=== freeflying_away is now known as freeflying [01:46] Can anyone point me towards a resource to put a lamp stack on a unbuntu server on ec2 [02:12] karamazov: ubuntu server guide? [02:12] jrwren: I've tried using it but I can't get apache to run... [02:13] karamazov: do you get an error? did you open port 80 in ec2 security group? [02:15] No error - does 80 need to be open for inbound or outbound? [02:17] inbound. [02:18] http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-network-security.html [02:19] jrwren: thank you [02:19] yw [03:30] hallyn_, one oddness with btrfs and lxc (using clone) that i havent figured out.. btrfs-progs won't report the subvolume or snapshot usage, looking over btrfs-progs src, the main delta seems to be quota group related wrt to subvol creation.. just curious if you've run into this before. [03:31] i'm trying to use send/receive against the vols, but without btrfs-progs being aware, thats problematic [03:39] stgraber, ^ [03:44] hazmat: nto sure what you mean by 'run into before'. i've not tried send/receive of vols. I use 'lxc-clone -B btrfs' a ton every day seems robust so far [03:45] (only caveat, in precise, is tha ti must use eatmydata to run dpkg and apt-get else it is dog-slow) [03:45] hallyn_, sure. but btrfs subvolume list /var/lib/lxc shows nothing.. with lxc-clone -s [03:48] hallyn_, i've got /var/lib/lxc setup as a btrfs mount, lxc-clone -s seems to implicitly pickup btrfs.. as the resulting container takes roughly 7m [03:50] explictly specifying -B btrfs actually results in rather odd behavior compared to the implicit, either full copy or error (http://pastebin.ubuntu.com/6435693/) [03:50] likely cause its not a subvolume [03:51] * hazmat tries with explicit btrfs on create [03:53] hallyn_, nevermind i think its user error. with dedup on the vol accounting for the minimal usage [03:55] hazmat: ok - sorry i went quiet, my laptop went berzerk [03:56] but why was precise-base not btrfs, if /var/lib/lxc is btrfs? [03:56] that sounds like an lxc-create but [03:56] bug [03:58] hallyn_, it was btrfs, just not a subvolume from lxc-create without the -B, what's also interesting is the implicit behavior with lxc-clone -s, its not the dedup, it is must be taking some form of snapshot i think as it completes in roughly the same time as a subvolume snapshot (0.09s) [04:05] but whatever form that is, doesn't get reported by btrfs-progs in saucy (currently building git head of those just to compare) [06:41] z [07:27] im trying to set up a ubuntu server for about 20 kids -- i want them to be able to ssh in, and make basic HTML pages. i have their usernames created -- how do i enable ssh for each one of them? [08:13] figuringout: if they've an account in the system, ssh should work out of the box, make sure to install openssh-server [08:49] Hi guys would anyone be able to help me troubleshoot using x11vnc to get access to a GUI on Unbuntu 12.04 from XenServer 6.2? [09:04] yolanda, well everytime you need to display that info you have todo an exec which is quite expensive on Java virtual machines [09:05] jamespage, a better alternative should be to parse the /etc/os-release file? [09:06] yolanda, I was trying to think of a way todo it in the init script and pass it in as a system property [09:06] "-Ddistribution=ubuntu" [09:06] that way its set dynamically on startup - but is then an in-memory read from that point onwards [09:07] so pass this var in the makefile? [09:21] yolanda, no - I'd look for a named system property in the Java code, and pass that in the tomcat7 init script [09:22] if it does not exist, then revert to current behaviour [09:24] well, we have the os.name but it should populate "Linux", setting that to Ubuntu could affect behaviour, right? [09:27] no - use a new property [09:28] yolanda - its trivial to set them on startup in the init script [09:28] other things might rely on os.name [09:28] and os is really owned by the JVM so you won't be able to change it anyway [09:28] ok, reading about that because i didn't know this about java, i see [09:31] yolanda, I'd probably add it to JAVA_OPTS in the catalina_sh function in the init script [09:34] jamespage, cool, i'll work on it [09:34] yolanda, great -thanks! [09:34] thx for the advice [10:07] jamespage, trying same build of couchdb for saucy [10:09] jamespage, it's problem with trusty, with saucy builds fine [10:16] utlemming, just looking at the walinuxagent proposals in the sponsoring queue - do they need to go to saucy still? or is trusty sufficient? [10:17] need todo the SRU dance if saucy is required. === freeflying is now known as freeflying_away === freeflying_away is now known as freeflying [12:27] jamespage, trying the tomcat patch. I pass to it properly: /usr/lib/jvm/default-java/bin/java -Djava.util.logging.config.file=/var/lib/tomcat7/conf/logging.properties -Djava.util.logging.manager=org.apache.juli.ClassLoaderLogManager -Djava.awt.headless=true -Xmx128m -XX:+UseConcMarkSweepGC -Dos.distribution_name=Ubuntu -Djava.endorsed.dirs=/usr/share/tomcat7/endorsed -classpath /usr/share/tomcat7/bin/bootstrap.jar:/usr/share/tomcat7/bin/tomcat-jul [12:27] i.jar -Dcatalina.base=/var/lib/tomcat7 -Dcatalina.home=/usr/share/tomcat7 -Djava.io.tmpdir=/tmp/tomcat7-tomcat7-tmp org.apache.catalina.startup.Bootstrap start [12:27] but then using System.getProperty("os.distribution_name") in code returns null [12:27] am I doing something wrong there? [12:27] yolanda, you can't use os prefix [12:27] its reserved I think [12:27] catalina should be ok? [12:27] or just use something not prefixed? [12:28] I'd do that - not prefixed [12:28] distribution.name [12:28] ok, let me try this [12:28] that's java-ish [12:29] annoying thing of tomcat is that it takes ages to rebuild [12:31] jamespage, also dealing with couchdb that's written on erlang. Do you know if os::getenv("DISTRIBUTION") should work, if passing that as env var in makefile? [12:31] yolanda, it depends when that call is made [12:31] yolanda, if its runtime then that won't work [12:31] if its build time that should be OK [12:38] yolanda: jamespage: is ^ your discussion any chance related to tomcat daemon not running in the trusty server images? and making the smoke failures? [12:39] psivaa, its not - but that sounds important as well! [12:39] jamespage: ok, then i'll report a bug for that [12:42] psivaa, please do [12:51] jamespage, build time === psivaa is now known as psivaa-lunch === gary_poster|away is now known as gary_poster === Yeti is now known as Yeti_ === Yeti_ is now known as yeti_ === yeti_ is now known as Yeti_Sno === Yeti is now known as Yeti_ === Yeti_ is now known as Yeti_Sno === psivaa-lunch is now known as psivaa [14:07] so jamespage, something is wrong with our approach for Tomcat:-Ddistribution.name=Ubuntu - but then System.getProperty("distribution.name") returns null [14:10] i think that this var needs to be defined in catalina.sh script [14:11] hazmat: ok, looking at the code, I have a vague recollection this was by design. If you say 'lxc-create -B best' then it will use btrfs if detected, but if you don't specify -B at all, it uses dir. [14:12] hazmat: note we *used* to always autodetect. I'd have to check mail archives to remember why it was changed [14:12] (Of course -B btrfs works too, but -B best lets you fall back to dir if you're not sure whether you'll be on btrfs) [14:30] jamespage: the reason for tomcat server issues in the smoke is http://bazaar.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/ubuntu.trusty/revision/2183 according to cjwatson [14:31] psivaa, hmm - I thought doko was holding off on doing that [14:33] jamespage: do you still need a bug? and if yes against what pkg? [14:33] tomat7 please [14:34] with a c tho [14:34] ack :) [14:41] jamespage: bug #1252290 [14:41] Launchpad bug 1252290 in tomcat7 "tomcat7 is disabled in the trusty server seeds starting from 20131115.1" [Undecided,New] https://launchpad.net/bugs/1252290 [14:41] bug 1252290 === Yeti_Sno_ is now known as Yetis [15:08] * jamespage waves at roaksoax [15:09] jamespage: \0/ [15:09] that was energetic [15:09] ok, so merging kombu and by default is using librabbitmq for amqp instead of python-amqp [15:09] so is this something we want to carry/make default [15:10] roaksoax, unless there is a good reason not to we should upstream align [15:11] jamespage: yeah, just making sure we are not gonna affect openstack for some reason [15:11] jamespage: i do think however, its use with kombu doesn't have ssl support [15:31] jamespage: well that sucks [15:34] jamespage: spent an hour with a broken nova testsuite on trusty (needed a newer migrate) [15:38] jamespage: https://code.launchpad.net/~zulcss/nova/nova-trusty/+merge/195620 [16:28] rbasak, zul, roaksoax, adam_g, smoser, yolanda: my find for the start of this cycle is dh-automake [16:29] jamespage: yeah you told me :) [16:29] zul, I like to go on about stuff I like [16:29] jamespage: hehe [16:29] jamespage: doc pointer? :) [16:29] zul, re the MP above - trying to understand the impact of it [16:30] jamespage: dropping of xcp? [16:30] roaksoax, apt-get install dh-autoreconf && man dh_autoreconf [16:30] roaksoax, http://manpages.ubuntu.com/manpages/trusty/en/man1/dh_autoreconf.1.html [16:30] jamespage: looks handy. Thanks! [16:31] * rbasak gave up with dh-make years ago [16:31] rbasak, save the manual 1000's of line of autoreconf patch that is sometimes required (I used this for openvswitch) [16:31] jamespage: yeah i meant more in the way of howto's or transition docs :) [16:31] jamespage: I use dh-autoreconf for that [16:31] but yeah I saw dh_autoreconf efore and I think at least one package i touch used it [16:31] jamespage: I'm sure you must have sponsored my use of dh_autoreconf before :) [16:32] rbasak, yeah - sorry - I mean dh_autoreconf - not dh-automake [16:32] doh! [16:32] rbasak, probably but that is details.... [16:32] :-) [16:32] :) [16:33] roaksoax, better guide - http://manpages.ubuntu.com/manpages/trusty/en/man7/dh-autoreconf.7.html [16:33] jamespage: i've seen it been used before [16:38] jacalvo, hey - zul is working on poking samba @ 4.x into trusty, but its making zentyal uninstallable right now which is blocking migration to the release pocket [16:42] jacalvo, are you guys still actively maintaining package in distro? if so could you take a look and see [16:43] jamespage, mmm, the problem is that the zentyal packages in universe are still using samba 3 [16:44] and the newer versions of zentyal that use samba4, are using a custom build of samba4 with bundled libraries... [16:44] so I'm not sure which is the proper solution here [16:45] jacalvo, so the samba4 package will be disappearing - samba is/has been bumped to 4.0.something already [16:46] jacalvo, we need to unblock the migration - is this something you can work out in the short term? or should we remove zentyal from archive until you come up with a fix [16:46] ? [16:46] jamespage, let me discuss this with the zentyal devs, I'll get back to you with news as soon as possible ;) [16:47] jacalvo, thanks - much appreciated! [17:11] jamespage, so, the removal would be only from the trusty archive, right? [17:11] jacalvo, yes - thats correct [17:11] and we can re-introduce later once you have stuff sorted [17:11] I think it's the safest option, we can guarantee a fix in the short term, because currently we depend not only on a custom samba4 package, we also have patched versions of bind9 and openldap :( [17:12] s/we can/we can't/ [17:13] first thing to do would be to try to introduce those patches upstream and make it work with the non-bundled version of samba [17:13] jacalvo, agreed - OK - I'll request removal [17:13] thanks!, and sorry for the inconvenience [17:14] jamespage, anyway, maybe removing zentyal-samba is enough [17:14] jacalvo, np - its all just part of the dev cycle :-) === Jikan is now known as Jikai [17:18] Hi folks, any CPU stressing or benchmarking tool that you could recommend? :) [17:26] zul, re the xcp removal stuff [17:26] I think running XCP on Ubuntu is not worth holding onto [17:26] but running against remote XenServer probably is [17:27] but thats the same model as VMware - for which we need a no-op hypervisor package for if that makes sense? [17:28] jamespage: right but it doesnt get tested and we dont really know what state the package is in [17:28] zul, does it rely on the xcp packages? [17:29] jamespage: it does [17:31] jamespage: im more risk adverse for ltses [17:31] Anyone know of a Growl server that doesn't require x11? I just want a central location to forward/receive all my notifications to/from and the box in question is headless. [17:31] zul, OK - I don't think we need todo anything explicit anyway [17:36] jamespage https://code.launchpad.net/~yolanda.robla/ubuntu/trusty/tomcat7/add_distribution_static/+merge/195641 [17:36] finally got it, i needed to take some fresh air and see it from another point of view [17:37] so i was testing the output locally with lynx, then i could not see the version in the footer, but with a normal browser i was able to test it [17:44] yolanda, \o/ [17:53] zerick: compiling the biggest packages you can get your hands on is a decent way to stress a machine. it'll involve disk IO, memory IO, CPU churning.. I've found a lot of machine check exceptions on hardware an hour or two into a large compile -- on hardware I thought was problem-free for a few years. [17:58] hello all. iwant to learn development of office purpose applications for ubuntu [17:58] such that they shud work in lan [17:58] well, you are suppost to change your thermalpaste and remove the dust every few years :) [17:58] which language shud i choose? [17:59] for developing? [18:00] patdk-wk: I never noticed any thermal correlation, and I didn't want to get my hands goopy :) [18:00] heh, you don't do it enough, if you get it on your hands :) [18:02] patdk-wk: indeed. thanks to you and jeffpc, I spent the weekend looking at used thumpers on ebay and trying my best to justify buying myself such a beast... luckily for me, it looks like they require 220 AC input, which is enough hassle to keep me from "testing it out" :) [18:03] I just got my home rack fully functional [18:04] harsh: depending upon what you're trying to accomplish, that might be progamming a web application in ruby on rails or django or tomcat (eww java) -- or it might be programming "native" desktop clients using Qt or GTK with e.g. QML or Python or C++ or C or whatever you like. There's plenty of choices, maybe too many. :) [18:04] have 10gbe and 5x1gbe going to it, and two 240v@30amp power outlets [18:05] patdk-wk: ooh, 10gbe *drool* [18:05] ya, looking at going 40gbe at the datacenter [18:05] pricing of 10gbe vs 40gbe really isn't an issue [18:05] seems driver support is more an issue [18:06] * patdk-wk wants sub .2ms network :) [18:06] or was that sub .2us [18:07] atleast I believe dropping network latency by tens of times, will make nfs work even better :) [18:07] patdk-wk: what kind of system do you use for your nas? [18:07] sarnold: ok [18:08] dual e5649, with 20x1gig disks and 4 ssd cache drives [18:08] patdk-wk: I looked int othe backblaze stuff a little bit but was shocked they used 5-port expanders. I'm sure it's a great tradeoff for them, but seems less good for a smaller, hotter system [18:09] well, they have no data turnover [18:09] mmm, hexacore :) [18:09] everything will go a max of 200MB/sec for them [18:09] right. most of their systems are going to be idle most of the time, as far as I can guess. [18:09] well, not idle, but processing like 10mbit/sec :) [18:09] and 8 gigs of ram feels like penny pinching :) [18:10] heh, not really [18:10] my enterprise windows backup software, will only use 4gigs of ram [18:10] I have installed 32gigs of ram, cause I couldn't actually but the system with less [18:10] granted on their scale, pennies need to be pinched. but it doesn't feel like much left over after the buffer cache eats most of it [18:10] the buffer cache is pointless for them [18:11] they won't be rereading data [18:11] patdk-wk: well, sure, but that's windows land, where 3.5 gigs is the limit. hehe. [18:11] normally for them, data comes in, data gets stored [18:11] the only reading that should be happening is fs metadata [18:11] and that should be pretty small [18:12] again anothr decision that works out okay for them.. :) [18:12] patdk-wk: did you build your machine from parts? or buy a pre-made or mostly-pre-made system from somewhere? :) [18:13] both [18:13] currently, I'm going premade case + mb [18:13] and the rest custom [18:13] having a hard time to pick what I want for my new esx servers [18:13] 2u server, or a twin 2u server [18:14] the twins lack ram slots :( [18:14] :( [18:14] but otherwise that seems impressive density [18:14] and power savings [18:14] and still gives me 3 slots [18:14] a 2u per server, while optimal, is more slots/disks/... than needed for esxi [18:15] but great for standalone workhorses [18:16] patdk-wk: sorry to interrup ..hve been following the discussion..2u? [18:16] 2 units [18:17] http://en.wikipedia.org/wiki/Rack_unit [18:17] patdk-wk: ok [18:18] patdk-wk: thx for the link..wat is the overall capacity of the NAS u have built [18:18] just 9tb at work, only using 3tb [18:18] at home, running 13tb usable [18:18] ok [18:19] patdk-wk: i have worked for 2 yrs in NAS support [18:19] other one is made more for pure speed [18:19] it's running 6tb [18:19] ok [18:19] so u make these NAS urself [18:19] 6tb full of 300g disks in 3way mirrors [18:19] ok [18:20] is EMC market leader in Eu as well [18:20] or its a different situation [18:21] and do u do the NAS provisioning also [18:21] for ur client [18:21] for my client? [18:21] ok [18:21] patdk-wk: mdraid or zfs? [18:21] hey RoyK :) [18:21] i thought u r hosting NAS boxes [18:21] roy, all zfs [18:22] harsh, yes, for *myself* [18:22] patdk-wk: thats gr8 [18:24] patdk-wk: zfs on linux? [18:24] defently not [18:25] hehe [18:25] bsd [18:26] fbsd or illumos? [18:26] illumos, my own custom builds [18:28] patdk-wk: did u modify the OS? [18:28] for ur use [18:29] the kernel, yes [18:29] ok [18:29] patdk-wk: i have to learn it how shud i go about it [18:29] how shud i start === matanya_ is now known as matanya [18:29] learn what? [18:29] learn to modify distros [18:30] heh? [18:30] that has nothing to remotely do with a distro [18:30] and hacking a kernel has nthing to do with a distro [18:30] ok [18:30] ok [18:33] harsh: I recommend reading Advanced Programming in the Unix Environment for everyone who wants to learn more about the internals. APUE is a fantastic introduction to the services provided by Unix-like operating systems. It would be difficult to learn kernel internals without knowing the interfaces exposed to userspace programs, and it is wonderful to know what is available easily vs what your programming environment -pretends- [18:34] sarnold: thx senior [18:40] sarnold: also today i dwnlded the pdf for ubuntu server [18:40] while configuring the server [18:40] do i have to work on DM-server as well [18:40] or i can skip it for now [18:41] I am trying to installl a local server [18:41] harsh: what's the DM-server? (keeping in mind that I almost never actually see our installer..) [18:41] ok [18:41] ok [18:42] * patdk-wk installs a lot, and dunno what dm-server is [18:45] patdk-wk: sarnold ok [18:45] patdk-wk: sarnold: ok [18:46] ok thanks to both [18:46] of u === marcoceppi_ is now known as marcoceppi [19:54] so i have a bash script thats run in xinetd on a port, its a simple echo scripts of heads, if i curl localhost:9999 is returns data, but if i curl it from outside, it just errors with no reply from host or connection reset by peer - here is the bash script, and xinetd script http://pastebin.com/Cyj3mayb - anyone have any insight? === gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster === Jikai is now known as Jikan [20:45] I set up an ftp server but when I try to ftp in I get connection refused [20:45] 21 is open [20:45] how do I define users [20:47] jParkton: depending on which ftpd you installed there are varying approaches, but generally you define users by giving people a unix account. [20:47] you mention "21 is open" ... what do you mean by that? [20:47] my ftp port [20:47] connection refused is generally "nothing is listening on this port, try again" [20:47] yes, I know what port 21 is. what I'm saying is what do you mean by it's "open" [20:48] if you're getting connection refused, chances are it's not "open" [20:48] my definition of "open" is "something is listening on it" [20:52] root 20444 18235 0 15:51 pts/2 00:00:00 grep --color=auto ftp [20:52] it is running but I dont see it in netstat -a [20:52] that's not an ftpd [20:52] that's grep. [20:53] what ftpd did you install, and how? [20:53] that is not an ftpd? [20:53] no, that's grep [20:53] 'ps aux | grep ftp' [20:53] hm [20:53] is presumably what you did [20:54] well, grep shows up in the process list [20:54] so that wont tell me if the process is running? [20:54] it will [20:54] but it will also tell you about grep [20:54] :) [20:54] so that did not show me the ftp process was running? [20:54] jkitchen: Did you install an ftp daemon, or are you assuming that Ubuntu came with one? [20:54] jParkton: ^ [20:55] I did apt-get install vsftpd [20:55] jParkton: no, all that was was grep. there may have been other output, but I'm gonna assume that was the only line [20:55] and then I edited the vsftpd.conf [20:55] jParkton: service vsftpd status [20:55] waiting [20:55] it's not running. try starting it: [20:56] but I have restarted it twice [20:56] service vsftpd start [20:56] then check your logs, chances are your config is broken [20:56] *or* maybe vsftpd has /etc/default/vsftpd and disables startup in there [20:56] (which is likely) [20:56] I don't have a scratch box handy or I'd install it to try [20:57] nothing in default [20:58] I'd check the logs first. [20:58] nothing in logs [20:59] well nothing in /var/log/vsftpd.log [21:57] in saucy is gssd a new service that is required for nfs clients? [21:57] or is it only if your using kerby? [22:00] hmm maybe my issue is gssd refuses to start [22:01] because its disabled in /etc/default/nfs [22:02] so maybe saucy just wants it on? === Jikan is now known as Jikai === Jikai is now known as Jikan === Jikan is now known as Jikai === freeflying is now known as freeflying_away === freeflying_away is now known as freeflying === Jikai is now known as Jikan === freeflying is now known as freeflying_away