[00:50] <mathiaz> kirkland: still around?
[00:50] <kirkland> mathiaz: mostly
[00:51] <mathiaz> kirkland: do you some time to help in debugging the stuck eucalyptus-cloud situation I'm running into?
[00:51] <kirkland> mathiaz: sure, i can try
[00:51] <mathiaz> kirkland: I can't get the credentials and thus can't access UEC
[00:52] <mathiaz> kirkland: so it seems that everything blocks when trying to access http://127.xx:8443/register
[00:52] <kirkland> mathiaz: should be https
[00:52] <kirkland> mathiaz: can you telnet to 127.0.0.1 8443
[00:52] <kirkland> mathiaz: is it actually listening?
[00:52] <mathiaz> kirkland: it can read the ssl certificate
[00:53] <mathiaz> kirkland: hm - well - not always apparently
[00:54] <kirkland> mathiaz: hmm, cannot reliably read the ssl cert?
[00:54] <mathiaz> kirkland: hm - wait - nothing listen on 8443
[00:54] <mathiaz> kirkland: there is something on 8773
[00:55] <kirkland> mathiaz: i think that's the CC
[00:55] <kirkland> mathiaz: the CLC is on 8443
[00:55] <mathiaz> kirkland: well eucalyptus-cloud is listening on 8773
[00:55] <mathiaz> kirkland: let me reboot the CC
[00:56] <kirkland> mathiaz: hrm, we have a real problem, if all this rebooting is required
[00:57] <mathiaz> kirkland: yeah - I'd like to get a UEC setup running though
[00:57] <mathiaz> kirkland: so that I can test my stress scripts (which I'm very proud of :) )
[00:57] <kirkland> mathiaz: i haven't tested today's ISO yet today
[00:57] <kirkland> mathiaz: i've been fighting wsdl stubs patch since about 10am
[00:57] <mathiaz> kirkland: ok - so now eucalyptus-cloud is listening on 8443
[00:58] <kirkland> mathiaz: i think i'm about done with it
[00:58] <kirkland> mathiaz: can you point a web browser to it?
[00:58] <mathiaz> kirkland: http://paste.ubuntu.com/351526/
[01:00] <kirkland> mathiaz: okay, let's look at the logs
[01:00] <kirkland> mathiaz: anything interesting in /var/log/eucalyptus?
[01:02] <mathiaz> kirkland: hm - cloud-error doesn't show anything new since the reboot
[01:02] <mathiaz> kirkland: it's full of jdbc errors though
[01:02] <kirkland> mathiaz: can you pastebin it?
[01:02] <kirkland> mathiaz: those could be the problem
[01:03] <mathiaz> kirkland: http://people.canonical.com/~mathiaz/cloud-error.log
[01:04] <kirkland> mathiaz: yeah, something is definitely wrong with the DB connection
[01:21] <mathiaz> kirkland: ok - so packages reinstalled
[01:22] <mathiaz> kirkland: I was able to register the NC
[01:22] <mathiaz> kirkland: euca_conf --get-credentials mycreds.zip failed though
[01:22] <mathiaz> kirkland: http://paste.ubuntu.com/351533/
[01:27] <kirkland> mathiaz: same error on the wget?
[01:27] <mathiaz> kirkland: yes
[01:27] <mathiaz> kirkland: it times out
[01:27] <mathiaz> kirkland: https://127.0.0.1:8443/ works well
[01:28] <mathiaz> kirkland: https://127.0.0.1:8443/register fails
[01:28] <kirkland> mathiaz: oh?  you can browse to it now?
[01:28] <mathiaz> kirkland: hm - you're right. I could try to go that way
[01:30] <mathiaz> kirkland: hm well - I can get to the first page
[01:31] <mathiaz> kirkland: but then it stays there: "Loading data from server..."
[01:31] <mathiaz> kirkland: something else is broken
[01:36] <kirkland> mathiaz: i'm planning on installing a fresh UEC tomorrow
[01:36] <mathiaz> kirkland: ok - I'm going to file a bug with my current setup
[01:37] <mathiaz> kirkland: you may run into the same issue tomorrow
[01:37] <kirkland> mathiaz: please do
[01:37] <kirkland> mathiaz: i likely will
[01:37] <kirkland> mathiaz: i need to dig through the logs more thoroughly
[01:41] <mathiaz> kirkland: https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/503180
[01:42] <kirkland> mathiaz: thx
[01:48] <sabgenton> !usb 3g modem
[01:48] <sabgenton> !3g modem
[01:48] <sabgenton> !3g
[01:49] <sabgenton> !wireless internet
[01:49] <sabgenton> anyone have some good ways to set up huawei 3g modems
[01:50] <sabgenton> I used to just make  a dialer and dial it
[01:50] <sabgenton> but maybe theres a nicer way in ubuntu
[01:50] <sabgenton> ?
[01:50] <sabgenton> :)
[02:03] <godsyn> please assist: I'd like to disable all auth for samba. How? I'd like everyone to be a guest.
[02:12] <bitprophet> Is there an easier way to obtain/rebuild debs from newer releases? right now I wget the dsc/tar.gz/diff.gz from packages.ubuntu.com and then dpkg-source + dpkg-buildpackage
[02:13] <bitprophet> but it'd be nice if there was something like apt-get source <release_name> <package_name> or something. hmm...maybe if I add the newer distro as a source line in my sources.list?
[02:20] <bitprophet> excellent, I can just add all the deb-src lines for the distro in question to sources.list and it works just fine. can't believe it took me this long to figure this out.
[02:20] <bitprophet> sorry for the spam :0
[02:20] <bitprophet> :)
[02:26] <orudie> how can I safely chown and/or chmod directory within /var/www/ so that the content are listed when accessed by the browser. Right now it says 403 Forbidden.
[02:27] <j416> orudie: the web user needs access to it. Either chown it to the web user, or chmod so that the web user has at least read access.
[02:28] <orudie> j416, the chown is set to www-data.www-data
[02:28] <twb> billybigrigger: apt-get --build source
[02:29] <orudie> j416, not sure what chmod should be set to , safely :)
[02:29] <twb> billybigrigger: to select a release, you can use -t as normal.
[02:29] <j416> orudie: if your web user is www-data, then it should be fine
[02:29] <twb> billybigrigger: bear in mind that backporting packages from a newer releases is something for experts to do, and only rarely.
[02:29] <j416> orudie: the only user that needs access to those files are your web user, supposedly
[02:30] <orudie> chmod 755 will be ok ?
[02:30] <twb> Oops, all of that was meant for bitprophet, who left.
[02:30] <j416> so you should be able to set directories to 700 and files to 600
[02:31] <j416> orudie: 755 for directories is ok, yes
[02:31] <j416> you can use 755 / 644
[02:31] <j416> it all depends on what you want your users to be able to see, of course.
[02:32] <j416> 700 / 600  ->  only the owner has access
[02:32] <j416> 755 / 644  ->  everyone can read, but only the owner can both read and write
[02:32] <orudie> just did 755 still forbidden j416
[02:33] <j416> what file are you trying to access?
[02:33] <j416> and what permissions does it have?
[02:38] <twb> j416: the httpd will need read access to it, too.
[02:38] <j416> twb: tell orudie :)
[02:39] <j416> if the httpd is run as the web user as it should be, it would have read access to it, though
[02:39] <twb> Not if the files are owned:grouped by j416:j416.
[02:39] <twb> Or orudie or whatever
[02:39] <j416> of course not
[02:39] <orudie> j416, its set to chown www-data.www-data
[02:39] <j416> twb: ^
[02:40] <orudie> :)
[02:40] <orudie> what should it be set to ?
[02:40] <orudie> www-data.root ?
[02:40] <twb> orudie: www-data:www-data is fine.
[02:40] <j416> orudie: are you sure you managed to chown it properly?
[02:40] <j416> (looking at your above syntax)
[02:40] <j416> (what does "ls -l" say?)
[02:41] <twb> I'm curious as to the details of the 403, too
[02:42] <orudie> this is from apache2 error.log client denied by server configuration: /var/www/server-status j416 twb
[02:42] <orudie> oops
[02:42] <orudie> Directory index forbidden by Options directive: /var/www/directory
[02:42] <orudie> that
[02:42] <j416> that's an entirely different error then
[02:42] <j416> it means you're trying to list the contents of a directory
[02:43] <j416> but you are not allowed to.
[02:43] <j416> (because your web server is set up not to allow it)
[02:44] <zul> whoop...freeradius finally has ssl support in lucid
[02:45] <orudie> how would I fix it ?
[02:45] <orudie> :)
[02:45] <sabgenton> any one know if there is some sort of management for 3g  modems in ubuntu server
[02:46] <sabgenton> or you just make your own script?
[02:46] <j416> orudie: normally, you don't want to allow directory listing, as it is a possible security risk
[02:47] <j416> if you really want to, the setting is in line 10 of /etc/apache2/sites-enabled/000-default
[02:47] <j416> in a default installation
[02:47] <j416> (I think)
[02:48] <j416> I might have modified mine a little, but thereabouts.
[02:49] <twb> sabgenton: IIRC they just act like any other modem.
[02:49] <twb> sabgenton: it's not really server gear.
[02:49] <sabgenton> yeah they do you just make a ppp conection
[02:49] <j416> orudie: if you're looking for more support on apache, perhaps #httpd is a good place.
[02:50] <sabgenton> ubuntu desktop has this gui thing that asks you what network they are with and sets up the settings for you
[02:50] <sabgenton> twb: ah found a script for my modem
[02:50] <sabgenton> https://help.ubuntu.com/community/DialupModemHowto/Huawei/E220
[02:51] <sabgenton> they could have had a terminal tool that set up a dialer like the desktop gui
[02:52] <sabgenton> the desktop setup sucks anyway cause it  relys on you having X open all the time
[02:54] <twb> Xvfb!  Har har!  (That's not a serious suggestion.)
[03:13] <darthanubis> is there a doc or faq somewhere that lays out how to setup wifi on my server. I don't want to run a DE on my server just to get it to connect to the WAP.
[03:15] <darthanubis> https://help.ubuntu.com/9.10/serverguide/C/network-configuration.html
[03:15] <darthanubis> that page does not cover wifi w/ wpa usage
[03:17] <manish> Hello
[03:18] <manish> Is it a good idea to change over to Ubuntu server 9.10 from existing Windows 2000 server based Network?
[03:19] <manish> Existing Network has PDC, Mail Server, Active Directory, File server, etc
[03:23] <darthanubis> manish, what do you think we'd say here?
[03:25] <manish> darthanubis, :) Yes
[03:28] <manish> But what I want to know is how complicated is setting up this and what are the resources I need to look for before getting hands on ubuntu server 9.10?
[03:30] <JanC> manish: it all depends on several things, like do you want to replace some servers, all servers, also replace desktops (some? all?), etc.
[03:31] <manish> to start with I want to replace PDC
[03:31] <manish> this server keeps failing. so looking for alternatives
[03:32] <manish> workstations are mix of Win2k and WinXP SP3 on LAN
[03:34] <genii> manish: I've found the official guide to be pretty useful as a starting point https://help.ubuntu.com/9.10/serverguide/C/index.html
[03:36] <manish> ok thanks will go though it during day.
[03:39] <genii> darthanubis: http://ubuntuforums.org/showthread.php?t=1205009 might be useful to you for the wpa from commandline
[03:40] <darthanubis> genii, ty
[03:45] <sabgenton> can i get back to the friendly menu of options when I installed the cd?
[03:46] <sabgenton> lamp etd
[03:46] <sabgenton> etc
[03:46] <jtaji> sabgenton: do 'tasksel --list-tasks'
[03:46] <jtaji> then you can install one with 'sudo tasksel install taskname'
[03:46] <sabgenton> awsome
[03:47] <sabgenton> will go try
[03:52] <sabgenton> jtaji: i just wanted open vdial but must be under other
[03:54] <sabgenton> wvdail
[03:55] <ScottK> manish: One of the key issues you'll run into in such a transition is the without Samba 4 (which is not yet released) you can't have a Linux AD PDC.  So consider a partial migration for now.
[04:06] <sabgenton> how do I rig apt aptutude or tasksel to the cd and not online
[04:22] <terinjokes> can i get a head-25 of /etc/init.d/networking seems I accidently wrote to that file
[04:24] <twb> terinjokes: I don't understand the question.
[04:25] <jtaji> terinjokes: http://paste.ubuntu.com/351600/
[04:25] <twb> Oh, you want a copy from an non-broken system?  It depends what release you're running.
[04:25] <terinjokes> 9.10
[04:25] <jtaji> top 25 lines he meant
[04:25] <terinjokes> jtaji: that looks right, thanks
[04:25] <terinjokes> jtaji: that's 93, give or take ;)
[04:25] <twb> Try aptitude download netbase; mkdir /tmp/x; dpkg -x netbase_*deb /tmp/x; cat /tmp/x/etc/init.d/networking
[04:26] <jtaji> interesting
[04:30] <twb> There's also simply "aptitude reinstall netbase", but I don't think it'll fix conffiles.
[04:30] <twb> (By design.)
[04:34] <terinjokes> yeah.... i forget when I have to connect directy to the box (it's normally headless) that ny keyboard is Programmer Dvorak, but the computer expects English US. i've got to rely on muscle memory (makes password prompts fun)
[04:36] <terinjokes> now, back to what i was doing, converting the system from router to bridge (due to some nasty problems with multicast, unless you guess know how to forward multicast)
[04:46] <terinjokes> ok, got a problem, i'm bridging wlan0 --> eth0, if the bridge isn't turn on (comment out br0 in interfaces) i can ping the internet from the box. if the bridge is on (br0 is uncomment, bridge_ports eth0 wlan0) the wlan0 interface gets an IP, but then it loses it
[05:33] <twb> teddymills: sudo dpkg-reconfigure console-setup
[05:33] <twb> Gah, bloody /quitters
[06:20] <terinjokes> i'm running two computers on a subnet with a linux box acting as a router, any ideas on how to forward multicast
[06:36] <danielck> can anyone help me with a locale problem I'm having on Hardy?
[06:37] <danielck> the locale command gives me "POSIX" for everything exept LANG which is empty
[06:37] <danielck> even though I've set sudo /usr/sbin/update-locale LANG=en_US.UTF-8
[06:38] <danielck> (rebooted and all)
[07:05] <terinjokes> danielck: it's not dead, it's just asleep
[07:05] <danielck> heh
[07:05] <jmarsden> danielck: What did you expect or want your locale to be?
[07:06] <danielck> en_US.UTF-8
[07:07] <jmarsden> OK, and does the output of   locale -a   include that locale?
[07:08] <danielck> hmm, now I'm catching on, it has en_US.utf8
[07:09] <jmarsden> danielck: OK, so try using that and see if it helps.
[07:10] <danielck> nope
[07:10] <danielck> after reboot, still POSIX
[07:10] <danielck> hmm, wait a minute...
[07:12] <jmarsden> OK...  Is the English language pack installed on your machine?  Does   dpkg -l | grep language-pack-en    tell you anything useful?
[07:13] <danielck> yes, seems to be installed like it should be, I actually already reinstalled that
[07:13] <jmarsden> OK.  And are the contents of /etc/default/locale    appropriate?
[07:13] <danielck> the contents of /etc/default/locale is now simply
[07:13] <danielck> LANG=en_US.utf8
[07:14] <jmarsden> Sounds like something is (re)setting your locale somewhere unexpected... did you add a command to .bashrc or /etc/profile or similar related to locale stuff at all?
[07:15] <danielck> no
[07:15] <jmarsden> Oh... and there is a dash in utf-8 ... check that LANG=   setting again?
[07:16] <danielck> well yes, that's what I had in the first place, UTF-8, but locale -a lists it without a dash
[07:18] <jmarsden> OK.  On my Karmic install here I have it as LANG="en_US.UTF-8" .. will check in a Hardy VM if I still have a hardy VM around to test with...
[07:18] <danielck> ok thanks
[07:19] <jmarsden> What you first installed the machine, was it set for some less common (non-English) locale?
[07:19] <jmarsden> s/What/When/
[07:20] <danielck> no, I should specify that it is a VPS
[07:21] <danielck> and I can't recall if I checked the locale when I installed
[07:21] <jmarsden> danielck: OK, but you installed it from an official Ubuntu Server ISO, yourself?  (Or did your VPS provider give you some customized thing?)
[07:22] <danielck> well, I didn't actually install it, since it came from an image, but anyway
[07:22] <danielck> customized
[07:22] <danielck> so it could have been posix to start with, but I'm not sure
[07:22] <jmarsden> ah, so that could possibly be one source of this issue... do you have a way to check (like run a local VM created from that same image)?
[07:23] <danielck> no, unfortunately my host doesn't provide tools for that
[07:24] <jmarsden> danielck: No, I meant download the image and then create a VM on a local PC you own, as a test :)  But OK...
[07:24] <jmarsden> I have a Hardy 8.04.3 Desktop VM  and it has LANG="en_US.UTF-8"  in /etc/default/locale and the locale command output is fine.
[07:25] <danielck> ok
[07:26] <danielck> is it possible to make an image internally from within hardy?
[07:28] <jmarsden> You won't be able to run a VM inside a VM (well, not as far as I know, that's still sort of a research topic, recursive virtual machines :)
[07:29] <danielck> oh I don't meen running it, just creating the image
[07:29] <jmarsden> OK, I have a Ubuntu 8.04 Server CD image (.iso) here, so I can create a new Ubuntu Hardy server VM and see about playing with locales in there...
[07:30] <jmarsden> danielck: That should be doable, but I'm not sure how it would help you
[07:30] <danielck> well... if that's your idea of a good time :) I'm very grateful though of course for your help!
[07:30] <jmarsden> danielck: It'll only take about 10 minutes, I have a fast desktop machine here... :)
[07:31] <jmarsden> And BTW I have had some interest in locale-related bugs, in fact my main Karmic host OS currently has 200+ locales installed (all avalable language packs!) because of some testing I was doing...
[07:31] <danielck> jmarsden: I mean, if I could create an image of my current setup, I could play around with a copy on my desktop in a vm, no?
[07:36] <jmarsden> danielck: Oh, I see.  Yes... for that you could probably use any disk image backup tool (or just dd piped to a compressor and then ssh, even) and then move the backup to your local PC, turn it into a VM disk image there and off you go.  There may be better ways than that, but off the top of my head that's how I'd do it.  I virtualized an ancient unmaintained Fedora Core Linux box that way a few months ago, although I was local so the transfer was
[07:36] <jmarsden>  over a fast LAN not over the Internet.
[07:37] <danielck> right
[07:39] <jmarsden> But if you could download the install image your VPS provider offers, you could start from that, which might be smaller and would let you go over the install process from it and see what happens.
[07:42] <danielck> jmarsden: indeed, that would be good
[07:43] <jmarsden> That's what I had in mind earlier when I suggested you could "run a local VM created from that same image" -- sorry if I was unclear.
[07:44] <jmarsden> Hardy server VM install in progress, BTW...
[07:46] <danielck> sorry, you weren't really unclear, rather my head is a bit fluffy inside
[07:46] <jmarsden> :) OK.
[07:50] <danielck> I'm going to reboot this machine, apparently it also needed some locale love, be back in a minute
[07:51] <jmarsden> OK.
[07:58] <danielck> hmm, setting the locale on this remote server worked just fine, and it's also a Hardy install
[07:58] <danielck> but from a different host and it's not a VPS
[08:03] <jmarsden> OK.  And my new hardy server VM set up its locale just fine "out of the ISO", too.
[08:04] <jmarsden> So it starts to look as though something in the VPS image or how it was installed could be responsible for the problem you are seeing.
[08:05] <danielck> yes... when googling about this I'm finding references to debian bugs which were fixed in 2005
[08:05] <danielck> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=330701
[08:06] <jmarsden> Seems unlikely those would still have been in Hardy in 2008.
[08:06] <danielck> indeed :)
[08:07] <jmarsden> As a cheap workaround, if you do   LANG=en_US.UTF-8 locale    # do you get output that looks "normal" (i.e. en_US.UTF-8 for all of the locale variables displayed)?
[08:08] <danielck> yes
[08:08] <danielck> except LC_ALL
[08:09] <jmarsden> OK.  So if you were to add a line     set LANG="en_US.UTF-8" in /etc/profile you might have a "mostly working" quick fix.
[08:09] <jmarsden> It's ugly and not at all what you are *supposed* to do...
[08:14] <danielck> right
[08:16] <jmarsden> Did that work for you?
[08:17] <jmarsden> Or "right" as it "right, it's ugly" :)
[08:18] <danielck> "right, it's ugly" :)
[08:20] <danielck> jmarsden: lets see if it works
[08:20] <jmarsden> Go for it.
[08:21] <danielck> now let's be clear, I just add this line to the end of /etc/profile?
[08:21] <danielck> LANG="en_US.UTF-8"
[08:21] <jmarsden> Yes.
[08:21] <danielck> didn't work
[08:22] <jmarsden> danielck: OK, if you . /etc/profile   # does that change the value of LANG in your shell?
[08:23] <danielck> no
[08:23] <danielck> even putting LANG="en_US.UTF-8" in my .bashrc won't work
[08:24] <jmarsden> Wow.  Then something is setting it back again within each shell?  Seems very weird.   Try    export LANG="en_US.UTF-8"
[08:24] <jmarsden> Maybe try first just at the shell prompt, then in ~/.bashrc, then in /etc/profile ?
[08:24] <danielck> within the file
[08:24] <danielck> ok
[08:25] <danielck> worked at the shell prompt
[08:25] <jmarsden> That's a start :)
[08:25] <danielck> also worked in bashrc after a re-login
[08:25] <jmarsden> Better still...
[08:27] <danielck> also worked in /etc/profile (removed it from .bashrc)
[08:27] <jmarsden> OK... so at least you now seem to have a working workaround :)
[08:29] <danielck> now I'm not much of a linux guru, so I have to ask, is the end result the same, or might I run into trouble with this approach?
[08:30] <jmarsden> danielck: There may be some services that won't "see" this approach and so won't have the new locale info.
[08:30] <jmarsden> And of course setting things up so different users have different locales is now awkward...
[08:30] <danielck> indeed
[08:30] <danielck> luckily I don't need multiple locales
[08:31] <danielck> my next trouble is getting php (running as cgi) load it's extensions...
[08:31] <danielck> :(
[08:32] <jmarsden> Is there a reason you run PHP as a CGI, it tends to have horrid performance that way.
[08:32] <danielck> I'm running nginx, and using php-fastcgi which I undestand should be pretty fast
[08:33] <jmarsden> Ah, OK.  I've only used Apache (well, and its predecessor NCSA httpd, a long time ago!).
[08:33] <danielck> got both running (apparently) fine, but php isn't loading any extensions
[08:33] <danielck> I gave nginx a try since apache was eating up all my memory on my 256 VPS
[08:34] <twb> I used to like thttpd
[08:35] <twb> Now I use busybox httpd
[08:35] <jmarsden> OK.  Is the user that nginx runs as the same one apache2 runs as?  Is there anything about the location or permissions of the php.ini file that changes when you use the CGI version?  I'm just guessing...
[08:36] <danielck> yes for both
[08:37] <twb> nginx doesn't Provides: httpd-cgi, so I wouldn't bet on it supporting non-static pages at all.
[08:37] <danielck> with cgi it uses /etc/php5/cgi/php.ini
[08:38] <danielck> twb: well it seems to work, i'm getting php working fine
[08:38] <danielck> the output of phpinfo just doens't show any extensions loading
[08:38] <twb> Hmm, the code itself refers to fastcgi, so I guess it's a packaging mistake
[08:38] <danielck> I have nginx installed from source
[08:39] <danielck> anyway, I'll be back, need a break from the computer :)
[08:39] <jmarsden> Why?  It's packaged, isn;t it?  OK...
[08:39] <danielck> jmarsden: huge thanks for your help!
[08:39] <danielck> jmarsden: the version supplied by hardy is pretty old
[08:40] <jmarsden> danielck: If you need a newer one, I'd look in hardy-backports or backport the one in karmic, rather than DIY from a source tarball...
[08:40] <jmarsden> Anyway, take a break first :)
[08:40] <danielck> pretty much followed this tutorial http://www.mensk.com/webmaster-toolbox/perfect-ubuntu-hardy-nginx-mysql5-php5-wordpress/
[08:40] <danielck> see ya in a bit
[08:55] <danielck> back
[08:57] <jmarsden> OK.  I'm sure how much I can help with nginx and compiling things from source tarballs and then wondering why they don't do what you want or need, though.  And it is approaching 1am here...
[08:59] <danielck> jmarsden: no worries, you've been a great help!
[09:00] <jmarsden> You might want to try using the Hardy packaged nginx and see whether php5-cgi then "just works" ?
[09:00] <danielck> I might do that :)
[09:01]  * jmarsden attempts to remove himself from chair, and move towards bed... goodnight :)
[09:07] <danielck> jmarsden: gnight
[09:15] <johe|work> maybe someone help me on http://pastebin.com/m2e2e1e1a its the strace output of an snmpd service, which dies with floating point exception,
[09:15] <johe|work> it affects amd64 systems with LTS but only real one, we have some of amd64 as vmware guests which dont die that way
[09:16] <johe|work> but also 6 real servers how does the some failure
[09:17] <jiboumans> good morning
[09:47] <_ruben> ok .. this is weird .. i have a server with (among other hdds) a 2G flash disk for the OS, 128M for /boot (ext2) and the rest LVM .. cfdisk/fdisk/parted all confirm this layout .. however when i do sudo mount /dev/sda1 /boot, it shows me an empty 2G partition
[09:49] <_ruben> the data most likely is still there, as the box does boot just fine (even though i dont seem to have a way to access any of the /boot/ files)
[11:45] <jla> morning all
[13:02] <Hellsheep> Hey Guys i'm having an issue installing linux-headers, i tried to use apt-get install linux-headers-`uname -r` except it gives me this error message: E: Couldn't find package linux-headers-2.6.18-128.2.1.el5.028stab064.4
[13:03] <Hellsheep> I tried typing uname -r and i get this: 2.6.18-128.2.1.el5.028stab064.4 - So which linux-headers do i install? >.<
[13:04] <pmatulis> Hellsheep: cloud stuff?
[13:04] <_ruben> you're not running a stock ubuntu kernel
[13:04] <Hellsheep> Hmm
[13:04] <_ruben> looks redhat'ish
[13:05] <Hellsheep> That would explain it then :P
[13:06] <Hellsheep> I was pretty sure i just installed Ubuntu 9.04 though >.<
[13:06] <Hellsheep> weird
[13:07] <_ruben> in Xen perhaps?
[13:08] <_ruben> 2.6.18 kernel is rather old for ubuntu .. think 6.06 or so ? :)
[13:09] <Hellsheep> Oh i see
[13:12] <Hellsheep> well thanks, ill try sus something out
[13:12] <Hellsheep> And see if i cant fix this :P
[13:17] <zul> morning
[13:31] <ttx> zul: yo
[13:32] <ttx> zul: you added "Using bzr for packaging (zulcss)" on tomorrow's meeting agenda, could you clarify ?
[13:36] <ScottK> Dapper was 2.6.15.
[13:37] <_ruben> close enough :)
[13:42] <zul> ttx: i want to get people starting using bzr for maintaining our server packages
[13:42] <ttx> zul: ok
[13:43] <ttx> zul: could this wait for the next week meeting ? This week's agenda is already overloaded
[13:43] <zul> ttx: sure
[13:43] <ttx> zul: ok, please move that line to a "next meeting" header on the same page
[13:45] <zul> ttx: done
[13:45] <zul> hey smoser
[13:45] <smoser> good morning
[13:45] <ttx> zul: thx
[13:45] <ttx> smoser: o/
[13:46] <smoser> good morning ttx, and happy new year
[13:46] <zul> ttx: stupid wikis
[13:46] <ttx> smoser: best wishes to you too !
[13:49] <ScottK> zul: Keep in mind that lamont already maintains postfix, bind, an who knows what else in Git.  We shouldn't mess with that.
[13:50] <zul> ScottK: yes of course
[13:54] <jdstrand> coffeedude: hi! so I added a preliminary debdiff to bug #274350. Since I am not a likewise user or have a way to test it, it may be too simplistic.... Would you mind reviewing and commenting on it?
[13:54] <coffeedude> jdstrand, Sure.
[13:56] <coffeedude> jdstrand, we are only looking at fixing this in lucid right?  So only in the new likewise-open packages.  The debdiff is correct for 4.1 but I've already fixed this in the likewise-open PPA.
[13:57] <smoser> zul, ttx where should issues 1 and 4 at http://groups.google.com/group/ec2ubuntu/browse_thread/thread/30f3af093528aad4 be raised ?
[13:57] <coffeedude> jdstrand, or is this a proposed change to backport to Hardy and later?
[13:57] <jdstrand> coffeedude: correct, only for Lucid
[13:58] <jdstrand> coffeedude: if someone want to SRU for hardy, I certainly wouldn't oppose it, but I am also not driving it ;)
[13:59] <coffeedude> jdstrand, K.  So you diff won't apply to the new packages.  I've already fixed it there like we discussed. pitti is reviewing the debs for me and they are planned for upload for alpha 2.
[13:59] <ttx> smoser: launchpad ?
[13:59] <jdstrand> coffeedude: that is great! Do you mind if I adjust the bug accordingly?
[14:00] <ttx> smoser: otherwise the ubuntu-cloud list sounds like a good place for discussion ?
[14:00] <coffeedude> jdstrand, please do.
[14:00] <smoser> neither is at all cloud related
[14:00] <smoser> at least i wouldn't understand why that would be
[14:01] <ttx> smoser: you mean every karmic has tons of console-kit-daemon warnings in logs ?
[14:01] <jdstrand> coffeedude: ok, done. thanks again :)
[14:02] <zul> smoser: they should open bugs in launchpad
[14:03] <smoser> fair enough
[14:08] <jdstrand> coffeedude: I was not aware of likewise-open *and* likewise-open5. Will both be in lucid? have you fixed both? (also, where are the new packages-- I can't seem to find the ppa)
[14:09] <ttx> jdstrand: both should update to likewise-open in lucid
[14:09] <jdstrand> ttx: so there will only be likewise-open in lucid, which is version 5?
[14:10] <ttx> jdstrand: that's the plan.
[14:10] <coffeedude> likewise-open5 is going away.  The new likewise-open packages replqace likewise-open5.
[14:10] <jdstrand> I see
[14:10] <jdstrand> gotcha
[14:10] <smoser> soren, around?
[14:10] <coffeedude> jdstrand, https://launchpad.net/~likewise-open/+archive/ppa/+packages?field.name_filter=&field.status_filter=published&field.series_filter=lucid
[14:10]  * jdstrand wonders why he couldn't find that...
[14:10] <jdstrand> coffeedude: thanks
[14:10] <Maleko> can anyone give me a better alternatve to the standard ping tool in linux? preferably one that can show jitter and hops in each ping reply
[14:17] <soren> smoser: Yup.
[14:18] <smoser> gonna send you an email realy quick. i'd like your comments
[14:21] <smoser> see inbox
[14:33] <zul> smoser: the initramfs tooks good to me
[15:08] <BBHoss> Hi, can someone explain the memory usage to me?  I have a server with 4GB ram.  MySQL is taking up 16% of that, but the RSS is around 600MB.  Tomcat is taking up about 300MB.  When I run free -m though, it only shows 105MB free, with a ton of memory in "buffers/cache".  Is this normal?
[15:08] <BBHoss> Monit keeps sending me warnings that it is over 85% memory usage
[15:08] <Jeeves_> BBHoss: Yes.
[15:08] <Jeeves_> The kernel keeps a lot of files in cache
[15:08] <_ruben> monit should "ignore" the cache/buffers
[15:08] <Jeeves_> so it doens't have to read all files from disk (which is slow)
[15:09] <BBHoss> _ruben: it doesn't appear to be
[15:10] <BBHoss> unless something  is using up all of the memory at that time
[15:10] <_ruben> then bug the monit authors to fix the check :)
[15:11] <BBHoss> _ruben: is it supposed to ignore that, or are you saying that it needs to?
[15:12] <_ruben> i never used monit, so i dont know its "proper" workings .. a *good* test would be to ignore it though (even though not a lot of apps actually do)
[15:17] <BBHoss> _ruben: why if it has plenty of memory in buggers and caches, would is be swapping to disk?
[15:17] <BBHoss> buffers
[15:20] <_ruben> swapping out unused (or hardly used) memory space in favor of more cache can speed up the system
[15:23] <BBHoss> _ruben: Here is my free -m output, It looks like only ~500MB is being cached/buffered, and at around 3500MB is actually used.  Is this correct? https://gist.github.com/c8c1bff98627edbe0d20
[15:24] <_ruben> correct
[15:24] <zul> ttx: ping
[15:24] <BBHoss> _ruben: well then where does the other 3500MB come from, the RSS from the two biggest processes sums up to under 1GB
[15:26] <BBHoss> _ruben: here is the ps aux output: https://gist.github.com/cc4e7a85e544533cdbc6
[15:27] <ttx> zul: pong
[15:27] <_ruben> you just reached the edge of my expertise .. the RSS versus VSZ etc stuff is still quite a mystery to me .. your virtual memory usage seems roughly 3.5G though
[15:27] <BBHoss> _ruben: heh ok
[15:39] <zul> ttx: do you want to me merge openvpn for you?
[15:41] <ttx> zul: you can
[15:41] <zul> ttx: thanks
[15:51] <kirkland> ttx: ping
[15:56] <ttx> kirkland: pong
[15:57] <kirkland> ttx: hey, i was just syncing today's ISO, hoping you might give me a 5-minute status update on where we are?
[15:57] <ttx> kirkland: sure
[15:57] <ttx> kirkland: I first tested the classic CLC+Walrus+CC+SC / NC topology
[15:58] <ttx> kirkland: ran into bug 503180, or something similar
[15:58] <ttx> got to stop / kill / start
[15:58] <ttx> otherwise started up and registered ok
[15:59] <ttx> then I tested CLC+Walrus / CC+SC / NC
[15:59] <kirkland> ttx: okay, yeah, mathiaz and I discussed that one yesterday
[15:59] <kirkland> ttx: I'll get Dan to ssh into my setup once it's up
[15:59] <ttx> CC+SC fails with several bugs
[15:59] <ttx> one of them is a bugfix that got lost in a merge
[16:00] <ttx> fix uploaded as rev825
[16:00] <ttx> one other is linked to the fact the publication shouldn't wait for the component to be started up
[16:00]  * soren pauses for dinner
[16:00] <ttx> Fix will be uploaedd in the next minutes
[16:01] <ttx> and the last one is due to the preseed file not being fetched from the CLC
[16:01] <kirkland> ttx: cool, man
[16:01] <ttx> (bug 503339, Colin will look at it)
[16:01] <kirkland> ttx: so i think I've got the wsdl stubs "handled"
[16:01] <ttx> I haven't looked into the fix, tbh
[16:02] <kirkland> ttx: i added a hook to the build that makes the build fail if the wsdl files have changed, printing an informative error message
[16:02] <ttx> was just wondering how you got rid of the patchsystem and kept a patch :)
[16:02] <kirkland> ttx: well, i wanted to discuss that ...
[16:02] <kirkland> ttx: we *could* apply that patch directly to the bzr source
[16:02] <kirkland> ttx: and my script could just update those files and debcommit
[16:02] <ttx> kirkland: that doesn't sound like a good idea :)
[16:02] <kirkland> ttx: but i wondered if it made sense to keep that one patch separate
[16:03] <kirkland> ttx: i also considered keeping a *separate* bzr repository for the generated code
[16:03] <kirkland> ttx: eucalyptus-generated
[16:03] <ttx> kirkland: it does make sense. however keeping the patch out of a patchsystem is probably less confusing
[16:03] <kirkland> ttx: and as part of our merge process, copy that in
[16:03] <kirkland> ttx: okay, so pull it out of debian/patches?
[16:03] <kirkland> ttx: and just put it in wsdl/* ?
[16:04] <ttx> kirkland: I think having a patchsystem arond seems to imply that patches shouold go there. Having it separate and directly applied in debian/rules would mitigate that
[16:04] <ttx> will make it more obvious to new contributors to the package
[16:04] <mathiaz> ttx: so what's the deal with https://bugs.launchpad.net/bugs/503180?
[16:05] <mathiaz> ttx: I've tried to restart eucalyptus-cloud, but it still fails
[16:05] <ttx> mathiaz: did you check for stale eucalyptus-cloud processes before starting again ?
[16:05] <ttx> I always had to kill -9 one
[16:05] <mathiaz> ttx: right - I had to do the same
[16:05] <ttx> and it still fails ?
[16:05] <mathiaz> ttx: that being said, I was actually rebooting the system IIRC
[16:06] <mathiaz> ttx: I'll give it another try
[16:06] <ttx> might be two separate bugs, with me only hitting the first layer
[16:06] <ScottK> ttx: I'm reasonably certain that isn't a package you want new contributors coming anywhere near.
[16:07] <kirkland> ttx: agreed
[16:07] <kirkland> ttx: okay, i'll affect that change today
[16:07] <ttx> ScottK: even experienced contributors might get tempted to use the existing patchsystem rather than apply patches directly to source, since that's the common rule
[16:08] <ScottK> True, but lintian will at least warn you about it.
[16:35] <mathiaz> ttx: hm well - I can get the credentials now - seems like the night was a good thing for the CLC
[17:00] <ttx> mathiaz: it looks a lot like the old db deadlock thing, though they are no longer using c3p0
[17:07] <ttx> kirkland: please release with whatever you have and spin a CD before the end of your day, so that I can pick up my changes tomorrow morning
[17:07] <kirkland> ttx: understood
[17:07] <ttx> kirkland: thanks !
[17:08]  * ttx grumbles as his parents crashed their ubuntu box
[17:09] <kirkland> ttx: did last nights work out well for you?
[17:09] <kirkland> ttx: i worked until about 8pm, uploaded, and asked slangasek to schedule a build for 4 hours later
[17:09] <ttx> kirkland: it was perfect.
[17:09] <kirkland> ttx: figured that should pick up my changes, and complete before your day started
[17:10] <kirkland> ttx: sweet, then i think we're in sync
[17:10] <ttx> we could have a regular build at that time
[18:11] <unit3> hmm... anyone know why, on boot, the device files in my lvm vg get assigned root:disk ownership, but when I create a new LV it gets assigned root:root ownership?
[18:12] <unit3> (on karmic)
[18:32] <turkeymonkey> in 8.04, i install xen-compatible kernels for my vps with the linux-image-xen package. i've noticed it doesn't exist in 9.10... what has replaced this packaged in 9.10 (or how can I install a xen-compatible kernel)?
[18:32] <unit3> nothing's replaced this. the problem is the dom0 patches don't exist in newer kernels.
[18:32] <unit3> you can install the old packages from hardy, but then you lose newer kernel functionality.
[18:33] <unit3> it's sort of a crappy situation, and is why people (like Redhat) are moving to kvm / libvirt instead.
[18:33] <unit3> but that's not totally production ready either, depending on your setup.
[18:33] <unit3> so... stick with hardy, I guess, for now.
[18:33] <turkeymonkey> thanks for the info, that explains alot.
[18:34] <turkeymonkey> hardy it is, then... still have around 3 years of support for server :)
[18:34] <unit3> yep. hopefully for 10.04 there'll be some new supported options, but it's hard to say, with xen in the state it is.
[18:38] <Hypnoz> anyone know if I can make an NFS export from an NFS mount? exportfs -r complains when the line is in there
[18:46] <zul> smoser: ping
[18:46] <smoser> here
[18:46] <zul> smoser: something like bzr branch lp:~zulcss/ec2-init/ec2-init-config ?
[18:50] <smoser> zul, yeah, basically. thats what we're looking for.
[18:50] <smoser> i wonder, why the differeing 'start on'
[18:52] <zul> smoser; i thought they would be more sensible
[18:53] <smoser> more sensible?
[18:54] <zul> that if cloud-config is started then the network should be started and local-filesystem should be started as well
[18:54] <zul> but ill defer to you though
[18:54] <smoser> cloud-config will not be emitted until / is mounted and network eth0 up
[18:55] <zul> smoser: ah ok so ill change those
[18:55] <smoser> so i dont knwo what would be the best practice..
[18:55] <zul> ditto
[18:55] <smoser> in some sense it makes sense to have each of those listed, the job is then more stand alone
[18:55] <smoser> but requires more maintainance
[19:00] <kirkland> mathiaz: howdy
[19:00] <kirkland> mathiaz: okay, fresh UEC installed today
[19:00] <kirkland> mathiaz: i am see your node-discovery problem
[19:00] <mathiaz> kirkland: o/ texas!
[19:00] <kirkland> mathiaz: first, no nodes are found
[19:00] <kirkland> mathiaz: eventually, after re-running, it does find the ipv6 address
[19:00] <kirkland> mathiaz: do you have an open bug for this one?
[19:01] <mathiaz> kirkland: not yet
[19:01] <mathiaz> kirkland: did you try to run avahi-browse from the command line?
[19:01] <kirkland> mathiaz: yup
[19:01] <mathiaz> kirkland: if you look at euca_conf you'll find the exact command used to find the nodes
[19:01] <mathiaz> kirkland: and so it's an issue with avahi?
[19:02] <mathiaz> kirkland: because when I tried to run avahi-browse from the command line I got the expected result
[19:02] <mathiaz> kirkland: that being said it seems non-deterministic
[19:02] <kirkland> mathiaz: i think it is probably an avahi issue
[19:02] <mathiaz> kirkland: since you first don't see the nodes, then the ipv6 address, etc..
[19:02] <kirkland> kees: howdy, are you around?
[19:02] <mathiaz> kirkland: I would say so as well
[19:03] <kirkland> kees: i seem to remember you sending me a note before xmas, about some new cool avahi improvement that you made, and we should take advantage of
[19:03] <kirkland> kees: i seem to have forgotten what that magic was :-)
[19:03] <kirkland> kees: jog a memory?
[19:06] <kirkland> mathiaz: hmm ...
[19:06] <kirkland> mathiaz: the shell command in euca_conf works correctly
[19:06] <kirkland> avahi-browse -prt _eucalyptus._tcp | grep '^=.*"type=node"' | cut -d\; -f8 | sort -u
[19:06] <kirkland> 10.1.1.71
[19:06] <mathiaz> kirkland: right
[19:07] <mathiaz> kirkland: how long did you wait?
[19:07] <mathiaz> kirkland: I'd try to reboot the CC, then run multiple times the avahi-browse command right after boot
[19:07] <mathiaz> kirkland: and see how the result evolves
[19:08] <kirkland> mathiaz: hmm, okay, i'll need to install a new node
[19:08] <kirkland> mathiaz: b/c this one is working perfectly
[19:08] <kirkland> mathiaz: i was able to grab my creds too
[19:08] <kirkland> AVAILABILITYZONE        |- m1.small     0002 / 0002   1    128     2
[19:09] <mathiaz> kirkland: so you didn't run into the stuck eucalyptus-cloud process?
[19:09] <kirkland> mathiaz: not yet
[19:09] <mathiaz> kirkland: if you can get your credentials you haven't run into it then
[19:10] <kirkland> mathiaz: okay, well, lucky me
[19:10] <kirkland> mathiaz: this is from an ISO install
[19:10] <kirkland> mathiaz: you do exclusively package installs, right?
[19:11] <mathiaz> kirkland: yes
[19:11] <mathiaz> kirkland: ttx run into the same problem
[19:11] <mathiaz> kirkland: and torsten as well
[19:25] <Hypnoz> anyone know if i can NFS export a directory that is an NFS mount?
[19:25] <guntbert> Hypnoz: why would you want to do that?
[19:26] <Hypnoz> to get around some subnet shit. A server (ServerA)can't see our nfs server, so i want to mount it on a server (ServerB) that can see both subnets. The ServerA can mount the NFS from ServerB
[19:29] <guntbert> Hypnoz: aha - understandable - the easiest way would be to try it?
[19:29] <guntbert> Hypnoz: but I really don't know
[19:29] <Hypnoz> # exportfs -r
[19:29] <Hypnoz> exportfs: Warning: /autohome does not support NFS export.
[19:31] <guntbert> Hypnoz: then it seems impossible - but try to read man exports
[19:40] <kees> kirkland: it's for registering a name without registering a reverse.
[19:40] <kees> i.e. being able to finally do the thing I had suggested as an option.
[19:40] <kees> http://git.0pointer.de/?p=avahi.git;a=commitdiff;h=63b105af928ab6027ce7c905b8ac051fc23a2880
[19:40] <kirkland> kees: ah
[19:41] <kirkland> kees: okay, so we would add -R to the options in the eucalyptus upstart scripts
[19:41] <kirkland> kees: that do the avahi publish
[19:42] <mathiaz> kirkland: well - I don't think we need to do that anymore
[19:42] <mathiaz> kirkland: IIRC the avahi-publish jobs are publishing the IP address to use
[19:42] <mathiaz> kirkland: the same way as they're publishing the type
[19:42] <kirkland> mathiaz: right, we encoded that into the publish string
[19:42] <mathiaz> kirkland: that way, euca_find_cluster doesn't rely on avahi name resolution
[19:43] <kirkland> mathiaz: i'm trying to figure out if it would make more sense to go back and do it kees' way
[19:43] <mathiaz> kirkland: we should actually do the same thing for the nodes
[19:43] <mathiaz> kirkland: well - I don't think so - in the case where the system has multiple IPs
[19:43] <mathiaz> kirkland: you wanna publish one specific IPs to be used
[19:44] <smoser> general "how does this work" question: given http://packages.qa.debian.org/p/python-boto.html (and presence of 1.9b-1 there) will ubuntu just pick that up ?
[19:44] <smoser> or does someone need to do something?
[19:46] <zul> smoser: it would have to be merged because there is ubuntu changes
[19:46] <kirkland> mathiaz: ah, right
[19:47] <zul> smoser: i could take care of that though
[20:00] <smoser> zul, you mind, i'll just put it together and you look at it ?
[20:01] <smoser> the diff is minimal.
[20:01] <zul> smoser: sure
[20:23] <smoser> zul, so should i open a bug to request the merge ?
[20:23] <zul> smoser: yep
[20:33] <zul> smoser: bug #?
[20:35] <smoser> its coming
[20:35] <smoser> i was learning requestsync
[20:35] <smoser> https://wiki.ubuntu.com/SyncRequestProcess
[20:43] <smoser> zul, https://bugs.edge.launchpad.net/ubuntu/+source/python-boto/+bug/503541 has branch merge request attached.
[20:48] <zul> smoser: you only changed the debian directory?
[20:48] <smoser> correct.
[20:49] <zul> but there...oh duh never mind
[20:49] <smoser> you can see the diff that i merged with:
[20:49] <smoser> bzr diff -rtag:1.8d-1..tag:1.8d-1ubuntu2
[20:49] <zul> smoser: yeah i wasnt paying attention for a sec
[20:49] <smoser> the 1ubuntu1 were merged upstream so no need
[20:51] <ewan_> Hey, can anybody point me in the right direction for setting up my ubuntu server as an IRC server?
[20:51] <unit3> irc's a little tricksy.
[20:52] <unit3> generally, you want to figure out which ircd you want to run, and then just use the ubuntu package.
[20:52] <unit3> the config docs will match the ircd you pick.
[20:52] <unit3> do you have one in mind?
[20:53] <zul> smoser: done
[20:53] <zul> and with that i bid you adieau (gotta go shovel)
[20:54] <ewan_> not really
[20:54] <ewan_> i just want something reasonably simple to set up
[20:54] <unit3> well... i can't recommend anything, I've always found irc servers to be sort of bizarre.
[20:55] <ewan_> fair enough. i've just read about ircd-hybrid so i'm going to give that a go
[20:55] <unit3> good luck. :)
[20:56] <joe-mac> is anybody using the HP management agents that say they are for jaunty only on hardy? we only use lts on our servers and i'm wondernig if i can throw these debvs in our custom repo...
[20:56] <nvme> does the ubuntu server ediition kernel have different set of modules than regular ?
[20:58] <genii> nvme: Yes
[20:58] <nvme> genii, where can i get a .deb kernel image for use with drbl ?
[20:59] <genii> nvme: No idea.
[21:03] <genii> nvme: If the particular module you need exists for the -server kernel, you can use the initramfs tools to make a new vmlinuz which contains it
[21:06] <genii> Apparently didn't care for that idea.
[21:09] <jla> off topic, but could some kind person tell me how to reduce the noise on this channel?
[21:14] <guntbert> jla: you can hide the joins/parts in your client
[21:17] <jla> guntbert I am using pidgin under windows, and it seems a remarkably sparse piece of software
[21:17] <guntbert> jla: look for something like "conference mode"
[21:18] <guntbert> jla: and there is #pidgin
[21:24] <jla> guntbert: thanks it appears to be in the plugins. Thanks
[21:25] <cemc> jla: I personally recommend some other app for irc (like xchat, mirc). they seem better suited than pidgin (imho)
[21:26] <jla> cemc I concur with yho.
[21:28] <ewan_> i'm using irssi. IRC in console is just too cool to resist
[21:29] <jla> cemc: do you know if xchat, mirc also support MSN messanger, I have to use it occasionally.
[21:29] <ewan_> I've set up ircd-hybrid and i can connect to the IRC server from the host (/connect 127.0.0.1) but i can't connect from any other machines on the LAN. i just get connection refused. any ideas?
[21:30] <jla> ewan: firewall?
[21:31] <cemc> jla: I don't think they do
[21:31] <Vog> anyone else having problems importing keys from the keyserver.ubuntu.com ?
[21:31] <jla> cemc: too much to hope for!
[21:32] <cemc> Vog: problems like how? it seems to work for me
[21:32] <jla> ScottK: did you get any answer on the amavis restart ?
[21:32] <guntbert> Vog: you can use most other gpg servers as well
[21:33] <cemc> jla: you mean https://bugs.launchpad.net/ubuntu/+source/spamassassin/+bug/502615 ?
[21:35] <ewan_> jla: could be. i don't know how to configure the ubuntu server's firewall though. or if i've even installed one
[21:36] <cemc> ewan_: maybe it's listening only on localhost? check with netstat -nlp
[21:38] <ewan_> cemc: you may be right. i see "127.0.0.1:6667" but no others ending with :6667
[21:39] <cemc> ewan_: it seems it's configured to listen only on localhost. you may want to check the config file (if it has one)
[21:42] <jla> ewan: I not sure how to config FW under ubuntu - I usually edit /etc/iptables, however I think ubuntu user User Friendly Firewall (UFW)
[21:43] <ewan_> cemc, jla: thankyou both, i got it working by editing the config file. could have sworn i already did that, but ah well :D
[21:47] <jla> cemc: .../502615 - yes, there was a brief conversation about whether amavis needs restar
[21:48] <jla> restart. my 2C was yes, to avoid stale sockets. But I am NOT authorative
[21:48] <cemc> jla: I was going to look into it, then I got distracted... I'll update the bugreport ASAP.
[21:50] <cemc> jla: I'm not sure it's the best thing to restart from crontab. not everybody uses sa-update, and even if... not sure if it's ok to restart amavis from cron. but I'll try to look into it. maybe a HUP if it supports it
[21:50] <cemc> I was going to try that first
[22:03] <jdstrand> jla: you can create an iptables script of course, but for a host-based firewall it is recommended you use ufw. See https://wiki.ubuntu.com/UncomplicatedFirewall. For a routing firewall, you might also consider something like shorewall
[22:04] <jdstrand> jla: that page has links to firewall docs as well (ufw et al)
[22:21] <Vog> cemc: Was trying to add a key for a launchpady ppa repository. It timed out twice and worked on the third try. Only after a abnormally long pause.
[22:22] <cemc> Vog: yeah, that happened to me a couple of times a while back. after some insisting it downloaded the key :)
[22:26] <marks256> Say if i have two NICs in a server, eth0 and eth1. eth0 goes to internet (on the network 192.168.0.0), and eth1 has a DHCP server (network 192.168.1.0). How do i specify where data goes? if i want to contact 192.168.0.50 (eth0) vs contacting 192.168.1.10 (eth1)
[22:26] <marks256> how does that even work?
[22:26] <marks256> In other words, is there a way to tell linux that all requests on the 192.168.1.x network get sent to eth1, but all other requests get sent to eth0?
[22:27] <Vog> iptables and static routes
[22:28] <marks256> can you point me to some reading?
[22:29] <Vog> marks256: https://help.ubuntu.com/community/IptablesHowTo
[22:30] <marks256> Vog, so the static routes are done using iptables?
[22:30] <Vog> it isn't easy, there are other methods to do what I *think* you are trying to do.
[22:30] <marks256> please share :)
[22:30] <Vog> no the routes are setup seperatly
[22:30] <marks256> oh ok
[22:31] <Vog> http://www.ubuntugeek.com/howto-add-permanent-static-routes-in-ubuntu.html
[22:31] <Vog> Honestly I'm just googling the terms and posting links
[22:32] <Vog> Google is your friend.
[22:32] <marks256> yes, i know. it's not your friend when you don't know what you want though. i've been googling these terms and have been coming up with what seems to be a lot of good information
[22:33] <marks256> Basically i want to setup a "private network" on a server of mine with glusterfs. That way all the FS traffic is on it's own private network (that can eventually be upgraded to a faster connection, without upgrading a whole building)
[22:33] <Vog> marks256: If you are trying to setup a easy to configure gateway you might want to look into ebox, pfsense or somethign like that.
[22:33] <Vog> Either way though you'll have a lot of reading ahead of you.
[22:33] <marks256> i don't need internet on the downstream
[22:34] <marks256> i expected lots of reading :)
[22:34] <Vog> Why not seup openvpn
[22:34] <Vog> setup
[22:34] <Vog> I'm not familiar with glusterfs at all...
[22:35] <marks256> openvpn?
[22:35] <Vog> Yeah it's a Open virtual private network
[22:35] <marks256> glusterfs is a distributed file system. It is scalable to large sizes, and seems to work very well.
[22:35] <marks256> what would i need vpn for?
[22:35] <Vog> essentially a private tunnel between 2 or more sites over public networks.
[22:36] <Vog> But as I said again I don't think I understand what you are trying to achieve :)
[22:36] <cemc> marks256: I'm not sure I understand what you're asking up there ;)
[22:36] <marks256> It's more from a bandwidth perspective. I don't want all that file system traffic running on an already bogged down network
[22:37] <Vog> So you want to do a multi wan?
[22:37] <Vog> some traffic over one interface the rest over another?
[22:37] <marks256> Vog, not really?
[22:37] <marks256> yeah sort of
[22:37] <marks256> but eth1 will be a dhcp server
[22:37] <stake> Hallo everybody.
[22:37] <marks256> actually no it wont. the gluster clients will be static ips
[22:38] <cemc> marks256: eth0 and eth1 will be connected to the same switch (the LAN switch) ?
[22:38] <benedikt> Can postfix and apache start with certificates with passphrases (without the need to enter them manually)
[22:38] <marks256> cemc, no. eth0 will be connected to the internet, and eth1 will be connected to another switch which connects the clients
[22:39] <marks256> i'll see if i can draw a picture of what i want real quick
[22:39] <cemc> benedikt: if you safe the passphrase in some file to load automatically, then the security of the passphrase is gone, don't you think? :)
[22:39] <stake> short question: I want to start a copy-job on my server to move a lot of data.
[22:39] <benedikt> cemc: good point. i didnt really think this though. ill stick to passphrase-less certificates
[22:40] <Vog> stake: ok
[22:41] <stake> When I want to use a ssh tunnel, can I start the job (from another pc) and then logoff the server?
[22:41] <stake> so it can work on it's own?
[22:42]  * stake is sorry for his poor English.
[22:42] <cemc> stake: what are you using for copyinh ?
[22:42] <cemc> like scp ?
[22:43] <cemc> copying locally or copying to another server?
[22:43] <Vog> stake: I suggest tar for the copy much better at lot's of files.
[22:43] <stake> cemc: Actually I wanted to move the data from disk to disk on the server. More something like cp or mv.
[22:43] <cemc> stake: use screen
[22:43] <Vog> good call dem
[22:44] <cemc> stake: http://www.gnu.org/software/screen/
[22:44] <stake> Vog: If I tar the files, the problem isn't solved. I still have to copy/move the tar. ;)
[22:44] <stake> cemc: ...was reading the man pages. :)
[22:44] <stake> thx
[22:44] <Vog> stake you can use tar to copy and move the files as well... :) let me get you a link...
[22:45] <Vog> http://www.cyberciti.biz/faq/howto-use-tar-command-through-network-over-ssh-session/ but this is over a network... not between drives sorry bout that.
[22:45] <cemc> stake: you start up screen and you start the copy process in one of the virtual terminals of that screen session. then you can detach from the screen session and logout
[22:46] <marks256> Vog, cemc here is a (crude) picture of what i want to do http://img10.yfrog.com/img10/4944/networkl.png
[22:46] <cemc> tomorrow you log back in then reattach to that running screen session and continue where you left off
[22:46] <Vog> marks looks like a basic network gateway.
[22:47] <stake> ejat:
[22:47] <cemc> screen is generally a good tool if you have a to do lenghty stuff (like copying for hours) and you have a shaky ssh connection to the server
[22:47] <stake> sorry.
[22:47] <marks256> Vog, ok. You may be right. i'll look into it. thanks :)
[22:47] <cemc> marks256: what's your question about this exactly?
[22:48] <marks256> my question is, how does pc1 tell the difference between the two address 192.168.0.5 and 192.168.1.5
[22:48] <stake> cemc: jap. That's what I'm looking for.
[22:49] <stake> cemc: But it has to run on the server right?
[22:49] <cemc> stake: yes
[22:50] <cemc> marks256: you have 1-1 address from 192.168.0.x and 192.168.1.x configured on eth0 and eth1 respectively. then it knows that those networks are directly connected to it, on those two nic's (if the netmasks are right I guess)
[22:51] <cemc> marks256: it will send packets to 192.168.1.x on eth1 'automagically', no need to worry about that. just have an IP from 192.168.1.x on eth1
[22:52] <marks256> so if i were to ping 192.168.0.10, it would know to use eth0, and if i were to ping 192.168.1.15, it would know to use eth1
[22:52] <benedikt> yes
[22:52] <benedikt> the routing tables take care of that.
[22:52] <marks256> huh. well now if only i can figure out how to set this up :)
[22:52] <cemc> yes, if everything's configured alright, it should
[22:52] <marks256> haha. IF everythign's configured right ;)
[22:52] <marks256> so basically i need routing tables?
[22:52] <cemc> well... :)
[22:53] <benedikt> marks256: no, it takes care of that itself. you can check then by typing "ip route"
[22:54] <marks256> so all i have to do it simply setup eth1 to be a host, and viola, everything will do it on it's own?
[22:54] <ScottK> jla: I've been offline most of the day.  I didn't hear back from ivoks yet.
[22:55] <benedikt> probably. depends on if you want the other pc's to also connect to the internet
[22:55] <marks256> benedikt, nah. if i need to do that, i've done it before, so it shouldn't be a problem.
[22:55] <marks256> Sweet. Thanks for the help all!
[22:56] <benedikt> then it will just be fine if you correct the right interface with the right ip address
[22:59] <stake> cemc: Ehm... I started screen. Started my job and logged off. Login again and looked in "top" if the job still runs. It does. ... How do I get back to the "screen"? :)
[22:59] <benedikt> screen -r <pi>
[22:59] <benedikt> pid*
[22:59] <benedikt> or just screen -r if you only have one
[22:59] <cemc> stake: I use screen -d -r
[22:59] <benedikt> then can be named somehow. i forget how.
[23:00] <stake> cemc: Great. Thanks man!
[23:00] <cemc> if you have more then one, a list will appear and you can choose
[23:00] <stake> You made my ... oh, a new day. ;)
[23:00] <cemc> benedikt: that's -S i think
[23:00] <stake> Really cool tool.
[23:01] <stake> Bye then!
[23:01] <cemc> bye
[23:01] <benedikt> cemc: ah, right. thanks.
[23:06] <mathiaz> kirkland: hey
[23:06] <kirkland> mathiaz: yo
[23:06] <mathiaz> kirkland: is there any reason why eucalyptus upstart job is shipped in eucalyptus-common?
[23:06] <mathiaz> kirkland: it relies on eucalyptus-cloud, which is not part of eucalyptus-common?
[23:06] <kirkland> mathiaz: are you questioning the existence of the job, or the location?
[23:07] <mathiaz> kirkland: on the NC there is an eucalyptus.conf file
[23:07] <mathiaz> kirkland: the location
[23:07] <mathiaz> kirkland: It seems that the upstart job should be part of eucalyptus-cloud
[23:07] <mathiaz> kirkland: rather than eucalyptus-common
[23:07] <kirkland> mathiaz: i think that's true ...
[23:07] <kirkland> mathiaz: i believe it was there so that you could just:
[23:07] <mathiaz> kirkland: as the job fails to start on NC
[23:07] <kirkland> sudo stop eucalyptus
[23:07] <kirkland> sudo start eucalyptus
[23:07] <kirkland> anywhere you are
[23:08] <kirkland> mathiaz: and it would automatically start/stop the right components for that system
[23:08] <mathiaz> kirkland: well - on an NC, sudo start eucalyptus doesn't work
[23:08] <mathiaz> kirkland: hmmm
[23:09] <mathiaz> kirkland: let me re-read the job
[23:09] <kirkland> mathiaz: correct ...
[23:09] <kirkland> mathiaz: i have a hack that makes that work
[23:09] <kirkland> mathiaz: but slangasek nacked it (softly)
[23:10] <mathiaz> kirkland: hm - well - the problem I see is that on an NC, the job doesn't do anything
[23:10] <mathiaz> kirkland: but it's still marked as respawn
[23:11] <mathiaz> kirkland: hm well - it's this line: [ -n "$services" ] || { stop; exit 0; }
[23:11] <mathiaz> kirkland: oh well - that works I suppose
[23:12] <mathiaz> kirkland: I've found out why I couldn't discover the nodes
[23:12] <mathiaz> kirkland: bug 446036
[23:12] <mathiaz> kirkland: if eucalyptus-nc fails to start because libvirt is not ready, then the eucalyptus-nc-publication job is not started either
[23:15] <mathiaz> kirkland: is automatic registration working?
[23:16] <mathiaz> kirkland: hm - seems so
[23:16] <mathiaz> kirkland: there is no need to run sudo euca_conf --no-rsync --discover-nodes anymore?
[23:19] <mathiaz> kirkland: yop - auto-registration works!
[23:19] <kirkland> mathiaz: autoreg is working for me too
[23:19] <mathiaz> kirkland: cool - I didn't know about that
[23:20] <mathiaz> kirkland: I'll update the test instructions then
[23:20] <mathiaz> kirkland: as there is not need for the sudo euca_conf --no-rsync --discover-nodes anymore
[23:25] <pilif12p> Can i dualboot desktop and server ?
[23:28] <kirkland> mathiaz: okay, so i'm looking at the nc upstart scripts ...
[23:32] <mathiaz> kirkland: IMO libvirt should be upstartified
[23:32] <mathiaz> kirkland: so that eucalyptus-nc can start on started libvirt
[23:33] <mathiaz> kirkland: the problem is that the pre-start script checks if libvirt is running
[23:33] <mathiaz> kirkland: euca_test_nc $HYPERVISOR > /var/log/eucalyptus/euca_test_nc.log 2>&1 || exit 1
[23:33]  * kirkland looks at libvirt's init script
[23:33] <mathiaz> kirkland: ^^ that fails - and then the job is just stopped
[23:34] <mathiaz> kirkland: even though it's marked respawn
[23:34] <mathiaz> kirkland: but respawn only applies to the exec part of the job (which is apache2)
[23:34] <kirkland> mathiaz: hmm, well if respawn isn't working properly we need to poke slangasek and keybuk
[23:34] <kirkland> mathiaz: hmm, should we move that to the exec?
[23:35] <kirkland> jdstrand: do you have any plans to upstartify libvirt-bin?
[23:35] <mathiaz> kirkland: well
[23:35] <mathiaz> kirkland: I think it makes sense to have apache2 executed here
[23:35] <mathiaz> kirkland: so that it's directly supervised by init
[23:36] <kirkland> mathiaz: yeah
[23:36] <mathiaz> kirkland: and the reason it fails is because of euca_test_nc $HYPERVISOR > /var/log/eucalyptus/euca_test_nc.log 2>&1 || exit 1
[23:36] <mathiaz> kirkland: ^^ that should be an upstart event *somehow*
[23:36] <mathiaz> kirkland: ie - don't start eucalyptus-nc if euca_test_nc is not succesful
[23:37] <kirkland> mathiaz: libvirt-bin looks pretty upstartable
[23:37] <mathiaz> kirkland: I wonder what euca_test_nc
[23:37] <mathiaz> kirkland: does?
[23:42] <mathiaz> kirkland: hm - seems like it mainly checks wether libvirt is running correclty