[02:41] <sarnold> arrrghhhAWAY: sorry, I don't know why the do-release-update thing isn't prompting for 14.04.1 upgrades from 12.04 yet, it's pretty far out of my experience
[03:40] <pmatulis> sarnold: ?
[03:43] <sarnold> pmatulis: ah, I had'nt read enough scrollback to realize that Nivex had found the core of the problem :)  see 13:57 here: http://irclogs.ubuntu.com/2014/07/27/%23ubuntu-server.html
[03:45] <pmatulis> sarnold: ok.  i thought things might have been postponed due to a nasty bug i saw a few days ago
[03:46] <sarnold> pmatulis: eek
[03:46] <pmatulis> (desktops not upgrading properly)
[03:48] <pmatulis> sarnold: https://bugs.launchpad.net/ubuntu/+source/ubuntu-release-upgrader/+bug/1347964
[03:51] <sarnold> pmatulis: wow. nice catch. that looks like reason enough to hold off recommending 14.04.1 upgrades :)
[04:29] <arrrghhh> pmatulis, sarnold just read back.  thx.  that would seem to only affect the -desktop version tho?
[04:30] <arrrghhh> or is it a kernel issue.  maybe I'm misreading this haha
[04:30] <sarnold> arrrghhh: yes and no; server vs desktop is just default package selections, there's not that much different between them
[04:31] <arrrghhh> ya but the issue seems to be related to the meta-package "ubuntu-desktop"
[04:31] <arrrghhh> which ofc the -server ed does not use/include
[04:32] <sarnold> right, but the same mechanism is used to upgrade them both :)
[04:32] <arrrghhh> ah.  hm.
[05:23] <MavKen> trying to send out mail using swiftmailer via amazon ses..
[05:23] <MavKen> Connection could not be established with host email-smtp.us-west-2.amazonaws.com [Unable to find the socket transport "SSL" - did you forget to enable it when you configured PHP? #32600]"
[05:23] <MavKen> ubuntu 14.04   openssl installed
[05:24] <MavKen> any ideas?
[07:30] <socketguru> hi
[07:30] <socketguru> what should I use for auto starting applications on lxde on ubuntu-server 14.04
[07:31] <socketguru> lxde has its own auto start script
[07:31] <socketguru> But I have to open 2 split windows on screen at startup
[07:31] <socketguru> So I can display 2 urls on same screen
[07:31] <socketguru> Does anyone have any idea about this
[07:31] <cfhowlett> !server | socketguru
[07:32] <socketguru> I tried for firefox plugin but could but could not find
[07:32] <socketguru> ok
[07:32] <socketguru> I was expecting some response from a real user
[07:34] <cfhowlett> socketguru you asked a server question.  experts are in #ubuntu-server.  expect to get ragged for installing a gui on a server.  signed /realperson
[07:56] <GTRsdk> can someone please tell me that default mysql doesn't store data in /var
[08:02] <Abhijit> GTRsdk, #mysql
[08:07] <GTRsdk> Abhijit, but this is on Ubuntu
[08:07] <GTRsdk> not RHEL, Gentoo, etc
[08:07] <GTRsdk> or even Windoze
[08:08] <GTRsdk> okay so it turns out that it wsa stored there :(
[08:08] <GTRsdk> now any idea on what would happen
[08:08] <GTRsdk> if I moved imprtant data to /tmp/ and rebooted?
[08:11] <ikonia> depends where /tmp is mounted from
[08:12] <GTRsdk> ikonia, I believe it is just tmpfs, however I do not see it in /etc/fstab
[08:13] <ikonia> so check it
[08:13] <ikonia> it may not even be mounted, it may just be a directory
[08:14] <ikonia> check where /tmp is mount/presented "from"
[08:14] <GTRsdk> A guy moved /var -> /tmp/var, rebooted, and now it is gone
[08:14] <ikonia> check where /tmp is mount/presented "from"
[08:15] <ikonia> chances are it's either tmpfs or disk that's set to clear down on boot as /tmp is normally transient data
[08:15] <ikonia> so it's gone
[08:17] <GTRsdk> bummer :(
[08:18] <GTRsdk> oh to check though wasn't there like /etc/mounts or something?
[08:18] <samba35> can you please tell me how do i restart network service in 14.04.1?
[08:19] <GTRsdk> samba35, service networking restart ?
[08:19] <Abhijit> samba35, service networking restart
[08:19] <Abhijit> :-z
[08:19] <samba35> yes ,its not working
[08:19] <Abhijit> samba35, what error it gives?
[08:20] <samba35> stop: Job failed while stopping
[08:20] <samba35> start: Job is already running: networking but some time out put is null but change are not effective
[08:21] <GTRsdk> ikonia, strangely it doesn't show up in my /proc/mounts
[08:36] <ikonia> GTRsdk: what is that strange
[08:36] <ikonia> why
[08:37] <GTRsdk> ikonia, I was expecting it to be either there or /etc/fstab
[08:37] <ikonia> why ?
[08:40] <GTRsdk> ikonia, I have no idea where else it would be
[08:40] <GTRsdk> But I can't exactly check anymore... just powered off the system
[08:40] <ikonia> GTRsdk: yeah bit late now after your messing around in other channels cross-posting as you don't want to say "I don't know"
[08:40] <ikonia> good luck
[08:40] <GTRsdk> I literally just said I have no idea where else it would be
[08:41] <ikonia> GTRsdk: no you didn't
[08:41] <GTRsdk> I am 90% sure it is tmpfs
[08:41] <ikonia> you just said it after I called you out in another channel for trying to dodge the question as you didn't know the answer
[08:41] <ikonia> and you are %0 certain
[11:44] <sorvin> Hi all. Anyone has any experience with preseeding? Whatever I try, I can't seem to generate a file that yields a completely automated installation process. It always pauses for hostname, for example. I tried using the preseed-example template from ubuntu docs, and it doesn't really work well. Any thoughts? thanks!
[11:58] <AtuM> sorvin, I have "some" experience.. http://linux.opm.si/programska-oprema/ubuntu-14-04-network-install  -> I've spent a lot of time testing different parameters.. mine is not a completely automated, but you might find some clues for your needs
[11:59] <AtuM> soren, this is only to give you some insight into where you might want to put parameters for install...
[12:01] <sorvin> thanks a lot!
[13:07] <Nivex> arrrghhh: https://bugs.launchpad.net/ubuntu/+source/update-notifier/+bug/1344762
[13:48] <Kunzem> Good day.
[14:05] <Kunzem1983> Good day Ubuntu gurus. I'm new to Ubuntu server and have stumbled on a wierd problem which i'm not finding help from google and don't know quite how to explain. i used nano to edit ssh_config to check if password authentication is set on yes. then used ctrl x to exit . I now have a > sign to type in commands but it doens't seem to accept any commands. if i enter a command it lists all the commands i tried ? how do i get back to entering commands
[14:06] <Nivex> Try Ctrl-D
[14:07] <Nivex> sounds like something got quoted oddly and that should break you back out to bash with a "Unexpected EOF"
[14:07] <Nivex> (except in this case we expect it since we just typed it :)
[14:09] <Kunzem1983> thanks i tried this on my virtual box with ubuntu server and it got me back login prompt
[14:09] <Kunzem1983> will give it a go on live server quick
[14:10] <Kunzem1983> Thanks Nivex , just what i needed :)
[15:31] <Nivex> bug was updated. The ambiguous "soon" :/
[15:31] <Nivex> so much for planning
[15:40] <weeb1e> I am trying to install cacti on the latest ubuntu server LTS, but everytime I use apt-get, it installs apache2 as well. If I remove apache2, it removes cacti. How can I install cacti without apache2?
[15:51] <jrwren_> that is called a dependency.
[15:51] <jrwren_> apparently, the cacti package depends on apache2
[15:52] <jrwren_> hrm, but its only listed as a recommends, not as a dependency: apache2 | lighttpd | nginx | httpd
[15:52] <jrwren_> when you run apt-get install, use the --no-install-recommends option?
[15:52] <ogra_> weeb1e, apt-cache show cacti|grep Recommends
[15:52] <weeb1e> jrwren_: It is a bug, which was supposed to be fixed 5 years ago, https://bugs.launchpad.net/ubuntu/+source/cacti/+bug/217591
[15:52] <weeb1e> Clearly it is not fixed
[15:52] <ogra_> there are other alternatives you can use but it seems to want a webserver
[15:53] <weeb1e> So the easiest solution would be to leave apache2 installed, but disable it from ever starting
[15:53] <ogra_> (or you can disable recommends altogether during install)(
[15:53] <ogra_> like jrwren_ said already
[15:53] <weeb1e> Oh I missed that, sorry
[15:53] <weeb1e> I did not know about "recommends"
[15:53] <weeb1e> I will try that now
[15:54] <jrwren_> weeb1e: oh yeah, there it is. an unfixed bug.  Be sure to click the "does this bug affect you?"
[15:54] <genii> weeb1e: apt-get -o APT::Install-Recommends=0        ...is the syntax
[15:54] <weeb1e> No wonder I didn't know about that, what an unintuitive syntax :P
[15:55] <jrwren_> its in the man page as --no-install-recommends which should do the same thing, but as I've never used that option, I cannot say if the manpage is correct.
[15:56] <genii> jrwren_: I usually use the syntax I gave above, this way you modify any variable which is reported by apt-config dump
[15:58] <ogra_> it is correct :)
[15:58]  * ogra_ uses --no-install-recommends quite often
[15:59] <ogra_> (and thinks like the ubuntu phones totally depend on it...)
[16:00] <lordievader> Good evening.
[16:01] <jrwren_> weeb1e: the requires is libapache2-mod-php5 | php5
[16:01] <jrwren_> that means OR
[16:01] <jrwren_> so if you install php5 package, that fulfils that dep, then it will not try to install libapache2-mod-php5
[16:01] <jrwren_> so... not a bug, but you have to know how to work around it.
[16:05] <weeb1e> Sorry, I'm going to try use that shortly
[16:05] <weeb1e> I just found out that my house got broken into today :(
[16:06] <weeb1e> --no-install-recommends does not appear to work
[16:07] <weeb1e> Neither does "apt-get -o APT::Install-Recommends=0 install cacti cacti-spine"
[16:08] <weeb1e> genii: Any other ideas?
[16:08] <weeb1e> Before I just disable apache and leave it installed
[16:08] <weeb1e> Oh, again I missed what jrwren_ said. I'll try install php5 first
[16:09]  * genii makes more coffee
[16:09] <weeb1e> Installing php5 installs apache2
[16:09] <weeb1e> So no, that is not an option either
[16:09] <weeb1e> It feels like I'm going round in circles now
[16:09] <weeb1e> I'm just going to find a way of disabling apache2 and be done with it
[16:11] <jrwren_> no wait! :)
[16:12] <jrwren_> what webserver are you using?
[16:12] <weeb1e> I've installed and removed apache2 and cacti more than 10 times now, I have a ton to install and cannot continue to waste time like this
[16:12] <weeb1e> jrwren_: Nginx
[16:13] <jrwren_> so apt-get install php5-fpm php5 cacti
[16:13] <jrwren_> that should meet the deps you need.
[16:14] <jrwren_> I did apt-cache show php5, to see what it Depends:, and the first listed is libapache2-php which is of course going to then pull apache, but its an or expression with a few ways to meet the dep.
[16:14] <jrwren_> you'll likely be using fpm?
[16:15] <weeb1e> Yeah I will need fpm installed too
[16:15] <weeb1e> I will try that, one last attempt
[16:17] <weeb1e> jrwren_: Now the apache2 package is not installed, but /usr/sbin/apache2 exists
[16:17] <weeb1e> Strange
[16:18] <weeb1e> I guess that is better than having the package installed, thanks
[16:21] <arrrghhh> Nivex, what's interesting about that bug you linked... is I got the same message, on a VM I built for testing 12.04.4 --> 14.04.1
[16:21] <arrrghhh> but I don't get that message on my physical "production" server if you will.
[16:24] <jrwren_> weeb1e: dpkg -S /usr/sbin/apache2 should tell you which package put that binary there.
[16:24] <jrwren_> weeb1e: apache-bin package. you can likely apt-get remove that.
[16:27] <weeb1e> Alright, yeah that does seem to be the case
[16:27] <weeb1e> I have not installed apache2 on a box in like 7 years, so all this is new to me
[16:28] <weeb1e> Which now seems like an accomplishment, considering how much stuff depends on it
[16:48] <omps> Hi
[16:50] <jrwren_> weeb1e: its too bad alt dep management isn't a bit easier, but i guess its a tiny learning curve
[17:18] <arrrghhh> Nivex, ok there's some issue which I'm not quite understanding.  My net new install 12.04.4 came with 3.11 kernel so it gave me the HWE message.  My physical 'production' server has always been using the 3.2 series kernel, which is not running HWE at all evidently.
[17:20] <Nivex> Yep. If you installed from .0, you have to add HWE manually. If you installed from .1 or later you get the HWE that was current at the time of release.
[17:21] <Nivex> You still have to manually walk it up from wherever you started. My server is running 3.8 (lts-raring)
[17:21] <arrrghhh> ok
[17:22] <arrrghhh> ya I haven't really touched the kernel on my server, so it's still on 3.2
[17:23] <Nivex> I guess we're right back to "I hate waiting."
[17:24] <arrrghhh> lol.  yep!
[20:59] <smoser> hallyn, would this be expected to work:
[20:59] <smoser>  lxc-create -t ubuntu-cloud --name=sm1 -- --release=trusty --arch=amd64
[21:00] <smoser>  http://paste.ubuntu.com/7888316/
[21:05] <hallyn> smoser: I don't remember whether ubuntu-cloud was properly converted in the end or not.  I thought it had been
[21:05] <smoser> hallyn, download doest work either there.
[21:06] <hallyn> smoser: you need to create your .config/lxc/default.conf
[21:06] <hallyn> ubuntu-cloud works fine for me unprivileged
[21:06] <hallyn> you need your lxc.id_map entries
[21:07] <smoser> right. i see that now.
[21:07] <smoser>  https://www.stgraber.org/2014/01/17/lxc-1-0-unprivileged-containers/
[21:07] <smoser> a DTRT util would be nice ther.
[21:07] <hallyn> I've thought about writing a shorter sipler howto
[21:07] <smoser> and a reference to such from the error.
[21:07] <smoser> :)
[21:07] <hallyn> well if we try to automate it too much we'll encourage unsafe reuse of idmaps
[21:08] <hallyn> i automated it that way for lxc-usernsexec bc there it doesnt' matter (it's a toy)
[21:08] <hallyn> but i don't want lxc-create using defaults if you didn't create any
[21:08] <hallyn> now maybe skel can do it?
[21:08] <hallyn> so when user is created, and homedir is filled in, it creates yoru lxc.conf according to your new subuid?
[21:08] <hallyn> eh, still too general
[21:09] <hallyn> i'm open to ideas
[21:09] <hallyn> perhaps a good thing to discuss at the sprint in august
[21:09] <stgraber> I haven't spent time trying to make this simpler because realistically we can't do it without any user interaction in a safe way
[21:10] <smoser> i dont know what is unsafe
[21:10] <hallyn> "it depends" :)
[21:10] <stgraber> I'm not too concerned about pulling the uid/gid ranges from /etc/sub{uid|gid}, that part is probably fine for 99.9% of our users, the concern is about giving veth quota by default to users
[21:10] <hallyn> which is why we haven't automated it
[21:10] <smoser> shadow added my user to /etc/sub[ug]id
[21:11] <smoser> ok. that seems quite rasonable.
[21:11] <stgraber> we could however have LXC detect those cases and spit out an helpful error message
[21:12] <hallyn> right, shadow gives you a arange - but we dont' know how you should use that range
[21:12] <hallyn> our error message is helpful IMO
[21:12] <stgraber> as we already did with a few of the other usual unprivileged containers problems (missing permissions and such)
[21:12] <hallyn> http://paste.ubuntu.com/7888316/
[21:13] <stgraber> hallyn: "Missing uid/gid map, to run a container as an unprivileged user you need to have lxc.id_map set for the container. Your user currently owns the following ranges: ... so you can set xyz in ~/.config/lxc/lxc.conf. Read more in lxc.container.conf" would be more useful :)
[21:13] <stgraber> hallyn: so you can do lxc-start, it fails, you edit .config/lxc/default.conf, copy/paste, done
[21:14] <stgraber> (and yeah, s/lxc.conf/default.conf/ above)
[21:14] <stgraber> with the suggested config being something like (but with the right ranges pulled from /etc):
[21:14] <stgraber> lxc.include = /etc/lxc/default.conf
[21:14] <stgraber> lxc.id_map = u 0 100000 65536
[21:14] <stgraber> lxc.id_map = g 0 100000 65536
[21:15] <stgraber> and we should also detect the missing lxc-usernet allocation case and give a better hint than "Quota exhausted" or whatever we return currently ;)
[21:19] <smoser> what logic was used to add default user 'ubuntu' to /etc/lxc/lxc-usernet ?
[21:19] <hallyn> that's not done automatically
[21:20] <hallyn> stgraber: I'll ack a patch liek that if you can get someone to write it :)  for today, i've finally got a usable list of subvolumes->parents so I can now impelment the actual recursive subvolume deletion
[21:21] <hallyn> the secret?  well, silly, sh->offset is actually the object id of the parent of a subvolume.
[21:41] <f-a-h-a-d> hi guys, for testing reasons i needed to deploy ubuntu maas+juju+Openstack on a network of some virtual and some real servers ... What is the best way to begin with?
[21:44] <sarnold> f-a-h-a-d: I've used this as a _guide_ when doing maas testing entirely in a virtual environment: http://bazaar.launchpad.net/~smoser/maas/maas-pkg-test/view/head:/maas-ephemeral-test-precise.txt -- it might also be a useful guide for you, but be aware that mixing virtual with physical may add more complications
[21:45] <f-a-h-a-d> sarnold: what complications?
[21:47] <sarnold> f-a-h-a-d: you'd have to understand the ethernet bridging a good deal better than I do :)
[21:51] <f-a-h-a-d> sarnold: i doubt that :)  but thanks bro
[21:51] <sarnold> f-a-h-a-d: hehe :) have fun