[00:33] <zul> well looky that i can connect to a running libvirt using minicom
[02:28] <matt___> i'm trying to install ubuntu-server 10.04 on this new box that has an onboard software raid. right before i get to the partition guide it detects the raid and asks me if i want to initialize it. i select yes. when it gets to the the actual partition guide it doesn't show any disks. i ctrl+alt+f2 to a new console and run "dmraid -ay" . if i hit  go back in the instal process and re-detect the disks it does see it. when it finall
[02:38] <twb> Don't use fakeraid
[02:39] <multiHYP> twb: reinstalled everything
[02:39] <multiHYP> now with more knowledge on how to install, remove, autoremove even dependencies etc. thanks to .deb!
[02:40] <matt___> how do i get it to detect the raid then?
[02:40] <multiHYP> twb: don't know if you remember me even, but you helped me alot the other day, thought i should tell the good news :)
[02:41] <twb> OK
[02:42] <qman__> matt___, simply don't use fakeraid, use software raid during ubuntu setup
[02:42] <qman__> it offers no real advantage, performance or otherwise, and it's unnecessarily difficult
[02:42] <matt___> so using the built in sw raid and the onboard sw raid is going to give the same performance?
[02:42] <qman__> a real raid controller can be a different story, but not onboard fakeraid
[02:43] <qman__> pretty much
[02:43] <qman__> in either case the CPU still does all the work
[02:44] <squidly> I'm having an issues getting ubuntu natty to mount an ocfs2 cluster with two 10.10 nodes. Does anyone know of a fix for that?
[06:01] <GeorgeJ> Good morning folks
[06:03] <GeorgeJ> I've setup apache with suexec, fcgi and php5. Everything works fine and dandy except for this one problem I'm having. .php files in the root directory are executed correctly, however, .php files in subdirectories are not for some reason.
[06:04] <GeorgeJ> The vhost configuration files looks like this: http://pastebin.com/DRBa25nJ
[07:51] <koolhead11> hi all
[08:07] <Daviey> hey koolhead11 !
[08:09] <Daviey> hey rbasak o/
[08:10] <rbasak> hello Daviey!
[08:15] <jamespage> morning all
[08:19] <Daviey> hey jamespage
[08:20] <GeorgeJ> How would one install php 5.2 on the 11.04? Or would you recommend 10.10?
[08:23] <jamespage> Daviey: thanks for sponsoring that jenkins upload for me
[08:26] <koolhead11> Daviey: hello sir
[08:26] <koolhead11> jamespage: hey
[08:26] <jamespage> morning koolhead11
[08:27] <koolhead11> very good morning indeed!! :)
[08:28] <koolhead11> Daviey: i had to leave the custom partition in middle, as i got to work on something new.
[08:29] <koolhead11> jamespage: playing  with/on jenkins is on my todo as well :D
[08:30] <Daviey> jamespage: No problem.
[08:30] <jamespage> koolhead11: great! let me know how you get on with it
[08:30] <koolhead11> sure  :)
[08:38] <jamespage> Daviey: think I need to raise an FFE for tomcat6/7 - we are a couple of versions behind upstream and latest fixes a CVE
[08:40] <Daviey> jamespage: we are currently = to Debian.. is anyone already working on it?
[08:40] <jamespage> I'll work on tomcat7 - I'm one of the maintainers
[08:40] <jamespage> tomcat6 with fix is not released yet but should be shortly
[08:40] <jamespage> again I can work on that
[08:41] <Daviey> jamespage: Groovy, if it's updated in Debian first, it will taste sweeter.
[08:41] <jamespage> Daviey: ack
[08:41] <just-a-visitor> GeorgeJ: you can try this: http://thejibe.com/blog/10/5/php-5210-debs-ubuntu-104-lucid
[08:42] <jamespage> I work on tomat7 this morning and RFS ASAP - should not be to much of an issue
[08:42] <Daviey> just-a-visitor: heh, we had this concern when we jumped to 5.3
[08:43] <Daviey> *surely* it is compatiable now?
[08:45] <just-a-visitor> I must be missed that part. Not sure, but seems easy enough to test.
[08:47] <Daviey> I really wouldn't recommend using third party php debs, even if the binaries can be trusted - there is no security fix update path.
[08:48] <Daviey> jamespage: with your recent lxc fun, did you encounter bug 832111?
[08:49] <jamespage> no I did not
[08:49] <jamespage> they either when 'terminated' or would not stop 'running'
[08:50] <Daviey> how odd.. i don't think that portion of code has changed recently.
[08:51] <Daviey> I wonder if smoser can still reproduce it.
[08:51] <Daviey> jamespage: Can you add a comment with what you found, and ask smoser to reproduce it?
[08:55] <just-a-visitor> I also think, that the latest stable branch should be used from the repositories. Although there should be a good reason to use 5.2. Other than compiling from source there are not much one can do.
[08:57] <Daviey> just-a-visitor: what is the good reason?
[08:58] <Daviey> I'm not saying you are wrong, just trying to understand it more.
[08:59] <Daviey> Drupal does seem to be 5.3 compatiable now?
[09:02] <Tm_T> FYI I updated the channel-specific ops call
[09:05] <Daviey> Tm_T: thanks.
[09:06] <just-a-visitor> Well he did not mention. For example I have a legacy server that cannot be upgraded for technical reasons, that requires some older packages. But in this case I would go with the latest stable and fix the app instead.
[09:07] <Tm_T> Daviey: ofcourse, it might not contain all expected names, so if anything seems to be missing... (:
[09:07] <Daviey> just-a-visitor: Yeah, i missed the question - got disconnected :)
[09:07] <Daviey> Tm_T: I assume you added everyone on the access list?
[09:08] <Tm_T> not everyone, I left out few names that were missing there previously, as I expected they were left out intentionally
[09:11] <Daviey> Tm_T: What you really mean is, that you left out poor infinity?
[09:11] <Daviey> :)
[09:12] <Tm_T> Daviey: pretty much, yes (:)
[09:20] <soren> Daviey: Did you have any intention of working on this yourself? https://bugs.launchpad.net/nova/+bug/838386
[09:27] <Daviey> soren: I do actually have a branch which makes the test suite pass if carrot isn't installed.
[09:27] <Daviey> Defaults to kombu for the magic, and falls back to carrot..
[09:29] <Daviey> *However* there seems to be a bug in the test.py that doesn't allow conditonal test skipping
[09:29] <Daviey> Ie, using the skip_if decorator on a class, rather than a function.
[09:34] <soren> Oh.
[09:35] <jamespage> Daviey: could you poke this for me please https://launchpad.net/ubuntu/+source/maven-plugin-tools/2.8-1/+build/2767384
[09:35] <jamespage> should be OK now
[09:35] <jamespage> sync ordering caused it to FTBFS
[09:37] <Daviey> jamespage: poked
[09:37] <jamespage> ta
[09:42] <n20> Hello world! I'm trying to setup postfix/dovecote on my server, but with no luck - here's what I'm getting in my .err-log: postfix/smtpd[29751]: fatal: no SASL authentication mechanisms
[09:48] <ersi> n20: Install sasl authentication :) (I'm looking for the package name atm)
[09:48] <ersi> libsasl perhaps
[09:49] <n20> Oh ok, I'll try that out!
[09:49] <n20> thanks ersi
[09:50] <ersi> since it seems like you're all setup with the config - looks like all your missing is the sasl auth files :)
[09:50] <ersi> np!
[09:57] <n20> ersi: Hmm... it seems like I have libsasl2-2
[10:23] <n20> Grrr downtime
[10:30] <linocisco> how to give two IP for a server ?
[10:30] <linocisco> how to give two IP address on one network card ?
[10:38] <ersi> linocisco: http://www.cyberciti.biz/tips/ubuntu-linux-creating-ethernet-alias-for-eth0-network-device.html
[10:42] <linocisco> ersi: so what is the idea ?
[10:42] <linocisco> ersi:  eth0 with one IP and eth0:0 with another IP ?
[10:43] <ersi> yeah, it's called alias and you get 'one more IP'
[10:43] <ersi> Replace eth0 with eth1 or eth2 or whatever your interface is called
[10:46] <linocisco> ersi: if i want to use internet , I have to put only one default gatway, right ?
[11:38] <soren> Daviey, zul: Do you guys have any fixes for LXC lined up? I got the impression there were some issues.
[11:38] <zul> soren: yeah there is a fix for lxc in the latest libvirt
[11:38] <soren> zul: Nothing for Nova itself?
[11:38] <zul> soren: and im working on some for nova
[11:39] <soren> Cool. What are they, and what's their ETA?
[11:39] <zul> soren: its mostly shutdown issues for lxc and i dont have an eta on them just yet
[11:39] <zul> still technically not bod (beginning of the day for me ;)
[11:41] <soren> ok :)
[11:59] <Daviey> soren: Yeah. smoser, zul and jamespage have all found various issues.. but it's ok because zul is ON THE CASE.
[11:59] <zul> hah
[11:59]  * zul punches in his ticket for the day
[12:12] <n20> I'm (still) trying to setup postfix/dovecote on my server, but with no luck - here's what I'm getting in my .err-log: postfix/smtpd[29751]: fatal: no SASL authentication mechanisms. I've set it up according to the guide, so I don't quite know what to do
[12:24] <aljosa> i've upgraded system "apt-get dist-upgrade" on a box with ibm db2 installed and now i have some issues with db2. anybody knows where i could discuss this since it's a package from "partner" repository?
[12:32] <zul> soren: yeah but if you want to use lxc with nova use the most current oneiric images and libvirt
[12:39] <soren> zul: What are the problems with earlier libvirts (if I wanted to backport the fixes)?
[12:41] <zul> soren: there is a couple of issues, it was using things like ip and ifconfig in the containers  and /dev/ptmx was a symlink to a host so, things like ssh would not work correctly
[12:43] <zul> soren: if you can figure out the lxc chardev stuff like you did for uml in libvirt that would be awesome
[12:54] <soren> zul: What's the problem exactly?
[12:54] <zul> well for one this is no serial support ;)
[12:55] <zul> so it doesnt wrte whats in the console to a file
[12:55] <soren> Of course there's no serial support. It's not virtualised :)
[13:01] <smoser> we should not need *anything* from any new libvirt to have functional lxc
[13:02] <hallyn> jdstrand: hey - on bug 840925, could the security team comment on whether you have any concerns?
[13:05] <zul> smoser: right
[13:07] <smoser> i dont want anyone spreading fud saying "we have to upgrade"
[13:07] <soren> zul: Err... so which is it?
[13:07] <smoser> bug 832123 is the one that was the biggest pain for us
[13:08] <koolhead11> will someone show some love to this bug? https://bugs.launchpad.net/ubuntu/+source/dbconfig-common/+bug/807038  This is becoming my show stopper in ensemble formula writing process
[13:08] <zul> soren: see above
[13:08] <smoser> soren, you were aware of bug 828211, but that did not affect hte ubuntu images, only my ttylinux images
[13:12] <soren> zul: I am looking above. Right where you say "if you want to use lxc with nova use the most current oneiric images and libvirt".
[13:12] <soren> But ok. ttylinux images. Got it.
[13:13] <smoser> but my latest ttylinux image has a work around for ip
[13:13] <zul> soren: correct thats what I would do use the most current oneiric images and libvirt
[13:13] <zul> smoser: what work around?
[13:14] <smoser> it has a suitable 'ip' command
[13:14] <zul> a ok
[13:32] <hallyn> jdstrand: thanks
[14:38] <hallyn> SpamapS: so bug 820675 (and its ilk), are they there to be discussed during UDS, or do you have a plan, or do we need to talk, or what?
[14:46] <idlemind324> so i have an openldap server the documentation indicates my uid's should start at 1000 in ldap but ubuntu's local users start at 1000. isn't having duplicate uids bad?
[14:54] <Daviey> Everyone having fun?
[14:55] <idlemind324> of course
[15:00] <lynxman> Daviey: yes \o/
[15:00] <Daviey> great!
[15:10] <robbiew> Daviey: am I right to assume that an 11.10 Orchestra server could install older client images, i.e. 10.04LTS or 11.04?
[15:11] <kirkland> robbiew: yes, it can
[15:11] <robbiew> kirkland: cool, thanks "Daviey" :P
[15:11] <kirkland> robbiew: it won't be a full "orchestra client", with rsyslog and nagios integration, etc.
[15:11] <robbiew> ack
[15:11] <robbiew> that's what I thought
[15:11] <kirkland> robbiew: because our "ubuntu-orchestra-client" package doesn't exist until 11.10
[15:11] <robbiew> exactly
[15:12] <kirkland> robbiew: but it can install any older ubuntu system
[15:12] <kirkland> robbiew: using a basic (or custom) preseed
[15:12] <robbiew> kirkland: I wonder if it makes sense to push for a microupdate (or whatever it's called) exception
[15:12] <robbiew> for SURs
[15:12] <robbiew> SRUs
[15:12] <robbiew> like landscape has for its' client
[15:13] <robbiew> Daviey: ^^^^thoughts?
[15:13] <kirkland> robbiew: personally, i don't think so
[15:13] <kirkland> robbiew: we'll have 12.04 LTS in a couple of months
[15:13] <kirkland> robbiew: and there are some other ways of going about doing it
[15:13] <kirkland> robbiew: like ppa:orchestra/ppa
[15:14] <kirkland> robbiew: but that's your team's call ;-)
[15:14] <robbiew> kirkland: yeah...and a ppa is less "support" hassle
[15:19]  * patdk-wk starts his ldap users at 100000, therefore skipping right over all known issues
[15:21] <idlemind324> In the Setting up ACL portion of Ubuntu Server Guide (OpenLDAP) I am trying to run: ldapsearch -xLLL -b cn=config -D cn=admin,cn=config -W olcDatabase=hdb olcAccess it executes and asks me for my password but when i enter my admin ldap password it says authetnication failed
[15:21] <idlemind324> any ideas?
[15:22] <idlemind324> the interesting part is i don't remember any user named cn=admin,cn=config ever being added as part of the guide but i may be interpreting the -D incorrectly
[15:31]  * Daviey read scrollback
[15:32] <SpamapS> hallyn: I do have a plan to use the network-services upstart job to control these "en masse" .. the start on for libvirt-bin is a little tricky (and, I think, not doing what you might think it does) but it probably just needs to be 'start on runlevel [2345]'
[15:32] <Daviey> robbiew: Yeah, i wouldn't like to SRU that.
[15:32] <Daviey> Adding a PPA makes sense if people /really/ want it.
[15:33] <SpamapS> As long as the PPA sees maintenance
[15:35] <Daviey> SpamapS: I don't think it would need to be blessed as such, and the burden would probably be low.
[15:35] <robbiew> yeah...I think that's best
[15:35] <Daviey> As in, not enabled by default, or anything daft.
[15:35] <SpamapS> No no no
[15:35] <SpamapS> just that, don't create a PPA for one user's request, and then never upload again
[15:35] <robbiew> yeah...just crack
[15:36] <robbiew> better to just say "upgrade!"
[15:36] <Daviey> feel the latest baby!
[15:36] <robbiew> then have them curse us for any bugs they encounter
[15:36] <memoryleak> Is there a programm to check how much RAM I can allocate from the VPS I'm using?
[15:36] <Daviey> lifeless: You rock my world.
[15:44] <melvincv> How may I start an app with Ubuntu server? Is it chkconfig command?
[15:45] <melvincv> !chkconfig
[15:45] <Daviey> melvincv: need more detail.
[15:45] <patdk-wk> this isn't redhat/centos
[15:46] <patdk-wk> update-rc.d
[15:46] <melvincv> Alright, suppose I need to start the samba service along with Ubuntu server. What do I need to do to ensure that it runs on startup?
[15:47] <Daviey> melvincv: it should start by default.
[15:48] <melvincv> Is it the same with all other services like apache, dovecot, postfix, nis, etc.
[15:48] <patdk-wk> normally, unless /etc/default/xxx says disabled
[15:49]  * melvincv has come over here from RHEL 5 and wants to know if Ubuntu server is better than CentOS.
[15:50] <Daviey> melvincv: If that is your primary benchmark, we are doing pretty well. :)
[15:51] <melvincv> Great! So I'll get a test server running. But why doesn't 'service network restart' work? I need to learn new commands. :(
[15:53]  * melvincv wonders if the guide in the channel topic is the best place to start?
[15:53] <Daviey> $ sudo /etc/init.d/networking restart
[15:54]  * melvincv wonders, "Yes, but why the extra typing???"
[15:55] <hallyn> SpamapS: sounds good, thx
[15:55] <melvincv> Well, long commands for a better server OS is actually a good bet.
[15:56] <SpamapS> Daviey: is that *actually* what we're recommending?
[15:56] <SpamapS> that seems.. very broken
[15:56] <SpamapS> We might want to think about renaming /etc/init/networking.conf .. it causes too much confusion
[15:57] <RoAkSoAx> Daviey: almost there for the demo
[15:58] <Daviey> SpamapS: That is how i do all my non-upstart jobs.
[15:58] <Daviey> Well, even some which are upstart i still do that :./
[15:58] <Daviey> sudo reboot, will also restart your networking :)
[15:59] <SpamapS> Daviey: we should be suggesting the use of the service command.
[15:59] <SpamapS> but even then..
[15:59] <SpamapS> service networking restart does the wrong thing. :-P
[16:00]  * melvincv thinks that's weird...
[16:00] <melvincv> Maybe I should file a bug?
[16:01]  * melvincv is checking out Getting Involved.
[16:03] <tumbleweed> anyone know who runs https://imagestor.canonical.com/ ? it appears that libcurl doesn't like it's cert
[16:03] <tumbleweed> *its
[16:03] <tumbleweed> *imagestore
[16:04] <patdk-wk> did you tell curl where the cert store is?
[16:04] <tumbleweed> oh I've given it the right CA cert
[16:04] <tumbleweed> it's not validating it
[16:05] <tumbleweed> openssl has no problem with it
[16:09] <patdk-wk> tumbleweed, ya, I can't get openssl to verify it, looks like it's not passing the intermediant certs
[16:10] <tumbleweed> yeah, but if you give curl the intermediate cert, that doesn't help. I'm getting the feeling gnutls doesnt' like wildcard certs
[16:10] <patdk-wk> doubt that is it
[16:11] <patdk-wk> how do I test with gnutls?
[16:15] <tumbleweed> patdk-wk: gnutls-cli
[16:16] <tumbleweed> no idea why gnutls doesn't like it
[16:17] <tumbleweed> "Fatal error: Error in the certificate"
[16:18] <patdk-wk> works for me, atleast with other wildcard certs
[16:18] <tumbleweed> yup
[16:19] <patdk-wk> but both gnutls and openssl have issues iwth that one
[16:23] <tumbleweed> from what I can tell, gnutls is far stricter about the intermediate certificate chain than openssl is (give openssl the certificate that signed this one, as a CA certifacet, and it's happy)
[16:24] <ersi> Maybe it's issued by a retard CA who has gotten it's trust revoked
[16:24] <tumbleweed> which really comes back to why I came in here. Anyone know who runs that service, it's clearly got a problematic configuration
[16:24] <patdk-wk> ersi, nope
[16:28] <patdk-wk> looks like that root is not trusted by ubuntu at all, but is via firefox
[16:37] <SpamapS> utlemming: ping re upload to oneiric...
[16:37] <SpamapS> utlemming: start on runlevel [2345] and local-filesystems and net-device-up IFACE!=lo
[16:38] <SpamapS> this is redundant in oneiric
[16:38] <SpamapS> runlevel 2 happens only after all interfaces in /etc/network/interfaces are up
[16:38] <SpamapS> zul: ^^
[16:38] <SpamapS> it also breaks runlevel 1
[16:38] <zul> dang it
[16:39] <utlemming> SpamapS: interesting...okay, redoing
[16:39] <SpamapS> local-filesystems, btw, *ALWAYS* happens before runlevel 2 , even going back to karmic
[16:39] <utlemming> SpamapS: https://bugs.launchpad.net/ubuntu/+source/squid/+bug/831628
[16:40] <SpamapS> Ahh, I'd have marked that Invalid. :) runlevel 2 happens after all automatic fsck's... so whatever he's seeing is more than a little surprising
[16:41] <SpamapS> I do actually think it needs to be just 'runlevel [2345]
[16:41] <SpamapS> hmm
[16:41] <SpamapS> I recall now that the reason for the 'or net-device-up IFACE!=lo' was to catch interfaces that come up after runlevel 2...
[16:44] <SpamapS> utlemming: so yeah, I think you can just do 'runlevel [2345]'
[16:46] <utlemming> SpamapS: for my education, do you have a link to where that is documented?
[16:46] <SpamapS> Heh.. http://upstart.ubuntu.com/cookbook/
[16:46] <SpamapS> utlemming: upstart can be a mess. Trying to simplify it so we don't have to think so hard for each upstart job.
[16:47] <zul> its like rocket science ;)
[16:47] <SpamapS> Yeah, it should be more like making cheeseburgers. :)
[16:49] <utlemming> as in "I can has cheezburgers"
[16:50] <SpamapS> ohaaaaiii
[16:54] <utlemming> SpamapS: how did the merge regress bug #561779? I was a good little boy and tested it. Also, I deleted the merge request -- so it didn't make it all the way.
[17:00] <SpamapS> utlemming: because on transition from runlevel 1 -> 2 , local-filesystems won't be emitted.
[18:01] <jetole> Hey guys. Does anyone know how I can force quit rmmod when I used the -w (wait option). I changed my mind on waiting for these modules but now rmmod won't respond to kill -9 (it's not even defunct) and I killed the parent process, bash, but now it's just been inherited by init
[18:02] <jetole> ... it's currently in a uninteruptable sleep
[18:18] <Daviey> smoser / utlemming: Does aufs not work in oneiric cloud images?
[18:19] <smoser> i dont know why it would not. other than possibly module not being present in -virtual
[18:20] <Daviey> smoser: seems that is the case
[18:20] <Daviey> CONFIG_AUFS_FS=m .. hmm
[18:22] <jetole> my question can be ignored. appearently a process in a uninteruptable sleep can't really be forced to quit. even kill -9 has no effect and no way around that other then rebooting. If the process changes it's state to not being in the sleep then it will receive the signal immediately and quit but while it's in that state it's untouchable
[18:22] <Daviey> smoser: we also don't have an admin group?
[18:22]  * jetole goos to the data center. bbiab
[18:25] <smoser> Daviey, i dont know that i care too much about admin... its likely not a regression (unless it is of the live-build tools)
[18:25] <smoser> if it is a regression, then we need it fixed.
[18:26] <smoser> please open a bug on aufs against kernel
[18:27] <Daviey> smoser: I just threw the question into -kernel.
[18:27] <Daviey> smoser: I just used a tool which expected the admin group to exist.
[18:27] <Daviey> Really, we should be mimicing the main flavour. :)
[18:27] <Daviey> IMO
[18:33] <GrueMaster> iscsi on arm is going to be a major headache.  I can get both target & initiator to work on arm, but I can't get iscsi-rootfs to work.  Not sure if iscsi is somehow corrupting initramfs.
[18:35] <koolhead17> hi guys
[18:38] <adam_g> GrueMaster: what are the symptoms?
[18:38] <GrueMaster> Garbage output on serial console when loading initramfs.
[18:41] <GrueMaster> sigh, now I'm fighting screen console output.
[18:42] <smoser> Daviey, also bug 844166
[18:43] <soren> jetole: Yeah, that's pretty much what "uninterruptible" means :)
[18:44] <soren> jetole: "defunct" means a process has terminated, but its parent hasn't called wait() on it. Quite different things.
[18:44] <smoser> and bug 844164
[18:45] <koolhead17> what/how will i proceed if i have some custom scripts initiated at first ubuntu deployment on laptop, like cloud-init ?
[18:46] <soren> jetole: All resources pertaining to defunct processes have been released except for one final thing: Their exit code. The kernel can't delete the process from its process table until someone has bothered to check its exit code (By calling wait*()).
[18:49] <Daviey> thanks smoser
[18:49] <koolhead17> hey Daviey
[18:49] <Daviey> hey koolhead17
[18:49] <koolhead17> Daviey: i needed answer of this question :)
[18:50] <Daviey> GrueMaster: you might be hitting bug 838809
[18:50] <koolhead17> shall i request a blueprint :D
[18:51] <Daviey> koolhead17: zul did do some work to get cloud-init working on normal hardware last cycle.
[18:51] <Daviey> Don't know if it got anywhere.
[18:51] <zul> hmmm?
[18:51] <koolhead17> ooh ok. it will be cool to have in end option coming in ISO install saying paste custom script, like cobbler, cloud-init or amazon :D
[18:51] <Daviey> zul: didn't you?
[18:52] <Daviey> koolhead17: well there is a late_command you can preseed whilst still in the installer.
[18:52] <koolhead17> Daviey: yeah am using that, i meant in general booting from ISO in hardware. if there is sumthing
[18:52] <koolhead17> :)
[18:52] <koolhead17> it will be really cool
[18:53] <GrueMaster> Daviey: I don't think this applies to me.  What I am seeing is far different.  Actual garbage on the console when the kernel jumps into the initramfs.
[18:53] <koolhead17> RoAkSoAx: hello
[18:53] <Daviey> GrueMaster: ok
[18:55]  * koolhead17 bows to robbiew
[18:55] <Daviey> koolhead17: You could throw something into /etc/rc.local and make sure it's flushed
[18:55] <koolhead17> aah, i did that only. :)
[18:56] <koolhead17> Daviey: whom should i run after for the bug i filled :P
[18:56] <Daviey> koolhead17: As i said, it really needs more debug info. :(
[18:56] <Daviey> It needs enough information for someone to be able to reproduce it
[18:57] <koolhead17> Daviey: i provided hardware detail and scenario as well :( and  syslog ceases after assigning the IP
[18:57] <jetole> soren: Yeah I understand what a zombie process is. Uninteruptable sleep is not a zombie or defunct process though and it can consume RAM and when that sleep ends then it can continue to use other resources
[18:57]  * jetole just got back from the data center
[18:57] <koolhead17> with the big Success! :P
[18:58]  * koolhead17 was inside there for past 2 days to get cobbler oneiric love working
[18:58] <jetole> soren: uninteruptable sleep can be, for example, when the process makes a sys call and is waiting on input from the kernel, in this case the kernel is blocking signals
[18:58] <magicblaze007> previously my apache wsgi apps were located in /var/www -- for some reason, now apache looks for one of those apps in /etc/apache2/htdocs -- any ideas how to fix this? I tried adding DocumentRoot in the virtualhost entry, but that doesnt help either
[18:59] <soren> jetole: Yes, exactly.
[18:59] <soren> jetole: And uninterruptible means exactly that: It cannot be interrupted (by, say, kill -9).
[18:59] <RoAkSoAx> koolhead17: howdy!! (sorry was out for lunch)
[18:59] <jetole> soren: take a process that calls a system function and the system function, the code withing the kernel itself is either waiting on something or processing something to complete the call. the process would be in uninteruptable sleep until the kernel / system function returns to it
[19:00] <koolhead17> Daviey: i will make video of it now and post it somewhere. haha :D
[19:00] <jetole> soren: well I understand that now but it's Linux. You never know who found a hack for what
[19:00] <koolhead17> RoAkSoAx: long time. how have you been.
[19:00] <Daviey> koolhead17: I mean, it needs a log showing failure.. or more debug info.
[19:01] <jetole> soren: in case it was say tweaking something in /proc or who knows. When I asked I thought there may have been a work around and when I said you can ignore my question because I now understand the answer, then what I meant was you can ignore my question because I now understand the answer :P
[19:01] <RoAkSoAx> koolhead17: pretty good, yourself?
[19:01] <jetole> lol
[19:01] <soren> jetole: I didn't see that until after I responded :)
[19:02] <koolhead17> Daviey: but i left the machine for over 1 hr and notthing happens
[19:02] <koolhead17> https://bugs.launchpad.net/ubuntu/+source/debian-installer/+bug/842509
[19:02] <jetole> soren: heh. It's cool. I'm justing teasing
[19:02] <koolhead17> i have added exact line the logfile leaves in syslong
[19:02] <koolhead17> *syslog
[19:03] <koolhead17> and yes i have passed "kopts" in my preseed file to listen to eth1
[19:03]  * jetole is now looking into a open-iscsi bug in 8.10 and how to resolve. Appearently open-iscsi in 8.10 is two different versions of open-iscsi wraped into one package that leads to errors
[19:03] <koolhead17> RoAkSoAx: fighting with custom partition via cobbler/preseed, all tricks been unsucessful
[19:03] <Daviey> koolhead17: Did you try updating your cobbler iso pool to the latest daily ISO?
[19:03] <RoAkSoAx> koolhead17: ah yeah, there are some limitations when doing so
[19:03] <jetole> https://bugs.launchpad.net/ubuntu/+source/open-iscsi/+bug/289470 yay!
[19:03] <RoAkSoAx> koolhead17: at least for no questions asked
[19:04] <jetole> soren: it just so happens there is someone mentioned in this bug by the name of soren
[19:04] <jetole> soren: don't suppose that's you?
[19:04] <koolhead17> Daviey: i did not unfortunately. :(
[19:05] <koolhead17> will get back to you after trying that
[19:05] <soren> jetole: sorry, which bug?
[19:05] <Daviey> koolhead17: Yeah, the kernel bumped - so it could be related to that
[19:05] <jetole> soren: https://bugs.launchpad.net/ubuntu/+source/open-iscsi/+bug/289470 - open-iscsi: two versions compiled to one package as a downstream from a debian error
[19:06] <koolhead17> Daviey: okey.no more troubling you until am done testing new ISO :)
[19:06] <RoAkSoAx> koolhead17: but are you trying to achieve anything using cloud-init and orchestra?
[19:06] <soren> jetole: Yup, that would be me.
[19:06] <Daviey> koolhead17: cool!
[19:06] <jetole> soren: well maybe you can save me sometime. so far I have read to where a resolution exists in a ppa. Should I go with that solution or should I continue reading to the end of the bug report?
[19:07] <koolhead17> RoAkSoAx: nopes working on some automation in datcentre, no orchestra
[19:07] <RoAkSoAx> koolhead17: oh ok
[19:07] <soren> jetole: No idea. It's been almost three years. :)
[19:07] <koolhead17> Daviey: i just loved the late command option in preseed, its been life saver for me. :D
[19:08] <RoAkSoAx> koolhead17: in ensemble/orchestra we are doing a late_command to pass cloud-init meta-data
[19:08] <koolhead17> i changed repo info/pushed interface static address and many more other stuff :D
[19:08] <koolhead17> RoAkSoAx: i have not touched orchestra, i have worked on ensemble though
[19:08] <Daviey> koolhead17: heh
[19:08] <koolhead17> and am stuck in ensemble because of a stupid bug
[19:09] <Daviey> :(
[19:09] <koolhead17> RoAkSoAx: https://bugs.launchpad.net/ubuntu/+source/dbconfig-common/+bug/807038
[19:09] <koolhead17> Daviey: am hoping to get the hostname bug issue fixed too in new image :)
[19:10] <Daviey> koolhead17: sweet
[19:11]  * koolhead17 wants to work/try out jenkins soon
[19:12] <RoAkSoAx> koolhead17: orchestra is just a wrapper for cobbler where we integrate various stuff
[19:12] <RoAkSoAx> koolhead17: so it is really cobbler when I'[m talkiing about preseeding cloud-init stuff
[19:12] <koolhead17> RoAkSoAx: ooh. ok
[19:13] <magicblaze007> on ubuntu, is the default directory for apache /etc/apache2/htdocs?
[19:13] <popey> What is the best supported method of installing Ubuntu from a 64-bit ISO onto a machine that has no optical drive?
[19:13] <koolhead17> magicblaze007: /var/www
[19:13] <popey> bonus points if you dont mention PXE
[19:13] <koolhead17> boot from usb :D
[19:13] <popey> how do I get the iso on the usb?
[19:14] <RoAkSoAx> popey: usb-creator
[19:14] <magicblaze007> koolhead17: something is wrong with my config -- http://paste.pocoo.org/show/471977/ -- for some reason -- myserver.com/chat goes and looks up /etc/apache2/htdocs...any ideas why?
[19:14] <popey> have you ever tested usb-creator with a server cd?
[19:14] <koolhead17> popey: how will it differ?
[19:15] <popey> usb creator is designed to make live cd usb images
[19:15] <popey> not alternate d-i ones
[19:15] <popey> AIUI
[19:16] <koolhead17> will give it a try tomorrow then
[19:18]  * popey tries it now
[19:18] <magicblaze007> anyone can help me debug my httpd.conf file please? http://paste.pocoo.org/show/471977/
[19:24] <popey> ..and the answer is no, it doesn't work
[19:24] <Daviey> koolhead17: How are you getting on with bug 795481?
[19:24] <popey> you get dropped to a boot: prompt with a vesamenu.c32: not a Com32R image error.
[19:25] <Daviey> Hmm, it worked in Natty.
[19:26] <koolhead17> Daviey: https://bugs.launchpad.net/ubuntu/+source/dbconfig-common/+bug/807038
[19:26] <koolhead17> stopped me from going there
[19:26] <koolhead17> and i moved on to cobbler :(
[19:26] <koolhead17> automation stuff
[19:26] <Daviey> RoAkSoAx: Have you tried creating usb sticks recently?
[19:27] <Daviey> (being related to testdrive, and all)
[19:27] <RoAkSoAx> Daviey: like couple weeks ago
[19:27] <RoAkSoAx> Daviey: what's wrong?
[19:28] <popey> unetbootin makes a usb stick that fails to find the CD ROM during install.
[19:28] <popey> usb creator fails with a boot: prompt
[19:29] <Daviey> popey: unetbootin is a PoS last time i tried it.
[19:29]  * popey tries dd
[19:30] <koolhead17> unetbootin is cool. i use that only
[19:30] <popey> for server CDs?
[19:31] <koolhead17> popey: for any ISO IMHO. So give it a try :)
[19:31] <popey> I did. As I said, it fails.
[19:32] <T3CHKOMMIE> hey guys, im having a kernel emergecny, can anyone help me figure out why my custom kernal is panicing when it tryes to mount the VFS /dev/sda1?
[19:32] <Daviey> Annoyingly i don't have a spare usb pendrive to hand, thanks to dendrobates :/
[19:32] <Daviey> T3CHKOMMIE: It sounds like the bug is "custom kernel"
[19:33] <Daviey> popey: Are you using usb-creator on natty or oneiric?
[19:33] <T3CHKOMMIE> Daviey, im not fan of this, but my insturctors apparently dont know what they are doing and ive read all the instructions, i think i just dont concetpually understand what is going on. i loads the kernal, trys to boot then panics when it tries to mount the vfs
[19:34] <Daviey> T3CHKOMMIE: Why are you using a custom kernel?
[19:35] <T3CHKOMMIE> Daviey, its a dumb class requirment.
[19:36] <T3CHKOMMIE> the require us to do this shit on modern oses with 3 year old lab instructions.... and when shit breaks they say "google it"
[19:36] <T3CHKOMMIE> so ive been here for 5 horus and cant get the GD thing to boot :@
[19:36] <popey> Daviey: oneiric
[19:37] <Daviey> popey: ah!
[19:38] <GrueMaster> T3CHKOMMIE: Sounds like it cant find initrd.
[19:38] <Daviey> popey: I suspect it's a bug in that. :(
[19:38] <T3CHKOMMIE> well, looks like with 10.4 i have to compile this orginally with some extra scripts.
[19:38] <T3CHKOMMIE> F word.
[19:39]  * popey reboot into natty and will test all three methods again
[19:39] <T3CHKOMMIE> GrueMaster, do you know of a way for me to get this kernel working on 10.04? if that is the problem?
[19:40] <Daviey> popey: cool, thanks
[19:40] <GrueMaster> T3CHKOMMIE: Not really.  I don't do custom kernel stuff.  But you should be able to boot to a standard kernel and rebuild the initramfs.
[19:41] <GrueMaster> Use "sudo update-initramfs -k <kernel>"
[19:41] <T3CHKOMMIE> awesome let me give that a shot.
[19:43] <T3CHKOMMIE> GrueMaster, if the kernel is already compiled is there a way to do this without recompiling?
[19:44] <GrueMaster> if the kernel & modules are installed (make install) then update-initramfs should build an initrd.
[19:49] <T3CHKOMMIE> ok, seems like im still having the problem, should i have something like initramfs=/dev/sda1 in the 40_custom file?
[19:51] <GrueMaster> Huh?  I don't even follow what you are trying to accomplish here.  what is your boot cmdline?  You should have grub (or other boot loader) loading initramfs=/boot/initrd.img or something like that.
[19:51] <GrueMaster> Look at a running system to get a better understanding.
[19:55] <Daviey> popey: how is it looking?
[19:56] <GrueMaster> Daviey: Here is the console output from booting my panda with iscsi rootfs.  http://paste.ubuntu.com/684686/  The only difference I can think of is that I added iscsi to /etc/initramfs-tools/modules and ran update-initramfs -u per http://etherboot.org/wiki/sanboot/ubuntu_iscsi
[19:56] <popey> dding Daviey
[19:58] <Daviey> popey: I don't have confidence that dd will work.. i think it will fail to find something early on during the installer.
[19:58] <Daviey> popey: did natty fail?
[19:59] <Daviey> GrueMaster: The kernel is on the sd?
[19:59] <GrueMaster> yes
[20:00] <Daviey> GrueMaster: do you have it connected to a network?
[20:00] <GrueMaster> Obviously.
[20:00] <popey> Daviey: dding to a partition
[20:01] <GrueMaster> Get beyond the basics, please.  I have been doing netinstalls on these systems since A2.
[20:01] <Daviey> GrueMaster: Can you try pinging?  I wonder if when the root fs comes up it is dropping the serial console.
[20:01] <popey> Daviey: which also fails
[20:01] <GrueMaster> No net.
[20:02] <GrueMaster> It isn't getting that far.
[20:02] <Daviey> GrueMaster: I'm trying to help here.  If you are going to be like that, it can wait until tomorrow.
[20:02] <popey> Daviey: trying usb creator
[20:03] <popey> and dbus errors in usb-creator
[20:03] <Daviey> popey: in natty?
[20:03] <popey> yes
[20:03] <Daviey> Gah
[20:03] <GrueMaster> Daviey: Sorry, but I have been doing system QA for the better part of 20 years, and I tend to get touchy when someone asks me noob questions (like "did you plug it in?")
[20:04] <Daviey> GrueMaster: TBH, my brain wasn't in gear.. If i had thought about it, i'd have realised that it must be connected for iscsi.  It was really a pre-question to see if you could ping it, confirming if it was still booting ok - and the serial console had just gone AWOL.
[20:04] <GrueMaster> Fair enough.  :P
[20:04] <popey> Daviey: formatted and trying usb creator
[20:05] <GrueMaster> From what it looks like, something in the initramfs is getting corrupted.
[20:05] <GrueMaster> The same system booted just fine to USB prior to updating the initramfs.
[20:06] <GrueMaster> That's why I find it so odd.
[20:07] <Daviey> GrueMaster: do you have a pastebin of a loca filesystem boot?  I don't want to set mine up at the moment.
[20:07] <Daviey> local*
[20:07] <GrueMaster> Sure, give me a sec.
[20:09] <Daviey> GrueMaster: hold fire, i have one
[20:09] <Daviey> GrueMaster: So.. http://pb.daviey.com/JOuM/
[20:09] <Daviey> GrueMaster: you get crap, where you should get - [ 2.430053] console [ttyO2] enabled
[20:09] <popey> Daviey: fails on natty too, usb creator that is
[20:09] <Daviey> Now i suspect it's a baud mismatch or something
[20:09] <popey> this is why I asked initially what the supported way was
[20:09] <Daviey> popey: The supported way is using usb-creator. :/
[20:10] <popey> hah
[20:10]  * popey keeps trying
[20:11] <Daviey> Would someone be able to reproduce using usb-creator-gtk to create a server pendrive, and check you can get to the installer.. Seems to be broken for popey, and it would be nice if it could be reproduced.
[20:11] <Daviey> kkthnx
[20:11] <popey> gonna try the netinst image
[20:11] <Daviey> mini.iso?
[20:12] <jetole> soren: ah thanks. Sorry for the long delay but got called into a meeting
[20:14] <Daviey> GrueMaster: Your /etc/init/ttyO2.conf settings are correct?
[20:14] <GrueMaster> Yes.  Verified.
[20:17] <Daviey> GrueMaster: So i have seen that crap on the output when i put wrong speed settings.. but if you are using the same settings as you normally use, i can't see how it could be that
[20:19] <GrueMaster> Ok, I checked something else, and apparently my SD had some corruption.  Now getting further.
[20:20] <GrueMaster> Yet another bug, it would seem.
[20:20] <Daviey> gah
[20:20] <GrueMaster> So now I appear to be stuck at the bug you mentioned earlier.
[20:21] <Daviey> GrueMaster: Does text input make more rubbish show on the console?
[20:22] <GrueMaster> Not after zeroing & reimaging the SD.
[20:22] <Daviey> GrueMaster: Ah!  You are getting good text now?
[20:22] <Daviey> and seeing bug 838809 ?
[20:22] <GrueMaster> Yes.  The initrd was indeed corrupt, but due to SD.
[20:23] <GrueMaster> I think so.  It just spewed a bunch of ext4 transport errors.
[20:24] <GrueMaster> But it has mounted the iscsi volume and is running through init.
[20:24] <Daviey> GrueMaster: hurray!
[20:24] <Daviey> I'd never have considered the SD being bad :/
[20:24] <GrueMaster> Well, until I see login, I am not calling this a full success.
[20:25] <GrueMaster> Yea.  The SD issue I have seen from time to time, but usually when reflashing daily images on it. These particular systems have dedicated SD cards that don't get reflashed, just reformatted.
[20:32] <Daviey> GrueMaster: ah
[20:34] <popey> Daviey: bug 645818
[20:36] <Daviey> popey: that is weird.. i know i used usb pendrives for server during the natty cycle.. :/
[20:41] <koolhead17> so popey was right?
[21:09] <Daviey> uvirtbot: did you see some issues we were seeing with the cloud images?
[21:09] <Daviey> err, utlemming
[21:10] <utlemming> Daviey: pong
[21:10] <utlemming> Daviey: the couple of issues -- admin user missing and build.info missing -- have been fixed
[21:11] <utlemming> Daviey: I'm finalizing the testing now
[21:12] <Daviey> utlemming: rocking!
[21:13] <utlemming> Daviey: you're in luck -- fixes confirmed. Tonights builds should have those included...unless you want them sooner. I could have new images ETA 4hours
[21:13] <Daviey> utlemming: i hope to be asleep by then :)
[21:14] <utlemming> Daviey: lol
[21:14] <Daviey> utlemming: what is the plan to diff the output of the old toolscheme to lb?
[21:15] <utlemming> Daviey: incdiently I was working on the image diff tool earlier today. I'll have an initial pass completed in a couple hours
[21:16] <utlemming> Daviey: I'm going to do side-by-side builds of old versus new for the next week and gets diffs on it.
[21:17] <Daviey> utlemming: cool!
[22:06] <talntid> had a full HDD.... deleted 9.1GB of logs, but it's still showing full.. any way to update without rebooting?
[22:12] <patdk-lap> it will still show as full, untill the files are closed
[22:12] <patdk-lap> sounds like those log files are still open by some program
[22:12] <talntid> hmm... I see...
[22:13] <talntid> lsof should be able to take care of that, right?
[22:13] <patdk-lap> I believe so
[22:13] <RyanP> It's quite possibly syslog keeping the space.
[22:24] <talntid> did a lsof | grep log, saw a lot of rsyslog entries with (file deleted) at the end.. figured, cool.. killed the rsyslogd process, still no space...
[22:24] <talntid> rsyslogd auto restarted, and no longer says deleted...
[22:31] <Daviey> If anyone wants to look at security fixing http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=640028 that would be great! :)
[23:16] <koolhead17> Daviey: thanks for updating the bug and assigning it to debain bug. :D
[23:25] <JRWR> On Ubuntu 11.04, Running OpenLDP and Apache2 using SVN/WebDav setup, In Firefox, I can auth fine to the server (LDAP Authz) in svn I can auth, but I cannot mount the WebDav as a drive in windows, It refuses to accept the login details, Site Config: http://paste.ubuntu.com/684801/ - Access.log http://paste.ubuntu.com/684802/ - error.log http://paste.ubuntu.com/684803/
[23:36] <bcessa> can someone here point me into some docs about how to setup certificate-based SSH auth? thanks, I don't mean plain priv/pub keys
[23:48] <jeeves_moss> what package has starutil in it?  I can't find it.