[01:17] <anthony__> hey all looking for a little ubuntu installation help
[01:25] <zxcvas> im trying to set up phpmyadmin on my ubuntu server that im connected to through ssh. the documentation keeps referring to my www fodler but i cant seem to find it.. should i put the phpmyadmin folder in /srv/ ? do i do the same for webpages?
[02:10] <samoangunner> Hi
[02:10] <samoangunner> I was hoping someone could help me.
[02:11] <samoangunner> I just installed ubuntu server and installed samba and swat
[02:11] <samoangunner> how do I launch swat?
[02:13] <jmarsden> samoangunner: Use a web browser to access it.  See https://help.ubuntu.com/community/Swat
[02:15] <samoangunner> thanx
[02:18] <samoangunner> do I run swat on the server or on a laptop
[02:19] <MTecknology> How can I figure out what OS I'm on?
[02:22] <samoangunner> in terminal type in uname -a
[02:23] <samoangunner> it will tell what your OS is
[03:11] <benedikt> would you say that it is normal for a vm guest (and host) to see relatively high load (and guest acting slow) under ~10-20 mbit datatransfer over ssh?
[03:12] <benedikt> Italian_Plumber: mario?
[03:13] <Italian_Plumber> speaking. :)
[03:13] <benedikt> just checking.
[03:14] <Italian_Plumber> actually, not-really-intending-to-login. :)
[03:16] <talcite> hey guys. I have something that's preventing my server from restarting. Can someone help me pin point it?
[03:17] <talcite> I issue the reboot command as root, and all the normal shutdown messages happen, but the machine itself never restarts
[03:19] <benedikt> how far does it go in the reboot process?
[03:21] <talcite> benedikt: sorry. multiple windows on a low res monitor. Please ping me next time. It goes as far as to say *Will now restart
[03:21] <talcite> except it never does
[03:27] <talcite> hmm a bit more info came to light. This problem is on all my servers that have an iscsi initiator. When I shut it down manually before issuing reboot, it can successfully reboot.
[03:51] <pnunn> Hi guys, can anyone give me an idea what the deal is with uploading an 8.04 image to an uec clound?  I've downloaded th uec image from uec-images.ubuntu.com but it doesn't have a kernel or rambisk image in the package, so when I try and use uec-publish-tarball it fails.  Where should I get the ramdisk and kernel images from? Ta Peter.
[03:56] <talcite> argh. I know what happened. The open-iscsi package is in rc6.d is S41, while networking is S35. What the heck guys =/
[03:56] <talcite> It'll never restart, ever
[03:56] <talcite> I moved networking to S50 and it works fine now
[06:23] <fluvvell> can anyone tell me the difference between inetd and xinetd?  And why can't I restart inetd when there is a service listed that isn't going?
[06:25] <joschi> fluvvell: xinetd is an advanced alternative to the old bsd inetd. it has more features
[06:25] <joschi> fluvvell: inetd and xinetd only start when the service definitions are ok
[06:25] <joschi> fluvvell: they can't know if it actually starts until someone opens a connection to the respective port
[06:26] <fluvvell> joschi, yeah well the ports aren't open and they're not accepting connections. The service is listed but unavailable
[06:27] <joschi> fluvvell: post your inetd.conf and relevant log entries
[06:28] <fluvvell> which log entries do you want ? its "        2000    stream  tcp    nowait  nobody /usr/sbin/tcpd /usr/sbin/nbdrootd /opt/ltsp/images/i386.img    " for the relevant line
[06:29] <twb> Anything running on top of inetd these days is probably wrong.
[06:30] <fluvvell> twb, why?
[06:30] <twb> Mainly because it's historical
[06:30] <joschi> fluvvell: 2000 should be a service name from /etc/services IMHO. when in doubt, add one yourself and use that name
[06:30] <twb> I mean, I suppose it'll still WORK
[06:31] <fluvvell> joschi, its the port I believe.
[06:31] <twb> inetd lost popularity because it creates a new process on every connection -- similar to traditional CGI.
[06:31] <joschi> fluvvell: yes. but inetd wants to have a textual description from /etc/services
[06:32] <joschi> fluvvell: you also need to correct your command line
[06:33] <joschi> fluvvell:  name-from-etc/services    stream    tcp    nowait  nobody /usr/sbin/tcpd tcpd /usr/sbin/nbdrootd /opt/ltsp/images/i386.img
[06:34] <fluvvell> so after editing, how does one restart inetd ?
[06:34] <joschi> twb: it's not wrong using inetd or xinetd for certain services. not every daemon needs to be high performance, connection and process pooling etc.
[06:34] <joschi> fluvvell: yes
[06:34] <joschi> fluvvell: argh, sorry. sudo /etc/init.d/openbsd-inetd restart
[06:35] <twb> joschi: I suppose so.  It's my knee-jerk reaction, though.
[06:35] <twb> I imagine it sucks when you get twenty classrooms with thirty hosts each all booting up at 9AM
[06:36] <fluvvell> That would be nice if I had /etc/init.d/openbsd-inetd   - Im running ubuntu 10.04
[06:36] <twb> fluvvell: it's probably /etc/init/inetd or so.
[06:36] <twb> (Note: no .d)
[06:36] <fluvvell> Ah, no thats exactly whats missing and why my system has stopped working!@
[06:36] <incidence> Hi, can I some how move postfix smtp tmp directory?
[06:37] <joschi> fluvvell: check which inetd implementation you have installed. there's more than one
[06:38] <joschi> incidence: sure. see options ending with "_directory" in postconf(5) (-> http://www.postfix.org/postconf.5.html)
[06:38] <fluvvell> joschi, yes I'm pretty sure its the openbsd one, I have another server online (not far) that *is* working , and that package is installed
[06:38] <fluvvell> finally, the port is open!
[06:40] <incidence> joschi: Thanks
[06:42] <quizme_> i have access to another computer on my LAN at 192.168.0.99 via SSH.  how can i mount it as "/vol" on my local (laptop's) filesystem
[06:42] <quizme_> basically i want to work on my desktop with my laptop
[06:43] <fluvvell> joschi, Thanks from me also - I find that being able to discuss a problem with other people at "server - level"  highly invaluable
[06:44] <joschi> quizme_: try sshfs
[06:44] <fluvvell> The answer was /etc/init.d/openbsd-inetd restart
[06:46] <MTecknology> Any of you happen to know how I coudl convert DATE_FORMAT(CURDATE(), '%Y%m%d070000') to pgsql?
[06:47] <quizme_> joschi thanks i will
[06:51] <joschi> MTecknology: `SELECT CURRENT_DATE + interval '7 hours';`
[06:55] <MTecknology> joschi: THANKS!
[07:32] <fross> as far as moving into the networking career field, what is the best distro to work with and get to know?
[07:32] <fross> I have become decent with ubuntu, but would I be at a loss if a company put me infront of a red hat or centOS?
[07:59] <joschi> fross: well, then install centOS inside a VM and play with it. but if you're into networking, you should make yourself comfortable with IOS, catalyst, JunOS and alike.
[08:00] <joschi> fross: also playing around with different networking daemons like quagga should be on your list.
[08:15] <glick> hey im trying to connect to a smtp server with my mail client and i get the error STARTTTLS command failed due to local problem
[08:15] <glick> anyone have any idea what that means?
[08:15] <glick> am i missing a program?
[08:16] <thesyko> hi there guys
[08:16] <twb> glick: STARTTLS is the signal to enable SSL.
[08:16] <twb> Check your SSL config
[08:17] <glick> twb, which ssl config?
[08:17] <glick> at the server or client?
[08:17] <thesyko> i would like to ask, why is example configuration file for ubuntu that is in /usr/share/doc/mysql-server-5.1/example cannot be used ?
[08:17] <thesyko> i tried the configuration file my-small.cnf and it is not working
[08:18] <Jeeves_> thesyko: Read the errors in /var/log/daemon.log.2.gz
[08:18] <thesyko> btw, i'm using ubuntu 10.04 desktop with mysql 5.1 server
[08:18] <Jeeves_> thesyko: Read the errors in /var/log/daemon.log
[08:18] <thesyko> ok
[08:18] <twb> I don't know.
[08:18] <twb> If you were using msmtp(1) you could run it with -d to test
[08:18] <twb> glick: msmtp is an smtp client
[08:19] <Jeeves_> There is no client configuration for starttls, afaik
[08:19] <Jeeves_> so it would be the server configuration
[08:20] <thesyko> now i'm using back the default configuration file that comes with mysql 5.1
[08:20] <thesyko> and it works
[08:21] <twb> Jeeves_: I was thinking something along the lines of the server using a self-signed cert
[08:22] <fross> joschi: Hey, thanks for the advice. Much appreciated
[08:22] <glick> http://dpaste.com/232699/
[08:22] <glick> thats my main.cf file
[08:22] <glick> i think ssl is set up correctly
[08:22] <twb> glick: you're using sendmail?
[08:22] <glick> no postfix
[08:22] <twb> Oops, nm
[08:23] <Jeeves_> glick: /var/log/mail.err would probably be able to tell you more
[08:24] <glick> Jeeves_, there are no error messages
[08:24] <glick> mail.log isnt much more helpful either
[08:27] <twb> Try #postfix
[09:17] <glick> i followed the ubuntu server guide to installing and configuring postfix
[09:18] <glick> i got certs from go daddy
[09:18] <glick> yet when when i try to connect to the smtp server using ssl or tls it wont connect
[09:18] <glick> yet if i connect without ssl it works
[09:18] <glick> i definately need to use ssl
[09:21] <joschi> glick: there definately should be some log entries about the issue. you can try to increase postfix's verbosity by adding "-v" to the processes in your master.cf (see master(5), http://www.postfix.org/master.5.html)
[09:21] <joschi> glick: also make sure that postfix can read the SSL certificate and the private key files and that these files are correct (e. g. don't use the CSR...)
[09:22] <glick> joschi, what do you mean dont use the csr?
[09:22] <joschi> glick: you've created a certificate request and sent it to go daddy to get your certificate.
[09:22] <glick> oh yeah
[09:22] <joschi> glick: don't accidentally use the CSR in your postfix configuration
[09:22] <glick> no im not using the csr
[09:23] <glick> i musing the crt and key files
[09:35] <glick> yayayyyy i finally got it working!
[09:35] <glick> i am the greatest!
[10:46] <xampart> how can i find out my machines power supply information?
[10:55] <twb> xampart: /sys/devices/class/BAT0 or so?
[10:55] <twb> xampart: or lshw
[11:18] <glick> in postfix sending an email to one of my email addresses gets rejected, i dont understand why
[11:18] <glick> in my main.cf i have this line which i suspect is to blame
[11:19] <glick> smtpd_recipient_restrictions = check_recipient_access hash:/etc/postfix/recipient_access,reject_unauth_destination
[11:22] <xampart> twb: is there a way to confirm that hdd's get enough power?
[11:23] <twb> I don't know
[11:35] <xampart> is it usual lshw gets stuck
[12:24] <lili> join/ #ubuntu-server
[12:24] <lili> . plz help me to config my apache2 for perl on ubuntu 10.4.!!!!!!!!!1
[12:24] <lili> hi
[12:25] <lili> plz help me
[12:26] <yaboo> just upgraded ubuntu server to 10.4 and seems my mac cannot see any of my samba shares now
[12:27] <yaboo> same for windows 7 machines also
[12:27] <yaboo> has samba changed again
[12:27] <lili>  plz help me to config my apache2 for perl on ubuntu 10.4.!!!!!!!!! i'm waiting for answering of frineds.
[12:28] <glick> anyone know why given this config file i cant send an email to customer_support@mydomain
[12:28] <glick> http://dpaste.com/232768/
[12:29] <glick> it comes back as rejected
[12:29] <glick> code 550
[12:32] <soren> glick: Where are you sending from? Where are you sending to? What's the exact error message? Does the user exist? Does DNS work? etc. etc.
[12:32] <glick> the user exists, dns works, im sending from myaddress@google.com to customer_support@mydomain.com
[12:34] <soren> You're running http://www.mydomain.com, but you haven't got a working MX?
[12:34] <soren> That seems hard to believe.
[12:34] <soren> Somehow.
[12:34] <glick> no im not running that domain
[12:34] <glick> lol
[12:34] <glick> that was a variable
[12:34] <glick> for my actual domain
[12:34] <soren> Then stop halfway anonymising your infor.
[12:35] <soren> Well, so you say.
[12:35] <soren> My current guess is: You're not using the right domain.
[12:36] <sanderj> Anyone knows about Ubuntu elastic cloud?
[12:36] <yaboo> also how do I setup ubuntu server to work with my monitor, the default does not work
[12:37] <sanderj> I'm wondring how can I install it into one huge machine.. instead of two ones.
[12:37] <soren> glick: It's simple, really. Your config file saysmyhostname = mail.boss.com
[12:37] <soren> whoops
[12:37] <soren> glick: It's simple, really. Your config file says myhostname = mail.boss.com, but you say you're sending to customer_support@mydomain.com. They don't match. You lose.
[12:42] <yaboo> cannot even get console on my server via keyboard, HOW DO I CHANGE THE VIDEO DISPLAY
[13:22] <zul> morning
[13:29] <hggdh> morning zul
[13:31] <zul> hey hggdh
[13:41] <hallyn> $(*&%$&*%)$&
[13:41] <hallyn> saturday i made backups to dmcrypt'ed usb disk.  sunday, repartitioned and installed maverick.  today, figured i'd recover some config files.  of COURSE, maverick wno't mount the backups.
[13:42] <hallyn> just as it did when i went from karmic to lucid
[13:42] <hallyn> maybe a lucid kvm partition will read them
[13:44] <Daviey> ;S
[13:44] <hallyn> Daviey: welcome back!
[13:45] <Daviey> thanks hallyn o/
[13:46] <hallyn> recon' this means i should find my mumble config info somewhere else so we can scrum in a bit :)
[13:55] <Omahn> Why oh why does the landscape client take up 937 megs of virtual memory......
[13:56] <Omahn> (as one example)
[13:58] <hggdh> Omahn: resident memory?
[13:58] <Omahn> hggdh: About 54.
[13:59] <Omahn> hggdh: But it does mean the swap file is pretty much full, just with landscape.
[14:04] <hggdh> Omahn: not necessarily. WHat does 'vmstat'show (this will be one single line now)?
[14:04] <Daviey> smoser, Are you online yet?
[14:04] <Omahn> hggdh: This is a different box, I've turned it off on the other box:
[14:05] <Omahn> procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa 2  0    104 170824  85436 357428    0    0     1     8    2    3  4  4 92  0
[14:05] <smoser> Daviey, here. whats up ?
[14:05] <smoser> and welcome back
[14:05] <Daviey> smoser, ta.  Wanted to ctach up with you, about the kernel upgrade in euca issue?
[14:05] <Omahn> hggdh: That vmstat is from a box only up for 97 days, the other was up nearly 250.
[14:06] <smoser> hm..i think that its fine now. you're talking about https://bugs.launchpad.net/ubuntu/+source/qemu-kvm/+bug/615529 ?
[14:07] <smoser> Daviey, ^
[14:07] <Daviey> smoser, Oui
[14:08] <Omahn> hggdh: Fresh startup on one of my boxes consumes over 350 megs of virtual memory. Just seems a tad excessive.
[14:08] <smoser> Daviey, i think that realistically, the solutoin i put in place is the right one.
[14:09] <Daviey> smoser, Ok.. so you have no concerns at the moment?
[14:09] <smoser> no, not really.
[14:09] <Daviey> smoser, Just want to understand if i can remove it from my list of things to watch :)
[14:09] <Daviey> awesome, top stuff smoser
[14:09] <smoser> basically i had forgotton that 'boot=on' was important
[14:10] <Daviey> ahh
[14:10] <Daviey> boot=on is now added somewhere?
[14:11] <Daviey> hallyn, Have you spotted bug 622588?
[14:11] <hggdh> Omahn: and you can tag this VM to landscape?
[14:12] <Omahn> hggdh: By 'tag' do you mean register?
[14:13] <zul> hallyn: ping....er...nevermind
[14:13] <Omahn> hggdh: If so, yes, it's registered.
[14:13] <hggdh> Omahn: I mean you are sure this VM usage is from landscape
[14:13] <Omahn> hggdh: Absolutely.
[14:13] <Omahn> hggdh: Top shows VIRT/SWAP usage for each of the landscape processes. It's also the same across all our servers.
[14:14] <hallyn> Daviey: i hadn't yet, no
[14:14] <Daviey> hallyn, That could have potential to be a humdinger.
[14:15] <hallyn> Daviey: isn't that the error it has always given, as long as i've used kvm in ubuntu?
[14:15] <hallyn> it proceeds to work no?
[14:16] <Daviey> hallyn, agreed, i've seen that warning lots... but he bug reporter suggests total malfunction. I suspect the error he stated is a red herring.
[14:16] <hggdh> Omahn: good enough for a bug, I guess. Could you open one please?
[14:16] <Daviey> hallyn, apport being less than helpful :/
[14:17] <hallyn> asking for clarification...
[14:17] <Daviey> hallyn, It was actually zul that spotted that bug
[14:17] <Omahn> hggdh: Will do.
[14:17] <kim0> hey folks .. we now have a "cloud" forum, and a first question :) http://ubuntuforums.org/showthread.php?t=1558900
[14:18] <Daviey> kim0, oh nioes
[14:18] <kim0> :)
[14:18] <kim0> Who wants the owner of answering the very first question hehe
[14:19] <Omahn> hggdh: Looks like a bug already exists.
[14:19] <Omahn> https://bugs.launchpad.net/landscape-client/+bug/459243
[14:19] <Daviey> kim0, Well volunteered :)
[14:19] <kim0> Yaay
[14:20] <kim0> Daviey: hey there .. shouldn't you be in #ubuntu-cloud as well :)
[14:20] <Daviey> kim0, indeed
[14:20] <kim0> jump in
[14:23] <hggdh> Omahn: yes indeed, describes your issue. Can you please add data there?
[14:24] <hggdh> Omahn: how long the client has been running, mem allocation, etc.
[14:24] <Omahn> hggdh: Done.
[14:24] <Omahn> hggdh: Ah, it doesn't show startup date. I'll add that.
[14:34] <a_ok> I do iscsi-boot. this works fine but the system never does a clean shutdown. It stops open-iscsi to early resulting in ext4 errors (quite logically since the root parition is gone)
[14:34] <a_ok> how do I fix this?
[14:37] <yaboo> is there a fix for getting a blank screen during boot up of server in 10.4?
[14:37] <twb> Yeah, roll back to 8.04
[14:37] <a_ok> ...
[14:38] <twb> If you've managed to install before that happened, you could try blacklisting the fucking KMS drivers
[14:40] <sherr> No need for swearing ...
[14:41] <a_ok> Is there anyone who had the same problem as I have?
[14:41] <sherr> Is iscsi-boot an upstart job or traditional init?
[14:42] <a_ok> upstart most likely
[14:42] <a_ok> rebooting
[14:43] <a_ok> eg turned power off and on as it hangs
[14:43] <sherr> Just wondering whether the order or startup/teardown can be modified, so the filesystems are unmounted before iscsi is shutdown/unloaded
[14:43] <a_ok> Yeah that would be helpfull
[14:44] <a_ok> I don't know if it needs write access to shutdown open-iscsi by the way
[14:46] <a_ok> perhaps I'm better off letting it die with the system
[14:50] <a_ok> what was the offical way of adding and removing services from runlevels agian?
[14:51] <sherr> update-rc.d
[14:51] <a_ok> thanks
[14:52] <a_ok> I always confuse it with the gentoo one. rc-update
[14:52] <twb> update-rc.d doesn't apply to upstart-based jobs, though (AFAIK)
[14:57] <a_ok> really I thought it used init's stucture to handle boot order etc
[14:57] <a_ok> and what to boot and what not
[15:00] <twb> Well, guess what, Ubuntu decided they didn't like things working
[15:05] <stittel> Hi! Is there a way to check on the web which packages of 10.04 LTS are considered as "server packages" and hence receive 5 years of support? The "supported software" link on http://www.ubuntu.com/server doesn't work.
[15:06] <twb> stittel: aptitude tells you nowadays
[15:06] <twb> I suspect that means it's in the Packages database and you can just use grep-dctrl
[15:07] <stittel> twb: I would like to check before installing Ubuntu, i.e. I am not running Ubuntu yet, so I can't check it with these command.
[15:07] <twb> Hmm, that's odd, now I can't see it
[15:07] <twb> stittel: there's also a Python script I was using before.
[15:07] <twb> ...re grep-dctrl, I guess it's in a secondary database that's only installed on desktops by default.
[15:08] <twb> stittel: bzr clone http://bazaar.launchpad.net/%7Enijaba/ubuntu-maintenance-check/trunk/
[15:10] <stittel> twb: As I said, I am not running Ubuntu yet and would like a way to check the packages supported as "server packages" on the web. http://packages.ubuntu.com doesn't seem to provide that information and on http://www.ubuntu.com/server the "supported software" link is dead.
[15:11] <stittel> twb: Isn't there another place to check it? Even a text file containing all "server" packages would make me happy. :)
[15:11] <jpds> stittel: Any server package that's in main.
[15:11] <smoser> kim0, what do i win for giving the first answer on forums ?
[15:11] <sherr> stittel: https://help.ubuntu.com/community/ServerFaq#What%20%28packages/repositories%29%20are%20maintained%20%28supported%29?
[15:12] <kim0> smoser: Yaaaay :)
[15:12] <sherr> Ugh - maybe just check here : https://help.ubuntu.com/community/ServerFaq
[15:12] <zul> smoser: a hug!
[15:12] <kim0> everyone .. please give smoser a hug
[15:12]  * kim0 hugs smoser
[15:12] <stittel> jpds: It's not always clear to me what is considered a server package or not. For instance, "SpamAssassin" can be run as a daemon or just as client as well es on a desktop for local or on a server for server-based spam filtering.
[15:13] <a_ok> I have said it once and I will say it again, apt should not mess with runlevels or start/restart services on its own
[15:15] <stittel> sherr: Well, OK, so it's just these packages http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.lucid/server-ship because http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.lucid/supported-common is empty.
[15:15] <stittel> And that also would mean that exim4 is not supported for 5 years, even though it is a server package from main. :-/
[15:16] <twb> stittel: yes
[15:16] <twb> The "five year support" thing is a crock
[15:16] <Omahn> twb: Hmm, I think our legal department would dispute that.
[15:16] <smoser> what is the hug : beer conversion rate ?
[15:17] <kklimonda> stittel: the list you have given looks short
[15:17] <Pici> exim4's 'supported' field claims '5y'
[15:18] <kklimonda> stittel: it contains less than 440 packages and dapper has support for over 780 - exim4 is one of them
[15:18] <stittel> kklimonda: Then something is wrong, because https://help.ubuntu.com/community/ServerFaq#What%20%28packages/repositories%29%20are%20maintained%20%28supported%29 states "Server packages are maintained for 5 years, and defined as all packages that are part of the the seeds "server-ship" or "supported-common" are supported for five years."
[15:18] <kklimonda> stittel: http://people.canonical.com/~ubuntu-security/dapper/supported.txt this list may give you an idea of sort what is supported.
[15:19] <kklimonda> stittel: but I think the only good way of checking support status at this time is to install Ubuntu yourself and run ubuntu-support-status..
[15:20] <kklimonda> but then I just got an exception trying to run it on my server :D
[15:20] <twb> kklimonda: last time I looked that python script I linked to worked
[15:20] <stittel> kklimonda: I could try it on my Ubuntu netbook if "ubuntu-support-status" works without installing the package.
[15:21] <stittel> kklimonda: According to your list some stuff I run like SquirrelMail or SpamAssassin are not supported.
[15:21] <stittel> Uh, actually SquirrelMail is just "universe".
[15:21] <kklimonda> stittel: that's the problem with this list - it's for dapper and spamassassin was still in universe in dapper
[15:23] <kklimonda> stittel: got moved to main in jaunty so it's supported for 5 years on lucid
[15:23] <sherr> The "ubuntu-support-status" program just aborts on Lucid for me - "Exception: No date tag found"
[15:23] <twb> Maybe that script isn't supported
[15:23] <kklimonda> sherr: yeah, it's reported already
[15:25] <stittel> Thank you everybody! Since a lot of packages (SquirrelMail, SpamAssassin, jabberd, openntpd) don't seem to be supported, I guess I will stick to Debian for the moment, even though I don't like their unpredictable release and support policy.
[15:28] <kklimonda> stittel: spamassassin is supported (other are not though)
[15:29] <hallyn> GAH!   a lucid kvm vm can mount the cryptsetup partition on external usb, the maverick host can't
[15:29] <twb> stittel: you mean "release when it's ready" as opposed to "release when it's not ready" ? ;-)
[15:29]  * twb glares upstart jobs for NFS
[15:31] <stittel> twb: Well, to know the release date of the next Debian release some time in advance, you need a crystal ball. And security support for the old release is "release date of the new release + 1". That makes the entire thing a bit unpredictable. :)
[15:32] <twb> stittel: I realize it's a problem
[15:32] <stittel> twb: And if you should choose to install Debian now, you can expect less than 2 or even 1 1/2 years of security support, because the next release can be expected to come out within the next 6 to 12 months.
[15:32] <twb> stittel: I just have't been particularly impressed since we switched to Ubuntu 8.04 back in late `07
[15:33] <Omahn> IMHO A lot of the problems with 10.04 is down to upstart and the lack of a complete migration to upstart. The current half attempt causes us many issues.
[15:34] <joschi> twb: I doubt that you switched to ubuntu 8.04 in '07 - unless you like using bleeding edge alpha/beta distributions ;)
[15:35] <twb> We don't start deploying a twelve-month project on a release that's going to be twelve months old
[15:35] <twb> The prisoners didn't actually get to use the system until well after April, but that doesn't mean we weren't targeting it and developing and testing on it.
[15:38] <sherr> Upstart seems to be the main cause of trouble for me in 10.04 i.e. LP bug 350936
[15:38] <twb> sherr: it and plymouth, perhaps
[15:38] <sherr> I got round that happily now using "molly-guard"
[15:38] <twb> molly-guard is a good Best Practice
[15:38] <sherr> Yes. I'm sticking it on everything now.
[15:39] <twb> I dunno about it fixing bugs, but I use it to avoid fat-fingering the wrong machine into oblivion
[15:39] <jdstrand> sherr: that shouldn't have anything to do with upstart. I don't think libvirt ever did that on shutdown, though I could be wrong
[15:41] <jdstrand> that bug is not easy to fix cause you can send an acpi shutdown event, but that doesn't mean the guest did shut down in a timely fashion, or even responds to acpi events
[15:41] <jdstrand> I'm not the one looking at that bug, but afaics, it isn't particularly an upstart issue
[15:43] <sherr> If you read through the bug report, it appeared very hard to use "virsh" bfore a shutdown, to even attempt a clean shutdown of guests. Stuff seems to be killed too quickly (/etc/init.d/sendsigs).
[15:43] <sherr> A lot of people, including me, wasted hours trying to sort it out. Many reboots ...
[15:44] <sherr> I learned quite a bit in the process though (including a bit about upstart) :-)
[15:58] <jdstrand> soren: hey. so I've noticed your uml work with libvirt, which is cool. however, will openstack by using uml primarily? will it still support kvm?
[16:01] <a_ok> sherr: for your information, upstart looks at the rd.* dirs to decicde what to start and probably also uses it to determen boot order. althought it might calculate it trough the dependencies (stated in the initscript)
[16:04] <sherr> a_ok: thanks. It's a pain trying to get a working "virsh list" during machine reboot or shutdown though. That's without modifying both libvirt-bin.conf and sendsigs (as per comment #8 in bug).
[16:05] <sherr> How do people shutdown guests on host shutdown or reboot with KVM/qemu? Is this stuff run in the "enterprise"? e.g. ubuntu-cloud? Or is the power just pulled from the guests? Seems wrong.
[16:11] <SpamapS> ttx: wb btw
[16:12] <ttx> SpamapS:  o/
[16:17] <webpragmatist> hey guys with ufw is it possible to allow a app rule through only an interface?
[16:18] <kpettit> I need to start looking at building out and managing a bunch of ubuntu servers. They will all have alot of the same code, but some different stuff as well.  Any suggestions for managing config's, deploying updates, etc?
[16:18] <webpragmatist> what i am trying to do is allow OpenSSH only through eth0
[16:20] <kpettit> webpragmatist: I haven't tried ufw, if it doesn't there are a ton that will.  That's a pretty common feature
[16:20] <webpragmatist> kpettit: well the question is specific to ufw
[16:20] <webpragmatist> ufw has profiles for apps
[16:21] <the_archit3ct> webpragmatist: just don't open net port 22 :P
[16:21] <webpragmatist> the_archit3ct: I understand the idea... i'm just trying to use ufw with the apps
 what i am trying to do is allow OpenSSH only through eth0
[16:21] <the_archit3ct> answering that
[16:22] <jdstrand> webpragmatist: yes
[16:22] <ssureshot> is there a prebuilt package for 389 directory server for lucid yet?
[16:22] <webpragmatist> jdstrand: can your provide the syntax for using allow on interface for app?
[16:23] <MTecknology> !search maria
[16:23] <MTecknology> !search mariadb maverick
[16:23] <webpragmatist> ufw allow OpenSSH on eth0 ?
[16:23] <kklimonda> ssureshot: I don't think so
[16:23] <jdstrand> webpragmatist: sudo ufw allow in on eth0 to any app OpenSSH
[16:24] <MTecknology> Is mariadb going to be available in Ubuntu 10.10?
[16:24] <jdstrand> webpragmatist: it requires the extended syntax when specifying an interface. see man ufw for details
[16:24] <ssureshot> kklimonda: bummer guess I need to downgrade my test box
[16:24] <webpragmatist> jdstrand: ya i was trying to piece it together from man wasn't working thanks
[16:25] <jdstrand> sure
[16:26] <webpragmatist> jdstrand: i kinda think that the app rules are annoying
[16:26] <webpragmatist> jdstrand: is there an extended status?
[16:26] <jdstrand> webpragmatist: they are totally optional
[16:26] <jdstrand> webpragmatist: yes. ufw status verbose
[16:26] <jdstrand> webpragmatist: also ufw status numbered
[16:27] <kpettit> is svn or similar good at keep config file histories?  Was thinking of using some sort of source code repository to manage config files on multiple linux machines.  Mainly for backup, cloning, testing, etc
[16:27] <webpragmatist> ah okay now it's no longer annoyying :)
[16:27] <jdstrand> kpettit: take a look at etckeeper
[16:27] <webpragmatist> jdstrand: I assume a limit is the same as an allow... so you wouldn't need both
[16:27] <kpettit> jdstrand: checking...
[16:28] <jdstrand> kpettit: oh, actually, that is for the local machine. scratch that
[16:28] <jdstrand> webpragmatist: you can't have both. you choose one of allow, deny, reject or limit
[16:28] <kpettit> yeah I have 6 or so servers now.  ANd could possible scale up to a hundred or so.  SO I need to setup a system to help me put out default configs and keep history or changes and such.
[16:29] <sherr> kpettit: does something like cfengine or puppet support that?
[16:29] <sherr> I'm thinking about that as well.
[16:29] <kpettit> cfengine scares me.  I haven't tried puppet
[16:29] <kpettit> Last place I worked at did cfengine and they all hated it with a passion.  I'll check on puppet
[16:30] <webpragmatist> kpettit: i use csync2
[16:30] <webpragmatist> but i only run two servers
[16:30] <webpragmatist> and it's hardly as powerful as puppet afaik
[16:30] <kpettit> All of my servers are on cloud.  Rackspace and GoGrid.  So I'll probally use there api's to manage cloning and backup's.  But I wanted to have something seperate to do config stuff
[16:30] <kpettit> webpragmatist: I'll check it out, thanks
[16:31] <kpettit> ah, puppet won't do the trick.  IT keeps things in sync and I need the ability to have things different.
[16:33] <sherr> http://www.google.com/url?sa=t&source=web&cd=13&ved=0CCEQFjACOAo&url=http%3A%2F%2Fen.wikipedia.org%2Fwiki%2FComparison_of_open_source_configuration_management_software&ei=0ZRyTMbWHNGk4QaviZTeCA&usg=AFQjCNFfJc8CNQCE8Io_6JPzvpf9pUtSfA
[16:34] <kpettit> sherr: Sweet, thanks
[17:04] <hggdh> Daviey: welcome back, and please have a look at bug 622818 ;-)
[17:04] <Daviey> hey hggdh !  Top work last week :)
[17:08] <Daviey> hggdh, Interesting!
[17:11] <mathiaz> Daviey: heya! welcome back!
[17:11] <Daviey> yo mathiaz o/
[17:11] <mathiaz> ttx: back welcome well as! heya!
[17:13] <hggdh> Daviey: I thought that you might need some incentive after a rather good time last week :-)
[17:14] <hggdh> Daviey: any news from your RT request for euca access?
[17:14] <Daviey> hggdh, no - none.. i think they are probably ignoring my RT tickets now :)
[17:16] <hggdh> heh
[17:37] <kpettit> have any of you guys tried amqp?  Was thinking of using that for admin type tasks to get tasks out to multiple servers
[17:37] <RoAkSoAx> smoser: ping?
[17:37] <smoser> RoAkSoAx, here
[17:39] <RoAkSoAx> smoser: are desktop images also to be testdrived?
[17:39] <smoser> they can be
[17:40] <smoser> i care less about that
[17:40] <smoser> and, they're not official releases (we only do dailies of them)
[17:40] <RoAkSoAx> smoser: ok then. I won't put them for now then. Now, I saw your comment on the bug report. So the only thing we need is the .img and nothing else from the tarball right?
[17:41] <smoser> -floppy
[17:41] <smoser> RoAkSoAx, you need the -floppy also: kvm -boot a -fda ${name}-floppy -drive file=disk.img,if=virtio
[17:43] <soren> jdstrand: oh, the uml stuff is just so that I can test stuff in the cloud, not for production use.
[17:43] <RoAkSoAx> smoser: ok so image and floppy. So, would it be possible to use http://uec-images.ubuntu.com/maverick/current/unpacked/ to download only the image and -floppy instead of the whole tarball or is that just a temporary folder?
[17:45] <smoser> the tarball that you want is as reported by http://uec-images.ubuntu.com/.manifest-daily
[17:45] <jdstrand> soren: cool. I was kinda concerned since you don't get MAC protections (ie, apparmor and selinux) when not using qemu
[17:46] <RoAkSoAx> smoser: yeah but wanted to save steps but only downloading the neccesary files. Anyways, I'm waiting for the tarballs to sync to work on the preparing process. SO should be matter of hours till I have it working
[17:47] <smoser> RoAkSoAx, well, that actually wouldn't save you a step
[17:47] <smoser> you still have to extract a tar file
[17:48] <smoser> i'd suggest using the .tar.gz . the unpacked will be there (ie, its not temporary).
[17:49] <smoser> but your saving of a step is at expense of a second download.
[17:49] <smoser> i dont have any strong feelings, but the .tar.gz is generally a more useful thing to download
[17:53] <RoAkSoAx> smoser: yeah well I guess I'd end up untarring .img and -floppy only instead of the whole tarball
[18:02] <c0dem4gnetic> how do i set up wireless networking during the install of ubuntu server? can i?
[18:33] <RoAkSoAx> smoser: btw, http://uec-images.ubuntu.com/.manifest-daily is not showing maverick tarballs anymore
[18:34] <smoser> hmm..
[18:34] <smoser> interesting
[18:38] <jono> hey all, just an FYI, UDS is now announced: http://uds.ubuntu.com/ - do go and apply for sponsorship
[19:09] <bogey> I have 1.8tb raid volume mounted at /var/iscsi and I want to make an iscsi target. What command could I use for this?
[19:10] <bogey> 600gb iscsi target with a u/p
[19:14] <sherr> bogey: probably similar to the way Falko describes on howtoforge
[19:15] <bogey> sherr, they are suggesting to use img files
[19:15] <sherr> similar convept though I would have thought?
[19:16] <bogey> well i dont know how to use /dev/sda but restrict it to 600gb
[19:18] <soren> jdstrand: Yet. :)
[19:18] <slug> hi guys, i'm looking for a tutorial to setup ldap on ubuntu 10.04 that actually works, any idea?
[19:18] <slug> ldap server i mean.
[19:19] <sherr> slug: there's an Ubuntu server guide - did you check that?
[19:19] <slug> sherr: yes, it doesn't work. if i follow it
[19:21] <slug> sherr: get strange invalid credentials issues
[19:23] <slug> my problem is that ubuntu puts the default configuration into a very strange point. i managed to configure for other distributions
[19:27] <sherr> slug: Ubuntu will follow Debian. Have a look on howtoforge (search) because there's often decent LDAP tutorials there.
[19:27] <slug> sherr: ok, i started there, but then i decided to take a look at what the server docs had to say and ran into problems. i'm trying that again, let's see how it goes. thanks :)
[19:28] <sherr> Good luck.
[19:31] <wastl> hi
[19:40] <slug> sherr: unfortunately ubuntu ldap/ldapd is fairly different from debian :(
[19:40] <wastl> indeed
[19:40] <wastl> ldap is a pain in the ass to setup anyways *g*
[19:42] <slug> wastl: did you manage to make it work with 10.04 ?
[19:43] <bogey> I have 1.8tb raid volume mounted at /var/iscsi and I want to make an iscsi target but only use 600gb. What command could I use for this?
[19:48] <wastl> slug,: no but i managed to make it work with debian lenny :)
[19:48] <wastl> slug: gave up on ubuntu
[19:49] <slug> wastl: great ;)
[19:49] <RoAkSoAx> smoser: it's almost ready, I just need the maverick images available in .manifest-daily to be able to test
[19:49] <wastl> slug: as far as ldap is concerned
[19:49] <smoser> yeah, i'll fix that really quick her.e
[19:49] <slug> wastl: i had debian before on that machine, but moved to ubuntu since i needed more recent software
[19:49] <raubvogel>  slug: ldap server or client?
[19:50] <slug> raubvogel: server
[19:52] <slug> raubvogel: client doesn't need much more than setting up ldap.conf and other basic stuff
[19:56] <smoser> RoAkSoAx, it shoudl bie fixed now.
[19:56] <smoser> unfortunately, i really dont know why it had gotten borked
[19:56] <RoAkSoAx> smoser: thank you :)
[19:57] <raubvogel> slug: I setup a ldap replication server (master is a gentoo box which will be replaced with an ubuntu one as soon as I have time to)
[19:58] <slug> raubvogel: hehe, that's funny. i have a gentoo server running ldap just fine. but we got a new machine and I installed ubuntu instead. but having trouble with ldap
[19:58] <raubvogel> Only thing *special* I did was to use the old style (slapd.conf) instead of ldif
[19:58] <slug> funny=coincidence
[19:58] <raubvogel> I thought so ;)
[20:00] <raubvogel> So, are you replacing the gentoo box compeltely or using the ubuntu one as slave server?
[20:00] <wastl> I thought that the latest versions of ldapd don't use slapd.conf the way it was used in earlier ones anymore
[20:00] <wastl> so due to this you _must_ use ldif instead
[20:02] <raubvogel> If you read the /etc/default/slapd file it says you can still specify the slapd.conf. They frown upon it but it can get you going :)
[20:02] <raubvogel> I need to do some errands and will be back. Need also to get my laptop charger and it is pouring rain
[20:03] <slug> raubvogel: replacing, hopefully
[20:03] <slug> raubvogel: if i can make it work
[20:06] <JamesHarrison> If I've got an ubuntu server that just kernel panic'd (Remote, so no idea what the panic was or what caused it), are core dumps written anywhere?
[20:15] <webpragmatist> if i am syncing two servers... with ufw can i just copy /etc/ufw between the two and restart ufw?
[20:15] <webpragmatist> i'm using csync2 to do this
[20:15] <jdstrand> webpragmatist: almost. you also need /lib/ufw/*rules
[20:16] <webpragmatist> those are the current enabled rules?
[20:16] <jdstrand> webpragmatist: oh, and /etc/default/ufw
[20:17] <jdstrand> webpragmatist: (the one in /etc/default/ufw only if you manually updated it)
[20:17] <jdstrand> webpragmatist: the ones in /lib/ufw/ are the ones you added with the cli command. eg 'sudo ufw allow OpenSSH' <-- that rule goes in /lib/ufw
[20:17] <webpragmatist> okay and the ones in /etc are?
[20:18] <jdstrand> webpragmatist: but you can also fine-tune things in /etc/ufw/*rules that aren't supported by the ufw cli command. see 'man ufw-framework' for details
[20:18] <webpragmatist> oh i see
[20:25] <webpragmatist> jdstrand: ufw status is not necessarily the inacted firewall? guess the only way is to sift through iptables -L?
[20:25] <webpragmatist> meaning if it hasn't been reloaded it wouldn't show you
[20:25] <Wintemute_> Hi all :)
[20:25] <Wintemute_> Any one knows where I can set the mount opitions of the system mounts?
[20:25] <Wintemute_> Options*
[20:26] <Wintemute_> Or I should write a script to change they?
[20:26] <jdstrand> webpragmatist: it should show you the running firewall, but it only shows you the rules you added via the cli command. you can use 'ufw show raw' for a complete iptables dump of all tables
[20:26] <guntbert> Wintemute_: /etc/fstab
[20:26] <Wintemute_> Not this
[20:26] <Wintemute_> none on /lib/init/rw type tmpfs (rw,nosuid,mode=0755)
[20:26] <Wintemute_> ;)
[20:27] <lamont> sigh... I hate that check almost as much as I hate that I didn't put it in in the first place.
[20:27] <smoser> SpamapS, just an fyi, http://www.mail-archive.com/grub-devel@gnu.org/msg16187.html was what i'm trying to mimick
[20:27] <Wintemute_> nor that proc on /proc type proc (rw,noexec,nosuid,nodev)
[20:28] <guntbert> Wintemute_: I have no idea about what you are talking :-(
 Are you on a ubuntu?
[20:28] <guntbert> Wintemute_: yes
[20:28] <Wintemute_> type mount on the term
[20:28] <Wintemute_> you will see your mountpoints
[20:28] <Wintemute_> plus the system ones
[20:28] <Wintemute_> (Like /dev)
[20:29] <Wintemute_> I want to change the mount options of the system ones
[20:29] <Wintemute_> mainly to ro or noexec
[20:29] <glen1> hey
[20:30] <guntbert> Wintemute_: I never heard of  "system" mount points, and no I don't know how to change them. Why do you want to do that?
[20:30] <Wintemute_> Security
[20:30] <glen1> if I have a webserver, how do I stop users from putting up a website on through their ftp accounts
[20:30] <Wintemute_> avoid file injection and execution
 disable ftp
[20:31] <Wintemute_> or remove the apache module
[20:31] <guntbert> Wintemute_: you don't change the options for the mount points but for the filesystems
[20:31] <smoser> Wintemute_, well, /etc/fstab or for mount points that don't appear there, /lib/init/fstab is where they come from (on recent ubuntu)
[20:31] <glen1> I have a server and serveral ftp accounts on it, im not sure how to stop them from creating a website on it
[20:31] <glen1> currently only used for css maps
 Thanks!!! :)
[20:31] <smoser> but guntbert is probably more being your friend, telling you the right way to do things.
[20:31] <webpragmatist> jdstrand: that's interesting because i never did ufw reload and a simple sync of thes conf files it showed the correct firewall
[20:31] <smoser> you probably dont' want to mess with /lib/init/fstab
[20:32] <webpragmatist> jdstrand: is it not necesssary to ufw reload ?
[20:32] <kim0> hi folks .. when /sys/kernel/mm/ksm/run == 1 .. that means ksm is active ?
[20:32] <Wintemute_> I want to see if is possible actually
[20:33] <jdstrand> webpragmatist: check with 'ufw show raw'. I did say 'should show' cause if you change stuff outside of the cli command out from under it, 'status' won't be accurate
[20:33] <Wintemute_> put a ro/noexec server to run
[20:33] <guntbert> Wintemute_: in /dev there are the devices, you mount them somewhere else, I don't know of any way to change options for the devices themselves
[20:33] <webpragmatist> jdstrand: i see okay... thanks
[20:34] <Wintemute_> they are not too much problem...
[20:34] <Wintemute_> owner root perms 755
[20:35] <smoser> kim0, it would appear that it means the kernel should be running ksm. not that it is being run "RIGHT NOW"
[20:35] <Wintemute_> But I don't like this: /lib/init/rw
[20:35] <nuno_> glen1: you can disable public_html
[20:35] <glen1> ahh, ill have a go :D
[20:35] <kim0> smoser: the thing is .. I have it "1" for a day now .. and pages_shared = 0
[20:35] <Wintemute_> or /var ones...
[20:35] <kim0> smoser: I'm not running VMs .. rather running 2 JVMs on a low ram instance
[20:35] <Wintemute_> no sense to have exec on then (in a hardened server)
[20:36] <smoser> kim0, hm..
[20:36] <kim0> smoser: I was kinda hoping to get some memory savings .. shared pages = 0 looks suspicious
[20:36] <smoser> yeah
[20:36] <smoser> and /sys/kernel/mm/ksm/sleep_millisecs has non-zero ?
[20:36] <kim0> smoser: 20
[20:37] <nuno_> glen1: /etc/apache2/mods-available/userdir.conf
[20:37] <smoser> kim0, sorry... i'd be stumped as you are.
[20:37] <kim0> smoser: ok no problem .. at least that doesn't make me feel too stupid :)
[20:38] <smoser> kim0, wait. what data did you think would be shared for your jvms ?
[20:38] <kim0> smoser: I kinda have no idea .. but I expected some core libs maybe to be shared .. just anything .. not zero
[20:38] <glen1> slug, cool
[20:39] <glen1> so they can only access the files, but not host a website?
[20:39] <kim0> smoser: the JVM executable itself maybe at least ?
[20:39] <kim0> smoser: I just echo 1 > run on my laptop .. and it's zero as well
[20:39] <smoser> yeah. i'd have thought something.  i dont know that the jvm executable itself would show up there.
[20:40] <smoser> it seems like i need to readm more about that
[20:40] <slug> glen1: i don't know, it depends what setup you have. you can disable directories to be served by apache
[20:40] <RoAkSoAx> smoser: http://imgur.com/st1Br.png
[20:40] <RoAkSoAx> smoser: is that expected?
[20:40] <smoser> eys
[20:40] <smoser> yes
[20:40] <glen1> thanks
[20:40] <smoser> RoAkSoAx, you should get a login prompt though
[20:41] <smoser> hit enter
[20:41] <RoAkSoAx> smoser: doing that but nothing
[20:41] <ideaman> i have a bug i cant't figure out. would appreciate anyone's help. After a flawless install of 10.04 or even trying 9, about 5/6 times on boot, it hangs at this message
[20:41] <ideaman> init: ureadahead-other main process(721) terminated with status 4
[20:41] <smoser> RoAkSoAx, how about in serial console (click i the window, then ctrl-alt-3)
[20:42] <smoser> RoAkSoAx, if youw ant to share a branch i'll take a quick look
[20:43] <RoAkSoAx> smoser: i get login prompt in serial console
[20:43] <RoAkSoAx> smoser: I'll share it in a bti
[20:49] <ideaman> i dont see you saying anything
[20:54] <ideaman>  i have a bug i cant't figure out. would appreciate anyone's help. After a flawless install of 10.04 or even trying 9, about 5/6 times on boot, it hangs at this message
[20:56] <sherr> ideaman: I see that occasionally/regularly but don't think it causes hangs/problems. Something else maybe? Last time someone asked about this "error" was because a network mount point (samba in /etcfstab) timed-out on boot. Causing a long wait.
[20:57] <ideaman> the only app i installed with the original install was ssh
[21:00] <hallyn> ideaman: googling, i see several bugs about it.  Most of the people who say they've fixed it for themselves seem to have cleared out extraneous entries from /etc/fstab
[21:01] <hallyn> ideaman: are you using lvm?
[21:01] <ideaman> yes
[21:01] <ideaman> been doing the same install on the last 10 boxes cookie cutter to this one
[21:01] <ideaman> this box is just a slightly new server
[21:01] <ideaman> same exact motherboard as the last servers though
[21:01] <ideaman> ill check fstab on it though
[21:02] <ideaman> perhaps a dban and try a clean install??
[21:10] <webpragmatist> jdstrand: this is weird.... the files are always dirty
[21:10] <webpragmatist> for /lib/ufw/*.rules and ufw.conf
[21:11] <RoAkSoAx> smoser: btw.. everytime you wan't to TestDrive that image, do you need it to untar every time so that it is a clean image, or would you like to run the image directly?
[21:11] <webpragmatist> anyone have any ideas?
[21:11] <RoAkSoAx> (which will already contain the assigned keys)
[21:12] <smoser> well, if you modify, you modify the file.
[21:12] <smoser> but i think testdrive's view of the world is throw away
[21:12] <jdstrand> webpragmatist: if you add rules or enable/disable the firewall, the files will change
[21:12] <smoser> but, as i suggested in the bug, you should create the backing store qcow file
[21:12] <smoser> then, you dont ever modify the original
[21:12] <webpragmatist> jdstrand: or reload?
[21:13] <webpragmatist> maybe that is it hrm
[21:13] <jdstrand> webpragmatist: I just tried with reload, it it does update the timestamp
[21:13] <webpragmatist> jdstrand: ya
[21:13] <webpragmatist> that was it thanks
[21:14] <webpragmatist> i have csync2 do-local which is not really necessary
[21:14] <RoAkSoAx> smoser: yes that's what I'm gonna do next. Right now I just wanted to have this working first, since creating a qcow2 is not a big deal to add
[21:14] <smoser> RoAkSoAx, what is the behavior of testdrive in other circumstances ?
[21:15] <smoser> if i run a live cd image, and then run it again, have my changes stuck ?
[21:15] <smoser> i'd just stay consistent with that behvior
[21:17] <RoAkSoAx> smoser: nop, changes won't stuck. But this case is different given that we using a disk image instead of an ISO. But, yes, I'll make it that way
[21:18] <smoser> oh, yeah.
[21:18] <smoser> that was a dumb question
[21:18] <smoser> i have not used a CD in so long
[21:18] <smoser> i was thinking of usb-creator-gtk made USB keys, which store data
[21:19] <RoAkSoAx> smoser: TestDrive creates a temporary disk image to run the ISO, if users installs, the image is kept, if no installation is being done, the images gets deleted
[21:21] <smoser> right. and how do you know if user installed ?
[21:21] <smoser> just curious
[21:24] <RoAkSoAx> smoser: well the image is kept because is not empty, so if users wants to run that image, he would just do "tesdrive -u file://name.img". However, everytime you run an ISO a disk image is create regardless there might have been a previous installation using the same ISO
[21:24] <smoser> how do you know if its empty ?
[21:25] <RoAkSoAx> smoser: http://pastebin.ubuntu.com/482583/
[21:26] <webpragmatist> any lsyncd users/
[21:26] <raubvogel> Back
[21:26] <smoser> fair enough
[21:27] <webpragmatist> I'm just wondering if on connection fail if it keeps the inotify actions on delay
[21:27] <raubvogel> smoser:  did you get your ldap thingie working?
[21:27] <webpragmatist> its hard to tell with the logs
[21:29] <smoser> raubvogel, i dont recal an ldap thingy
[21:29] <smoser> that doesn't mean i didn't ask about something like that.
[21:29] <raubvogel> smoser: sorry, it was slug
[21:30] <raubvogel> My cut-n-paste skills are weak this rainy day
[21:30] <slug> raubvogel: getting there :)
[21:30] <raubvogel> We should not start talking about my typing skills though...
[21:31] <slug> raubvogel: now i have phpldapadmin talking to it and able to create/edit accounts/groups. trying to setup dns now
[21:31] <slug> raubvogel: followed this: http://tuxnetworks.blogspot.com/2010/06/howto-ldap-server-on-1004-lucid-lynx.html
[21:31] <slug> raubvogel: created a small .sh to do it for me
[21:31] <slug> i'll put it online or as a comment on that page.
[21:32] <raubvogel> I will do it ldif style maybe by the end of the week
[21:35] <raubvogel> slug: so you used ldif?
[21:35] <slug> raubvogel: yeah
[21:36] <raubvogel> It does not seem to be bad but, as you said, it makes much more sense doing all that mucking around through a .sh script
[21:38] <slug> raubvogel: yeah, i just didn't want to type everything again if something didn't work
[21:48] <kim0> smoser: The KSM doesn't really work, because it's not magical. The user space app needs to register a memory range with the kernel using the madvise syscall to hint the region being "mergeable" .. qemu-kvm probably does this, but without it nothing is gonna happen
[21:48] <smoser> kim0, that makes sense.
[21:49] <kim0> I wonder if one can write a c-app to mark all private pages of a process as mergeable
[21:49] <kim0> memmerge -p 1234 -p 2345
[21:49] <kim0> something to think about :)
[21:53] <RoAkSoAx> smoser: I'll finish this later since I gotta run. It will be ready by tomorrow
[21:54] <RoAkSoAx> for testdrive-cli only. Next week I'll work on the -gtk
[21:54] <smoser> k
[22:20] <slug> raubvogel: any luck setting up dns+ldap?
[23:00] <Poizon> hey everyone, im having problems winth euc
[23:00] <Poizon> someone could give me a hand ?
[23:00] <Poizon> with'
[23:02] <nealmcb> !ask
[23:02] <binBASH> uec you mean
[23:02] <binBASH> not euc
[23:03] <Poizon> yes binBASH, sorry
[23:04] <Poizon> i've been trying to put a ubuntu image for lucid, but it fails with "socket.error: [Errno 113] No route to host
[23:05] <nealmcb> Poizon: what command did you run to get that error?
[23:08] <Poizon> nealmcb: uec-publish-tarball ubuntu-10.04-server-uec-i386.tar.gz lucid-$(date +%Y%m%d%H%M%S)
[23:08] <Poizon> oh i got what i was...
[23:08] <Poizon> zomg
[23:09] <Poizon> the dns configuration on uec
[23:09] <nealmcb> Poizon: Good job!  Isn't it cool how formulating a good question often leads to the answer without outside help?
[23:10] <Poizon> yeah well, i thought i had named running
[23:10] <Poizon> so i didnt expect it to dns problems
[23:11] <Poizon> thanks anyway =)
[23:11] <nealmcb> :)
[23:12] <hallyn> all right - does anyone here have experience with multipath?  I'm seeing what *appears* to be a race condition between the initrd's attempt to mount root, and udev having updated the /dev/disk/by-uuid/ links to point to /dev/mapper/*
[23:12] <hallyn> Just wondering whether I'm on crack before I start trying to "fix" it
[23:13] <Poizon> the new udev is a headache
[23:13] <hallyn> how has it changed?
[23:13] <hallyn> (this is in lucid server 10.04.1)
[23:13] <hallyn> (so maverick-only changes would not be my problem, at least not yet :)
[23:14] <Poizon> i got problems in a gentoo server, i had to just dont upgrade udev
[23:14] <hallyn> hm, i'll go look at the package source then, thanksf or the tip
[23:15] <Poizon> look at udev "rules", udev does recoignize your disks but it makes different path
[23:15] <Poizon> in a ubuntu server i just had to boot it under a live cd, look the udev rules, and change the fstab
[23:17] <hallyn> Poizon: problem here is that multipath locks /dev/sda1 and offers it as /dev/mapper/something, then udev updates /dev/disk/by-uuid/ to point into /dev/mapper (iiuc), but init is trying to mount /dev/disk/by-uuid before udev updates the links
[23:18] <hallyn> (on some systems, not all)
[23:18] <Takyoji> So I have an old desktop motherboard that's certainly functional; and has a "network boot" option (not sure which type of network boot however. Is it possible to have an installation on a desktop stored on another system (a system that would act like a PXE server)?
[23:18] <Poizon> yah thats what i tell you, when udev makes the recoignition it writes udev rules somewhere in /etc/
[23:18] <Poizon> you can look there, read them and update your fstab in order to get your system ounted
[23:18] <hallyn> Poizon: but i'm sayin gthe rules are right, they are just being acted on a split-second too late
[23:19] <Poizon> mounted'
[23:19] <Poizon> oh
[23:19] <Poizon> i see
[23:19] <hggdh> hallyn: you do virt, do you not?
[23:19] <hallyn> Takyoji: not sure where the simplest howto is (and i don't do it), but it certainly is possible
[23:19] <hallyn> hggdh: what do you mean?
[23:20] <hggdh> hallyn: I got bug 619843 where we found a problem on Maverick
[23:20] <hggdh> hallyn: it deals with console.log being owned by root:, 660
[23:21] <hallyn> hggdh: i'll go look at that if you promise to read up and tellme if you have any hints on udev.vs.initrd.vs.multipath races  :)
[23:21]  * hallyn goes to look
[23:21] <hggdh> hallyn: heh. I am trying to find one myself... on mounted-tmp
[23:22] <hggdh> hallyn: and it *is* a race, since rbooting with init=/sbin/init --debug causes the bug to vanish
[23:24] <hallyn> hggdh: and, since the link which it is trying to mount is correct by the time you get the busybox shell
[23:25] <hallyn> hggdh: well i don't think bug 619843 would be apparmor, as that is concerned with MAC...
[23:26] <hallyn> hggdh: libvirt update maybe...
[23:26] <hggdh> hallyn: I am suspecting so also... I will try to hunt it down, thanks
[23:27] <hggdh> hallyn: BTW -- IS told me the CanonicalCloud is closed to the external world
[23:28] <hallyn> hggdh: very inconvenient :(
[23:29] <hallyn> hggdh: where is that console located by default?
[23:29] <hallyn> oh, it's just the logfile at least?
[23:30] <Takyoji> So is PXE de facto as of network boot and pretty much supported by systems made at least after 1998 or so?
[23:30] <hallyn> Takyoji: that's my understanding
[23:30] <hallyn> hggdh: standard virt-viewer+libvirt+kvm on my brand-spanking-new maverick laptop makes it:
[23:30] <hallyn> '-rw------- 1 root root 9733 2010-08-23 17:26 lucid1.log
[23:31] <Takyoji> I suppose this is the suitable wiki page I'm looking for: https://help.ubuntu.com/community/DisklessUbuntuHowto
[23:33] <hallyn> Takyoji: looks nice
[23:39] <SpamapS> smoser: you could totally use a regex for that btw