[00:37] so I made a mistake. I tried to do an upgrade while using an ssh connection. the connection quit, now when I try to do the upgrade on the system itself, it says that the administration directory can't be unlocked and another process might be using it. what should I do>? [00:39] MasterZuFu: what kind of upgrade is it? do-release-upgrade or similar? [00:41] do-release-upgrade [00:42] check if there's a screen process running, it usually starts one up if you do the upgrade over ssh [00:45] how do you do that bradm? [00:52] MasterZuFu: sudo screen -list [00:53] MasterZuFu: and then probably sudo screen -d -r if there is something [00:53] hello guys, why /var/log/messages does not exist in ubuntu 12.04? [00:56] adam_g: still around? https://code.launchpad.net/~zulcss/cinder/cinder-fix-deps-testr/+merge/170724 [00:57] zul, ya [00:58] zul, you already added that earlier [00:58] r112 [00:58] adam_g: oops.. [00:58] adam_g: nm go enjoy the rest of your day ;) [00:59] zul, also, a lot of our havana jenkins builds were blocked from building b/c python-pbr because a pre-requisite of building tarball with sdist. that should be fixed now [00:59] adam_g: ok cool [00:59] because == became [00:59] * zul goes wonder off [01:00] cya [01:28] Anyone who knows how to set in /etc/network/interfaces that i want to receive a IPv6 prefix using DHCPv6 Prefix Delegation? [01:35] hey halvors1, no responses yet.. [01:43] bradm, so I let it finish it's upgrade, it rebooted...now the entire server is corrupted. i can't login, i'm getting crazy characters and boxes, and nothing is displaying properly....what do I do? :( [01:46] to install a lamp stack, is it best to use tasksel, or install the lamp meta package? [01:47] or more importantly is there no difference, and tasksel is merely installing said meta package? [01:47] any person of awesomeness able to assist with this? how exactly can I reinstall ubuntu-desktop? That's what looks all funky [01:47] MasterZuFu: try running dpkg-reconfigure -a [01:47] MasterZuFu: burn the ISO to a USB drive and boot into it?? [01:47] MasterZuFu: maybe you're interrupted upgrade process left a pile of packages in half-installed state... [01:47] O.o that's no fun [01:48] is there any way to not have ubuntu-desktop load up upon boot so that I can get to the command line? [01:49] MasterZuFu: yes ^_^ [01:49] MasterZuFu: echo manual >> /etc/init/lightdm.override [01:49] MasterZuFu: details here: http://upstart.ubuntu.com/cookbook/#override-files [01:50] cancel my Q, i rtfm :) [01:50] dv81: ah, what's the answer? :) [01:51] according to https://help.ubuntu.com/community/Tasksel, its just installing the meta packages you already see in such as aptitude :) [01:51] aha :) === dv81 is now known as wo0f [01:53] apparently theyre define here: /usr/share/tasksel/ubuntu-tasks.desc === RobbyF_ is now known as RobbyF === wo0f is now known as wo0fPC [03:39] does 'nload' draw a graph? [03:51] sarnold, you asked me if my 10.04 install was relying on bsdgroups mount option. I confess I don't really know what you mean. [04:40] fluvvell: bsdgroups is a way to propogate group ownership of files and directories -- if a directory is setgid to a group, and the filesystem is mounted with bsdgroups, all files created in that directory are group owned by the directory group -- and directories are created with group owner _and_ setgid, to make sure the subdirs also get it... [04:48] sarnold: isn't that just the normal way setgid directories work? [04:50] bradm: hrm, seems I've confused bsdgroups and sysvgroups :) [04:51] bradm: I thought the setgid only worked with bsdgroup mount option.. [04:51] sarnold: nope, definately not [04:52] sarnold: I think bsdgroup means any newly created files / directories takes the group id of the directory it is in, but with sysvgroups it needs the setgid bit to do that [04:53] bradm: thanks man :) I've assumed the wrong thing for like a decade. never paid enough attention to the bit in the manpage... [04:54] I've never needed to bother with bsdgroups, I just usually do the setgid bit if I need that. [04:55] fluvvell: so :) there we go -- did you use bsdgroups _or_ did you have the setgid bit set on your directories? :) [04:59] sounds like something to make people migrating from bsd happy [04:59] *nod* [05:09] tmux Q: why is ctrl-b-c not creating a new window [05:09] using putty [05:10] wo0f: is caps-lock on? [05:10] NOPE [05:11] caps* [05:11] :) [05:11] wo0f: does ctrl-b ? show c as being bound to new-window? [05:12] as in 'ctrl-b (shift)?' ? [05:12] wo0f: yes. [05:12] wo0f: that shows key bindings [05:12] bradm: showing nothing :( [05:13] does ctrl-b anything work? [05:13] doesnt seem to [05:13] maybe its remapped [05:13] mine is ctrl-a instead [05:13] check ~/.tmux.conf to see if there's any remapping been done [05:15] no tmux.conf in /root/ [05:15] need to not be root? :S [05:16] .tmux.conf - note the leading . [05:16] not shown with ls-a [05:16] ls -a [05:17] not sure then [05:17] maybe its in the global config [05:20] roaksoax, was thinkin about the context rendering + exceptions earlier and not sure the solution i had in mind would work 100%. gonna work on finishing off that cinder one tomorrow. hopefully figure somethin out then. [05:22] bradm: oh, there is no default config, and have to build your own?- exciting! [05:22] depends if you like those garish colors :) [05:23] wo0f: well, there should be some kind of basic config [05:24] bradm: you would deffo have thought [05:25] surely it doesnt matter that it was installed as a dependancy of byobu? [05:26] I wouldn't think so [05:28] OMG [05:28] PANIC OVER [05:29] thanks for help sarnold bradm [05:29] oh? what fixed it? [05:29] wo0f: what was it? :) [05:29] while googling tfm, i came across: "I hope you are not just pressing Ctrl+A. Basically, you have to press Ctrl+A and then release keys, then press "?" " [05:29] ohh... [05:30] *noob face* [05:30] aha [05:31] sure enough, smashing all of ^BC at once doesn't work, but ^B? did work. [05:31] screen/byobu has been too lenient with me ;P === marlinc is now known as marlinc|away [06:00] sarnold: what does the "^" represent ? [06:00] wo0f: control === smb` is now known as smb [06:53] odd one I haven't seen before. host returns two lines; the first is a valid answer, the second is NXDOMAIN. any suggestions why it might give conflicting answers to the same query? === marlinc|away is now known as marlinc [07:27] I'm reading about how to optimize my apache settings for a ridiculous amount of concurrent users, and this one article said "The default amount of "maximum" number of threads is usually enough; however, there can be conditions where this number may have to be increased. The threads usage can be monitored." I've really tried to figure out how they can be monitored but I think I'm asking google the wrong questions. Any ideas? [07:27] nilli: server-status is the magic phrase, that'll show you what apache is doing. [07:28] is that an apache module? [07:30] yeah, mod_status [07:30] thanks :) [07:55] Hello. I installed awstats to my Ubuntu Server. I followed the instructions on the Ubuntu how-to page but I am getting errors emailed each time crontab is processed. I "su root" and then "crontab -e" and put the awstats update command for each domain in the root crontab. What may have went wrong? [07:56] The error I am getting in the log file is that it could not open server log. [08:01] guys, is this new ? passwd: Authentication token manipulation error [08:01] I try to change a password of a user [08:02] Yamakasi_ That means when you re-typed the password, you didn't type it the same the 2nd time. [08:03] riz0n: sure thing ? [08:03] joe@HMCS-Server:~$ sudo passwd joe [08:03] Enter new UNIX password: [08:03] Retype new UNIX password: [08:03] Sorry, passwords do not match [08:03] passwd: Authentication token manipulation error [08:03] passwd: password unchanged [08:03] riz0n: no, I don't get the message it's not typed right [08:04] riz0n: Sorry, passwords do not match is what I don't get [08:05] Check out http://askubuntu.com/questions/91188/authentication-token-manipulation-error and see if any of the situations outlined on that page may be causing you the issue. [08:06] riz0n: will reboot that machine first [08:13] riz0n: no Solution there [08:14] so i would sudo passwd -d username [08:14] then passwd to set a new one [08:16] riz0n: did that too, didn't work [08:16] never had this [08:23] can you change passwords on other accounts? [08:25] riz0n: for root the same issue... and the rights on the files are OK [08:29] riz0n: dunno as I never, really never seen this [08:29] Sounds like filesystem is only getting mounted read-only [08:30] but I don't know a whole lot about what could be going on. [08:30] riz0n: yes it looks like it, but tge strange thing is that I can just do an update [08:31] and just did [08:46] when i log into my server with ssh, it warns that my partitions will be checked next boot. how can i stop that? [08:57] Ben64: look at the tune2fs manpage. You can examine the mount count, max mount count before check, last check time, max time before check etc. options and change them. [08:58] but will changing that unmark it for fscking on boot? [08:58] I think so but am not sure. [08:59] Look out for the two ways an automatic fsck is triggered: mount counts, and time-based. [08:59] ok [08:59] i'm scared of my server not being responsive for that long [08:59] its like 8000mi away, i can't get to it to fix it either [08:59] Then you should test and experiment on a different machine first. [08:59] good idea [09:03] riz0n: copied a new 12.04 image and that is working OK, soing a release upgrade now [09:22] good morning everyone, i have a tiny huge issue with my apache2 [09:23] as in: it does not serve any pages, neither static nor dynamic [09:26] Drag0nflamez, start with the obvious. Is is started, is it listening on port 80, is there any error in the logs ? [09:27] it listens on port 80, if i do tail -f on the logs i see this error [09:27] moment pls [09:28] http://pastebin.com/nRc5qDS8 [09:28] Which is a strange error because it didn´t happen yesterday [09:28] and in the meanwhile the configs weren´t changed [09:28] hmm, something is already listening to port 80 [09:28] already tried reloading the apache2 service, destroying my precious uptime by doing sudo reboot [09:29] Drag0nflamez: lsof -i :80 [09:29] Drag0nflamez: get that output [09:29] or sudo fuser -n tcp 80 [09:29] doesnt output anything [09:29] also nothing [09:29] Drag0nflamez: you are starting Apache as root right? [09:29] ye [09:30] Drag0nflamez: $ sudo netstat -ltnp | grep 80 [09:30] also nothing [09:30] i am puzzled... [09:30] Weird. [09:31] if you telnet on port 80 , do you get a prompt of some sort ? [09:31] Drag0nflamez: And if that returns nothing: sudo apache2ctl restart [09:31] same error but now only for ipv6 [09:32] but v4 still doesnt serve anything [09:34] also why is it trying to bind to 0.0.0.0? [09:34] which is a bogus IP... [09:34] it s listening to _any_ adresses [09:35] i know that (explicitly told to listen to *:80) [09:42] so what could be causing this... [09:42] hello there [09:42] Hello. [09:43] anyone tried to configure proxy dot pac on ubuntu server ? [09:43] to get have access with apt over internet [09:43] ? [09:44] I'm not sure apt supports that, so I could be mistaken. But a proxy.pac will generally only specify a single proxy for access to an external apt repository, so you can just set Acquire::http::Proxy to that. [09:46] how ? [09:56] Hi [09:57] #virtualbox [09:57] I'm looking for a web hosting solution, I'm currently using bluehost, which is not bad, do you have any other suggestion, or should I just go again for bluehost ? [09:57] maxagaz: not really something for #ubuntu-server [09:58] ^ [10:17] Daviey: I think the impact of that is a bit beyond my experience of ovs [10:17] I'll ping the upstream ML for guidance [10:19] jamespage: Yeah, i started trying to hack it in based on similar projects solving the same problem.. but i wasn't convinced i was doing a good job, so droppe dit [10:19] thanks [10:22] ah, fixed my issue [10:22] Apparently 2 conflicting listen directives in sites-availalbe [10:22] available* [10:22] which should have been in ports.conf [10:23] Thank you Apache2 sites examples [10:23] anyways, goodbye. [10:32] Daviey, OK _ upstream emailed for guidance [10:33] Daviey, it might be a big job to enable the 802.1ad feature in ovs, but a smaller job to make it compatible with the 3.10 kernel if that makes sense. [10:36] Daviey, zul, adam_g: http://people.canonical.com/~jamespage/ca/havana/ - specifically openvswitch [10:36] Yeah, i don't think we need 802.1ad for anything this release.. so just making the thing work without it, might be suitable [10:36] if we can compat it [10:38] Daviey, I think we can [10:39] but it won't be accepted upstream I think [10:40] jamespage: it would be enough for the short-term. As of next week, we could be in a pickle [10:41] Daviey, OK - I'll look at a hack compat for then [10:41] jamespage: that sounds super, thanks [10:48] I have a bunch of custom upstart jobs. Is there an easy way to have a command/job that I can start, stop or restart to start, stop or restart them all? [11:09] jamespage: +1 [11:24] Pupeno: you can create a "dummy" job that can just be stopped or started manually. upstart will keep track of its state but otherwise do nothing. [11:24] Pupeno: then your bunch of custom jobs can each have the changing status of this job in their start and stop conditions. [11:24] Pupeno: so you can do it, but whether or not that's a good idea depends on your situation [11:37] zul, ta - uploaded [11:38] jamespage: with luck i should have qemu ready today [11:38] Daviey, once openvswitch is built do you want to sync havana-staging -> havana-proposed? [11:38] or should we wait for zul's qemu as well [12:09] jamespage: makes sense to do them together i guess. [12:21] jamespage: the openvswitch is tested right? [12:22] because i need to test the qemu stuff after it builds [13:01] lifeless: yes please, a bug is probably best === railsraider_ is now known as railsraider [13:06] /win 8 [13:07] roaksoax: dude... [13:08] zul: what? [13:08] lol [13:08] roaksoax: hehe [13:08] ?/ [13:08] time to switch irc clients ;) [13:09] zul: lol... nah... I don't like xchat like you do [13:09] :P [13:11] jdstrand: hey, just wondering, have you looked at all at the libvirt fine grained access control for libvirt APIs thread? [13:12] no, sorry [13:12] (I haven't really, just noticed it today. it has polkit and selinux drivers - well, will have. sounds like we may want to add apparmor at some point) [13:12] ok [13:12] for now i'll add it to my list of apparmor stuff i need to look at for libvirt [13:12] cool, thanks [13:13] it's not something I have a lot of time for atm, but will add it to my todo list to review at some point [13:13] cool, ttyl [13:14] hallyn: spice question for you [13:14] zul, yes - I ran the smoke tests we have for dep-8 on 12.04 as well [13:14] jamespage: cool just checking === klaas_ is now known as klaas [13:15] hallyn: libspice-server-dev depends on libspice-protocol-dev (>= 0.12.0~) but its not built at all (libspaice-protocol-dev) [13:15] zul, only problem I have noticed is that the upgrade from 1.4.0 -> 1.10.0 is not working [13:15] whats not working? [13:16] zul, the 1.10.0 restart hangs trying to get the version from the old running userspace daemons [13:16] jamespage: oh [13:17] zul: not built at all? [13:17] hallyn: never mind i read the changelog [13:17] ok === wedgwood_away is now known as wedgwood [13:40] hi all [13:41] so I want to install wordpress on sub.domain.com but I want it to be used through www.domain/wp/ [13:41] can this be done and if so how? [13:42] i am on ubuntu 12 + nginx [13:43] yousaf_: google nginx rewriting dowmains [13:43] ok but do you get what i am looking for? [13:43] its not just a simple redirect [13:44] i don't know if "virtual path" is the right word for this but that is what i need [13:44] yousaf_: yes i do [13:45] yousaf_: there are plenty of examples on the net that tell you how to do it [13:45] this http://forum.nginx.org/read.php?11,2531 ? [13:46] yousaf_: sure === Jikan is now known as Jikai === Jikai is now known as Jikan [14:11] I have ubuntu server 12 I think on one computer, and I talk from this main ubuntu computer, I use ssh on the command line to talk to use commands onthe server, I want to add a small file, I am pretty much sortof beginner with command line, i use it some, but what of a text editor that I can dump text into, so it comes out as a file over the other end ? [14:12] So now i use ssh and can do some things through the network, but what is a simple editor i could use? [14:13] nano [14:18] thank you [14:18] smb: so you're discussing bug 1191782 on the m-l? do you want to assign it to yourself? [14:19] apparently i missed someone having nginx issues? [14:19] hallyn, Hm, yeah. Could have done that already [14:19] * TheLordOfTime shrugs === Jikan is now known as Jikai [14:32] does the apache2 directive Timeout affect slow file uploads? if it takes a user 65 seconds to upload a file and Timeout is set to 60, will the upload fail? [14:35] Hey guys - is there a way I can apt-get install mysql and set the root password in the same line ( or via bash script ) ? [14:35] nvr mind: http://stackoverflow.com/questions/7739645/install-mysql-on-ubuntu-without-password-prompt :) === marlinc is now known as marlinc|away [16:01] adam_g: howdy! Why do you think the Exceptions wont be enough? [16:01] adam_g: i keep getting this erro: http://pastebin.ubuntu.com/5787162/ [16:02] with this: http://paste.ubuntu.com/5787165/ [16:04] marcoceppi: ping [16:04] marcoceppi: any ideas? http://paste.ubuntu.com/5787130/ [16:04] (the trace at the end) [16:04] * marcoceppi looks === marlinc|away is now known as marlinc === marlinc is now known as marlinc|away [16:14] roaksoax: it looks like either database_host isn't being sent to the templating, or there's an error in the template itself causing it to bail === marlinc|away is now known as Marlinc [16:15] roaksoax: I can take a deeper look in a little later, but I think adam_g or jamespage would be faster at helping [16:17] marcoceppi: cool thanks for the input. I was thinking somehwere in between those lines too [16:19] roaksoax, can I see the full branch please? [16:21] jamespage: lp:~andreserl/charms/precise/glance/port [16:34] jamespage, checking that branch now. [16:34] adam_g, yeah - me to [16:35] something wonky is going on with the hookenv serializable stuff [16:35] yeah [16:35] the context is not empty though [16:35] roaksoax, the exceptions wont work because for every write() multiple contexts can raise an incomplete exception, when we're only interested in a single exception. eg, context.amqp() would be incomplete on a shared-db write() if no amqp relation exists [16:37] adam_g: I see, yeah definitely that would be a problem [16:39] adam_g: that render_configs() would have to be smarter then and only render the config its being asked to render [16:40] that way the Exceptions would work [16:40] without mayor change [16:40] to other than how we do the CONFIGS.write() [16:43] roaksoax, adam_g: hmm - the {% if database_host %} is calling __getattr__ with the key __nonzero__ [16:43] that feels like a bug in Serializable to me [16:43] adding that function to the class fixes the problem [16:44] jamespage, to which? [16:44] Serializalble [16:44] " def __nonzero__(self): [16:44] return self.data is not None [16:44] " [16:46] hmm. i haven't seen this but i havne't been using the @cached stuff yet [16:47] adam_g, not sure that 'cached' is the issue - the underlying function still returns a serializable [16:48] adam_g, I think its the way that jinja2 is interrogating the object [16:48] Serializable(UserDict.IterableUserDict) [16:49] but for a simple, scoped relation_get its not a dict, its just a value. [16:52] i set my ubuntu installation to static but for some reason it is not able to access the internet === Marlinc is now known as Marlinc|away [17:02] why is it saying invalid syntax(21)??? [17:02] http://ircshare.com/i/bp [17:03] I am on the process of configuring an LDAP server on a ubuntu server through the CLI - >https://help.ubuntu.com/10.04/serverguide/openldap-server.html [17:03] help pls [17:06] when replacing this olcRootDN: cn=admin,dc=example,dc=com with my own configuration for a domain as procopio.co.uk should I add olcRootDN: cn=admin,dc=procopio,dc=co,dc=uk ???? [17:07] hello friends === Marlinc|away is now known as marlinc [17:09] I'm not sure what kind of conclusion I can draw up here, but I figured I'd ask as I'm grasping for ideas. I have Ubuntu Server 12.04.2 running. I was having a truckload of issues with my server recently, particularly when I'd rsync a lot of data to it. Last night it went into read only mode and reboots did nothing. I swiched the SATA3 cable for the SSD (OS drive) and it boots up fine. But what's weird is my monitor just goes black during boot. [17:09] I see no login screen. [17:09] I can SSH into it fine and I'm not seeing any errors, but uh... any ideas? I just found it weird the monitor went black with 0 login screen @ CLI... [17:13] help pls [17:15] hello Everyone! [17:17] i'm a newbie to the world of managing servers [17:17] jamespage/adam_g: fun times [17:17] jamespage/adam_g: http://people.canonical.com/~chucks/ca/ [17:19] i find myself managing 5 servers (4 Ubuntu Server 12.04 LTS + 1 ARM Debian). I would wondering if any of you have any recommendations for a server-monitoring program/platform? === marlinc is now known as marlinc|away [17:20] vDevices: nagios === marlinc|away is now known as marlinc === goddard is now known as Guest49563 [17:37] zul, those look okay. one nit, 'New upstream release for the Ubuntu Cloud Archive' is a bit misleading. they're just no-change rebuilds [17:39] adam_g: what do you suggest? [17:40] zul, "New update for..." its not that big a deal. i've been modifying it to "New security update..." when its a security update rebuild [17:41] adam_g: okies i can do it for the next round [17:49] jamespage, these ceph-osd log errors look familiar to you? http://paste.ubuntu.com/5787475 [17:49] adam_g, thats a new one for me [17:50] zul, note that cloud-archive-backport has a --changelog option [17:50] jamespage/adam_g: should i re-do it? [17:51] zul, nah, just a nit [17:51] I'd not block on that basis [17:51] adam_g: ok so +1? [17:51] i'd be surprised if anyone other than us look at these changelogs anyway :) [17:51] zul, ya [17:51] cool [18:01] :D [18:01] being super lazy [18:01] really have 0 time to spare [18:01] the thought of getting up is something I cannot handle [18:31] When you restart networkign with the init script, you get "Running /etc/init.d/networking restart is deprecated because it may not enable again some interfaces" [18:31] what is the non-deprecated way? [18:46] jkyle: sudo start/stop/restart servicename [18:46] Alternately: sudo service servicename start/stop/etc [18:48] genii: but will that Do The Right Thing for the other interfaces? [18:48] sarnold: Hopefully! [18:49] If the /etc/network/interfaces is set up correctly don't see why it would be a problem [18:50] genii: some networking services are provided elsewhere; lxc-net, for instance.. [18:52] * genii sips and ponders [18:53] yeah. I'd like to think it would all just work :) but not knowing makes me wonder.. [18:56] hello [18:58] adam_g: the other option would be to have 1 exception per type of context... though that would probably be non-generic and many exception errors [19:00] roaksoax, yea. i was thinking of adding an attribute to the class(s) that can be checked to see which contexts have been completely satisifed [19:00] roaksoax, so the calling hook can do something like: if 'shared-db' not in CONFIGS.context_complete: return else: CONFIGS.write(...) [19:01] roaksoax, im gonna have a go at that this afternoon. still bangin on ceph atm [19:02] adam_g: ok cool! === roasted_ is now known as roasted === marlinc is now known as marlinc|away === marlinc|away is now known as marlinc === bean__ is now known as bean [20:36] the restricted repos contain closed-source stuff right? [20:41] Yes, but it's limited to things like firmware and drivers to get hardware working. [20:45] this is more of a java question, but how do i run "ulimit -c unlimited" at startup, before all other things? [20:45] I've got a crazy question :^) I'm trying to get an Ubuntu 12 server downgraded to an Ubuntu 10 server (app requirement) remotely. I have console access via IPMI, but no PXE environment setup on that network. Any ideas how to get the net-install to boot from Ubuntu 12's GRUB menu? I've tried copying the initrd and kernel image and creating a menu entry for it.. but it doesn't seem to come up. No error, just a blank [20:45] screen after the Grub menu [20:47] LargePrime: can you stuff that into whatever shell script starts your application? [20:48] Can I add it to that user? there is an application user defined [20:49] do i need any rights to modify ulimit [20:49] LargePrime: raising limits does require superuser powers.. [20:49] LargePrime: see also /etc/security/limits.conf [20:52] i have a user 1003. can i just add that user to the file? [20:52] and do i really want unlimited? [20:55] LargePrime: since I can never remember the measurement units, I tend to use unlimited :) hehe [20:55] LargePrime: .. and you'd have to lose a core file because it was that little bit too big.. [20:58] so i can just add the line "1003 soft core unlimited" to /etc/security/limits.conf and save sarnold ? how do i get that loaded? [20:59] do i have to restart the server? [20:59] LargePrime: you might have to use the username [20:59] cant use number? [20:59] ok [21:00] dunno, I don't see any examples in the file :) [21:00] LargePrime: the limits are applied when the user logs in through the PAM stack, any service that has pam_limits.so in the stack will enforce them.. [21:00] how do i know if my user has a pam stack [21:00] and this use does not "login" [21:01] it is an application [21:01] and Wehn i say "does not" i mean " i think" [21:01] LargePrime: hrm, then maybe editing an upstart config file? [21:01] LargePrime: how does it start? [21:02] witha a /multicraft -v start [21:04] LargePrime: what runs the command? cron? upstart at boot? [21:05] upstart at boot. but there is weirdness. this start is one user, but the application has 10 other users it has to keep parts of it secure to themselves [21:06] so i guess i need all these users to have a raised ulimit [21:06] I guess I should contact the app dev [21:07] LargePrime: here we go: http://upstart.ubuntu.com/cookbook/#limit [21:07] LargePrime: I assume the first process runs as root, so it can start the other processes as other users? [21:08] adam_g: so I was looking at the context for keystone, and in glance we have 'auth_uri', which is not in the context. Is there a way to easily add extra var names to the context? [21:08] genii: yeah, that' doesn't work for networking. I just scripted a ifdown -a && ifup -a [21:09] jkyle: Good to know for future. [21:10] maybe start/stop might work. no restart....I'll test the start stop after this pxe run, but I think it has the same prob as the init.d [21:10] roaksoax, that doens't come from the relation, its made up of values from the relation: (from the bash version) set_or_update "auth_uri" "http://$keystone_host:$service_port/" [21:11] adam_g: yeah, but we still need in the context [21:11] roaksoax, you dont necessarily, just http://{{ keystone_Host }}:{{ service_port }} in the template [21:12] adam_g: ahh yeah you are right [21:12] * roaksoax needs coffee a [21:12] roaksoax, if you wanted to inject it into the context you could, after the check to context_complete() but its not necessary IMHO [21:12] adam_g: yeah I see your point. it is indeed not necessary [21:13] * genii slides roaksoax a fresh mug of coffee just right for drinking [21:14] genii: lol thanks! [21:15] roaksoax: np [22:04] anyone setup a static ip with comcast business class? [22:33] adam_g: the other option would be to return True/False from write() [22:34] to see whether it was written successfully or not [22:34] roaksoax, well, write() is inspecting any number of contexts. i'd expect write() to return true in shared-db hook even if its amqp context is incomplete [22:35]