[00:01] (The rest of that firewall doesn't really concern you.) [00:01] no problems. [01:09] jdstrand: did you have any libvirt bugfixes you were still planning to try to push for o? [01:52] SpamapS: bug 865686 - is it true that upstart will kill any pre-stop job after 10 seconds? I didn't observe that in my testing fwiw [01:59] hallyn_: might depend on which release you're on? === kentb_ is now known as kentb-out [02:08] twb: it might, but i *think* everyone is talking in terms of oneiric here. But maybe not [02:09] yeah bug submitter at least is [02:09] hopefully that timeout isn't a new upstart feature [02:29] hallyn_: ignore me, I missed the bug # === ScottSanbar is now known as SanbarComputing === SanbarComputing is now known as ScottSanbar === ScottSanbar is now known as SanbarComputing === EricJ_ is now known as EricJ [04:51] I noticed a mistake in this tutorial: https://help.ubuntu.com/community/OpenLDAPServer#Creating_a_DIT_with_the_RTC_System [04:51] The last line of the section, "olcDbIndex: objectClass eq", it's missing the object class to index. [04:51] and I don't know what one it would be. [04:55] or I'm misreading the error message. "ldap_add: Invalid syntax (21) - additional info: objectClass: value #1 invalid per syntax" [04:57] Takyoji: you ran ldapmodify instead of ldapadd, maybe [04:57] I actually did run ldapadd [04:57] That attribute is perfectly valid and noraml === RudyValencia- is now known as RudyValencia [05:06] So then where is the syntax error? :P [05:09] I don't know [05:09] I would guess that you already had that object in the database, or you made a transcription error [05:20] actually I think I know what it is, it's probably not described since cosine.ldif, nis.ldif, nor inetorgperson.ldif was imported. [05:20] of which isn't mentioned on that link [05:28] nis.. Ugh [05:34] SpamapS: it's not actually nis [05:35] SpamapS: you'd do better to complain about cosine, anyway, they are like Teh InterNet 0.1b2 [05:35] The only reson they're still around is because there are accidental dependencies in the schemas === rickspencer3__ is now known as rickspencer3 [06:10] twb: I really don't like ldap either.. but it seems we're stuck with that crapfest [06:11] LDAP is better than NIS+ or whatever [06:12] Certainly better than putting account objects in mysql [06:13] I think LDAP itself is OK, it's just that nobody has made it intelligible to small-timers yet, so you have a huge ramp-up cost [06:14] All the OpenLDAP people have a huge hard-on for compiling stuff in /opt against the latest openssl and building in support for enterprise bidirectional sync between databases and shit, but if you just want "the Linux equivalent of AD on SBS", then you don't care about that. [06:14] And of course kerberos is even worse [06:19] twb: well kerberos is sort of complimentary to LDAP [06:19] twb: IIRC, 389 server was started to try and make this easier [06:20] * SpamapS prepares to board his red-eye [06:20] *complementary [06:20] "complimentary" would be if LDAP said "hey krb, you have a nice arse" [06:33] twb: lol. [06:33] twb: I'd say that of LDAP, not kerb ;) [06:34] And then LDAP *maces krb in the face* [06:34] Rarr! === skrewler_ is now known as skrewler === skrewler_ is now known as skrewler === skrewler_ is now known as skrewler === smb` is now known as smb [07:53] morning all [07:53] rbasak: around? [08:29] morning everyone [08:30] morning lynxman [08:30] jamespage \o/ [08:37] lynxman: and how are you on this fine day? [08:37] jamespage: enjoying the mini summer although slightly jet lag still, and you? :) [08:37] lynxman: all good here - enjoying a nice cup of fresh coffee! [08:39] jamespage: brilliant :) [08:43] rbasak: question re nova - should we be using patchset 1 or 2 from https://review.openstack.org/#change,706? [08:43] I think its 2 - but we currently have 1 in the archive and 2 in the packaging branch which I need to resolve [09:08] How are things looking? [09:10] Daviey: meh [09:10] endeavouring to sort out the nova break this morning [09:10] seems we might not have the right patchset for the ringbuffer/libvirt stuff in the version in the archive [09:15] jamespage: What breakage are you seeing? [09:16] Daviey: same break from yesterday - nova-common postinst and permissions on /var/lib/nova [09:16] ah [09:16] I have that fixed - however the patch for libvirt console in the archive is patchset 1 from upstream [09:17] find being over excited [09:17] whereas the patchset in the packaging branch is patchset 2 [09:17] any idea which we should be using? I've not found rbasak this morning yet [09:19] jamespage: patchset 2 [09:19] \o/ [09:20] I'm also thinking about trying a process-based solution instead of a thread-based one [09:20] great - we are not at the moment - I'll update the changelog entry for 0ubuntu5 to reflect that we used the wrong patch in 0ubuntu4 and re-upload [09:21] super [09:21] How are we doing about timing at the moment? Is it worth me trying a process based improvement, or would that not hit the archive in time? There doesn't seem to be any conclusion in the review at the moment for the upstream submission though I suppose ODS is slowing that down [09:26] Daviey, rbasak: uploaded [09:26] let see if we can not annoy the release team today :-) [09:27] rbasak: Yeah, I think we'll have to continue with the patchset #2.. I'll bug vish today, and see if he can comment. [09:31] Daviey: hey sir o/ [09:31] hey ly [09:31] :) [09:39] Daviey: I applied the security patches to my puppet FFe, was wondering if it was worth it also to provide packages for 2.7.5 [09:45] lynxman: I really don't think 2.7.5 can go in Oneiric. [09:46] Weare a week before release, the amount of testing, potential fallout is too high. [09:46] Daviey: I know :) was just asking [09:47] It's /because/ it's a mian package that we need to be more careful. [09:47] Daviey: sure, we ran all the unit tests and it was fine though [09:47] (thankfully puppet provides them) [09:47] If we ddiidn't care about it, then sure.. but because we do.. [09:47] Daviey: of course, more than understandable [09:47] oh, if unit tests pass it must be solid.. lets ship it :) [09:48] Daviey: still 2.7.3 should still be considered, specially now that all has been patched [09:48] Daviey: sorry, didn't hear you through the sarcasm ;) [09:48] lol [09:48] heh [09:51] Daviey: sorry for being so insistant on it, but I think it's important :) [09:51] * lynxman praises Davieys infintite patience [09:52] * jamespage biab [09:56] lynxman: I think if we look to drop it to universe, we can bump it :) [09:57] can't have everything :) [09:57] Daviey: *sad panda* but understandable [09:58] https://bugs.launchpad.net/ubuntu/+source/puppet/+bug/865462 is a report of a regression in a security update. What should I do with it? [09:59] puppet to universe? [09:59] a small part of me just died. [10:00] TREllis: he was joking ;) [10:01] lynxman: whoops, I'll promise to replace the kittens that just died :) [10:01] TREllis: sacrificing kittens again in your basement? ;) [10:02] lynxman: :-) [10:02] has anyone packaged netcf? [10:03] TREllis: I'm saying that a new upstream release of puppet this close to release, is too dangerous for us to support. [10:04] The fact that it is in main, means we have to care about supportability. [10:05] Daviey: aye, I agree with no shipping the latest & greatest right this second anyway [10:06] hallyn_: did you get anywhere with netcf packaging? I see you piped up on http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=573072 === PleXT is now known as PleXs === cloakable_ is now known as cloakable [12:24] rbasak: that issue is only on lucid, and I'll fix it [12:27] jdstrand: thanks! [12:32] TREllis: no, haven't foudn the time. Note that it's not *packaging* netcf. It's implementing it for debian networking [12:33] hi all [12:33] TREllis: i'm hoping the next cycle will nice and stable, and there'll be time to do that :) [12:34] soren: i never got your reply ysterday :( [12:34] *yesterday [12:37] hallyn_: ah cool, dberrange posted about it the other day. He's done the implementation for debian & ubuntu [12:37] hallyn_: http://berrange.com/posts/2011/09/28/porting-netcf-to-debianubuntu-suse-and-windows/ [12:37] hallyn_: so I started some packaging [12:41] TREllis: awesome! [13:05] TREllis: Wow, that's really good news! [13:05] soren: yes, other than making me feel like a lazy heel :) [13:06] hallyn_: Get in line, man. That's been on my todo list since 2008 :) [13:07] i'm trying to set up a pxe boot menu. i've followed some instructions that i've found here: http://www.serenux.com/2010/05/howto-setup-your-own-pxe-boot-server-using-ubuntu-server/ when boot up my client it says "could not find kernel image: ubuntu-installer/i386/boot-screens/vesamenu.c32" [13:08] * hallyn_ takes his spot in line [13:14] soren: indeed, perhaps openstack can use it ;-) [13:15] well, nova specifically [13:16] TREllis: Yep. [13:19] I'm looking for a way with orchestra / cobbler to override the automatically generated pxelinux.cfg/default file - I don't want to include hostname or domain n the append line. [13:19] These get setup automatically when I run 'cobbler sync' [13:32] sorry, my dumb ass put ubuntu-installer directory in pxelinux.cfg instead of the root tftp directory =\ [13:32] SpamapS, around? [13:33] sidnei: yes, at conference so my responses may be interrupted. whats up? [13:34] SpamapS, got pinged by a pypy developer about getting a package into ubuntu (again, it was in hardy but got removed right after apparently). no idea who or how to ask about that. [13:36] sidnei: https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages [13:37] awesome, thanks! [13:37] hey all [13:38] I'm trying to setup an LDAP server, and its way more overhead than I'd prefer to manage [13:38] are there any out of the box solutions you would recommend [13:38] something like the Fedora Directory Server, or Mandriva Directory Server? [13:43] Hello, I am now beside my ubuntu-server that is a webserver and hosts a website [13:44] I want to configure it to that way where I can just type my website's name instead of the local ip address, what do I do? [13:44] boxybrown: try the Ubuntu serverguide [13:45] atdprhs: set up name resolution (/etc/hosts or DNS) [13:45] I went through and I configured it,but I reached to that point in my programming, I need it to be sitename.com instead of 192.168.1.X [13:45] dns is empty [13:45] /etc/dns is empty [13:46] hosts: 127.0.0.1 sitename.com localhost [13:46] 192.168.1.2 sitename.com localhost [13:46] pmatulis: I already tried the ubuntu server guide, and followed the directions for installing OpenLDAP [13:47] pmatulis: anything I should do? [13:47] I'd rather not have to deal with schema setup and writing my own scripts for adding users etc etc, I was wondering if there were any 'out of the box' packages like Mandriva Directory Server [13:47] atdprhs: yes, google for 'linux hostname resolution' [13:48] okayz, thank you, give me 15 minutes and I will tell you if I got any progress or not [13:49] boxybrown: what do you mean by 'out of the box' packages? [13:50] pmatulis: it already has schemas setup, they already have scripts and/or a user interface for managing users [13:50] whats the general rule of thumb for swap space? if im assigning say 15GB, how much swap space should I assign? [13:50] boxybrown: schemas depend on what the purpose of the directory is [13:51] boxybrown: you can alwyas install a graphical client to administer it [13:51] boxybrown: such as 'Apache Directory Studio' [13:51] pmatulis: yes I'm aware, but when you are just trying to setup a simple LDAP to do user management, theres a lot of cruft to try to understand with regards to schema setups [13:52] I want that to be abstracted as and have reasonable defaults [13:52] boxybrown: LDAP by nature is a very versatile tool, you cannot expect a 'one size fits all' type arrangement [13:52] if it doesnt exist on ubuntu, it doesn't exist, I was just wondering if there were suggestions [13:53] pmatulis: this is an unreasonable argument because this tool exists on other distros. whether you agree with those packages or not is irrelevant, i'm simply asking if theres such a solution for Ubuntu [13:53] boxybrown: i'm not arguing [13:53] thanks for your help [13:54] boxybrown: you're welcome === kentb-out is now known as kentb [13:57] anybody ever tinker with LTSP using PPC clients? [14:11] soren: is openstack code still in bzr right now? === mendel__ is now known as mendel_ [14:26] whats the general rule of thumb for swap space? if im assigning say 15GB, how much swap space should I assign? [14:28] you dont technically *need* swap [14:29] my rule of thumb is 1.5x or more the amount of RAM you have, but only if you intend to use more than the amount of physical memory [14:29] AdvoWork: as much as you need [14:29] (commits as of 10/1 so i'll assume so) [14:33] jamespage: ping [14:33] robbiew: call time? [14:34] aye [14:34] jamespage: g+? [14:34] robbiew, good with me [14:35] jamespage: one sec...will send an invite [14:38] jamespage: invite sent...I think [14:50] anybody ever tinker with PPC clients on an LTSP server? [15:11] hallyn_: No. [15:12] hallyn_: Well, the trunks are in git, but the stable branches are still in bzr, I think. [15:12] oh [15:12] the wiki page linked which was supposed to tell me where to get the code, did not exist :) [15:13] feh, now it is. wiki glitch earlier today? [15:16] Not that I know of. [15:17] hallyn_: root 3557 0.0 1.8 408844 298044 ? Ss Feb24 0:41 /usr/sbin/apache2 -k start [15:17] hallyn_: Seems to have been running untouched for a while :) [15:24] SpamapS: there? [15:25] soren: it pointed me to a 'do you want to creae this page' page (it being the link from http://wiki.openstack.org/HowToContribute) [15:25] oh well [16:12] RoAkSoAx: you'll be updating https://wiki.ubuntu.com/ServerTeam/Orchestra with instructions and such, right? I **think** you said "yes", but my memory is crap during the month of the release. [16:16] robbiew: hehe yeah I will [16:16] robbiew: im actually still writing the document from the sprint, which it will also serve as base to update the wiki [16:18] cool [16:18] thnx! [16:29] hallyn_: You are most confident that the libvirt changes only impact lxc? [16:31] Daviey: it passed qa-regression-testing at any rate [16:32] hallyn_: If it blows up in my face, can i blame you? :) [16:33] of course [16:35] nah, i wouldn't do that. [16:35] (he would) [16:36] hallyn_: Just to confirm, it doesn't touch the kvm (or other) interfaces? It doesn't look like it, but can you confirm? [16:36] Daviey: fyi: should have my last bug fixed later tonight, I think this ones a fixer [16:36] ikonia: rocking, what bug number? [16:36] it doesn't touch those interfaces. It does however moving virExec* from util/util.c to util/command.c, but nothing broke due to it [16:37] not got number here in office hence "tonight" [16:37] ikonia: Just ake sure it's not one already fixed.. [16:37] make* [16:37] I learnt that lesson :) [16:37] cool [17:01] hallyn_: you around? [17:01] utlemming: yes [17:02] trying to make sense of kernel signal delivery code [17:02] hallyn_: qemu arm images are working again [17:02] last nights build boots all the way to a login prompt (albeit slowly) [17:03] utlemming: cool [17:03] i assume that's arm VM on x86 host? [17:04] yes [17:05] qemu-system-arm -M beagle -sd [17:46] Hi, ive installed apache on ubuntu server everything seems to be working fine but I added a directory to /var/www and I get a 403 and 508 error when I try access those files froma browser can anyone help please? === guampa_ is now known as guampa [17:50] dnyrgr: Do those directories belong to www-data ? [17:57] guampa: Im new to apache (im trying to put up some doc. generated by doxygen on a local dev. server) so I dont know what that is. [17:58] tabfail, but, permissions aren't related specifically to apache, moreso to the OS [17:59] or should i say "ownership" [18:03] no I don't seem to have a 'www-data' group [18:03] guampa: ^ [18:03] what OS ? [18:03] ubuntu server [18:04] "getent group www-data" doesn't return an entry? [18:04] yes it does sorry [18:04] If you installed apache it should have made www-data entries [18:05] give that group ownership to the resources [18:07] still get the same errors [18:08] I can add files to /var/www the problem only seems to occur in the 'doc' subdir. of www [18:08] Im also generating the files and then cp the directory to doc [18:08] no sure if that helps [18:10] I Installed and configured Samba via webmin, but for some reason why I attempt to login it rejects it and the log shows the line "Can't become connected user!". Any ideas on how to fix this? [18:11] guampa: any more ideas? [18:11] dnyrgr: idk know what errors you got [18:12] but you might check the permissions are correct besides ownership [18:12] hi all [18:12] meaning the group (if you gave www-data group-ownership) has to have rx permissions on everything you want apache to access to [18:13] guampa: 403 in the browser 505 in access log and 'client deied by server configuration' in the error log [18:13] i don't think that's related to perms/ownership but still check those [18:15] dnyrgr: also you might want to try in #httpd too, apache's channel [18:17] guampa: lol thats where I started they sent me to debian and ubuntu which sent me here [18:17] hehehe, "irc soccer" it's called ^^ [18:19] guampa: ok well thanks for all the help thanks [18:19] from the error it looks like it should be something in the apache config, but i can't help beyond that, sorry [18:39] SpamapS: is it a good time to talk about squid & maverick? === med_out is now known as medberry [19:46] hggdh: probably not, at ODS [19:55] SpamapS: if you see adam_g, tell him I'm back [19:55] nevermind [19:55] lol [19:56] SpamapS: roj. For the record, there is no FS recovery after applying the maverick-proposed squid; but still the system states the FS is in use, and remounts it RO === medberry is now known as med_out [20:25] I'm having some confusion with orchestra. I'm adding a system with "cobbler system add --name=test-vm-1 --profile=lucid-x86_64 --mac=52:54:00:af:2c:9f" [20:25] The issue is that the pxe menu for this system, the profile is listed as 'test-vm1' instead of 'lucid-x86_64' [20:26] As seen in http://pastebin.com/xPuvUimy [20:27] Daviey: oh, we never finished our discussion [20:27] Daviey: I vote let's go with the longer patchset - it's the one i tested with qa-regression-testing [20:44] fixxxermet: so the profile you have selected in your systems file will be used for that mac address :) [20:44] lynxman: hey [20:45] * koolhead17 finds channel in deep silence today!! [20:45] koolhead17: Sorry, systems file? I thought the profile was selected by the --profile switch [20:46] fixxxermet: i meant that particular system will use profile as mentioned/provided by systems file [20:47] What is the systems file, though? [20:48] From the pxelinux.cfg file you can see "url=http://192.168.10.1/cblr/svc/op/ks/system/test-vm-1" is being based off the --name and not the --profile [20:50] I'm adding a system with "cobbler system add --name=test-vm-1 --profile=lucid-x86_64 --mac=52:54:00:af:2c:9f" [20:51] right === med_out is now known as medberry [20:52] Honestly I'm not seeing what you're getting at. [20:54] fixxxermet: what i meant by that is, in you cobbler/orchestra configuration, you can have N number of profiles [20:54] right [20:54] now when you want to use one particular profile for a specific system on basis of MAC address you use cobbler system add [20:54] and there you can specifically provide it whiich [20:55] profile it should use [20:55] am i much clear now? [20:55] Isn't that what I did? I want a particular system, based on mac address, to use a particular profile, from the --profile switch? [20:56] hmm so what is the the problem? [20:57] is isn't using the profile that I'm telling it to. [21:00] fixxxermet: can you inspect the entry of your system file once [21:03] fixxxermet: cobbler system report --name= > [21:09] Profile : lucid-x86_64 [21:09] @ koolhead17 [21:10] fixxxermet: interesting: so this profile is overwritten via orchestra? [21:13] koolhead17: No, the profile isn't touched. When I run 'cobbler system add' and then 'cobbler sync', it makes the pxelinux.cfg/Mac-add-res-s file. In that file, instead of pointing the url to the profile, it points the url to the system name [21:14] fixxxermet: you dont need to use cobbler-sync after adding the system i suppose. [21:15] koolhead17: running cobbler-sync sets up the tftp pxe files [21:15] otherwise it boots using the default pxe file [21:16] fixxxermet: i never did this sync think with standalone cobbler afaik [21:17] koolhead17: have you used orchestra? I haven't used standalone cobbler myself [21:18] i have use cobbler on ubuntu. installing orchestra as am having conversation with you on oneiric [21:20] I am experiencing a lot of I/O wait on my system. Is there any way to see which processes/users are causing the I/O wait? I have tried top and htop, but they do not seem to reveal this. [21:24] iotop? [21:25] hallyn_, great idea. I will try that next time. The load is currently over. [21:26] I've run into a reproducible issue when performing an apt-get dist-upgrade on ubuntu-10.10 server [21:26] it basically bonks my install [21:26] shutdown: error while loading shared libraries: libdbus-1.so.3: wrong ELF class: ELFCLASS32 [21:27] any ideas? === medberry is now known as med_out [21:31] :( [22:26] hallyn_: yeah, that was settled. :) === jjohansen is now known as jj-afk [23:05] is there any way to get a list of packages that shows which are 32bit and which are 64bit? [23:26] <|G0LTaR|> hello, i got a problem :/ i try to perl script and i got this error: Can't locate HTTP/Request.pm in @INC (@INC contains: /etc/perl /usr/local/lib/perl/5.10.1 /usr/local/share/perl/5.10.1 /usr/lib/perl5 /usr/share/perl5 /usr/lib/perl/5.10 /usr/share/perl/5.10 /usr/local/lib/site_perl .) [23:26] <|G0LTaR|> some help? :( [23:52] Where would I find the list of upstart events triggered by the new cloud-init software in Ubuntu 11.04 Natty? They apparently have changed from some older versions. [23:54] ...and now I try a different Google search and up pops smoser's alerting me that it was going to change and what I should use: http://groups.google.com/group/ec2ubuntu/browse_thread/thread/3a0ccf40a45e7857 [23:54] I'll give this a try.