[01:01] <a7ndrew>  
[01:01] <RoyK>  
[01:18] <failover>  
[01:54] <arooni> how do i upgrade my ubuntu server packages (specifically security ones)
[01:55] <qman__> arooni, sudo apt-get update && sudo apt-get upgrade
[01:55] <arooni> thank you
[01:55] <qman__> use sudo apt-get dist-upgrade to install updates which require installing new packages
[01:55] <arooni> also i'm looking for somethign to keep track of why i seem to run out of memory
[01:55] <qman__> such as kernel updates
[01:55] <arooni> i.e. when my app crashes its too late to figure out the culprit
[01:56] <qman__> use top, or script something with free and ps to log
[01:56] <qman__> so next time it happens, you have an idea of what was going on
[02:04] <arooni> is there a package that does that nicely?
[02:04] <arooni> how do i make bash my default shell
[02:10] <RoyK> arooni: chsh -s /bin/bash [username]
[02:33] <qman__> top is likely capable of doing what you want, but you'll have to dig deep in TFM to find the right switches
[02:34] <qman__> I barely know the basics of it
[02:37] <arooni> ERROR: APACHE_PID_FILE needs to be defined in /etc/apache2/envvars
[02:47] <hallyn_> SpamapS: collectd lxc plugin - https://github.com/dotcloud/collectd/tree/lxc-plugin   seems right up your alley :)
[02:50] <arooni> how do i recreate the /etc/apache2 folder?
[02:50] <qman__> you could purge apache, then reinstall
[02:50] <qman__> make sure you back up any important configs first
[02:50] <julian_c> Would <dpkg-reconfigure apache2> work?
[02:51] <qman__> I don't think so, but I haven't tried
[02:51] <arooni> reinstalling i dont see /etc/apache2
[02:51] <qman__> that's because you have to purge it first, not just remove
[02:51] <qman__> apt-get remove --purge or just apt-get purge
[02:51] <qman__> and the package is probably apache2-common
[02:52] <arooni> i manually removed the /etc/apache2 directory
[02:52] <qman__> yes, but you still have to purge the package
[02:52] <qman__> or it won't install new configuration
[02:52] <arooni> qman__, i purged it first (both apache2-common and apache2
[02:53] <arooni> and reinstalled but i'm getting no /etc/apache2
[02:53] <qman__> must be in a different package then
[02:55] <qman__> ah
[02:55] <qman__> apache2.2-common
[02:55] <qman__> arooni, ^
[02:55] <qman__> purge that, then reinstall, and it will recreate the configuration
[02:55] <arooni> yup after you said that i purged every installed apache2 package
[02:55] <arooni> and it worked
[02:56] <qman__> I found that by using dpkg -L and some tab completion
[03:08] <zul> hallyn_: http://zulcss.wordpress.com/2011/05/16/the-big-uds-o/
[03:24] <hallyn_> zul: cool, thx
[03:38] <maxagaz> hi
[03:38] <maxagaz> how to chose which one to adopt between svn and git ?
[03:39] <Corey> maxagaz: Git.
[03:39] <Corey> It solves the branching / merging headache.
[03:40] <Guest43057> bazaar
[03:44] <Corey> Guest43057: Unless you're the Ubuntu project, bad idea. :-)
[03:44] <maxagaz> Guest43057: can I use bazaar over git ?
[03:44] <fallous> bazaar is pretty craptacular in my experience
[05:40] <AlexMax> The last time I restarted my server, SSHD did not come up.  I'm not really familiar with upstart, how can I tell if openssh is actually attempting to start at boot?
[05:40] <AlexMax> I have no idea if it's not being started at all, or if it's being started and not working
[05:40] <AlexMax> My host was able to start ssh manually from the terminal
[05:51] <Proz01d> noob question.....need to transfer files to my server and from my server. nothing is installed yet... all i have is a web based vnc access.  I'm trying to setup openssh as we speak  but need to get the keys transfered so i can connect..
[05:55] <Proz01d> any ideas....
[06:06] <koolhead11|afk> Proz01d, do you have password for the same server? So in order to transfer key you can log in there 4 once?
[06:07] <jmarsden> Proz01d: Or, use ssh the other around ... ssh from the server command line *to* an sshd on your local machine?
[06:30] <greppy> Proz01d: you could put your public key on the remote server with just a copy & paste, if you can do that over webvnc that is.  the ssh key files are plain text.
[06:35] <Proz01d> <koolhead11|afk: yes i have the password
[06:36] <Proz01d> greppy: the web interface doesn't let me cut and paste... :(
[06:36] <Proz01d> it's java based Xen
[06:36] <Proz01d> trying to test my openssh install / config and i get..
[06:37] <Proz01d> connection closed by UNKNOWN when i do "ssh -v localhost"
[06:40] <jmarsden> Proz01d: What did you do to your server??  sudo apt-get install openssh-server   # should just work.
[06:40] <Proz01d> damn it.
[06:41] <Proz01d> well i was follwing a guide that was telling me to change the configurations etc.
[06:41] <jmarsden> Lesson #1: Quit following guides.
[06:41] <Proz01d> lol
[06:41] <jmarsden> I am 100% serious.
[06:42] <jmarsden> They may be old, they may be inaccurate... if you do not understand them and agree with their ideas, don't follow them.
[06:43] <jmarsden> So now you get to   sudo apt-get purge openssh-server   and then   sudo apt-get install openssh-server    # and see if that fixes it.
[06:43] <Proz01d> will that remove all the configs?
[06:43] <jmarsden> Yes.
[06:43] <jmarsden> apt-get remove leaves the configs, apt-get purge ... purges the config files .
[06:44] <Proz01d> i did the remove last time
[06:45] <arooni> ok i think i finally have an apache question.  i want to test that my apache/app config is working before i make mydomain.com => newly set up server.  right now i have ServerName and ServerAlias set... is there a way for me to directly address this particular virtualhost by ip/port?  and how would i add that to the apache config?  ps there are two rails apps running on this box
[06:45] <Proz01d> same error
[06:46] <jmarsden> arooni: Make that virtual host the first one and it should be the default one.
[06:46] <arooni> jmarsden, how do i know which is first/second.  all i know are the two sites are already enabled via a2ensite
[06:47] <jmarsden> Proz01d: OK... so does    sudo netstat -ntl | grep :22     # show you that something is listening on port 22 ?
[06:48] <Proz01d> i'm purging the client as well
[06:48] <koolhead11|afk> arooni, in the virualhost file you can specifically mention which port it has to run
[06:48] <arooni> koolhead11|afk, awesome; do you know which config option that is
[06:48] <koolhead11|afk> namevirtualhost and port
[06:49] <koolhead11|afk> arooni, lemme recheck :P
[06:49] <jmarsden> arooni: I think you can just make the filename under sites-available/ be alphabetically first ... I've not done it for a while :)
[06:50] <Proz01d> jmarsden: cleaned everything up and i do have listeners on 22 for ip4/6
[06:50] <Proz01d> there we go
[06:50] <Proz01d> !
[06:51] <koolhead11|afk> arooni, Listen and port number :)
[06:51] <koolhead11|afk> Proz01d, seems that will solve your issue :D
[06:52] <jmarsden> Proz01d: OK... next time... don't mess with config files you don't understand, that do not need to be messed with :)
[06:52] <Proz01d> trying remotely.... from my local machien.... hope this works
[06:52] <Proz01d> :)
[06:53] <koolhead11|afk> arooni, http://ubuntuforums.org/showthread.php?t=1440839 :D
[06:54] <jmarsden> if you need to add your public key to the remote server's authorized_keys file, use  ssh-copy-id user@remoteserver.example.com   on your local machine to do that.
[06:58] <Proz01d> my local machine is windows
[06:58] <Proz01d> and i'm trying to connect via putty
[06:58] <arooni> its working now!
[06:58] <arooni> man it took longer than i thought to get my ubuntu server running like my debian 4.0 one was
[06:58] <arooni> always longer than you think
[07:00] <jmarsden> Proz01d: ick.  OK, use pscp to copy the files you need, if you have to do it the hard way :)
[07:01] <Proz01d> wow... so much easier when i take your advice... no guides!
[07:02] <jmarsden> :)
[07:02] <Proz01d> i just want to secure this a little
[07:02] <jmarsden> Against what?
[07:03] <Proz01d> anything..
[07:03] <Proz01d> i'm just using defaults right now..
[07:04] <jmarsden> The only way to secure a machine against "anything" is to disconnect it from the Internet and bury it in a concrete box :) :)
[07:04] <jmarsden> What are you worried about, in the realm of "security"?
[07:04] <Proz01d> shuoldn't i use non standard ports.. disable password access and use keys instead...
[07:05] <Proz01d> those were some readings from guides
[07:05] <jmarsden> Random script kiddies?  Your government agencies?
[07:05] <Proz01d> both
[07:05] <Proz01d> paranoid :)
[07:06] <Proz01d> it's my first server...
[07:06] <Proz01d> lol
[07:06] <jmarsden> Using keypairs is a good idea... that is what I thought you were copying the public key file for.  Did that work?
[07:06] <jmarsden> The non-standard port thing is... well, it's your choice.  adding denyhosts is also one way to reduce the clutter from automated attacks on port 22.
[07:07] <Proz01d> i'm a little lost on PKI....
[07:08] <Proz01d> the pub / private keys generated on the server.  i should grab the pub key and copy it to my local machine and i should be okay to connect right?
[07:09] <jmarsden> Other way around... generate a keypair locally on your desktop PC, then copy the public key from it to ~/.ssh/authorized_keys and set the permissions correctly.
[07:09] <jmarsden> Oh, and if you use Putty I think you have to convert the key from Putty's format to normal ssh format...
[07:14] <koolhead11|afk> hi RoyK
[07:20] <Proz01d> by normal you mean openssh key?
[07:21] <jmarsden> Proz01d: yes.
[07:21] <jmarsden> I've not used putty for a long time, though.
[07:22] <jmarsden> When I need to use Windows machines for this sort of thing, I tend to install Cygwin (bigger, but much more unix-like shell environment, so more useful)
[07:39] <Proz01d> this is going to be the death of me.... the authorizedkeys files is supposed to be located in "%h/.ssh/authorizedkeys
[07:39] <Proz01d> according to my config but no such dir or file exists..
[07:39] <Proz01d> i need to create this manually?
[07:41] <Proz01d> i'm guessing %h means /home/[user]?
[07:57] <joschi> Proz01d: yes. but the file is named 'authorized_keys' by default (mind the underscore)
[08:04] <arooni> how do i make it so that a user still has to sudo ; but doesnt have to type password
[08:04] <arooni> each time
[08:04] <Bashew> Hi everyone, I have a bit of an issue. I have a server I'm going to colo but I'm experimenting right now. My server has only 1 network card eth0, with ips 192.168.1.[30-34]. I want to selectively forward ports from individual IPs to certain VMs, and have certain VM's send their traffic out of certain IPs
[08:05] <Bashew> like on 192.168.1.32:80, go to one VM whereas 192.168.1.30:80 goes to a completely different one
[08:07] <Bashew> arooni, if i remember correctly, sudo shouldn't bother you for a password for 15 minutes after you first use it
[08:09] <Bashew> arooni, i haven't tried this myself but you could give http://www.webupd8.org/2010/04/how-to-change-sudo-password-time-out-in.html a try. just give the timeout a big number
[08:10] <jmarsden> arooni, Bashew: You can set that to whatever you want in /etc/sudoers, the default is 15 minutes.  The setting is called timestamp_timeout
[08:10] <arooni> you would do that over not prompting for a pw?
[08:10] <jmarsden> If you make it too big and go to lunch with your terminal unlocked, you are asking for trouble :)
[08:11] <jmarsden> arooni: Once every 15 minutes, for sudo priviledges, is pretty reasonable... why is it a problem for you?
[08:17] <arooni> i am just typing the pw a lot
[08:17] <arooni> i guess from differnet terminals
[08:17] <arooni> each time i have a different session open
[08:17] <arooni> have to reauth
[08:17] <Bashew> you could try using screen
[08:17] <arooni> ok thats a cool idea
[08:17] <arooni> i've heard of it before
[08:17] <Bashew> open screen up on your first terminal, then on subsequent terminals, use screen -x
[08:18] <Bashew> a side benefit is, if you ever drop your local connection, your stuff will still be up :)
[08:18] <w00> arooni, or you could add something like this in /etc/sudoers: %wheel  ALL=(ALL)       NOPASSWD: ALL
[08:18] <w00> arooni, so if you are in group wheel you don't have to type pass to sudo
[08:29] <maxagaz> when I run rm /tmp/*, is there somewhere a log of all what has been removed ?
[08:29] <maxagaz> if not, is it possible to set it ?
[08:30] <Bashew> you could make a log every time you run it
[08:30] <Bashew> rm -v /tmp/*  > /path/to/output.txt
[08:30] <Bashew> but there are no logs
[09:16] <missil> hi all
[09:56] <airtonix> I have a ubuntu 10.04 server running apache2 containing two virtualhosts... same domain one is http the other https. currently they both serve files out of /home/ubuntu/vhostname/public_html but i want to move it to /var/www/vhostname/public_html . i can make this happen in the http vhost def file but not the https. when i make the relevant changes in the https vhost file and restart it fails and doesn't ask for the ssl cert password.
[10:09] <remix_tj> airtonix: what's the error?
[10:09] <remix_tj> check the error.log of apache
[10:13] <airtonix> remix_tj: i have to check something to confirm... #httpd are being nazis and unhelpful
[10:13] <remix_tj> :-)
[10:13] <remix_tj> check the log
[10:13] <remix_tj> and let me know
[10:13] <airtonix> equiv of Oh hoho you're supposed to spell it "you're" not ur
[10:14] <airtonix> anyway testing...
[12:48] <koolhead11|afk> why i can`t ask a question at https://launchpad.net/cobbler  :(
[12:49]  * koolhead11|afk looks at Daviey, 
[12:50]  * Daviey hides
[12:50] <Daviey> koolhead11|afk, So the launchpad cobbler project is not related to the upstream project management...  Do you have a question specific to ubuntu server?
[12:51] <koolhead11|afk> Daviey, i have a question regarding cobbler on natty.
[12:51] <Daviey> koolhead11|afk, The best place to throw that would be ubuntu-server mailing list tbh.
[12:53] <koolhead11|afk> Daviey, i just wanted to know if its a good way to hardcode my local repo intp preseed file ?
[12:54] <koolhead11|afk> *into
[12:56] <Daviey> koolhead11|afk, it is 'a way'... if you are using cobbler purely for that, then it is purely overkill.
[12:58] <koolhead11|afk> Daviey, cobbler import --mirror is not an option it seems in case of ubuntu :D
[12:59] <Daviey> koolhead11|afk, If something isn't working that you expect it to, please raise a bug.
[12:59] <Daviey> I know i haven't used that feature.
[12:59] <koolhead11|afk> Daviey, but i cannot in launchpad
[13:00] <koolhead11|afk>    /o.0\
[13:03] <koolhead11|afk> Daviey, thanks.
[13:08] <Daviey> koolhead11|afk: Hmm
[13:08] <Daviey> koolhead11|afk: You can't in launchpad?
[13:09] <Daviey> koolhead11|afk: Oh, raise a bug - do that here: https://bugs.launchpad.net/ubuntu/+source/cobbler/+filebug
[13:09] <koolhead11|afk> Daviey, yay!! got it
[13:10] <koolhead11|afk> Daviey, i was looking at https://launchpad.net/cobbler all this time
[13:22] <bencer_> anyone with experience on multipath stuff? i've a lucid box attached to an emc san, dmesg reports qlogic driver load and i can see the 4 scsi devices created by the hba on dmesg, but multipath -l reports nothing
[13:22] <bencer_> i've filtered out these scsi devices con lvm.conf but still nothing
[13:25] <bencer_> running multipath on debug shows the scsi devices, but can't manage them: http://paste.ubuntu.com/608968/
[13:41] <tyreza> hello there
[13:41] <tyreza> how to check disk on ubuntu?
[13:45] <patdk-wk> how do you mean?
[13:45] <patdk-wk> df? smartctl? badblocks? ....
[13:47] <Pici> fsck?
[13:47] <tyreza> fsck
[13:48] <a7ndrew> fsck checks filesystems, not disks. smartctl might be what you are after
[13:51] <tyreza> i simply need to check filesystem on drbd
[13:51] <tyreza> what i have to do ?
[13:52] <Tensibai> Hi there
[13:53] <Tensibai> May someone help me with preseed files ?
[14:02] <tyreza> how to find drbd partition on my system ?
[14:03] <patdk-wk> blkid?
[14:07] <ppetraki> tyreza, shoudn't they be defined in your drbd config?
[14:07] <tyreza> correct blkid
[14:07] <tyreza> i find drbd  partition
[14:08] <tyreza> normally it is in /dev/mapper
[14:08] <tyreza> now how to fsck on that partitioN ?
[14:09] <soren> drbd is not a filesystem.
[14:09] <soren> fsck is for checking filesystem integrity.
[14:14] <tyreza> what have to do to make a complete check on drbd ?
[14:16] <dawolf123> Does anyone know where I can get a list of the maximums for ubuntu server?
[14:18] <pmatulis> dawolf123: maximums?
[14:18] <dawolf123> such as the maximum amount of memory, cpus, disk etc...
[14:23] <RoAkSoAx> morning all
[14:23] <RoAkSoAx> lynxman: pin
[14:23] <RoAkSoAx> lynxman: ping
[14:24] <lynxman> RoAkSoAx: pong :)
[14:24] <pmatulis> RoAkSoAx: o/
[14:25] <highvoltage> zul: something seems to be wrong with your blog configuration on planet, your post is summarized and links back to planet, so it's hard to find the actual entry to read it
[14:26] <highvoltage> (been a problem for a while, not sure if it's intentional)
[14:26] <jamespage> hey RoAkSoAx
[14:26] <zul> highvoltage: yeah i know about it i havent figured out how to fix it yet
[14:26] <RoAkSoAx> pmatulis: o/
[14:26]  * zul kicks wordpress
[14:26] <glen1> hey
[14:27] <RoAkSoAx> zul: /me still waiting for endorsement :
[14:27] <RoAkSoAx> :P
[14:27] <zul> RoAkSoAx: everytime you ask ill kill a kitten ;)
[14:28] <highvoltage> zul: mind if I look at it?
[14:28] <RoAkSoAx> zul: ehehe
[14:28]  * lborda hi folks
[14:28] <lynxman> jamespage: James Page! o/
[14:28] <jamespage> hey lynxman
[14:29] <glen1> when I read about websites buying new hardware for their visitors, how do they know how much hardware to purchase?
[14:29] <pmatulis> dawolf123: that is kernel-specific
[14:29] <lynxman> glen1: simple logistics and previewing demand :)
[14:30] <glen1> I see. So they simply say will have this many users so we need to deal with them
[14:30] <dawolf123> k, thx
[14:30] <lynxman> glen1: exactly
[14:31] <glen1> So if I have 100 new customers a month, how would I calculate how much storage, cpu, ram they need. its something that i never knew thanks lynxman haha
[14:31] <lynxman> glen1: you either allocate limited resources per user, otherwise you oversell the available resources and try to keep up with demand
[14:32] <glen1> is there a comparison like say 1000 users per cpu?
[14:32] <lynxman> glen1: your mileage may vary a lot, so there's no strict rule, it depends on what kind of services your users will be running
[14:33] <glen1> ahh :) So kinda extrapolate a perfect situation. thanks lynxman
[14:34] <lynxman> zul: you up for some packaging crack hell :)
[14:34] <lynxman> glen1: no prob :)
[14:34] <zul> lynxman: depends
[14:34] <lynxman> zul: just wanted to poke your brain slightly with a couple issues I'm having
[14:34] <zul> lynxman: sure ask away but im not very alert yet
[14:35] <lynxman> zul: no worries, your half awakeness packaging expertise is way better than my fully aware one ;)
[14:37] <RoAkSoAx> jamespage: James Page \o/
[14:38]  * jamespage takes a bow...
[14:38] <lynxman> RoAkSoAx: I was sitting one next to jamespage, he's such a gentleman and scholar
[14:38] <lynxman> s/one/once/
[14:38] <RoAkSoAx> how's everybody doing post-uds hangovers :)?
[14:39] <lynxman> RoAkSoAx: I barely cope
[14:39] <RoAkSoAx> lynxman: hehe I had to move and tide up my new place :S
[14:39] <lynxman> RoAkSoAx: that'll keep you awake through jet lag
[14:40] <RoAkSoAx> lynxman: I'm way past jet-lag by now :):D
[14:40] <lynxman> RoAkSoAx: me too, going through London and back.. crossed the timezone twice :)
[14:46] <lynxman> zul: is there any way to determine package install order? afaik there isn't
[14:47] <zul> lynxman: not really an acceptible way why?
[14:47] <lynxman> zul: I have a postinst script that fixes squid-deb-proxy config for unknown enviroments
[14:47] <lynxman> zul: I was thinking of just creating a metapackage that gets the package and does the postinst config change, would that be kosher?
[14:47] <zul> lynxman: you can use a pre-depends but thats frowned upon
[14:48] <zul> lynxman: whats the problem you are trying to solve?
[14:48] <MTecknology> I'm trying to use XIV on a Linux box. A storage admin gave me two LUN's. I used xiv_attach to get them seen my the disk. Now apparently they're being seen out of order. (/dev/mapper/mpath15 needs to be assigned to VolID 595). Wondering if by chance maybe someone in here has a clue about something like that...
[14:49] <lynxman> zul: I need to change squid-deb-proxy config file, to allow ppa import and also open it to more nets than the default value
[14:49] <zul> lynxman: why not talk to mvo about it then?
[14:49] <lynxman> zul: I was also thinking of looking at the squid-deb-proxy code and just dotdee the config, but that's a bit more complicated and I would rather find a way around it for the moment
[14:49] <lynxman> zul: did talk with him, I know the changed I need to do :)
[14:50] <lynxman> zul: it's just the question of how can I modify the config so the metapackage installs fine
[14:50] <lynxman> zul: another way would be creating our squid-deb-proxy package with debconf hooks attached
[14:51] <zul> lynxman: i think you might want to do the debconf hooks and then we can preseed it with puppet or something
[14:52] <lynxman> zul: can even preseed it from the metapackage, and proceed as we agreed with other default configs and have a puppet module as well for future "modification" from the sysadmin side
[14:52] <zul> lynxman: that makes sense for me
[14:52] <lynxman> zul: cool, I'll go that way then, thanks
[14:53] <Daviey> lynxman: Have you spoken to mvo?
[14:55] <Daviey> lynxman: If squid-deb-proxy introduces a documented interface for config management, then Orchestra can edit it as it sees fit.
[14:55] <lynxman> Daviey: yeah, before UDS, he told me where to touch and how and such, but there's no debconf hooks to do so in squid-deb-proxy
[14:55] <lynxman> Daviey: that's the issue, it's just a plain config file
[14:55] <lynxman> Daviey: actually several of them
[14:58] <Daviey> lynxman: Yeah..  do you have a diff of chanegs you want?
[14:58] <lynxman> Daviey: I'll have in a few, was just agreeing with zul the best way to handle it
[14:59] <Daviey> lynxman: Well seeing that changes you want to introduce, largely determines how best to handle it :)
[14:59] <lynxman> Daviey: I see your train of thought :)
[14:59] <Daviey> choooo chooo. :)
[15:00] <lynxman> Daviey: I'll add debconf hooks in order to be able to a) add or remove access to different things (updates, ppas, etc) and also be able to add/remove networks to the ACL or just lift the acl completely
[15:00] <lynxman> Daviey: after speaking with mvo the default acl makes not much sense in heterogeneous networks scenario
[15:01] <lynxman> Daviey: it's limited to only private IP ranges
[15:01] <Daviey> ahh
[15:01] <lynxman> Daviey: and lifting the ACL will only mean that if someone knows the proxy they'll be able to proxy packages only
[15:01] <lynxman> Daviey: and also it comes with access to ppa's disabled by default, and we want that for orchestra
[15:01] <zul> why?
[15:02] <lynxman> zul: because orchestra is in a ppa
[15:02] <Daviey> i thought ppa.launchpad.net was added last cycle?
[15:02] <Daviey> (to squid-deb-proxy)
[15:02] <lynxman> Daviey: it's added but commented by default
[15:02] <lynxman> Daviey: so not active
[15:03] <Daviey> oh
[15:03] <zul> i would leave it as it is and ask the user if they want to enable it or not
[15:03] <lynxman> zul: hence why I want to add a debconf hook :)
[15:03] <zul> lynxman: k
[15:04] <lynxman> zul: it'll ask the user if you reconfig the package and we'll preseed the selection from the orchestra-client package
[15:05] <Daviey> lynxman: Less yapping, more patching :P
[15:05] <lynxman> Daviey: aye aye captain!
[15:05] <zul> lynxman: what daviey said
[15:05] <lynxman> zul: aye aye coach!
[15:05] <lynxman> (adapts the aye aye to different cultures)
[15:06] <Daviey> lol
[15:46] <hggdh> jamespage: good morning/afternoon/something dear sir. When would super-james have a bit of time to get in details on the amazingly profound advanced testing BP?
[15:46]  * hggdh hopes that, with so much fluffery, an answer will be forthcoming
[15:49] <hggdh> Daviey: I think we have some unfinished business, do we not?
[15:49] <hggdh> something about Lucid
[15:50] <RoyK> good localtime();
[15:51] <hallyn_> mdeslaur: meeting-filled morning (relatively), so will ping you this afternoon
[15:52] <mdeslaur> hallyn_: sure, np
[16:05] <jamespage> hggdh: yeah - have been thinking that one through over the last day or so.
[16:11] <hggdh> jamespage: perfect
[16:12] <jamespage> hggdh: I think we need an approach to each type of testing; trying to build a one-size fits all probably won't work that well
[16:12] <jamespage> and might end up quite brittle at the end of the day
[16:13] <hggdh> jamespage: I agree 100%. There is *NO* such thing...
[16:13] <jamespage> hggdh: this is my current thinking
[16:14] <jamespage> 1) Re-use existing ISO testing to ensure first install works OK - this should cover LVM, RAID etc....
[16:14] <lwhalen42> hey all, I'm trying to install a KVM instance of Ubuntu Server 11.04, but I can't seem to get the 'virsh console' feature to work
[16:14] <jamespage> 1b) Re-use existing ec2 testing to ensure AMI's are OK
[16:14] <lwhalen42> here's my virt-install: http://pastebin.com/SSYs5RbC
[16:15] <jamespage> 2) Use orchestra to create complex deployments either on bare metal or vm's
[16:15] <jamespage> or maybe ensemble
[16:15] <hggdh> to be tried and seen
[16:16] <jamespage> 3) Build out a testing framework to support testing complex deployment scenarios (probably based on mcollective which is already part of orchestra)
[16:17] <jamespage> hggdh: agree that orchestra and ensemble need work; but I don't want to re-invent the wheel
[16:17] <stetho> I'm having a problem setting up an Ubuntu server as a router between two LANs (192.168.55.0 and 192.168.56.0). I've done the /proc/sys/net/ipv4/ip_forward = 1 and the sysctl stuff but I still can't get any further than pinging the other networks interface on the on the ubuntu box (eg, 192.168.55.x can ping 192.168.56.253 but not 192.168.56.x) Anyone suggest what I might have missed?
[16:17] <jamespage> 4) Upgrade testing - we need to review with mvo to see how we deliver this
[16:18] <lwhalen42> stetho: I don't know offhand what you may have missed, but can you hang wireshark off either interface and see what's making it through?
[16:18] <jamespage> 5) boot testing - no idea at the moment
[16:18] <lwhalen42> also, could you possibly have iptables filtering traffic?
[16:18] <hggdh> jamespage: (3) could be split into (3.1) complex server deployment and (3.2) AMI/UEC instance testing
[16:18] <jamespage> yeah - I was thinking along those lines.
[16:20] <hggdh> jamespage: I am a bit worried this is too much for one single cycle -- so we need (if I am correct) to prioritise some items, and accept that not all will be there
[16:20] <hggdh> right now
[16:21] <jamespage> hggdh: agreed - far to much for one cycle
[16:21] <hggdh> jamespage: and there is smb's request to facto in also
[16:22] <jamespage> hggdh: yeah - although I think that should be fairly easy to achieve
[16:23] <hggdh> jamespage: oh yes, the framework is already in place, just expanding the test suite. But I do not want to do it and have something that does not quite align with the rest
[16:23] <jamespage> I think it fits well into the testing AMI images use case
[16:24] <hggdh> I would like to have it compatible with AWS or UEC (openstack/euca)
[16:25] <hggdh> so we could run on either with just some conf file changes
[16:28] <jamespage> hggdh: hmmm - that should be possible
[16:32] <lwhalen42> does anyone even do virsh-serial-like installs with Ubuntu?  Or is that more of a Xen/Redhat workflow?
[16:38] <mouseclone> GRRRR... I don't understand why when I purge a program using aptitude and then reinstall I do not get all of the .conf files back for that package.  I'm missing radiusd.conf and I"m unsure how to get the defalut file back.
[16:39] <Daviey> hggdh: we do!
[16:40] <Daviey> hggdh: i'm tied to the meeting right now, but following that.
[16:40] <hggdh> Daviey: certainly
[16:42] <mouseclone> anyone know how to get the default files back in the /etc/freeradius directory?
[16:43] <mouseclone> apt-get install doesn't put them all back
[16:44] <lwhalen42> mouseclone: maybe "dpkg-reconfigure"?
[16:46] <mouseclone> dolwhalen42: doesn't work.  Tries to start the services, which fails because htere is no radiusd.conf file.
[16:58] <smb> hggdh, jamespage What I was looking for sounded rather like making use of the existing framework for most parts. My actual testing may or may not be different but the main use was the launching and access to instances in a larger scale than ... err one or two.
[16:58] <hallyn_> zul: did the new lxcguest work for you?
[16:59] <zul> hallyn_:  not yet will do so now
[16:59] <hallyn_> k
[17:00] <mouseclone> anyone know how to reprocess packages so that they fully reinstall?
[17:00] <hallyn_> zul: http://people.canonical.com/~serge/lxcguest_0.7.4-0ubuntu9_amd64.deb
[17:01] <zul> hallyn_: what did you fix anyways?
[17:01] <hallyn_> jdstrand: in case you fetched it last week, there had been a bug in my libvirt package last thu (forget what), so as of last friday the proposed oneiric libvirt package is at http://people.canonical.com/~serge/libvirt-0.9.1-sync.2.tar.gz
[17:02] <hallyn_> zul: it now allows for LIBVIRT_LXC_UUID=xyz as indicator of running under libvirt
[17:02] <hallyn_> zul: then the next libvirt sync will drop your patch, and make use of just that
[17:02] <jdstrand> hallyn_: thanks. I had not yet. I hope to look at it today or possibly tomorrow
[17:02] <hallyn_> (it's a new boot argument which libvirt added for containers as of 0.9.0 or so)
[17:02] <hallyn_> jdstrand: thanks
[17:03] <hallyn_> oops, mtg
[17:03] <lwhalen42> mouseclone: is there a /etc/radius.conf.dpkg or similar?
[17:03] <zul> hallyn_: okie dokie
[17:03] <Thanatos__> Hello . Since I upgraded to Natty , I have to do a manual poff / pon after reboot for ddclient to update my dyndns properly .. does anyone else have the same issue ?
[17:04] <mouseclone> lwhalen42: no the files i need are in the freeradius-common packages or freeradius-common_2.1.8+dfsg-1ubuntu1_all.deb package in the repoes
[17:05] <lwhalen42> when you uninstalled them, did you just 'remove' or did you 'purge'?
[17:05] <mouseclone> lwhalen42: aptitude purge freeradius
[17:06] <lwhalen42> hrm
[17:07] <mouseclone> lwhalen42: I have downloaded the packages, ar vx to get the tar.gz files and have put some of them back.
[17:07] <mouseclone> I just really really do not understand why adding a package back doens't over write everything
[17:09] <lwhalen42> the packager thinks its smarter than you ;-)
[17:13] <mouseclone> think i got it.. stupid stupid hobbit.. aptitude purge freeradius freradius-common freeradius-utils
[17:14] <mouseclone> now it is reinstalling itself and now it has started back up
[17:14] <lwhalen42> ahh, missed a package?
[17:15] <lwhalen42> I'm still trying to figure out how I can do the install dialog on a VM without having to run VNC
[17:15] <mouseclone> lwhalen42: didn't know that you had to purge 3 of them to make it work correctly again.
[17:15] <lwhalen42> dpkg -l | grep radius can be helpful for that
[17:16] <mouseclone> what do you mean by install dialog on a VM (virtual machine)?
[17:22] <lwhalen42> mouseclone: I mean, when I currently do a virt-install against an ISO (or any install source, really), in order to say "Partition my drives thusly, install these packages", etc, I have to hook up a VNC viewer from my desktop
[17:22] <lwhalen42> I'd like to be able to say "virsh console <vmname>" and do the install like that
[17:23] <mouseclone> ah
[17:36] <hallyn_> zul: if that lxcguest works all right, then please grab http://people.canonical.com/~serge/lxc_0.7.4-0ubuntu9-package.tgz   and dput it :)  (I need to request the lxc upload rights, i know)
[17:37] <zul> hallyn_: okie dokie
[17:48] <lwhalen42> silly question: is there an Ubuntu Server "netboot" or similar ISO?
[17:50] <lwhalen42> booyah, Google is my friend: http://cdimage.ubuntu.com/netboot/natty/
[18:02] <SpamapS> lwhalen42: we've been doing a ton of work getting the Cobbler provisioning system working too btw.
[18:02] <SpamapS> lwhalen42: can help with managing a whole bunch of server installs.
[18:32] <lwhalen42> SpamapS: wait, I can rock Natty installs via Cobbler/koan?
[18:50] <SpamapS> lwhalen42: yes in the latest cobbler
[18:50] <SpamapS> lwhalen42: its also available in Natty and later
[18:50] <lwhalen42> most excellent
[18:51] <SpamapS> lwhalen42: I believe we tested installing lucid too
[18:59] <rcaskey> has anyone had problems with samba adding machine user accounts in the wrong case?
[19:26] <MTecknology> Anyone know anything about multipath? I tried issuing mpath_ctl add path mpath16 but apparently that syntax is wrong?
[19:28] <hggdh> Daviey: so. Back to Lucid and SRU... what do you need to test?
[19:31] <ppetraki> MTecknology, linux multipath or some proprietary vendor MP util? That invocation doesn't look like a native MP client
[19:33] <MTecknology> ppetraki: the invocation could also be me not having a clue what i'm doing...
[19:34] <ppetraki> MTecknology, no worries
[19:40] <Daviey> hggdh: I just need to do something, will then upload a bunch of new packages to the PPA.
[19:41] <Daviey> hggdh: build time dependant.
[19:44] <hggdh> zul: can I use the test rig?
[19:44] <zul> hggdh: yep
[19:44] <hggdh> thanks
[20:00] <SpaceBass> hey folks
[20:00] <SpaceBass> anyone know of a good package to create RSS feeds of files in a directory structure?
[20:02] <CyrusB_1> where is the dir where services / processes are started on startup? ubuntu 10.04. I remember there might be a few .. rc.d is ringing a bell
[20:05] <pmatulis> CyrusB_1: /etc/init.d and /etc/init
[20:07] <CyrusB_1> automaticlly get run on startup?
[20:08] <CyrusB_1> thought that was just for services, and places to place scripts
[20:14] <pmatulis> CyrusB_1: then look into /etc/rc2.d for sysvinit scripts and look in each and every script under /etc/init for upstart scripts/jobs (grep 'start on runlevel \[2' /etc/init/*)
[20:14] <CyrusB_1> yep, got it
[20:14] <CyrusB_1> thank you pmatulis
[20:34] <pryorda> Upgrading from 32bit to 64bit?
[20:44] <hggdh> Daviey: interesting... installed NTP, and -- on start, about 20 min from boot -- got a time correction of 0.2 seconds
[20:45] <mama21mama> hi
[20:46] <mama21mama> "solution to this?"
[20:46] <mama21mama> open /dev/null failed: No such file or directory
[20:46] <mama21mama> ubuntu server 10.04
[20:46] <guntbert> hggdh: it takes some time for the daemon to find the correct values for the "hardware clock"
[20:46] <mama21mama> "not find a way"
[20:47] <hggdh> guntbert: not if you are running with -g -- which is what we default to
[20:49] <guntbert> hggdh: no, I meant that ntpd tries to get a "feel" for the hardware clock (is it slow, does it wobble?) - and that takes some time so its estimates will not be too good from the start
[20:50] <guntbert> *it's
[20:52] <hggdh> guntbert: oh, OK, yes. In this case, since this was a brand new server isntall, there was no drift file
[20:53] <hggdh> guntbert: what surprised me is that we sync (right now) with ntpdate on interface up time; in my case, we are using the same NTP server for both ntpdate and ntp (after I installed NTP)
[20:53] <guntbert> hggdh: ah . "drift" was the word I was looking for - yes and that seems to explain the necessary correction
[20:54] <adam_g> ls/win 2
[20:55] <guntbert> hggdh: I'm sure the hardware clock might be off by that amount and without a drift estimate...
[21:01] <hggdh> guntbert: it might be. But to drift 0.2 seconds in 20 min is sort of surprising
[21:12]  * guntbert cannot be surprised by hardware glitches an more :)