[00:39] hi all, sorry to bother here but I'm unable to find the answer anywhere else: i'm looking for a clean way to close the session (logout) from the command line. gnome-session-quit wont work, nor the dbus method shown here: http://askubuntu.com/questions/15795/how-can-you-log-out-via-the-terminal. Also, I don't want to just 'sudo reboot' but a nicier, more polite method so all running application can [00:39] safely end it's ongoing stuff and then end. Any idea? === Lcawte is now known as Lcawte|Away [00:44] msx: terminal applications should do something sane when they receive a SIGHUP signal [00:44] msx: since HUP doesn't necessarily make sense to guis, they are less likely to do something sane when receiving it. [00:45] I've never tried logging out while applications are 'open' -- can linux apps pause or stop a gui-initiated logout or shutdown process? no idea... [00:46] sarnold: ahh, I didn't kno GUI apps didn't follow SIGHUP :P good to know! [00:47] msx: one hopes they either handle it or block it, but a great many programmers go out of their way to avoid learning about signals altogether. (I can't say I blame em. :) [00:48] but, i.e. with Unity, when you close a session and there are GUI apps active (like a web browser) they will cleanly shutdown, if you just quit the X server they will most likely crash. So me thinks that there should be a clean, polite way to ask the apps to close and then quit the X session - I could be totally wrong oc! [00:48] sarnold: humm, okay, wasn't aware of that, thank you very much for the heads-up :) [00:50] msx: hrm, that dbus answer looked promising [00:50] sarnold: already checked it out but won't work on 14.04 :S [00:51] sarnold: at first i think it didn't work because i called the method from within tmux but I have the same error from a plain bash shell [00:53] msx: hey, I think I found something worth trying [00:53] sarnold: weee! Tell me! [00:53] :D [00:53] msx: install the d-feet package, run d-feet, and poke aroud in the org.freedesktop.login1 namespace [00:54] roger that, on my way [00:54] msx: there's an /org/freedesktop/login1/seat/seat0 path in there, with a Terminate() method [00:54] great, sounds promising [00:54] hi guys i have installed vsftpd on my ubuntu server and add user but when i access by ftp client i get login incorrect...i sepcified all teh details on the user i added already- any idea guys..? [00:54] msx: and under the org.freedesktop.login1.Session interface, there's a Kill() method.. wonder what that does :) [00:55] sarnold: haha :D [00:55] msx: I hope that does the trick :) [00:56] i'm on that, i'll keep you informed! [00:56] thanks :D [01:04] sarnold: thank YOU! However this will take a while as I recently started to learn dbus once and for all... as recently as 6 minutos ago =_= [01:04] msx: hehe, similarly, I've done my best to ignore dbus, too. hahahaha. :) [01:04] and indeed this D-Feet browser is very useful [01:05] yeah, dbus was 100% baffling before someone pointed me to d-feet. now it's just 70% baffling. :) [01:05] xD well, seems we all do the same thing until the tides finally catch us, lol [01:05] next up: systemd [01:07] sarnold: well, i have already a foot on it, i've been using arch for a while - and fedora at work - until i returned to my first love (yep, ubuntu) [01:07] sarnold: you will eaither hate it or love it, but for sure there's no space for any other feeling between ;) [01:08] s/eaither/either/g [01:09] msx: makes sense. I've wanted a real service manager for ages and cgroups feels less ghetto than ptrace, but i'm not loving how it consumes everything else.. ntp? network configuration? sigh :) [01:12] sarnold: that's my *exact* same gripe on systemd: it started as a much needed upgrade for the PID 1 but since then it has been taking over so much that in a beginning was out of its scope. I don't know if this is a real thing or not but someone told me Poettering want an OS fully based on systemd as its core, most fundamental part and 'sane' API... [01:12] msx: hehe, I don't know his intentions, but it sure has grown significantly.. [01:12] *wants [01:13] absolutely [04:37] Hi, is this a good place to ask for help about UVTool? [04:38] radbasa: yeah, but might be a bit late (or early, depending upon your timezone :) -- anyway, can't hurt to try [04:47] hi all, can anyone suggest regarding my problem... I have installed ubuntu server 14.04 and want to run systemd for init services...is systemd stable enough for this? I was using archlinux before and systemd was great.. but now we have to use production server which is ubuntu [04:49] socketguru: http://www.piware.de/2014/04/booting-ubuntu-with-systemd-now-in-utopic/ [04:50] thanks :) [04:53] so you saying systemd will work without any problem? [04:53] no :) [04:53] has anyone here used systemd on ubuntu server? [04:53] sarnold: no ? [04:53] it looks like it'd be difficult to get it to go on 14.04; and on the devel release, might be decent enough to try [04:54] but for production server, we have to give a product which should be stable ...So, what is safe way? [04:54] stick with upstart [04:54] Do you think upstart is a safe choice? [04:54] yeah... okay [04:55] systemd may be ready for 14.10 or 15.04, but not yet [04:55] the thing is i haven't used upstart.. and I dont know if upstart will smooth journalctl log service [04:55] it might be fine for you to play with, but I wouldn't want to run a business on it yet [04:55] thanks a lot sarnold [04:56] I have successfully create a VM using uvt-kvm with the default usermode networking. I am trying to create a VM with bridged networking. … wait a minute… sarnold, in the process of collecting everything (command-line commands and results, config files, and log files) for posting at pastebin, I discovered that the bridged network VM is up and working, but uvt-kvm wait and uvt-kvm ip isn't seeing it. http://pastebin.com/ZkRKDFta [04:58] radbasa: yikes, that's way beyond my experiences with it :/ sorry [04:59] it's ok. maybe i'll just post it at launchpad [04:59] talking about network [05:00] let's say you are using vagrant using virtualbox as a provider, and into the vm, you have some lxc containers with private ip addresses [05:01] but i want to access from the guest directly to the linux container private address (which is into the vm) [05:04] sebas5384: I think you've got two options; (a) set up the guest to do IP forwarding and run it as a router/gateway (b) use brctl to add an ethernet bridge in the guest, bind it to the NIC, and bind the lxc containers' IPs to the bridge. [05:04] sebas5384: I don't know either approach well enough to actually do it myself :) but I hope it gives you some reading.. [05:05] time to bail :) good luck [05:05] sarnold: thanks man! [05:05] sarnold: really helped :) [05:10] i used brctl to create an ethernet bridge on the host OS and bound the guest OS to it [05:40] radbasa: thanks [05:41] radbasa: did you sow it in some article? i'm lookin for something for guiding me [05:59] sebas5384, I followed step three of this http://xmodulo.com/2014/02/use-kvm-command-line-debian-ubuntu.html [05:59] radbasa: awesome!! === taikat_ is now known as taikat [07:24] I have a KVM host with 7 Windows VM's running inside ... the reachability of the VM's inside using ping is pretty unstable whereas to the host via either hostname or IP address is fine. The Host is configured with a bond and a bridge ... any ideas what could be causing this behavior? [08:12] zetheroo1: ubuntu 14.04? [08:12] yes [08:13] zetheroo1: check the conversations on https://lists.ubuntu.com/archives/ubuntu-server/2014-July/thread.html [08:13] zetheroo1: disabling KSM might help you [08:13] or upgrading to a newer kernel. 3.13 which comes with 14.04 hase some troubles [08:15] Host is completely upgraded [08:17] zetheroo1: check the archives on ubuntu-server maillinglist. url above [08:17] there are some serious problems with 3.13 kernel [08:17] try installing 3.15 [08:19] this is a productive system ... I cannot be shutting it off and on again or trying to install kernels etc ... [08:19] :( [08:21] 10 to 15% packetloss on production systems is also bad [08:21] henkjan: Then I'm lucky I'm not affected. :) [08:22] zetheroo1: https://lists.ubuntu.com/archives/ubuntu-server/2014-July/006940.html [08:23] I have to get the VM's migrated then [09:46] what is the downside of disabling KSM? [09:48] zetheroo1: google points me to http://www.linux-kvm.org/page/KSM, http://en.wikipedia.org/wiki/Kernel_SamePage_Merging_%28KSM%29 and https://www.kernel.org/doc/Documentation/vm/ksm.txt [09:48] zetheroo1: so I don't know, but I guess it just costs you some more memory. === Lcawte|Away is now known as Lcawte [11:01] smoser: fixed up https://code.launchpad.net/~xnox/cloud-init/fix-systemd-install-paths/+merge/227918 to be truly generic [11:02] smoser: and all tests pass now in the modernised debian packaging https://code.launchpad.net/~xnox/cloud-init/fix-packaging/+merge/227931 [11:39] I started getting this recently on one ubuntu system (12.04 LTS) [via the daily cron email] "run-parts: /etc/cron.daily/mlocate exited with return code 1". There is nothing else (about a lock), and there is no lock. There are no stale NFS shares or out of space mounts. I've even purged the DB and re-created it. Also, if I run it manually, it works without any problems. I can't figure out what's causing this for the life of me. Anyone run into anyth [11:41] I normally just uninstall mlocate [11:42] see no point in letting it abuse my server every night [11:42] if you don't use the locate command, that would be a fine thing to do [11:42] It's convenient for when you want to lookup a file, but then again, I guess you grind the entire system for the benefit of not grinding it once when you actually need it [11:42] that's the thing - i do use it every once in a while [11:43] what's strange is I don't see anything about a lock in the cron email. And what's even more strange, it works when I run it by hand. I did a dump of the process table at the time it runs by the cron.daily, and there is nothing else running at the same time/blocking it [12:20] xnox, thanks. i'll list [12:20] list == look [12:28] hi - anyone used to apt-cacher-ng? I am running a virtual server here which acts as apt-proxy for all our ubuntu-servers. this works great so far. now i am realizing that apt-cacher-ng might be as well able to serve other distributions. is that true? experiences are welcome [13:09] fidel_: I use apt-cacher-ng. I like it. And without any new config, it handles debian too. [13:09] fidel_: I haven't tried it with openSUSE, CentOS, etc. [13:10] apt-cacher-ng brings the install of the cluster down to under 15 min :) [13:15] hi peetaur2 [13:16] so you are handling with 1 apt-cacher-ng install both ubuntu and debian? [13:16] I doubt apt-cacher-ng will work for yum [13:16] do i understand that right? no issues cause of simular filenames etc as it has a proper folder-structure in its cache-folder i assume [13:17] why would simular filenames matter? [13:18] well i am unsure. i could imagine it could be critical if packages of ubuntu and debian would be mixes. example: debian client gets paket MC - so its stored in the apt-cacher cache. now an ubuntu client is asking for that package aswell [13:18] fidel, please think [13:18] they exist on DIFFERENT SERVERS [13:18] ubuntu packages are not on the same servers as debian ones [13:18] and if they where, well, wouldn't be an issue [13:19] i do know - but we are talking about the apt-proxy apt-cacher-ng offers to both clients in that case [13:19] the url for each is different [13:19] yes, it STORES the hostname too [13:19] or did you think it should strip it? [13:19] as mentioned above - i am/was unsure === Mogwai is now known as Guest89336 [13:20] fidel_: yes, one Ubuntu apt-cacher-ng handles any random version of Debian or Ubuntu. [13:20] for me it looks like it crates a clean folder-structure inside its cache folder- like that it shouldnt be critical at all [13:20] *creates* [13:20] k [13:20] apt-cacher-ng uses some convention where the client tells the server which hostname to use, so it can go anywhere without server side config ... [13:21] if that convention was "http proxy" or "socks5" then it would work on any distro that uses those standards. [13:21] but if it is some debian-only thing, then nope [13:21] in my apt-cacher-ng folder, the first item is the server name [13:21] now, to help the cacher work better, it has some regex you can define, that will merge multible server names into one [13:22] like for sourceforge [13:22] aptproxy is junk and needs config. apt-cacher and -ng just work out of the box, with any urls. [13:22] the real issue wouldn't be ubuntu vs debian, but using ppa's and 3rd party sources though [13:22] any apt ones that is [13:22] ppas work fine too... [13:23] thanks to both of you [13:23] peetaur2, even if it didn't store the hostname/paths? [13:23] I have several machines using apt-cacher-ng with zfs ZoL PPA for example [13:23] peetaur2, are you familiar with 'UNetbootin'? [13:23] yes no config.... just "apt-get install apt-cacher-ng" and assuming firewall, etc. are set up, it'll work [13:23] jeffreylevesque: that hack to make usb sticks boot CDROM isos? only aware of its existence, and have not used it. [13:24] ok, thank you! [13:25] in fact, every debian and ubuntu machine I know of in the office uses my apt-cacher-ng (excluding private VMs on people's windows that I don't control), [13:25] and it's nearly flawless. apt-cacher would crash often and needed a watchdog. apt-cacher-ng works great, and only once ever was not responding and a simple restart fixed it. [13:28] so..... I tested urls like http://aptproxy:3142/startpage.com/ and tried by setting my proxy in the proxy settings, but it just gives me an error page saying to use the apt setup. [13:28] so my guess is it won't work for yum, zypper, etc. [13:28] now to try removing such filters in the config ;) [13:54] rbasak, I have raised a bug with debdiffs for merging upstream nagios-nrpe (Bug#1348142) but it's not clear to me who I should subscribe as the sponsor. [13:54] Looking [13:54] thanks [13:56] gnuoy: normally you need to subscribe ~ubuntu-sponsors. That gets it into the sponsoring queue at http://reqorts.qa.ubuntu.com/reports/sponsoring/index.html. [13:56] gnuoy: but I'll stick this on my todo to review/sponsor so no need. [13:56] rbasak, perfect, thank you [14:41] jamespage: newer pycadf is needed according to the requirements.txt for now [14:41] for nvoa [14:41] zul, +1 [14:42] zul, sqlalchemy +1 [14:43] zul, I think that general if a minor version bump is involved (rather than patch) its probably worth updating for Juno CA [14:43] jamespage: ack...i think new depencies we should agree on automatic +1 === xtriz is now known as fifo [15:47] hi [15:47] I'm trying to use winscp to login as admin then change session to root. I changed sftp serer to "sudo su -c /usr/lib/openssh/sftp-server" When I have this line in visudo everything works fine: admin ALL=(ALL) NOPASSWD: ALL [15:48] but I want to limit that NOPASSWD so I tried using admin ALL=(ALL) NOPASSWD: /usr/lib/openssh/sftp-server [15:48] and now winscp cant connect. how to solve this? [15:49] sudo isn't calling sftp-server. It's calling su. [15:49] Why are you calling su anyway? [15:50] Create a dedicated user, give it a specific group and assign the group the necessary permissions. Never use root for such tasks [15:56] what if I need to work on some files in /root (via winscp)? [15:59] Maybe you should reconsider how you have set up your workflow and the access to it [16:01] well there's nothing in /root yet, I was just looking for a general way. Thanks === mjohnson151 is now known as mjohnson15 === mjohnson151 is now known as mjohnson15 === unixoid is now known as unixoid|afk [18:01] having trouble on trusty with vnc server - can anyone confirm this as a problem: https://bugs.launchpad.net/ubuntu/+source/gnome-session/+bug/1274013 [18:01] Launchpad bug 1274013 in gnome-session "gnome-session is broken without 3d acceleration, breaking gnome-flashback on non-3D H/W and cloud environments (Forwarded-X, VNC and NX/X2GO) (dup-of: 1251281)" [Critical,Confirmed] [18:01] Launchpad bug 1251281 in gnome-session "xrdp from gnome-session-fallback shows static gray screen via remmina or vinagre" [High,Triaged] [18:01] patrick_M: You'll probably have better luck asking in #ubuntu, as ubuntu-server does not ship with a GUI. [18:02] ok, I'll try there also - but just as an fyi, this is on server [18:07] Good evening. === JanC_ is now known as JanC [19:27] hi I am looking for refind package, any one knows where I can download it from? thx [19:45] I am trying to install and configure rsnapshot. I am a bit overhwlmed by certain parts though. Any help? [20:37] do any of you guys know how to switch sessions on ubuntu server. I was doing an update and my laptop died and it got stuck mid update and now the dpkg resource is busy being used. [20:38] I wanna try connect to the session and finish the update I think its stuck at a prompt [20:44] bitbyte, I dont know how to do it after the fact, but I generally use the screen utility to run such things, since it does allow reconnecting. [21:30] I am building apparmor profiles for an ubuntu 14.04 server, I wonder if I should expect 'usr.lib.postfix.master' to pop up in future revisions of the apparmor-profiles package [21:30] hey guys im trying to rename alot fo files and they all are like “OP-41 L@mBerT” im trying to remove the L@mBerT any ideas where I can start on getting this achived theres around 800 files to renmae [21:46] The backup/restore guide for LDAP on the server guide doesn't appear to work https://help.ubuntu.com/lts/serverguide/openldap-server.html#ldap-backup [21:46] I'm failing at the first slapadd for config.ldif [21:47] slapadd: could not add entry dn="cn=config" (line=1) === solarfly_ is now known as solarfly === Lcawte is now known as Lcawte|Away [22:49] . [22:49] .² [22:54] hallyn, stgraber: have tracked down a juju lxc issue to the change of meaning of the "-c" flag on lxc-start in 0.9 to mean device instead of filename [22:54] any suggestion on the best way to support 0.8 and newer versions? [22:56] does 0.8 barf if we give it -L (or --console-log) ? [22:58] so after hours of debugging, i finally found the bug which was causing a lot of issues i was getting [22:58] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1346917 -- it looks like it was commited but how do i know if it was built with it? [22:58] Launchpad bug 1346917 in linux "Using KSM on NUMA capable machines can cause KVM guest performance and stability issues" [High,Fix committed] [22:59] or rather, how do i know if/when a kernel comes out with that updated [23:01] mnaser: when it's published, a notice will be appended to that bug with the changelog entry; subscribe to that bug for an email [23:01] sarnold: ah okay, is there any usual "timeframe" or "schedule" for kernel compiles? [23:01] yeah that was a hellofabug [23:02] thumper: that change has caused some problems... I'll (or the list will) have to think through it. Could you please post an email to lxc-devel, or open a bug? [23:03] mnaser: I think they happen every three weeks... [23:03] ah i see [23:04] phew, think i've uncovered another annoying bug in libnih-dbus-assumptions... converted cgm from a dbus-send wrapper script to a c program; now the GetTasks method on an empty cgroup claims to return an error. Sigh. [23:06] hallyn: opened this bug for juju https://bugs.launchpad.net/juju-core/+bug/1348386 [23:06] Launchpad bug 1348386 in juju-core/1.20 "lxc template fails to stop" [High,Triaged] [23:09] thumper: stuck it in my list to look at tomorrow. disappearing shortly - ttyl [23:10] hallyn: ack [23:11] (i think stgraber will be back tomorrow, he may actually remember something about this other than that others have brought it up :) [23:14] thumper: so to be clear, waht you want is what is now the "-L" option? [23:14] I think having lxc detect that the console device is not a device, and falling back to -L, would be reasonable [23:15] hallyn: what I want is really just a way to be able to determine which flags to pass [23:15] since we want to support centos, can't even use dpkg to figure out version [23:16] actually, that would be reasonable... [23:16] the fallback that is [23:17] problem is, that supporting a fallback option for lxc is a big problem, as you have many version to tweak [23:17] if juju could work it out, it would be easier for us [23:17] you say 0.8 was when this happened? [23:18] the docs for 0.8 say -c is a filename, 0.9 -c is a device [23:18] oh i think that was commit 596a818d4b8b55586d36af518b745cd96b24c67a [23:18] just looking at docs to determine this [23:18] "separate console device from console log" [23:18] sounds like that could be it [23:19] so you do have a simple way to detect then, [23:19] try "-L", if it fails, you can use -c for -L [23:19] we could [23:19] if using -L on 0.8 fails in a way we can detect [23:20] it's kinda icky, but would work [23:20] it should fail giving you usage output [23:20] or i can do the fallback for -c, but that wont' help with all the version already out there [23:22] right [23:24] I must dash - if the bad-args-detect works for you, pls comment in the bug? I'll check it tonight. [23:24] * hallyn out [23:24] kk