[01:27] What permissions are needed on a public samba share? I have nobody:nogroup 755. When I set it this way, I can create objects but not delete them (I receive permission denied errors). force user = nobody is set on the share. My smb.conf: http://paste.ubuntu.com/956471/ Permissions of directory and parents: http://paste.ubuntu.com/956505/ [01:28] there is also #samba [01:31] flaccid: I know. I've asked there as well. [01:35] samba gets annoying, so im kind of avoiding this one sorry [02:31] * txomon|home is away: Estoy ocupado === railsraider_ is now known as railsraider === jtv1 is now known as jtv === railsraider_ is now known as railsraider [07:51] i can;t get apache to install and work with the config a i copied from another server, says No apache MPM package installed [08:11] railsraider: install it :) === smw_ is now known as smw [08:17] morning o/ === lynxman- is now known as lynxman [08:21] railsraider: apt-get install apache2-mpm-prefork [08:21] i tired that it didn't work, complained about MPM is not installed [08:22] just use the default config and move the custom parts [08:22] did this config come from another os/distro? [08:22] i am setting up 2 new servers based on running servers [08:23] yeah but - same os/distro? [08:23] i copied the config [08:23] yeah but - same os/distro? [08:23] let me double check [08:23] i believe i did [08:24] railsraider: have you read the Ubuntu Server Guide? [08:24] !docs [08:24] documentation is to be found at http://help.ubuntu.com and http://wiki.ubuntu.com - General linux documentation: http://www.tldp.org - http://rute.2038bug.com [08:24] its the same distro 10.04 [08:24] railsraider: you're probably doing something funny, then [08:25] how is apache configured on the old machine? [08:25] what do you mean? [08:25] apache2-mpm-prefork is already the newest version. [08:25] sudo service apache2 start [08:25] No apache MPM package installed [08:26] well, what parts of apache is it you need? what modules? lots of vhosts? [08:26] You should also install etckeeper last week, so that you can see what difference have been introduced before it broke [08:26] it's probably easier to reconfigure the default install than to just move an existing config [08:26] *AND* you should configure it by hand so that you *understand* what happens [08:27] i don't know the config - i mean i understand it but afraid I might skip something [08:28] i can pastie the config [08:29] basically i have 3 MPM modules mpm_event_module mpm_worker_module mpm_prefork_module === almaisan-away is now known as al-maisan [08:31] every time i try to install one of those one of those that are already installed gets removed [08:31] why? [08:31] railsraider: because they're mutually exclusive [08:32] You should use worker unless you need PHP [08:32] This should be explained in the Ubuntu Server Guide -- have you read that yet? [08:32] no [08:32] i'll read that [08:35] here's my apache.conf http://pastebin.com/rge6UhZr [08:36] Looks OK to me [08:39] from what i understand it needs all 3 modules of that MPM but I'm unable to install all 3 [08:45] railsraider: wrong [08:51] railsraider: those are enabled once mpm is installed [08:52] Setting up apache2-mpm-prefork (2.2.14-5ubuntu8.9) ... [08:52] No apache MPM package installed [08:53] that's after i did sudo apt-get install apache2-mpm-prefork [09:07] railsraider: erm - you need to use the default config [09:07] railsraider: your current config is probably broken [09:07] railsraider: so fix it... === al-maisan is now known as almaisan-away [09:53] how do i only upgrade the security packages on my ubuntu 10.04 lts? === daker__ is now known as daker [09:57] spurkis: the easy way is to edit /etc/apt/sources.list and everything in /etc/apt/sources.list.d and make sure that only security is listed. The advanced way is to use pinning. See https://help.ubuntu.com/11.10/serverguide/package-management.html for a general guide, and the apt_preferences manpage for details on pinning. [09:58] im thinking of doing the following: sudo aptitude update && aptitude safe-upgrade [09:59] spurkis: are you just trying to avoid upgrading to 12.04, or do you want to also avoid any stable release updates to 10.04? [09:59] spurkis: if you're just trying to avoid upgrading to 12.04, you don't need to do anything different. It won't happen unless you edit your sources.list or run do-release-upgrade. [10:00] I wouldn't upgrade to 12.04 yet [10:00] iirc do-release-upgrade won't work on 10.04 LTS until 12.04.1 is released [10:00] indeed [10:01] rbasak, hey, thanks for fixing the server install instructions (though its sad that you have to do michaels work) [10:01] and editing sources.list isn't recommended [10:01] why will it not work until 12.04.1? [10:01] You can run "do-release-upgrade -d" if you wish [10:02] ogra_: no problem! [10:03] (though I'm not sure when "-d" will start pointing to quantal as opposed to precise) [10:11] rbasak, any thoughts on what we might do with openmpi this cycle? [11:02] I need a directory on my machine to be a mount point to server via http, is there a command for that? [11:09] jamespage: I think we should sync 1.5 from debian when they have it in sid. And keep syncing 1.4 from debian. Apart from that, I'm not sure that there's anything else we need to do. [11:10] rbasak, do you know when 1.6 is planned for? [11:10] "soon", IIRC, but I'm not sure if that'll be in time for Q [11:10] * rbasak checks [11:13] jamespage: I can't find anything on expected 1.6 upstream release date. But I don't think we can really do anything further without input from stakeholders. I'll get ARM and some of the hardware vendors we're speaking to when we next speak to them. [11:14] lol - sounds like a plan rbasak === Gallomimia_ is now known as Gallomimia [11:43] hi all [11:43] zul, around? [11:44] https://answers.launchpad.net/nova/+question/195073 [11:44] how can i report a bug keeping/mentioning only ubuntu-package? [11:45] this seems to be issue with ubuntu packaging [11:46] koolhead17: too early for zul to be around ;) [11:46] lynxman, i 4got [11:46] gosh [11:47] so the weekend has gone by. any huge issues with the new release? all going fine? [11:47] lynxman, i need help in filing bug, how can i explicitly mention that bug is with ubuntu-packaging not upstream [11:47] koolhead17: fill it in the ubuntu package [11:47] hey why is update-apt-xapi running at 78% cpu suddenly? [11:47] i didn't say it should run [11:47] koolhead17: https://launchpad.net/ubuntu/+source/nova [11:48] lynxman, you rock. thanks a lot :) [11:48] koolhead17: np [12:12] hallyn: I'm currently working on the LXC plenary for UDS, I'll ping you later when most of the slides are ready so you can have a look and catch what I forgot to mention :) [12:25] https://launchpad.net/bugs/798462 [12:25] Launchpad bug 798462 in update-manager "update-manager's free space calculator for /boot seems insufficient" [Medium,Fix released] [12:25] sad panda :( [12:31] I installed a fresh copy of precise and copied my (lucid) .ssh to my home directory but I can't login using keys. I found out (with -vvv) that it expects the key now to be a DSA key so i changed the ssh_config to read id_rsa and the key gets sent but the server is not accepting it (And I can't get a password now to check what's wrong on the server) .. Any ideas? [12:32] OpenSSH version on Lucid is 5.3p1 while on Precise it's 5.9p1 [12:35] good morning [12:37] Never mind. I was using the wrong user! Cheers~ === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates === Gallomimia_ is now known as Gallomimia === Gallomimia_ is now known as Gallomimia [13:47] stgraber: sounds good === Gallomimia_ is now known as Gallomimia [13:57] m_3: good morning! Do you have a blueprint for charm testing? I thought charm testing on ARM using MAAS might be a work item for you. What do you think, and do you have a blueprint it could go into? There's servercloud-q-arm-deployment to cover juju on ARM and MAAS at the moment. [13:58] rbasak: hey man [14:00] rbasak: I suggested a bunch to the list last week... following up with the actual blueprints later this morning. I'll ping you when I find or submit the charmtesting one [14:02] m_3: OK, thanks. Would you be willing to take a charm testing on ARM work item, and if so, do you think it should go into the charm testing blueprint or into the arm deployment blueprint? [14:03] rbasak: yeah, I'm happy to help with it... probablyl won't have the bandwidth to own the whole thing [14:05] m_3: great, thanks! [14:08] rbasak: please make a note that we should get together over uds and talk about the hw available for charm testing [14:08] rbasak: that's the biggest unknown for me [14:08] m_3: sure === Gallomimia_ is now known as Gallomimia [14:45] hallyn: just saw your e-mail about mounting something into a running container (and matching blog post). Do you think we can/should implement something similar to OpenVZ's way of dealing with that? [14:46] hallyn: they basically have the equivalent of /var/lib/lxc//rootfs and /var/lib/lxc//mount [14:46] hallyn: where /mount is only populated while the container is running and shows the actual running rootfs (with all the mounts that happened there) [14:47] hallyn: mounting anything in /var/lib/lxc//mount would make it appear in the container [14:48] so happy, neither euca-bundle-vol nor ec2-bundle-vol seem to work now [14:49] hey guys im using ntop for monitoring a single box. It works really nice...and its easy to setup. I have a question though about a few different protocols it monitors [14:49] is anyone familiar with Ntop? [14:50] stgraber: yeah, that sounds good. we can discuss that vs the /shared option at uds. [14:51] hallyn: yep [14:51] some time back i changed the way some of my scripts work. (im doing alot of heavy moving of files) after this the over all ip decreased. However the Bytes sent stayed the same [14:51] hallyn: will add a note to the spec so I don't forget about it ;) [14:51] stgraber: we'd also be getting rid of the hack where we currently build mounts unde r/var/lib/lxc [14:51] make tha t/usr/lib/lxc [14:52] so i like it [14:52] hallyn: yeah, that'd be a lot less confusing ;) I'm sure some people must have been wondering wth was happening when apparmor would block the mount and the path would show /usr/lib/lxc/... :) [14:55] hallyn: done. I also moved all the topics into the whiteboard, some were still in the blueprint description. Looks like that'll be a busy session (and I'm sure the list will get bigger) [14:58] stgraber: I wonder if we can pick an evening for a lxc hackfest, so that during the blueprint session we can just discuss what we want to do, and defer implementation discussions (and bikeshedding) to the hackfest/deepdive [14:59] GAH! I just found out dlezcano won't be at uds [14:59] hallyn: sounds good. Who else do we have at UDS who usually hacks on LXC? [15:00] yeah, I figured as much when I saw it wasn't a shared Linaro event this time around... [15:01] I see we have a lot less activities planned in the evenings this time around too, we basically seem to be free Tuesday, Wednesday and Thursday evening (though I guess we'll have a team dinner one of these) === nonotza_ is now known as nonotza [15:12] stgraber: I'm not sure there'll be anyone else, in that case. Unless we can entice zul :) [15:13] hallyn: is there going to be alcohol involved? === Gallomimia_ is now known as Gallomimia [15:17] zul: evening hackfest? yes. [15:17] hallyn: i could be presuaded yes.. [15:17] unfortunately i'm not international so i can't pick up a bottle fo scotch duty-free [15:18] dont they make duty filters? [15:18] i would hope so === dendrobates is now known as dendro-afk [15:20] i updated my ubuntu server from 11.10 to 12.04 and now I get emails from apticron showing 144 updates are available but they have already been installed. Any suggestions? [16:02] Hey, does anyone know of a "VirtualBox" type program (or a way for VirtualBox) to be set up to work for dual boot and be accessed by both? [16:03] virtualbox or vmware workstations should both do that fine [16:05] I installed 12.04 server on a virtual box to mess around with, but for work (I work from home) I have to run Windows, I'll have to take a look. [16:06] Didn't have time to go through settings or anything, so I figured I'd ask in here, everyone seems pretty knwoladgeable (except me :P) [16:14] Hi [16:16] i need a short consultation :) basically, I am setting up hosting for a webshop. I have nginx set up with varnish, mysql and all that. Now I need a mail server. So went with this tutorial: http://flurdy.com/docs/postfix/ which basically makes me install postfix, courier and a lot of AV / spamfilters etc [16:16] however, I came across this section [16:17] where I need to provide my smtp server details. and now I wonder whether I should make my own smtp server [16:19] if you're having to ask whether you should create your SMTP server, you're probably not in a position to run an smtp server [16:22] thanks for the advice man === dendro-afk is now known as dendrobates [16:54] after upgrading to 12.04 vsftpd stops - respawning too fast [16:54] I kept my old configuration file === Lcawte|Away is now known as Lcawte === matsubara is now known as matsubara-lunch [17:06] ok [17:10] any idea where to start on this mess - http://pastebin.com/M0Ddxydd [17:25] tarvid: check your config file against the packaged version perhaps? [17:27] I have zillions of errors in configuration [17:28] is it possible to upgrade to 12.04 LTS from a 10.04 LTS without having to go through the intermediate upgrades? [17:29] yes. LTS TO LTS is supported. [17:29] med_ I tried a do-release-upgrade and "No new release was found" [17:30] drPoO, did you apt-get update first? [17:30] My /etc/update-manager/release-upgrades has as Propmt=lts [17:30] yup [17:30] ill try again [17:30] I did a direct upgrade just prior to release. I'll see if I can still do that. [17:30] I think do-release-upgrade isn't intended to work by default until 12.04.1 is released [17:30] (prior to release I had to add the -d flag, shouldn't be necessary now.) [17:31] I'm not sure how to force it. Will "do-release-upgrade -d" still work, or would that upgrade to Q? [17:31] rbasak, really? I was not aware of that. [17:31] * med_ tests on a vm. [17:31] * rbasak could be wrong [17:32] so should I wait for 12.04.1? [17:33] drPoO: that's entirely up to you. You should be able to upgrade now though, and the only changes to the system will be security updates and only bugfixes for bugs deemed severe enough. === Gallomimia_ is now known as Gallomimia [17:35] https://wiki.ubuntu.com/PrecisePangolin/ReleaseNotes/UbuntuServer#From_10.04_to_12.04 [17:35] does recommend the first point release. [17:35] and it requires the -d flag. [17:37] rbasak, so I would have to run do-release-upgrade -d? [17:40] drPoO: yes. see med_'s link for instructions [17:40] thank you [17:52] utlemming: do you have the bug number handy for http://summit.ubuntu.com/uds-q/meeting/20357/servercloud-q-cloud-imgages/ not getting scheduled correctly? [17:53] jcastro: was going to take a look and see if that can be moved to Mon/tuesday per attendees availability [17:57] utlemming: does tuesday, 16:15 work for servercloud-q-cloud-imgages/ [17:58] arosales: no, it needs to be earlier...I think their plane leaves at 5:00pm [17:59] drPoO: if it's a production machine and there's nothing you really need in 12.04, just wait [17:59] jcastro: looks like it may need to be earlier [17:59] utlemming: what date/time would be optimal? [17:59] that is, if there's nothing you really need in 12.04, never mind upgrading at all :P [18:00] arosales: 10 or so? [18:02] utlemming: ok on it [18:02] jcastro: thanks [18:03] utlemming: tuesday 0900? [18:03] jcastro: that works [18:03] sound good? that'll give people plenty of slack [18:10] want to upgrade oneiric to precise. Can I change the suite name in sources.list, then apt-gte distro-upgrade? === dendrobates is now known as dendro-afk [18:13] ok, here's one... I have an 11.10 x86_64 guest running on a Windows 08 R2 Hyper-V host, everything works OK except I have duplicate drive /dev/sda and /dev/hda with identical uuids. how do I get ubuntu to stop recognizing one of the drives? === dendro-afk is now known as dendrobates [18:15] here's the disk info http://paste.scsys.co.uk/195440 === matsubara-lunch is now known as matsubara [18:28] hello [18:28] hello. [18:28] o/ [18:28] anyone using postfix on centos???? [18:28] wrong channel. but yes. [18:29] thanx, have a lot of questions, administering postfix but have a mailq defunct issue and I always solve that by doing a shutdown [18:30] what channel must I join to get knowledge in how to administer the right way my server? [18:30] #centos? [18:30] or #postfix [18:30] or something [18:30] I would look for a #centos [18:30] channel [18:32] ok, got it [18:32] thank you so much,.. [18:34] === RoyK^ is now known as RoyK [18:44] hallyn, is https://blueprints.launchpad.net/ubuntu/+spec/servercloud-q-dnsmasq-lxc-libvirt really worth a session? [18:44] it seems fairly clear to me that dnsmasq should bind to localhost only. [18:45] smoser: clear to you, and if stgraber agrees with you then we don't need a session [18:45] but there did seem to be some people disagreeing [18:45] clear to me, clear to you, clear to upstream libvirt .. [18:45] i did not see disagreemtn. [18:45] libvirt != dnsmasq packagers [18:46] someone, i don't remember where, said it was not the right answer [18:46] well, knowing who and what the complaint was would be useful [18:46] :) [18:46] it's in one of the relevant bugs [18:48] smoser: the libvirt ppl aren't just saying bind dnsmasq to localhost only. they're also saying to not run a separate dnsmasq for libvirt (and lxc) [18:49] ah. i see. [18:49] that would seem more entailed to me. [18:49] but that does seem pretty smooth in dnsmasq. [18:57] smoser: I still don't mind skipping a whole session if I can get yourself and stgraber to meet with me for 10 mins to chat about it :) === Gallomimia_ is now known as Gallomimia [19:05] hallyn: so what's the current state of the discussion? if it's still to ship a dnsmasq.d config for lxc and depend on dnsmasq instead of dnsmasq-base, I'm still against it but cyphermox and I have been discussing some other ways of working around the current conflicts with multiple resolvers installed on the same machine [19:05] hallyn: the easiest being to bind 127.0.1.1 or 127.0.2.1 instead of 127.0.0.1 [19:07] stgraber: doing that in main dnsmasq package you mean? [19:08] hallyn: still catching up on the bugs, I've heard so many ways around it that I'm not sure what's currently proposed as a fix :) [19:09] hallyn: is the current proposed fix to restrict dnsmasq spawned by /etc/init.d/dnsmasq to only listen on 127.0.0.1? if so, I'm not against it as I don't really care about that package but dealing with upgrades will be fun (if anyone actually uses dnsmasq for external queries) [19:15] stgraber: exactly (so many ways around it) that's why I wanted input in a session :) [19:16] 127.0.0.1 is smoser's proposed fix, yes. i'm happy with it, but that means little [19:16] hallyn: I'd suggest dropping that blueprint and I'll add a generic dnsmasq entry to the networking blueprint I'll register later today/tomorrow [19:16] sounds good. i wonder how i drop a blueprint :) [19:17] because that problem affects LXC/libvirt but also other things like LTSP users using dnsmasq for dns/dhcp/tftp and desktop users who since 12.04 have dnsmasq for their local resolver [19:17] so, i do use dnsmasq for remote queries (when my laptop is a bridge for a wired-only server laptop), but not sure it would affect me as i'm pretty sure i specify the nic by hand [19:17] we need to either find a way to have all of these to play nicely or find a way to get a single dnsmasq covering all our use cases and controlled by the various daemons (not very likely) [19:18] stgraber: can you subscribe me to the blueprint you're adding it to? [19:18] sure [19:18] i don't have much to say, just don't want to be surprised if/when there are changes to lxc/libvirt needed [19:18] thanks [19:18] Daviey: are you around? [19:32] Bah, now I'm pissed! What's WRONG with Oracle to stop permitting OS distribution? Are they just trying to kill Java? [19:32] Anyone here used apache2-mpm-worker ? I'm getting hit really hard on a box and realized it's running prefork - which, to my understanding, is slower than worker. Right? [19:32] Correct [19:34] FunnyLookinHat: Is Baidu spider hitting you? [19:35] Psi-Jack, yeah [19:35] you too? [19:35] I can't nail it down... [19:36] http://www.netnuisance.net/ip/se.php [19:36] Also - I've heard worker can screw up mod_php - is that true? [19:36] Firewall out each of those, and you're golden. [19:36] FunnyLookinHat: It doesn't work with mod_php, at all. [19:36] YOu have to run PHP with worker through fcgid. [19:37] ah [19:37] Ok - soon as I get into my server I'm going to just iptables them out [19:37] lol [19:37] Too bad I can't login right now... :-/ [19:37] Yep. [19:38] Thanks a bunch Psi-Jack :D === Gallomimia__ is now known as Gallomimia [19:47] Psi-Jack, just to check... I should be doing this for the IP ranges provided: iptables -A INPUT -s 61.135.192.0/18 -j DROP [19:47] been a while since I did iptables :-/ [19:47] And will those persist through reboot if I add them that way ? [19:47] hallyn: I marked your blueprint as superseded by the foundations one. That way we also get a link to the new one for free :) [19:48] yea [19:50] err - -I === dendrobates is now known as dendro-afk [20:15] stgraber: cool, thanks === Gallomimia_ is now known as Gallomimia [21:04] Is there any major benefit running MAAS if only about 20% of system are going to be Ubuntu. The rest are Genoo [21:05] depends [21:05] if you have 5 machines, no :) [21:05] if you have 5000, definitely. [21:05] the more the merrier [21:08] yeah. It will probably only be 5 Ubuntu boxes running OpenStack. The rest will be pre-existing Gentoo boxes and new virtual boxes in OpeStack === railsraider_ is now known as railsraider === flaccid_ is now known as flaccid [21:28] RamJett: AIUI, MAAS+juju is the easiest way of deploying OpenStack. It'd probably be worth it just for that. === dendro-afk is now known as dendrobates [21:32] rbasak: Yeah I took a look at that, but it want you to install 11 nodes. I only really need 3 Ubuntu boxes. [21:32] I second rbasak, however, it requires some patience. [21:32] RamJett, minimum OpenStack install is 6 servers. [21:32] Anyone here have a sec to check my saved iptables? I'm restoring this on reboot, but wanted to make sure these drops will actually get hit and not be redirected prematurely [21:33] http://pastebin.com/BrUgPm3h [21:34] I have it running on 1 but plan on 3. I have manually set it up. The boxes have plenty of horse power to run well over 100 KVM instances [21:49] gents! === matsubara is now known as matsubara-afk === dendrobates is now known as dendro-afk [21:53] I really hate IIS... === dendro-afk is now known as dendrobates [22:44] arst [22:57] I tried to upgrade to precise and now I get this error during upgrade :( [22:57] slapd (2.4.28-1.1ubuntu4) wird eingerichtet ... [22:57] Backing up /etc/ldap/slapd.d in /var/backups/slapd-2.4.25-1.1ubuntu4.1... done. [22:57] Moving old database directories to /var/backups: [22:57] Loading from /var/backups/slapd-2.4.25-1.1ubuntu4.1: [22:57] - directory dc=radsport-thalmann,dc=ch... failed. === classix_ is now known as classix [23:32] damn it, ubuntu update crashed my ldap database [23:35] holy crap === Lcawte is now known as Lcawte|Away [23:45] <_KaszpiR_> welcome to backps. === dendrobates is now known as dendro-afk [23:54] _KaszpiR_, i've got backups, but it's a crap if the server is fucked up and not local [23:54] <_KaszpiR_> cough * production server * cough [23:54] <_KaszpiR_> ;)