=== ScottK2 is now known as ScottK [00:20] keithzg, did you add a pass phrase to your keys ? [00:27] K17MT: Nope. Hence my surprise. [00:27] It works fine with just normal ssh, and via ansible. [00:29] KI7MT I mean. [00:46] keithzg, so you not using pw-less log in with keys .. I know on 12.04, each session, where I had a pass-phrase on a key, Im asked for it the first time, not after, so was wondering if ssh was seeing the vm's as a new session each login attempt. [00:47] KI7MT, nope, I most definitely am using keys without pw, and yet virt-manager is prompting me for passwords nonetheless. [00:48] keithzg: any idea what it is using the passwords -for-? [00:49] if the keys aren't encrypted, I'm curious what it intends to do with the password you might give it :) hehe [00:49] sarnold: Oh, when I then use the respective passwords for the remote accounts I'm using, they work. So it's basically just ignoring that I have keys defined in .ssh/config and just asking for my credentials instead. [00:49] oh, I see. hrm. that's not cool. [00:50] And it doesn't remember them either, of course, so I have to type them repeatedly. [00:53] (I imagine it'd play nicer that way if I had the gnome keyring installed; KDE and Qt applications always seem to play nicely with GNOME keyring, but GNOME and GTK applications never seem to bother using KWallet, the lazy buggers). But of course that's besides the point! [00:53] Hello All -- I just built apache 2.4.6 -- I have it up and running but PHP is not working [00:53] I keep getting an error /opt/httpd/modules/libphp5.so: cannot open shared object file: No such file or directory [00:54] how do I get php to work with apache 2.4.6 [00:55] * keithzg finds it somewhat perplexing that all the virt-manager documentation seems to just say "you can use ssh keys, it's all cool!" without any technical discussion of where it's pulling what keys it should be using from or what might go wrong...it's all supposed to "just work", which as usual is a recipe for it NOT doing so for me ;) [00:55] SiliconG: did you compile and install php? [00:55] I did not - I used apt-get to install php [00:55] do I have to build that from source also? [00:55] I suspect yes [00:56] hmm -- there has to be a way to get it to use the other package [00:56] SiliconG: if you built apache from source, you'll have to have either built PHP from source into where it expects it to be, or to have compiled Apache against the existing PHP installation in the first place. [00:56] apache modules are built for specific versions of apache, right? I think even if it worked, it'd be bound to break somewhere or else.. [00:58] SiliconG: Out of curiosity, why did you build apache from scratch yourself anyways, instead of just using the packages for it too? [00:59] It's great fun to compile stuff from source, but for something like Apache I'd worry about keeping it easily up to date for security reasons . . . [00:59] I could not find packages for apache 2.4.6 or greater [00:59] SiliconG: What version of *buntu are you on? [01:00] 12.04 LTS [01:01] Ah. One release before apache2 2.4.6 entered the repos. [01:02] In that case, you're definitely better off just compiling the PHP module yourself, you've got a radically newer stack you're trying to use. [01:03] ok then I will do that [01:04] I'm a bit confused that you're using 2.4.6 instead of 2.4.7, but ;) [01:04] do you think I should also compile my own version of mysql? [01:04] I am using 2.4.7 [01:05] (I just compiled that myself to replace the apache version on a server that was, terrifyingly, exposed to the internet...while running 2.0.59, yikes) [01:08] did you remove the other version of apache? how did you setup your box? [01:08] In this case it was a replacement, but it's really quite another story. [01:09] The box was running a Linux distro that hasn't existed for about 6 years ;) [01:09] I am wondering did you put apache in the opt folder or what? [01:09] wow - that is awesome [01:09] keithzg: hah, cool, which one if you do'nt mind? :) [01:09] Trustix! [01:09] ah the loyal opposition :) hehe [01:10] (I came from Immunix, back in the day, hehe :) [01:10] ahhh, haha, I see I see [01:15] that is impressive -- ok so I am building libxml first -- I guess I have to build that also before building php [01:16] do I need to do ./configure --prefix=/opt/ for everything I build -- what is the best practice for building custom stacks? [01:17] --prefix=/opt/ or /usr/local is nice. /usr/local/ won't be touched by packages (well, shouldn't be) but whatever works for you [01:19] so will I have to rebuild apache? [01:19] or just build php? how should I deal with the old versions that are installed via apt-get [01:22] SiliconG: I think you build php5 after you've built apache [01:22] yea that is what I thought [01:23] should I remove apache using apt-get the other version? [01:23] SiliconG: from http://packages.ubuntu.com/apache2 and .../php5 you can find the .dsc files, which have Build-Depends: lines. That's a decent indicator of which order things ought to be compiled. [01:23] SiliconG: if it makes it easier for you to keep track of rather than having two installed versions side-by-side, sure [02:04] anyone out there do a ultra new stack and document it ? === peter is now known as Guest34187 === petey_ is now known as `petey` === `petey` is now known as petey === petey is now known as petey_ === petey_ is now known as `petey` [07:13] hi. I want to setup moinmoin wiki. I'm following this guide: http://moinmo.in/HowTo/UbuntuQuick#Nginx_Installation [07:14] I've been convinced that Python virtualenv's aren't a good idea, so I'm not using them [07:14] installing to /srv/moin/ [07:15] but when I do `start moin` then error 'unknown job' [07:15] `service moin start` same error [07:17] how does upstart identify services? It looks like its just by the name of the conf file [07:45] in another tutorial, uwsgi is made the service: http://www.projectlense.com/LowEndBox/MoinMoin/ [08:14] I'm running into this issue when I'm trying to setup Bind9. http://pastebin.com/NRgEkuQE what am I doing wrong? [08:23] I can't authenticate myself in posfix from a remote machine. Ithink there is some problem with sasl. what should I do? [09:22] jamespage: https://code.launchpad.net/~yolanda.robla/charm-helpers/rabbitmq_ha/+merge/199453 [09:23] yolanda, not quite [09:24] i misunderstood that? [09:24] yolanda, I can't remember tbh [09:24] lemme grab the branch and look at the entire code path [09:25] ok, just saw your comment about the MP and executed it [09:25] yolanda, yeah - I'm not always right [09:26] you are! [09:26] yolanda, don't we want to set rabbitmq_hosts if clustered is not true? [09:31] jamespage, actually we need to set it for active-active [09:32] yolanda, yeah - but that's not clustered [09:32] clustered == active/passive with pacemaker et al [09:32] yes, so that's not ok [09:33] i think it makes sense to check for the reverse condition, and also send that if there are more than one [09:33] it doesn't make sense to send if we only have one rabbit_host anyway [10:57] yolanda, sounds good [10:57] I think [10:57] (brain still in dormant mode after christmas break) [10:57] i'm just testing [10:57] jamespage, that also happens to me, hard to be back from so long holiday [11:03] jamespage, tested with cinder: https://code.launchpad.net/~yolanda.robla/charm-helpers/rabbitmq_ha/+merge/199453 [11:03] yolanda, OK - that looks better [11:03] is this covered by existing unit tests? [11:04] no [11:04] i need to add some test cases [11:05] let me do it before merging then [11:05] yolanda, please do - I pasted test failures into the MP [11:05] I think they just need updating tbh [11:07] ok, we talked also about using persistence for rabbit, there had been some work in a helper, do you know if there has been any update on this? [11:13] what would be the best tool to graph traffic from apache logs? === Ursinha is now known as Ursinha-afk [11:31] RoyK: awstats ? mrtg ? [11:55] jamespage, i updated tests to match our results [12:11] yolanda, OK _ merged - just for future reference 'make lint' is a good one to check as well [12:12] I tidied that up on merge :-) [12:12] oh, sorry [12:12] i feel very slow today [12:12] i'm just looking at unison for rabbitmq-server now [12:23] hi. we've got suspicious traffic on our network. I want to setup an Ubuntu server between the gateway and the outside world. No squid or anything. It must just receive any connections, iptraf must be able to see them, and then they should be forwarded on. Can this be done? [12:44] makara: yes, and it is straightforward. A bridge will just transparently forward everything, and tcpdump or anything else that uses pcap will be able to see all the traffic. But using a server like that seems overkill. Why not just a switch that can give you a mirror port? [12:45] I say transparently, but of course LAN machines will see a MAC address change. With a switch mirror port, that wouldn't happen. [12:46] rbasak, I'm unsure of the topology here. We don't have an admin and I've got to sell this idea to the monkey in the control room [12:46] it's all Windows server here [12:47] You'll need to understand the topology either way. [12:47] A switch is more transparent. [12:48] You'll still want a machine to take and analyse the data, of course. Ubuntu can do that as well as anything else. [12:49] reading https://en.wikipedia.org/wiki/Port_mirroring [12:49] sounds good [12:50] rbasak, how would ubuntu be setup to do that, without interfering with it's own connections [12:51] i mean, do you send all traffic to a certain port on the listening server? [12:52] makara: you'll either need to filter out its own connections on any analysis that you do, or you could use a second port on the server. === gary_poster|away is now known as gary_poster === Ursinha-afk is now known as Ursinha [13:32] Hi all [13:33] people, how I do to create a openvpn tunnel for each ppp interface? I already has one tun0 running, but I have ppp0 and ppp1, So I would like one tun1 using ppp1, how I do? [13:42] anyone? [14:10] hi [14:12] I am trying to install !xen, but I get errors 'socket:/org/xen/xend/udev_event': No such file or directory [14:12] any ideas? [14:13] jamespage, do you know if something has changed in serverstack? i'm using the serverstack_dnsmasq script but it isn't recognizing my names anyway [14:13] Kartagis, That is harmless. The systemd udevd does not support socket as a target [14:13] not that I know of [14:14] Kartagis, In doubt you can comment out those rules in /lib/udev/rules.d/xend [14:14] smb: I also don't think Xen is installed properly [14:14] smb: I even updated grub, but still I can't boot to it [14:15] Kartagis, Did you use the xen-system-amd64 meta package to install? [14:15] smb: I used xen-hypervisor-4.1-amd64 [14:16] was I wrong? [14:16] 4.1.. erm which release are you basing on? [14:16] 13.10 [14:17] xen-hypervisor-4.3-amd64 [14:17] sorry [14:17] Kartagis: out of curiosity - why not KVM? [14:17] no reason [14:19] oh [14:19] I remember now [14:19] I couldn't find out what replaced python-virtinst [14:23] Kartagis, Ah Yeah, 4.3 sounds better [14:24] Kartagis, Have you configured your system to show the grub screen? That should show a section with Xen hypervisor on reboot [14:24] anyone can help me to say how I create multiple openvpns tunnels in ubuntu? I already has the tun0 running between two ubuntu boxes [14:25] I not understand. I need one client.conf for each tun (tun0, tun1, .. tun10)? or in just the client.conf I put all? [14:28] smb: http://paste.debian.net/75017/ [14:28] these are my options [14:28] brb [14:32] Kartagis, You should check the default line. It most unlikely contains "Xen 4.1". Check /boot/grub/grub.cfg [14:34] Kartagis, "Ubuntu GNU/Linux, with Xen hypervisor" [14:36] I have just realised [14:36] I have 4.3 [14:36] that line contains 4.1 [14:36] :S [14:36] brb [14:38] Kartagis, This may help to fine-tune grub options: https://wiki.ubuntu.com/Kernel/Reference/Xen [14:39] Hm, I guess I should make a note about the default line [14:41] http://paste.debian.net/75022/ [14:42] finally === NCommander is now known as Guest55180 [14:50] jamespage: https://code.launchpad.net/~zulcss/nova/ftbfs-jan1/+merge/200847 [14:51] zul, I don't think that makes much sense - maybe I'm being dumb [14:54] jamespage: yeah i agree... [14:54] jamespage: im going to wait until the kombu issues are sorted and resubmit [16:01] Hello, I'm in charge of instantiating a LAMP stack on 20-30 virgin Linode servers in the coming months [16:01] I'd like to perhaps make an image of a single installation and copy it onto other servers [16:01] What would be the best way to go about something like that? Is there a better way? [16:05] garbagegod: look into juju [16:06] garbagegod: but whichever way you do it, don't do it by hand. Creating an image is going in the right direction, but really you want to automate the entire setup of each instance and the connections between them. Then apply that automation to each instance you start. No need for an intermediate image then. [16:07] Thank you [16:15] Hi, Does anybody know how can I change the current sent print job's pritner? I want to change it from a real printer to 'print to a file' === negronjl_ is now known as negronjl === mjohnson15_2 is now known as mjohnson15 [17:07] gaughen: you can see all server bug triage activity here: https://lists.ubuntu.com/archives/ubuntu-server-bugs/ [17:07] This covers all packages that the Ubuntu Server Team is subscribed to, which should include everything in main that our team has committed to look after. [17:08] i knew there was a bug i forgot to bookmark xD [17:08] bug list* [17:26] Can I use Juju with ubuntu server as opposed to ubuntu cloud? :/ [17:39] garbagegod: Yes. [17:42] What if my host doesn't appear to be on the initial list of supported hosting environments for juju? [17:46] garbagegod: You have to use the maas provider. [17:47] garbagegod: https://juju.ubuntu.com/docs/ [17:47] garbagegod: maas/local/manual are your options. [17:50] Okey dokey… Is this preferable to say, having a shell script that installs everything? [17:50] Taking into account that the servers don't need to be connected in any way [17:51] garbagegod: Choice is entirely yours. [17:53] In your opinion, which approach is more efficient / flexible / easy to manage? [17:53] I'm assuming Juju === medberry is now known as med_ === lfaraone_ is now known as lfaraone === _KaszpiR_ is now known as _KaszpiR___ === _KaszpiR__ is now known as _KaszpiR_ [19:04] Hello [19:06] I'm wondering is it worth upgrading my server from 13.04 to 13.10, considering my main use is serving dynamic sites? [19:12] is there a way to jail users when they connect via SSH so that they can't `cd` out of their /home/$USER/ directory? [19:25] TheLordOfTime: have fun! :) http://wiki.apparmor.net/index.php/Pam_apparmor_example [19:25] ewww.... >.< [19:25] ironhalik: it is worth making sure you are on a supported OS release, beyond that, it's up to you... [19:25] * TheLordOfTime gives up without an attempt :P [19:25] TheLordOfTime: aww. :~( [19:26] sarnold: i don't have the energy just yet to jail users, the only users I'm concerned about are the others on this network, but none of them have sudo so i'm not sure I should be scared. [19:26] sarnold: nginx security stuff and annoying people from askubuntu have drained my energy for the day :/ [19:27] TheLordOfTime: d'oh :/ [19:27] sarnold: I was mainly wondering if there were any significant changes to performance - nginx, php, mysql [19:27] ironhalik: oh. I'd assume more or less inconsequential changes. [19:27] ironhalik: as compared to what [19:28] TheLordOfTime: to packages in 13.04 [19:28] ironhalik: there shouldn't be any substantial speed changes in terms of server responsiveness [19:28] 13.04 -> 13.10 wasn't a *huge* delta for nginx, php, or mysql AFAICT [19:28] ironhalik: maybe two milliseconds difference which is negligible [19:29] * TheLordOfTime does, however, recommend you rely on the nginx PPAs instead of the packages in Ubuntu for nginx, as the PPAs are newer and usually kept in sync with what Debian has [19:29] and sarnold knows why i push those PPAs so hard :p [19:29] I'm having some performance issues and I plan on upgrading my server (hardware wise). I can do it from a snapshot, or, what the hell, install 13.10 on it :> [19:30] ironhalik: are the performance issues from the hardware or the software? [19:30] if they're hardware then you'll have the same issues with 13.10 or others [19:30] if they're software, then we need to do some diagnostics [19:31] TheLordOfTime: it's mainly CPU load due to PHP hits [19:31] funnily enough [19:32] ironhalik: I'd set up caching if it's a big issue, should lower CPU load a little, however if it's solely because of network traffic hits, then you should consider doing some tweaking of PHP's memory usage and settings [19:32] and restrict what it actually can use. [19:32] * TheLordOfTime doesn't have a guide for that though, he always uses pretty-much-default settings because they just work for the hardware he has around [19:33] ironhalik: TL;DR, though, 13.04 -> 13.10 won't resolve your "number of site hits" issues. [19:33] I tried that, but considering I'm not that proficient with administration, with all the tweaks, it would quickly become unmaintainable for me [19:33] sarnold: how well do you know PHP's settings, perhaps you can point them at suggestions for limiting PHP's mem usage [19:34] ironhalik: IIRC, the settings are set-and-forget [19:34] TheLordOfTime: not at all. I detest php and everything about it. :) hehe. [19:34] sarnold: heheh [19:34] sarnold: explains why everyone hates php :p [19:34] sarnold: their packages are PAINFUL to revise >.< [19:34] my patch changed the php5-fpm socket handling but meh, that was painful to create too [19:34] TheLordOfTime: owww.. thanks for the warning. :) [19:35] sarnold: in 5.4 i think it was (old patch, already in Ubuntu) [19:37] sarnold: here's the old debian bugreport that caused nginx users a headache for a while :p http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=650204 [19:38] sarnold: but yeah that's basically all i did, some bug closures because php5 fixes got in but were never autoclosed [19:38] TheLordOfTime: groan, that looks horrible ;) [19:38] :p [19:39] sarnold: FINDING that file was a pain >.> [19:39] HI for 12.04 update 3, I need to get vmlinux with debug info and Linux source tree. I want to use it for jtag debugging. Do you know where I can get those? thx [19:39] it still haunts me >.< [19:40] ironhalik: usually the php settings changes are set-and-forget, same for nginx caching unless your sysadmins were crazy enough to use the default conf file and web root path [19:41] rostam: I think the linux-source package ought to suffice for the source; don't know about the debug info, though, sorry [19:44] sarnold, thanks a lot, [19:44] TheLordOfTime: I am root! :P [19:44] Hi, Does anybody know how can I change the current sent print job's pritner? I want to change it from a real printer to 'print to a file' [19:45] TheLordOfTime: I already use pagespeed for caching, and I tweaked PHPs memory usage a bit [19:45] but I dont trust myself, i prefer to use defaults for the settings I dont entirely understand :P [19:46] then ask questions :P [19:46] 'tis how one learns :P [19:47] Im more of a coding guy :P I lack the background in networking and stuff [19:47] I maintain the server out of necessity [20:04] on a server with multiple nics/network interfaces, can apparmor limit process access to a subset of network interfaces? [20:05] hazmat: not atm [20:05] jjohansen, ack, thanks [20:05] jjohansen, re not atm, does that mean roadmap for trusty? or backburner? [20:06] Hi, Does anybody know how can I change the current sent print job's pritner? I want to change it from a real printer to 'print to a file' [20:06] hazmat: its not scheduled for trusty, but likely the release afterwards [20:07] likely u or v at earliest.. [20:27] hallyn, heya, http://pastebin.ubuntu.com/6716934/ [20:27] so I ran into this problem with juju, and I tried just running lxc-ls [20:27] sudo lxc-ls seems to work [20:28] is this normal? [20:28] jcastro: juju runs lxc as root [20:29] http://pastebin.ubuntu.com/6716390/ [20:29] is the juju error [20:29] jcastro: does 'sudo lxc-lx' work ? what does it return [20:33] it returns nothing [20:33] but appears to execute correctly [20:33] let me spawn a container to make sure ls works [20:40] thumper, sudo lxc-ls lists the container [20:53] jcastro: you ahve to run lxc-ls as root now [20:53] jcastro: however, it shouldn't give you a backtrace like that, just a nice error msg [20:53] so that's bug-worthy [20:53] jcastro: if you have unprivileged containers you can list those yourslef :) the problem otherwise is that there may be setuid-root binaries in the container which we don't want host users to use to exploit the system [20:54] so /var/lib/lxc is not world-readable [21:02] hallyn: I had a quick look and I'm not sure how to better handle errors from list_containers... the exception currently being emitted seems reasonable to me. [21:02] hello what is the default ftp server on ubuntu 12.04 ? [21:03] rtyuiop: there is no default, just install the one you want. [21:03] unfortunatelly i can't [21:03] hallyn: arguably lxc-ls should catch it and print an error instead of crashing but that issue will vanish with the next milestone as it'll instead list the user's containers [21:03] rtyuiop: Why not? [21:04] stgraber: cool, it can wait then, certainly not high prio [21:04] stgraber, I filed a bug on that [21:04] wrt printing an error [21:04] hallyn, is there any other changes in lxc that you think might affect juju? [21:04] since we have thumper here [21:05] this is the error : http://paste.ubuntu.com/6717135/ i got [21:05] jcastro: ok, thanks [21:05] rtyuiop: vsftpd is in main [21:05] so what to do , [21:05] ? [21:05] i won't install IT [21:05] http://askubuntu.com/questions/343047/installing-skype-on-ubuntu-13-04-64-bit [21:06] skype has broken your package manager so you can't really install anything [21:06] rtyuiop: vsftpd is the one we support (sarnold beat me to it ;)). That installation error isn't related to vsftpd at all but to you having force installed a broken package earlier on [21:06] jcastro: I can't think of anything [21:06] rtyuiop: Read line 5 of your paste :) [21:06] http://askubuntu.com/questions/140246/how-do-i-resolve-unmet-dependencies should help [21:17] sarnold: ping [21:20] ok perfect [21:20] working [21:20] but can't system can't access on ftp server [21:21] have i need to create ftp list of user access ? [21:21] rtyuiop: First, please define "access system". [21:22] have i need to create a file called allow user where i have to put user can access on ftp server ? [21:23] rtyuiop: No, you dont have to do that. What do you mean by "access system"? [21:24] rtyuiop, you can setup anonymous or user restricted / pw authentication: https://help.ubuntu.com/community/FtpServer [21:24] GA bekks .. hows goes it [22:33] zul: hello :) [23:17] sarnold: sorry to nag can you have a look at the beanstalkd MIR, the openstack builds in the lab are broken because of it [23:18] zul: sure :) I suspect I'll start it tomorrow afternoon, likely conclusion on friday [23:18] sarnold: k [23:32] Hello again [23:32] so... Im fighting with postfix. And during some blurry moments of enlighment, I managed to, more or less, set it up [23:33] I can properly send mail from the box, the aliases I needed work properly and get forwarded to their reciepents [23:33] but no incoming mail is forwarded [23:33] from outside the box [23:35] my domain is configured with MX 10 MX.MYDOMAIN.TLD. and MX.MYDOMAIN.TLD has an A record with my IP [23:35] from what I understand, this should work [23:36] postfix has virtual_alias_domains mydomain.tld [23:36] any ideas? [23:38] ironhalik: anything in the log messages? can you send messages to it by hand by just typing smtp? [23:39] hmm, sec [23:40] interesting [23:40] I tried to send mail to root@mydomain.tld [23:40] from my personal gmail account [23:41] since I had a catch-all on the domain, the logs say it smtp'd back to my account [23:42] nothing in mail.err [23:42] sarnold: and mail.log says it was sent [23:43] ironhalik: so it -looks- like the mail made it to the system correctly, but it just wasn't delivered correctly? o_O odd :) [23:44] now I sent mail to different domain (that seems to work from the box), and therwes no trace of it in the logs