [15:29] "user_friendly_names yes" is set in /etc/multipath.conf, maybe it barks because it doesn't just use dm0? [15:29] I think that is the option that creates that name. [15:29] That's possible. [15:29] Maybe I should reboot without that. [15:30] Is that the default or did you change it? [15:30] multipath.conf doesn't exist, but it is mentioned as a common default in the example file. [15:31] I'll try removing that file, remaking, rebooting. [15:32] Removing the example file? [15:33] baffle: Ah, changing that shouldn't matter. [15:36] soren: Ah. Instead of /dev/mapper/mpath0 I now get /dev/mapper/1676a6767b76767c76767d767ee4211243 (The UUIDs I guess). .Yay! :) [15:36] baffle: Heh. [15:36] baffle: Do you have any links to that in /dev/disk/by-id ? [15:37] soren: Yes. [15:38] soren: dm-uuid-mpath-36006016027d01e004c41e4e1e7f4dc11 -> ../../mapper/36006016027d01e004c41e4e1e7f4dc11 [15:38] soren: dm-name-36006016027d01e004c41e4e1e7f4dc11 -> ../../mapper/36006016027d01e004c41e4e1e7f4dc11 [15:38] soren: scsi-36006016027d01e004c41e4e1e7f4dc11 -> ../../mapper/36006016027d01e004c41e4e1e7f4dc11 [15:38] (This is after boot, I didn't check that inside the initrd) [15:39] That's fine. [15:41] Oh! [15:41] Are there not /dev/dm-* device nodes? [15:42] Hmm. [15:43] soren: No? [15:43] That would exlain it. Hah! [15:43] Ok, grab the kpartx.rules, change the following: [15:44] At the bottom there are a few calls to kpartx. [15:44] "kpartx -a -p -part /dev/$kernel" [15:44] should read [15:44] "kpartx -a -p -part /dev/%k" [15:44] Please try that [15:44] Should I rename them back to 95 as well? [15:47] That would be lovely. [15:47] ...so as to only test one fix at a time. [15:47] soren: is there any issue with mentioning open-vm-tools in the documentation? [15:48] sommer: No, please do. [15:48] soren: cool thanks [15:49] server in the blogs: http://www.bmighty.com/blog/main/archives/2008/03/ubuntu_eyes_the.html [15:51] Oh, I used the wrong name. [15:51] $k, not %k. :) [15:55] Nopes, no partitions in /dev/mapper still... [15:58] I wonder how it compares to centos udev rules.. [16:00] baffle: h.. [16:00] hm.. [16:00] baffle: You won't get much out of that. [16:01] Probably not. [16:02] baffle: CentOS uses a completely different approach and an ancient kernel (AFAIK). [16:02] soren: What does %k represent? [16:02] I belive it's the kernel's name for the device. [16:03] And $kernel is? [16:03] unknown [16:06] * soren thinkgs some more [16:07] soren: The options to kpartx.. "-a -p -part ".. I guess -p -part just means the devices will be named mpath0-part1 -part5 etc.. So that's OK I guess.. Hmm.. [16:07] Right. [16:08] hello all, when i usind nfs server it work sometime and then not work again i get permission message " mount.nfs: homeserver:/downloads failed, reason given by server: Permission denied " any help ? [16:10] soren: Maybe I (you?) should create a bug... [16:10] bugreport. [16:10] I guess the bug allready is there. :) [16:18] baffle: Ah.. $kernel is little-used synomym for %k. [16:18] Heh. [16:22] hm [16:22] dovecot -a [16:22] dupms configuration [16:23] idea! :) [16:24] YEah, we've used that before :) [16:24] baffle: Hmm... I think I know how I can fix it. I don't like it, though. [16:24] check if socket listen { client } is set; if not, add configuration to dovecot.conf; if yes, echo 'you already have configured sasl, you need to do bla bla...' [16:25] I wrote code to do that while back. [16:25] ivoks: Hang on. [16:26] ivoks: Doh! I put it on pastebin, and it's gone now. [16:26] never mind... [16:26] baffle: Are you in a hurry, or do you have time for me to read a few pages of stuff? [16:27] baffle: While you wait, maybe you could run "udevtest /block/" and paste the output somewhere? [16:37] baffle: Hahah! [16:37] baffle: Ok, here we go. [16:38] Instead of %k, put either %N or $tempnode (they're equivalent, so whichever one you think is prettier). [16:40] ivoks: December 3rd, this channel. Check the logs :) [16:43] soren: lol [16:43] i already droped that idea... [16:43] we need conffile management system :D [16:44] Alright :) [16:45] we can't create package that would patch configuration file [16:45] that's insane [16:45] if someone alters configuration file, patching fails [16:46] maybe we could alter original dovecot.conf and replace # with #UBUNTU-SASL# [16:46] and then, after installation of dovecot-postfix-sasl package, just sed 's/#UBUNTU-SASL#//g' :) [16:48] anyone with better idea? :) [16:51] ivoks: why don't we ship a default dovecot configuration with sasl enable ? [16:52] cause maybe postfix isn't installed [16:52] dovecot doesn't depend on postfix [16:52] ivoks: and when the postfix-sasl package is installed we setup the sasl socket to the right path [16:52] ivoks: like linking to it or something like that [16:52] and if there's no postfix, dovecot fails to start cause of missing direcotry in which it should create socket [16:53] we could try linking [16:53] ivoks: ah yes - so make dovecot not fail it the directory is missing [16:53] ivoks: /it/if/ [16:53] why not create socket in /var/run/dovecot/auth-socket [16:53] and then, after installation of postfix-sasl, just link it to /var/spool/postfix/auth/whatever [16:54] that way we wouldn't touch dovecot's config and other programs could use it too... [16:54] ah... i know why... [16:54] that file should be writable by postfix user [16:55] soren: Rinsing and repeating. [16:55] mathiaz: i don't think it would be easy to make dovecot not to fail if there's no directory [16:56] right... and softlink wouldn't work, since postfix is jailed [16:57] bindmount [16:58] we've been through this :) [16:58] we were, yes [16:58] * soren afj fir a little bit [16:59] soren: %N = no luck. [16:59] soren: Uhh, should I add something like touch /tmp/tried_%N in there and retry? :) [17:00] soren: Can I have more than one RUN+ lines? :) [17:00] mathiaz: why not sed deovcot.conf? it's not conffile :) [17:00] ivoks: it's control by ucf IIRC [17:01] so, we don't want to ask user anything on upgrade... [17:02] but we already do some seding on dovecot.conf in postinst scripts [17:03] ivoks: in the postinst dovecot pkg ? [17:03] ivoks: that's allowed. [17:03] no, in package that doesn't provide dovecot.conf [17:03] dovecot.conf is owned by dovecot-common [17:03] and we sed it with dovecot-imapd|pop3d === \sh is now known as \sh_away [17:03] ivoks: well - it's the same source package [17:04] ok, so from same source is ok? [17:04] then why not have another binary in dovecot source? [17:04] which would just sed dovecot.conf and add postfix hooks over postconf [17:04] exactly the same thing as -imapd|pop3d + postfix stuff [17:06] if it's ok for imapd, i don't see why it wouldn't be for another package [17:07] baffle: You need the conditions there too, so copy both lines and change the PROGRAM+ bit. [17:08] omg... i have /usr/shareFeisty directory :) [17:08] So? [17:12] eh... see you later === leonel_ is now known as leonel [17:18] baffle: I'm dying to hear about your tests. [17:18] :) [17:28] soren: A friend came by. [17:28] baffle: Quite alright. [17:29] soren: What "PROGRAM+" bit? [17:29] soren: You mean the RUN+ bit? [17:29] baffle: Yes. My bad. [17:35] baffle: brb [17:36] soren: Hmm. Didn't work. [17:39] baffle: Could you pastebin the rules file as it is now? [17:43] soren: http://www.pastebin.ca/949190 [17:45] baffle: Oh, it should just read %N. Not dev/%N. [17:48] Oh. [17:49] Sorry if I misled you. [17:50] Well.. I'm excited at least. :) [17:52] soren: Woooooohooo! [17:52] I like the sound of taht. [17:52] soren: I think it worked. I'll just remove my debugstash. [17:52] baffle: Coolness. [17:52] soren: Or, well, I'm quite sure it worked actually. [17:57] soren: soren Yep, it booted up now. But it had to wait for the root device to be available for 10 secs or so, due to timeouts on the FC paths etc.. [17:58] soren: So, what we did was change 95-kpartx, and add copying of udev rules (multipath & kpartx) to hooks/multipath [18:02] baffle: Yup. [18:02] soren: Package uploaded yet? ;-P [18:02] baffle: Hehe.. [18:03] baffle: Is "Dag Stenstad" the right name to send a greeting to in the changelog? :) [18:03] soren: Yeah. === ewook_ is now known as ewook [18:04] soren: Ah, I actually had set an ircname. :) [18:08] keescook: did you take a look at Bug #203449 [18:08] Launchpad bug 203449 in dovecot "[dovecot] [CVE-2008-1199, CVE-2008-1218] privilege escalation" [Undecided,Confirmed] https://launchpad.net/bugs/203449 [18:09] nxvl: sure, working on a few other things currently. [18:10] keescook: oh! ok ok, i was wondering why wasn't it uploaded or marked as high importance :P [18:11] is Bug #129630 still needed on hardy? [18:11] Launchpad bug 129630 in dovecot "[feature-request-gutsy] Include managesieve patch in dovecot" [Wishlist,New] https://launchpad.net/bugs/129630 [18:13] nxvl: your debdiff for but 203449 seems to be for something else? === \sh_away is now known as \sh [18:38] keescook: you are right, i upload the incorrect debdiff === RB2_ is now known as RB2 [18:39] keescook: fixed [18:41] nxvl: thx :) [18:43] I seem to be having a problem starting Squid. I returns no feeback/error codes at all, it just does not start? [18:51] Ignore that request! My simple error, working fine now! [18:57] what package is cups-config in? [19:39] yis there a way to configure a printer class in cups to only use the second printer if the first is down? [20:00] LiENUS: I would go ask in #cups [20:01] LiENUS: hmm.. seem you are there already... [20:01] already asked [20:01] no ones talkjing [20:01] yep, not too many people there [20:02] be patient then, someone may come aroud and see your question [20:45] its also not the official cups room [20:57] LiENUS: really? which one is the official one? [20:57] there is no official cups irc channel afaik [20:57] mmmh too bad [21:12] damnit this printer just stoped working entirely [21:13] i hate vista :/ [21:14] LiENUS: best info I've found. My 15 minutes of googling are up......http://www.cups.org/newsgroups.php?s1+gcups.general+v2+T+Qfailover [21:15] sommer: fyi-- the fix for apparmor/slapcat et al will be in post-beta [21:15] jdstrand: sweetness [21:16] sommer: see bug #203898 [21:16] Launchpad bug 203898 in openldap2.3 "slapcat broken when default apparmor profile is enabled" [Undecided,In progress] https://launchpad.net/bugs/203898 [21:20] sommer: btw, the fix is simply to make slapcat and friends hard links rather than symlinks. so you can do that locally and continue your testing [21:23] jdstrand: that's cool, my main concern was how to document the process :-) [21:25] sommer: they will work as expected once they are hardlinks [21:44] nijaba: that bug has been assigned to jelias by himself and i have a moral crisis right now [21:44] nijaba: at one hand he's to new to take care of something that important as his first assignement and we can't wait until he understands and fix it [21:45] nxvl: that's what i think [21:45] nijaba: on the other hand is kind of rude just to work on this and assing it to me [21:45] nxvl: so should we ask him if he needs a hand and tell him the importance of it? [21:46] mm [21:46] he doesn't even has a nickname or confirmed e-mail on his LP [21:46] meh, sometimes we need to be rude [21:46] mathiaz: for your viewing pleasure: http://www.youtube.com/watch?v=jF9jEJg-K0Q&feature=related [21:47] nijaba: but if you want to do the bad part and comment on the bug report that this is to important to be done by him and that we need it quick due BF i will be really thankfull [21:47] nxvl: I'll do that [21:49] sommer: :) [21:49] * nxvl HUGS nijaba [21:52] nxvl: what do you think of: [21:52] This bug just got highlighted a bit too much and beta is right at the door so I am going to ask nxvl to fix it quickly. We tried to contact you before deciding this but could not find your details on Launchpad. [21:52] Please do not feel frustrated about this, there is nothing against you here and there are certainly other bugs where your help will be very appreciated and we'll have the time to mentor you. [21:57] nijaba: :D [21:57] mm [21:57] mathiaz: kirkland: Apologies for my absence at today's meeting. I spent most of the last hour and a half watching an fsck screen while my computer decided to check the file system after 30 reboots. [21:57] i don't even get a promt asking for password [21:57] sent [21:58] nxvl: isn't mysql already installed on your test machine? [21:58] owh: yuck [21:58] owh: you should be able to find the logs [21:58] owh: wrt us, the init script work for Hardy is pretty much dead [21:58] kirkland: Yeah, I'm waiting for the cron job to update: http://irclogs.ubuntu.com/2008/03/19/%23ubuntu-meeting.txt [21:59] kirkland: I figured that :) [21:59] ah [21:59] owh: it's not a totally bad thing [21:59] owh: they're right, ultimately [21:59] kirkland: C'est la vie. [21:59] kirkland: Who do I send the bill to :) [21:59] * owh is kidding. [21:59] ;-) [21:59] owh: well, it's just delayed reallly [22:00] kirkland: So, yeah, where to next? [22:00] owh: there will be a pile of work to do when the Intrepid trees open [22:00] kirkland: So, was there agreement on how we should proceed? [22:00] owh: we will want to work closely with Keybuk, though, on getting this stuff working in upstart [22:00] nijaba: clean pbuilder [22:00] * owh is still waiting for log update :( [22:00] owh: in the short term, we need to focus back on Hardy bugs [22:00] looking at the ones that are open, solving those [22:00] hi there [22:01] nijaba: i'm installing mysql-server-5.0, is that binary package, didn't it? [22:01] kirkland: Cool. Any I should know about right now? [22:01] owh: http://blackbird.kaarsemaker.net/mootbot/meeting/ubuntu-meeting.log.20080319_2200.html [22:01] that's the hard core focus of the beta period we're in now [22:01] mathiaz: Tah [22:01] owh: i'd say just have a look at them, grab a couple that you think you can solve [22:01] can anyone give me a bit of support for postfix/ dovecot sasl? [22:01] owh: ping me or anyone else here, if you want some more direction [22:01] nxvl: really weird. I reproduced it installing "sudo tasksel install lamp-server" on a clean install this afternoon [22:01] not testing that really, but as a side effect [22:01] kirkland: Any specific server bugs around? [22:02] mathiaz: you have any favorites? [22:02] owh: not off of the top of my head [22:02] nijaba: trying [22:03] kirkland: Email me if you think of any, meanwhile I'll keep my eyes open (in my "spare" time) [22:03] nijaba: i tested using apt-get install mysql-server-5.0 [22:03] kirkland: I'm fighting an iridium short burst message data modem :| [22:04] owh: kirkland: bug 203696 seems new - although related to dapper [22:04] nxvl: should be the same [22:04] kirkland: It knows what it wants to do, but refuses to tell me :) [22:04] Launchpad bug 203696 in dovecot ""undefined symbol: quota" on dapper while trying to use imap_quota" [Undecided,New] https://launchpad.net/bugs/203696 [22:04] nxvl: it is a d-i question [22:04] owh: kirkland: to confirm for dapper, we first need to make sure it's fixed in hardy [22:06] nxvl: errr a debconf question [22:06] nijaba: so i need to test the iso? [22:06] oh [22:06] kirkland: there is also bug 189616 that should be investigated [22:06] Launchpad bug 189616 in dovecot "connection problems under load with hardy dovecot" [Undecided,New] https://launchpad.net/bugs/189616 [22:06] i will tri dpkg-reconfiguring it [22:06] kirkland: but it may be hard to diagnose [22:06] try* [22:07] mathiaz: Yeah, I was just going to say, how do you load test without load ? [22:07] nxvl: it will only be asked if the password was not there before I think [22:07] mathiaz: I suppose you could write some script that hammers away at it. [22:07] owh: well - that's the interesting part of it ;) [22:07] Heh [22:07] owh: how can you stress-test dovecot [22:08] mathiaz: just give it to elmo ;) [22:10] mathiaz: Hmm, just read that bug, and James already tried that :) [22:17] sommer: Did you need anyone to proof the server guide? [22:17] s/proof/proof read/ [22:19] nijaba: i confirmed it, it was a problem with pbuilder [22:23] owh: kirkland: there is also bug 197606 and 199144 that look interesting [22:23] Launchpad bug 197606 in mysql-dfsg-5.0 "Invalid db.table name 'g2_AccessMap' at /usr/bin/mysqlhotcopy line 845" [Undecided,New] https://launchpad.net/bugs/197606 [22:23] Launchpad bug 199144 in apache2 "Apache2 with mpm_worker times out with many concurrent requests" [Undecided,New] https://launchpad.net/bugs/199144 [22:24] mathiaz: I've just taken initial ownership of bug 203696. [22:24] Launchpad bug 203696 in dovecot ""undefined symbol: quota" on dapper while trying to use imap_quota" [Undecided,New] https://launchpad.net/bugs/203696 [22:24] owh: proof-reading the server guide would also be welcomed - the deadline for big changes is tomorrow [22:24] mathiaz: Yeah, I'm just looking at typo's, spelling, etc. No content. [22:24] owh: you can read the current version of the server guide here: http://doc.ubuntu.com/ubuntu/serverguide/C/ [22:25] mathiaz: Tah, just found the link myself :) [22:25] mathiaz: Is there a way to download the lot, or do I just mirror it with wget? [22:26] owh: hum... mirror with wget would be great [22:26] As nijaba I think pointed out, this should be in the server MOTD. [22:26] owh: you should also consider branch from bzr [22:26] owh: If you don't mind reading the docbook format [22:27] mathiaz: Don't care either way, what ever is easiest to get to my local HDD. [22:27] owh: that you way you can fix the grammatical errors right away, push your branch and sommer can merge it back in trunk [22:27] mathiaz: And simplest to give updates/ [22:27] s/\//./ [22:27] owh: well if you're not afraid for bzr the most helpfull way is to branch the source code [22:28] mathiaz: I'm not afraid of things I don't know about :) [22:28] How do I find out where the bzr source is? [22:29] * owh is wandering through ~ubuntu-core-doc [22:29] owh: check out https://wiki.ubuntu.com/DocumentationTeam/Repository [22:31] owh: I think the branch you wanna checkout is https://code.launchpad.net/~ubuntu-core-doc/ubuntu-doc/ubuntu-hardy [22:31] owh: ask sommer about it to make sure [22:31] nijaba: if it fails 2 times i need to reask for password or just rise an error? [22:31] mathiaz: Yeah, I was trying to figure out how to just get the server guide. [22:32] owh: apparently the server guide is burried in the ubuntu-hardy branch [22:32] nxvl: re-ask, after an error [22:32] nxvl: as we do for the user password on install [22:32] mathiaz: Not immediately obvious to me either: http://bazaar.launchpad.net/~ubuntu-core-doc/ubuntu-doc/ubuntu-hardy/files/asommer70%40gmail.com-20080319202309-2nfmhrdj9vufcvge?file_id=3487%401e033ed4-87e8-0310-8703-d68c7bbd86c9%3Atrunk%3Aubuntu [22:32] That seems to be the root of that. [22:33] nijaba: heh, i was just downloading the debian-installer source to copy the structure :P [22:33] nxvl: ;) [22:33] owh: it's in generic/serverguide/ [22:34] owh: generic/serverguide/C/ to be exact [22:34] Ah, I was already in ubuntu :) [22:34] mathiaz: Yeah, now all I need to do is figure out the bzr command, but I think I have it licked :) [22:35] owh: bzr branch http://bazaar.launchpad.net/~ubuntu-core-doc/ubuntu-doc/ubuntu-hardy [22:35] owh: once you've started to make modification you push your own branch [22:35] mathiaz: I was going with bzr checkout --lightweight http://bazaar.launchpad.net/~ubuntu-core-doc/ubuntu-doc/ubuntu-hardy/files/asommer70%40gmail.com-20080319202309-2nfmhrdj9vufcvge?file_id=2104%401e033ed4-87e8-0310-8703-d68c7bbd86c9%3Atrunk%3Ageneric%252Fserverguide%252FC [22:36] How does one get information/release notes for a given package it says it is going to update via apt-get update? [22:36] owh: I'm not sure it'd work [22:37] mathiaz: It didn't :) [22:38] owh: yes - you need to branch http://bazaar.launchpad.net/~ubuntu-core-doc/ubuntu-doc/ubuntu-hardy [22:38] owh: and pull down the whole documentation src :/ [22:39] mathiaz: I'm trying to avoid that. [22:39] mathiaz: This was close: bzr checkout --lightweight lp:ubuntu-doc/ubuntu-hardy/generic/serverguide/C [22:40] mathiaz: What about just downloading the 20 files individually, making patch files and adding them to a bug report or emailing them to sommer? [22:41] owh: well - that's another option but it seems more work [22:41] owh: on sommer's side [22:41] mathiaz: Yeah, I'm conscious of that too. [22:41] * owh is beginning to fear bzr :) [22:45] mathiaz: Seems I'm not alone in my quest: http://www.mail-archive.com/ubuntu-server@lists.ubuntu.com/msg00830.html [22:46] owh: yes [22:46] i'm having a weird problem. I just removed hpijs and now when i go to add a printer it gets to the point where i give it a URI and when i click continue it hangs [22:47] mathiaz: I've given in and am now checking out the lightweight version of the whole guide :( [22:47] owh: I think it makes sense - however it was never implemented [22:47] owh: hopefully for the next cycle we can sort this out [22:48] mathiaz: The emails start a long discussion which indicate some staunch opposition to the idea as far as I can see. [22:49] mathiaz: This all works really well if you've got gobs of bandwidth. Not so much if you're not quite so well equipped. [22:50] mathiaz: I have a very high latency connection, 1second round trip, so all this interactive stuff, that is, the get something, parse it, get something else, parse it, really takes time. [22:53] owh: I started a "bzr checkout --lightweight bzr+ssh://bazaar.launchpad.net/~ubuntu-core-doc/ubuntu-doc/ubuntu-hardy/" and it's already at 100M [22:54] mathiaz: Mine started at a similar time and it's 2.2M, its getting each individual file and not ever getting to 1Mbit. [22:54] mathiaz: As I said, high latency, this interactive stuff is just rubbish. [22:55] s/interactive/iterative/ [22:55] owh: yeah - it will take a looongg time to checkout the documentation [22:56] mathiaz: So much for taking it with me on my phone to have a read. Ah well. I'll kill it and start again when I return :( [22:56] Later all. Gotta run. [23:01] anyone here got any idea why cups is hanging when i hit continue after inputing a device uri? === kloeri_ is now known as kloeri [23:22] nijaba: i think i have it [23:22] nijaba: building for testing === \sh is now known as \sh_away