[00:55] <exutux> hi all
[00:55] <exutux> how can I see wich mirror does apt uses?
[00:56] <patdk-lap> it gives you the ip when it runs
[00:57] <exutux> patdk-lap: uhm well but if I want to change it?
[00:59] <patdk-lap> than change it
[00:59] <exutux> patdk-lap: where?
[01:03] <exutux> if I want to use main mirror or other on APT where I can change it?
[01:10] <exutux> there is an easy way or I need to change repos on sources.list?
[01:32] <PerfM> yesssssssssssss
[01:32] <PerfM> I love ubuntu channels
[02:18] <lunaphyte_> hi.  i'm having some trouble installing bind9 via apt-get.  http://dpaste.com/662518/
[02:19] <lunaphyte_> as you can see in the pastebin, it complains that the post-installation script exited with status 1, but when i run the script manually, it exits with status 0.  how can i further troubleshoot?
[02:25] <twb> lunaphyte_: dpkg --configure?  Why are you doing that?
[02:25] <lunaphyte_> because it broke during the initial install process.  it's half installed.
[02:25] <lunaphyte_> apt-get install bind9 does effectively the same thing.
[02:26] <twb> Purge and reinstall it
[02:26] <lunaphyte_> i've done that, but i'll try again
[02:26] <twb> If you're running /var/lib/dpkg/info/foo.postinst by hand, you should analyse it and probably pass some args
[02:30] <lunaphyte_> twb: for reference: http://dpaste.com/662521/
[02:30] <lunaphyte_> twb: yes, i'm trying to figure out which argument i might pass to it.
[02:30] <lunaphyte_> ah, 'configure', perhaps
[02:31] <twb> It's something like "configure 1.9 2.0"
[02:31] <lunaphyte_> aha, yes
[02:31] <twb> the debian policy has exact details
[02:31] <lunaphyte_> ok, thanks.  this should get me a bit further.
[02:32] <lunaphyte_> http://dpaste.com/662522/
[02:34] <twb> If you get to the point where you are looking in /var/lib/dpkg/info to fix a problem, you are into Deep Magic
[02:34] <twb> If you were a newbie I'd say "forget it, just reinstall from scratch"
[02:37] <patdk-lap> heh, so far my issues have been easy to fix for that
[02:37] <patdk-lap> normally conflicting file, or just manually stopping a service, cause auto stop failed
[02:38] <patdk-lap> wonder if that might be a dash vs bind issue
[02:43] <twb> If bind is assuming /bin/sh is bash it needs to be shit-canned
[02:43] <twb> I would be surprised if that's the case, though.  Maybe in 2007 that might've happened, not today
[02:44] <patdk-lap> ya, it hsouldn't not in a package like that
[02:44] <patdk-lap> but from that error, first thing that comes to mind
[02:44] <patdk-lap> but then, I haven't used bind since pre-2005
[02:45] <patdk-lap> or maybe longer
[03:28] <lunaphyte_> aha.  it's rndc-confgen
[03:33] <lunaphyte_> which is actually an openssl error
[03:37] <lunaphyte_> there's the deep magic :)
[03:38] <lunaphyte_> so the post install shell script for bind9 runs rndc-confgen, which of course uses openssl, since it does some cryptographic hash stuff.
[03:41] <twb> +1 for nsd3
[03:41] <lunaphyte_> when making changes to openssl.cnf, i'd overlooked a particular setting, leaving an invalid reference to a nonexistent file.  this caused the openssl engine to choke [which was what the actual message was referring to], and so rndc-confgen choked.
[03:42] <lunaphyte_> heh
[03:42] <twb> lunaphyte_: good to know that you fixed it, tho
[03:42] <lunaphyte_> a little bit of a goose chase, but that's always the case with openssl, it seems.
[03:43] <lunaphyte_> i should have know better.  i've seen those sort of message types enough to know it was openssl.  at least it was my fault though.
[03:43] <lunaphyte_> *known
[03:43] <twb> So how is that migration to pki going? ;-P
[03:43] <lunaphyte_> migration to pki?
[03:43] <twb> That new ssl implementation that isn't openssl nor gnutls
[03:43] <twb> moco and gnome stuff already use it
[03:44] <lunaphyte_> oh.  that's a new one to me.
[03:45] <lunaphyte_> do you have an url to share?  my googling isn't returning many clues
[03:46] <twb> it's libnspr or libnss or somethign stupid like that
[03:46] <lunaphyte_> oh - nss?
[03:46] <twb> Not nss as in nsswitch tho
[03:46] <lunaphyte_> yeah, there are a few folks out there jumping on that bandwagon
[03:46] <lunaphyte_> i'm not super thrilled about it, to be honest.
[03:47] <twb> AFAICT gnutls doesn't work and openssl can't be used by binary distros for licensing reasons.
[03:47] <twb> If nss can be both legal and reliable, then IMO it wins
[03:47] <lunaphyte_> i'd prefer to have a larger group be able to get along and all contribute to existing projects, rather than these seemingly non stop iterations of yet another choice.
[03:48] <twb> Assuming it doesn't add in e.g. "insecure" or "unproven" into the mix ;-)
[03:48] <twb> lunaphyte_: nss has been around for a while; didn't it come out of netscrape?
[03:48] <lunaphyte_> funny thing about gnutls, there are lots of people who have "lots of problems" with it, but when you really start digging into things, you find out that the vast majority of the problems are the user's fault.
[03:49] <twb> I'll tell you what my problem is
[03:50] <twb> Using openldap on both ends only TLS port (not 389), and slapo-ppolicy, it completely fails to work until I relink it against openssl.
[03:50] <lunaphyte_> in fact, a lot of the time, the problems that people have are because gnutls is much more strict about security in certain contexts then openssl
[03:50] <lunaphyte_> oh - that's where i recognize the nick from :)
[03:50] <twb> And I complained to the openldap people about this and they basically said "fuck you.  PS: it is TOTALLY legal to link against OpenSSL, Debian are retards"
[03:50] <lunaphyte_> yeah, i know the openldap folks don't like gnutls.
[03:50] <twb> I grant you this is not necessarily gnutls' fault
[03:51] <lunaphyte_> i use openldap with gnutls, and have no issues.
[03:51] <twb> You probably are using starttls
[03:51] <twb> Or not using ppolicy
[03:51] <twb> Or you're using a non-LTS release so the whatever bug has been fixed :-)
[03:51] <lunaphyte_> there were some hurdles to overcome, but not insurmountable.  i use starttls, ldaps, and ppolicy.
[03:52] <lunaphyte_> although i which i didn't have to use ldaps.
[03:52] <lunaphyte_> *wish
[03:52] <twb> Oh and #openldap also wanted me to use something other than padl on the client side
[03:52] <lunaphyte_> yeah, that's was probably me bitching about that
[03:52] <lunaphyte_> padl stuff sucks.
[03:52] <twb> No it was hyc
[03:52] <twb> He wanted me to buy a support contract from him or something
[03:52] <lunaphyte_> oh, heh.  yeah, i think we both fels the same way.
[03:52] <lunaphyte_> *feel
[03:52] <lunaphyte_> oh, hmm.
[03:54] <twb> I'm not keen on pulling third-party libraries in when it comes to security
[03:54] <lunaphyte_> he's not always real interested in spending time supporting things he doesn't like
[03:54] <twb> Yeah I understand that
[03:54] <twb> Frustrating for me as a user
 ok. i am having issues getting the transmission-daemon to keep its edited json settings file after a restart. i followed to seperate guides to set it up(http://1000umbrellas.com/2010/10/04/updated-transmission-installationconfiguration-on-ubuntu-server)and(https://trac.transmissionbt.com/wiki/EditConfigFiles) neither worked
 i am running ubuntu-server 11.10
[03:57] <lunaphyte_> but that's probably where the support stuff came in.  symas pays the bills mostly with the support stuff, as far as i know.
[04:08] <i3luefire> http://pastebin.com/SaZvGHYW                                      http://pastebin.com/UafzH6j2
[04:09] <lunaphyte_> twb, patdk-lap: thanks for the help.
[04:09] <lunaphyte_> time to run.  long drive tomorrow.  have a good night.
[06:11] <pppurple> is there any one that would be able to answer a few questions about ubuntu/linux before i try switching to that OS blind?
[06:13] <Resistance> pppurple, for server or desktop use?
[06:19] <pppurple> hmmm guess not
[06:19] <SpamapS> pppurple: somebody responded
[06:19] <SpamapS> 22:13 < Resistance> pppurple, for server or desktop use?
[06:19] <pppurple> desktop
[06:19] <pppurple> sorry i guesss i didnt see it. i got d/c
[06:19] <SpamapS> #ubuntu is a better place to ask about Ubuntu desktop
[06:19] <pppurple> aww yeah, i guess this is the server channel
[06:19] <SpamapS> tho many of us are users of it
[06:21] <pppurple> thanks
[06:35] <pawan_tejwani> I am not able to boot into ubuntu server after it has been installed on the system
[06:35] <pawan_tejwani> I guess there is some problem with the hardware BIOS, can anybody help me with hardware BIOS configuration ?
[06:49] <twb> SpamapS: haha, I thought pppurple was some obscene XMPP-over-dialup thing until I saw the nick in scrollbacl
[06:55] <mtaylor> Daviey: you awake yet?
[06:57] <mtaylor> SpamapS: how about you?
[07:04] <SpamapS> mtaylor: yeah I'm up fixing libmysqlclient B.S. still :)
[07:04] <mtaylor> SpamapS: wow. lovely
[07:04] <SpamapS> mtaylor: down to about 25 problem children from 111 on Thursday
[07:11] <lifeless> SpamapS: thats quite some familty
[07:11] <lifeless> SpamapS: hey, do you know of an existing API service (self hostable :P) similar to what I just described on the lp dev list ?
[07:11] <SpamapS> Yeah, and I thought Octomom with her 14 kids was bad enough
[07:11] <lifeless> mtaylor: ^
[07:12] <mtaylor> lifeless: um. I may not be following the lp dev list at the moment- archive link? (or summary)
[07:13] <lifeless> https://lists.launchpad.net/launchpad-dev/msg08512.html
[07:13] <SpamapS> lifeless: reading
[07:16] <mtaylor> lifeless: still reading - but just as a thought completely out of left-field ... have you considered co-opting diaspora?
[07:18] <SpamapS> lifeless: my mind jumps to graph databases.. really isn't that whats needed.. "is this person in the interest graph of that event?"
[07:18] <lifeless> SpamapS: perhaps
[07:18] <lifeless> SpamapS: for a foaf traversal query, much more so
[07:18] <lifeless> SpamapS: but this isn't really a graph heavy problem
[07:21] <mtaylor> lifeless: other than diaspora (which is much larger than what you're looking for, although ultimately also what you're looking for)
[07:21] <mtaylor> lifeless: I do not know of an existing API service like that
[07:23] <lifeless> I struggle to think of diaspora as a service TBH :)
[07:39] <mtaylor> lifeless: well sure.
[07:40] <mtaylor> lifeless: but a diaspora server by its nature considers timelines, events, and event subscriptions
[07:40] <mtaylor> lifeless: so the probably not correct thought here is merely that, if the overall design of the thing you're talking about looks a lot like a social network ...
[07:41] <mtaylor> lifeless: then perhaps using a server written to be a node in a distributed social network wouldn't be ridiculous (with specific data sources implemented as 'apps' or whatnot)
[07:41] <SpamapS> and it quacks like a social network..
[07:41] <mtaylor> SpamapS: quack quack
[07:43] <SpamapS> ok.. enough fighting autoconf
[07:43] <lifeless> mtaylor: perhaps :P
[07:43] <mtaylor> SpamapS: bwahahaha
[07:43] <lifeless> mtaylor: OTOH I suspect we're still at least 1 may be 2 OOM's higher volume than diaspora
[07:44] <mtaylor> lifeless: entirely probable
[07:44] <SpamapS> I don't know diaspora's underlying design, but just the fact that its made to be sharded makes me think it should scale incrementally..
[07:45] <mtaylor> lifeless: although that's diaspora the services vs. diaspora the server
[07:45] <lifeless> mtaylor: you can't know that one scales will till the other has scaled successfully
[07:45] <mtaylor> lifeless: ++
[07:46] <lifeless> I have a deep and abiding trust that folk I've never heard of can write a highly scalable efficient and robust service...
[07:46] <mtaylor> hehe
[07:46] <lifeless> e.g. not something I'll take on faith :)
[07:46] <mtaylor> I did consider replacing my blog with a diaspora server
[07:47] <mtaylor> I could do that and then post a few more inflamatory blog posts and we could test scaling :)
[07:47] <lifeless> hahah
[07:48] <SpamapS> mtaylor: we're all still rubbing salve on the burns from your heinous crime of asking to have the gnome 2 clock brought back. ;)
[07:48] <SpamapS> ok, autoconf vanquished.. time to slep
[07:48] <SpamapS> sleep too
[07:49] <mtaylor> perhaps I could do one pointing out the differences between text fonts and display fonts
[07:50] <mtaylor> and why you don't want a the characteristics of a display font in a text font
[07:50] <mtaylor> I'm sure that would stomp on the last remaining shred of goodwill anyone has for me :)
[07:55] <jamespage> morning all
[07:58] <mtaylor> morning jamespage
[07:58] <jamespage> hey mtaylor - hows things?
[07:59] <mtaylor> jamespage: good! I'm causing trouble as usual
[07:59] <mtaylor> jamespage: you?
[07:59] <jamespage> mtaylor, so I see :-)
[07:59] <jamespage> mtaylor: all good my end!
[08:24] <koolhead11> hi all
[09:24] <lynxman> morning o/
[09:44] <Daviey> mtaylor: i'm always awake
[09:57] <koolhead11> hey lynxman Daviey :)
[10:06] <Daviey> hey koolhead11
[10:07] <koolhead11> Daviey: making list title whats new in "12.04" for discussion. I am following this http://status.ubuntu.com/ubuntu-precise/group/topic-precise-servercloud-workloads.html
[10:07] <koolhead11> along with ubuntu cloud blog fridge
[10:08] <koolhead11> is there any other place i can look for same
[10:16] <Daviey> http://status.ubuntu.com/ubuntu-precise/group/topic-precise-servercloud-infrastructure-deployment.html
[10:16] <Daviey> http://status.ubuntu.com/ubuntu-precise/group/topic-precise-servercloud-service-orchestration.html
[10:20] <koolhead11> thanks Daviey
[10:20] <Daviey> Anyone seen Arnaud on irc?
[10:20] <Daviey> koolhead11: np
[11:01] <koolhead11> spice spice
[11:19] <koolhead11> Daviey: https://blueprints.launchpad.net/ubuntu/+spec/servercloud-p-glusterfs-mir  let me know if i can help with this :)
[11:20]  * koolhead11 has some good friend over there :D
[11:24] <Daviey> koolhead11: If you want to help with that, i'd be most pleased :)
[11:24] <koolhead11> Daviey: tell me how can i
[11:24] <Daviey> koolhead11: check the package works, and see what depends it has for us to consider putting it in main.
[11:24] <koolhead11> ok
[12:00] <tefx> at the risk of lookign and sounding like a noob its bveen years since i did it last nbtu how the hell do you set up private ns servers and then bind the domain usign bind lols
[12:00] <tefx> cos i spent the last 16 hours on it and my head about to explode if i dotn get somewere with it
[13:06] <koolhead11> jcastro: around
[13:24] <Daviey> smoser: are you alive yet?
[13:45] <zul> morning
[14:27] <stgraber> hallyn: good morning. Just wondering why ubuntu-sponsors is subscribed to bug 869590 as you apparently have upload rights for libvirt in lucid
[14:35] <hallyn> stgraber: i think that got auto-subscribed.  What I was actually waiting on is bug 869553 to get fix released, as without that this fix is not sufficient.
[14:35] <stgraber> hallyn: ok, sounds good. Should I unsubscribe sponsors then?
[14:35] <hallyn> and that one is waiting on jdstrand to have some time to verify if the fix is ok, or if we somehow need to be more specific.
[14:36] <hallyn> stgraber: yes, thanks.
[14:36] <robo> hello: i need to install a php pecl extension and found a RedHat RPM for it. What's the cleanest way to install this pecl extension? Use alient and convert the RPM or just use PECL to install it?
[14:36]  * jdstrand is trying to catch up... deluge of irc backlog and email is crazy :)
[14:36] <hallyn> stgraber: yeah comment #5 is where it happened.
[14:37] <hallyn> jdstrand: i understand, this isn't by any means urgent i don't think, not many ppl using tunneled migration, else they'd be yelling :)
[14:37] <stgraber> hallyn: argh, that's bdmurray's bug bot, again :) though I think it now adds some logic to detect upload rights and not subscribe sponsors/reviewers for no reason :)
[14:38] <hallyn> stgraber: maybe there's some way we can tag a patch to have the bot ignore it?
[14:39] <stgraber> hallyn: yeah, I remember us discussing a magic tag, something along the lines of leave-this-bug-alone, mostly for the kernel bugs though (where the bug bot comments every time a new kernel is uploaded)
[14:40] <robo> n/m. dh-make-pecl seems to be the proper way to install pecl packages :-)
[14:43] <RoyK> which pecl extension is this?
[14:44] <RoyK> oh
[14:44] <RoyK> pecl == pear?
[14:45] <robo> RoyK, it's mongo
[14:45] <robo> and pecl != pear
[14:46] <robo> pecl is written in C, pear is written in php
[14:46] <robo> so a rule of thumb is use the pecl package. If it doesn't exist then use pear
[14:46] <RoyK> k
[14:46] <robo> but the tricky part is getting pecl to work with the package manager
[14:46] <robo> that's where dh-make-pecl comes in
[14:47] <RoyK> same applie to pear or mixing any other package managers like perl CPAN etc
[14:48] <robo> yeup
[15:14] <lynxman> jamespage: ping
[15:15] <jamespage> lynxman: pong
[15:15] <jamespage> wassup
[15:16] <lynxman> jamespage: got the latest debdiff for you, before I created it just wanted to check that the changelog looks good :)
[15:16] <lynxman> jamespage: http://pastebin.ubuntu.com/752558/
[15:18] <jamespage> 1.2.1-0ubuntu3 -> 1.2.1-0ubuntu2.1 (see https://wiki.ubuntu.com/SecurityTeam/UpdatePreparation#Update_the_packaging)
[15:18] <jamespage> lynxman: ^^
[15:18] <lynxman> jamespage: aah cool
[15:20] <lynxman> jamespage: does the rest look sane?
[15:21] <jamespage> lynxman: I think so
[15:21] <jamespage> lynxman: stick it on the bug and I'll take another look
[15:22] <lynxman> jamespage: cool, ty!
[15:24] <remix_tj> sono gay
[15:25] <remix_tj> fucked remote assistence with assholes friends -_-
[15:25] <Tm_T> remix_tj: ...
[15:35] <lynxman> jamespage: it's in the bug now, #884908
[15:36] <jamespage> bug 884908
[15:50] <raubvogel> Let's say I create a iptable rule in my firewall (ubuntu 11.04) to forward traffic from anywhere port 5432 to 10.0.0.24 port 22
[15:50] <jamespage> lynxman, uploaded :-)
[15:50] <raubvogel> iptables -L -t nat shows me the rule
[15:51] <lynxman> jamespage: \o/
[15:51] <lynxman> jamespage: thank you sir *bows*
[15:51] <raubvogel> Should I also see it in, say netstat -apn | grep 10.0.0.24 ?
[15:53] <SpamapS> raubvogel: no
[15:53] <SpamapS> raubvogel: netstat is at a layer below iptables
[15:53] <raubvogel> SpamapS: Aha. Any suggestions for finding out why it is not working?
[15:53] <SpamapS> raubvogel: or, actually.. above it.. depending on how you look at it. Either way, iptables gets in front of netstat
[15:54] <SpamapS> raubvogel: /proc/net/ip_conntrack might prove helpful
[15:55] <SpamapS> raubvogel: iptables -L -t nat --verbose  might also prove useful
[16:06] <raubvogel> SpamapS: /proc/net/ip_conntrack should provide connections that are alive/taking place, right?
[16:06] <hallyn> stgraber: am i supposed to have upload rights to libcgroup?
[16:08] <stgraber> hallyn: in precise, yes, are you trying to upload to something else?
[16:08] <RoyK> raubvogel: iirc, yes
[16:08] <hallyn> stgraber: yeah, oneiric-proposed
[16:08] <stgraber> hallyn: ok, let me change oneiric too then (I need to run the command once per release ...)
[16:08] <hallyn> stgraber: thanks!
[16:09] <stgraber> hallyn: should work now
[16:09] <hallyn> i'd  consider applying for coredev to stop this sillyness, but last few days have proven i have some things to learn yet.
[16:09] <hallyn> thanks, i'll repush :)
[16:10] <stgraber> having packagesets be per release is useful sometimes but also a pain in cases like this :) In most cases if someone can upload to the current release, I'm also fine with them uploading an SRU (as SRU need additional review anyway)...
[16:11] <hallyn> yeah that was my thought - sru has extra approvals anyway
[16:33] <hallyn> ppetraki: could you take a look at the dmesg output in https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/893450/comments/5 and tell me if it looks kosher?
[16:34] <ppetraki> hallyn, sure
[16:34] <hallyn> ppetraki: thanks!
[16:59] <raubvogel> SpamapS: No seeing anything there. And in /var/log/messages.log I see Nov 28 11:11:36 vpn kernel: [2166490.865348] IN=eth0 OUT=eth1.64 SRC=66.172.33.139 DST=10.0.0.24 LEN=60 TOS=0x00 PREC=0x00 TTL=63 ID=6960 DF PROTO=TCP SPT=55688 DPT=22 WINDOW=14600 RES=0x00 SYN URGP=0
[17:03] <SpamapS> raubvogel: you also need to allow those natted packets in your FORWARD rules
[17:11] <raubvogel> SpamapS: this is how it looks like : http://pastebin.com/PB4TzQEM
[17:11] <SpamapS> raubvogel: iptables -L FORWARD -v
[17:12] <ppetraki> hallyn, so that snippet looks fine. With regards to configuration, scanning the sd devices in LVM would be unnecessary, though it should't cause any harm either. To checkout the disks you can use the smartutils, run short and extended tests.
[17:12] <SpamapS> raubvogel: FORWARD is not in the nat table
[17:14] <ppetraki> hallyn, using smart tools: http://askubuntu.com/questions/78600/disk-utility-reports-slower-initial-spinup-time-via-s-m-a-r-t/81369#81369
[17:17] <raubvogel> SpamapS: I forgot the prerouting entry, http://pastebin.com/hcV0djZv
[17:19] <hallyn> ppetraki: thanks.  too bad :)  it sure seems slow coming up
[17:22] <ppetraki> hallyn, well, SATA is slow (14-20 sec spinup), and the block devices are stacked. Could be a UDEV problem or storage stack advertising that it's ready before it really is. I'd need to see the logs from the actual failure to comment intelligently.
[17:24] <hallyn> ppetraki: here https://launchpadlibrarian.net/85667014/syslog.txt   is the actual error that i'm trying to figure out  (i.e. '/devices/virtual/block/dm-2')
[17:28] <ppetraki> hallyn, hmm, would need udev verbose logging and a better understanding of how many LVs the system actually has
[17:28] <ppetraki> hallyn, that includes snapshots
[17:30] <ppetraki> hallyn, if it is a scan issue, than making the rootdelay really big would help, like 3-4 mins
[17:33] <ppetraki> hallyn,*updated title*
[17:38] <hallyn> ppetraki: ok, so i need to go back and find the recommended way to ask users for udev logs
[17:38] <hallyn> verbose ones, that is :)
[17:39] <ppetraki> we have a troubleshooting section on UDEV, but that involves killing it and starting it in the foreground with super logging iirc
[17:40] <hallyn> maybe https://wiki.ubuntu.com/DebuggingLvm
[17:41] <hallyn> i'd better try that out in a kvm first
[17:43] <smoser> adam_g, is there some document that discusses '<type>_orchestra_provisioning_server._tcp</type>' in avahi ?
[17:43] <ppetraki> hallyn, wish we had a ramdisk switch for that
[17:44] <adam_g> smoser: what do you mean?
[17:44] <smoser> is there somthing that discusses the naming convention ?
[17:45] <hallyn> well it's out of date for certain.  /scripts/init-premount should be scripts/init-top
[17:45] <smoser> ie, why the _
[17:45] <ppetraki> hallyn, changing /etc/udev.conf to udev_log="debug" and then rebuilding the initrd might also work
[17:45] <smoser> you had "_orchestra_cobbler._tcp"
[17:45] <hallyn> ppetraki lemme try
[17:46] <ppetraki> hallyn, cool
[17:47] <smoser> adam_g, following your lead i'd pick "_orchestra_provisioning_server._tcp", but i dont really know why
[17:47] <adam_g> smoser: looking now
[17:47] <smoser> https://bugs.launchpad.net/ubuntu/+source/cobbler/+bug/893189 comment 1
[17:47] <adam_g> smoser: for something specific, all man pages and other uses use that convention so...
[17:50] <hallyn> ppetraki: well that sends the debug output to /var/log/boot.log, but not from initramfs' udev.  not sure if that's needed for this bug.
[17:57] <hallyn> but i think we can just have him update /usr/share/initramfs-tools/scripts/init-top/udev and then update-initramfs...
[17:57] <adam_g> smoser: i believe the convention is part of the dns-sd specs http://www.dns-sd.org/ and, looking at it, arbitrary service types might be abusive. perhaps keeping the type _http._tcp and filtering by the service name may be better? i'll take a look how other services use it
[17:57] <hallyn> think i'll update https://wiki.ubuntu.com/DebuggingUdev
[17:58] <smoser> adam_g, squid-deb-proxy uses: http://paste.ubuntu.com/
[17:58] <smoser> so we're no worse.
[17:59] <adam_g> smoser: yeah, ive got a few services insatlled here that are using "non-standard" service type
[18:07] <storrgie> I'm installing ubuntu server 11.10 here, I have / on a md volume. When I get to the grub portion of the install I see 'executing grub-install /dev/sda failed this is a fatal error' and I really cant work beyond this
[18:07] <SpamapS> smoser: I take issue with s-d-p since it can't use PPA's or external sources
[18:07] <smoser> unrelated, SpamapS
[18:08] <SpamapS> smoser: I guess my point is that it advertises something it cannot doe
[18:08] <SpamapS> do even
[18:11] <SpamapS> smoser: if all we're doing is logging into cobbler, why are we advertising orchestra and not cobbler?
[18:13] <smoser> cobbler explicitly disabled their avahi advertising. it wouldn't make much sense imho to advertise a servie that is not something we control and the upstream explicitly *stopped* advertising.
[18:14] <storrgie> with 11.10 can you install to a software raid1?
[18:14] <smoser> and it seems to me to make more sense to me as an orchestra thing, as the one thing we're expecting right now to use it for is registration, which is not a cobbler (specific) thing.
[18:18] <smoser> SpamapS, ^
[18:20] <smoser> adam_g, i realized my pastebin fail above.
[18:21] <smoser> it seems pastebinit is broken
[18:22] <smoser> it seems pastebin.ubuntu.com doesn't like stuff that starts with '<?xml"
[18:22] <smoser> http://paste.ubuntu.com/752758/ is better.
[18:22] <smoser> anyway.
[18:22] <adam_g> smoser: thats cool, was looking at squid-deb-proxy.service when you pasted
[18:23] <smoser> i just blindly run pastebinit and copy and paste the output.
[18:23] <SpamapS> smoser: meh, did upstream give a reason for disabling their avahi?
[18:24] <smoser> bug 893189 has a link to the commit.
[18:24] <smoser> i dont really care either way, and we can put it in cobbler, but the one thing we're going to use this for is not cobbler.
[18:26] <SpamapS> smoser: oh? We're not doing direct cobbler xml-rpc anymore?
[18:26] <smoser> well, we are. but the cobbler-enlist stuff isn't from cobbler.
[18:27] <smoser> so i dont' know.
[18:27] <SpamapS> ... maybe it should be. ;)
[18:27] <smoser> i dont feel strongly, but i personally feel "orchestra" is a better place for that.
[18:28] <smoser> well....
[18:28] <smoser> interesting
[18:28] <SpamapS> for enlist, it makes sense that it would be its own thing.. thats why we have the API. But for avahi.. just finding "the cobbler server" .. seems like that would be advertising *cobbler*. :P
[18:28] <smoser> it would look to me jus from that diff and nothing eles, that cobbler probably advertises '_http._tcp'
[18:28] <smoser> which is clearly not sufficient.
[18:29] <SpamapS> yeah, we don't want _http .. we want _cobbler_on_http .. :)
[18:29] <SpamapS> Anyway, I prefer loose coupling during heavy innovation, so this probably makese sense to have outside cobbler until we know it works.
[18:30] <smoser> ok.
[18:32] <RoAkSoAx> smoser: I agree with you, a better place might be orchestra itself
[18:32] <smoser> RoAkSoAx, so how would you think i should land this.
[18:32] <smoser> do you want an upstream release for it ?
[18:33] <smoser> or commit to upstream
[18:33] <smoser> and then release a new ubuntu with a patches patch
[18:39] <smoser> RoAkSoAx, ^
[18:39] <smoser> and https://code.launchpad.net/~smoser/orchestra/add-avahi/+merge/83664
[18:46] <adam_g> RoAkSoAx: ping
[18:48] <RoAkSoAx> smoser: this should probably be upstream orchestra
[18:48] <RoAkSoAx> adam_g: pong
[18:48] <smoser> yes.
[18:48] <smoser> but to get a fix in today
[18:49] <smoser> do you want re-release an upstream orchestra
[18:49] <smoser> versus just an ubuntu
[18:49] <smoser> the merge proposal above is against upstream.
[18:49] <RoAkSoAx> smoser: i can release upstream orchestra
[18:49] <Daviey> hey o/
[18:50] <smoser> RoAkSoAx, ok then, you want to just pull that change and release ?
[18:50] <adam_g> RoAkSoAx: nvm..  orchestra-import-isos now requires an argument to do the imports where it did them by default previously. ill fix and push to you soon
[18:50] <RoAkSoAx> smoser: yp will od
[18:50] <Daviey> smoser: if this is in the cobbler package, advertise it as cobbler - if it is in the orchestra package, orchestra-cobbler IMO
[18:51] <RoAkSoAx> adam_g: `yeah it does, check that it wont break the update-settings nor the check for correct commands and so on
[18:51] <RoAkSoAx> Daviey: i think this should be orchestra side
[18:51] <RoAkSoAx> Daviey: as this are features for orchestra rather than features we would upstream cobbler
[18:51] <smoser> Daviey, yes. its going into orchestra package at the moment.
[18:52] <adam_g> RoAkSoAx: huh? the cronjob that gets installed doesn't do anything now.
[18:52] <smoser> and advertised as orchestra.
[18:52] <Daviey> okay, great
[18:53] <RoAkSoAx> adam_g: err cause of the no argument I pressume
[18:53] <RoAkSoAx> err i mean cause of the argument we need now
[18:57] <smoser> RoAkSoAx, i've verified that adding that file to /etc/avahi/services/ results in 'avahi-browser --all' listing
[18:57] <smoser>  eth0 IPv4 Orchestra Provisioning Server on brickies     _orchestra_provisioning_server._tcp local
[18:57] <smoser> so if you could get a new package out sooner than later, that would be very nice.
[18:59] <Daviey> Should we look to rename cobbler-enlist to orchestra-client-enlist ?
[18:59] <Daviey> smoser: Does it expose ipv6 out of interest?
[18:59] <smoser> i did not.
[19:01] <smoser> but i can make it
[19:01] <smoser> http://paste.ubuntu.com/752819/
[19:01] <Daviey> great
[19:01] <smoser> Daviey, do you want it to ?
[19:02] <Daviey> I think we should..
[19:02] <SpamapS> Seems like it should stay named cobbler-enlist
[19:02] <smoser> RoAkSoAx, if you're that merge proposal, please re-pull it. i just added ipv6
[19:02] <SpamapS> Unless there is a plan to have it enlist in some super-service?
[19:03] <SpamapS> (which is why I also am puzzled why we aren't just advertising cobbler on avahi)
[19:05] <dkn> anyone catch that mad sale on the intel 320 SSD's? $1/GB!
[19:05] <RoAkSoAx> smoser: ok, will take care of it in a bit
[19:20] <RoAkSoAx> adam_g: is yout fix gonna be ready anytime soon?
[19:21] <adam_g> RoAkSoAx: i will
[19:21] <RoAkSoAx> adam_g: so that I just release once
[19:22] <adam_g> RoAkSoAx: yeah, wait on me. ill let you know when its good to go
[19:22] <RoAkSoAx> adam_g: cool thanks
[19:24] <RoAkSoAx> smoser: merged. will be released once adam_g sends his fix over
[19:25] <smoser> k
[19:25] <smoser> gracias
[19:25] <RoAkSoAx> smoser: do you want a new binary package for it? or should it be shipped with the provisioning server?
[19:26] <smoser> the merge proposal shipped it with the provisioning server
[19:26] <smoser> did you not see that ?
[19:26] <smoser> or did i not do it right
[19:26] <RoAkSoAx> smoser: no it is fine, i was just double checking
[19:26] <RoAkSoAx> whether you want it where it is, or a new -publish package
[19:26] <RoAkSoAx> Daviey: here's what koan --replace-self does: http://pastebin.ubuntu.com/752851/
[19:27] <smoser> i think it makes most sense to be in provisioning-server
[19:27] <RoAkSoAx> smoser: ok
[19:27] <Daviey> RoAkSoAx: is that what we want?
[19:30] <RoAkSoAx> Daviey: if we easily want to re-install a system i think it would be desirable
[19:35] <Daviey> RoAkSoAx: right, but if we can work out a sane way to hard code a url to ipxe.. (other than embdedding a script), that seems to have a cleaner solution
[19:35] <Daviey> Ie, every reboot will recided to reinstall or boot local
[19:36] <Daviey> regardless of firmware pxe.
[19:36] <Daviey> and dhcp next-server.
[19:36] <RoAkSoAx> Daviey: basically, it just downloads the initrd,linux palces it /boot, updates grub, and when it restarts, it does the installation
[19:36] <RoAkSoAx> Daviey: yeah its 2 different approaches
[19:36] <RoAkSoAx> Daviey: i'll look into ipxe now
[19:36] <Daviey> RoAkSoAx: Do you think koan is superior?
[19:37] <RoAkSoAx> Daviey: idk if superior, but koan definitely been made to be a helper tool for cobbler
[19:37] <RoAkSoAx> Daviey: I'm gonna look into ipxe-grub
[19:38] <RoAkSoAx> Daviey: cause, you wanna use ipxe-grub *after* the system has been registered, but not installed right?
[19:38] <RoAkSoAx> Daviey: while koan, is for when the system has already been installed
[19:49] <Daviey> RoAkSoAx: no, this satisfies the reinstal scenario, where we don't have dhcpd access.
[19:49] <RoAkSoAx> Daviey: ok, so why use ipxe when koan is already in place then
[19:49] <RoAkSoAx> Daviey: i mean, we already have the way of doing so
[19:50] <Daviey> RoAkSoAx: Using koan you need to access the client machine, right?
[19:50] <RoAkSoAx> Daviey: we can either use koan an adapt it to automatically look for the orchestra server
[19:50] <RoAkSoAx> Daviey: yes you do
[19:50] <Daviey> if we can do this via pxe, we don't need to access the machine
[19:50] <Daviey> just tell cobbler to reinstall, and powercycle using power control
[19:51] <Daviey> right?
[19:51] <RoAkSoAx> Daviey: so you mean a "remote" command?
[19:51] <Daviey> yeah
[19:51] <RoAkSoAx> Daviey: right, we can tell the machjine that the remote command is koan
[19:51] <RoAkSoAx> Daviey: or ipxe already supports something similar
[19:51] <RoAkSoAx> Daviey: IIRC, cobbler had a feature on which would allow remote control for reinstallas and stuff but was removed
[19:52] <Daviey> RoAkSoAx: right, but if we can use ipxe scripting - we get this for free..
[19:53] <Daviey> The same model works if it is hardware pxe boot, or ipxe then
[19:53] <RoAkSoAx> Daviey: ok let me look into ipxe first
[19:53] <RoAkSoAx> and then I'll be able to give a better opinion
[19:53] <Daviey> great!
[20:49] <adam_g> RoAkSoAx: expected behavior if i  run  'orchestra-import-isos -u -i' ?
[21:00] <RoAkSoAx> adam_g: show show error, either one of them should work
[21:01] <RoAkSoAx> or usage
[21:11] <vasosanitario> alguem me passa lista de regras de iptables
[21:23] <kirkland> could someone nudge Bug #887344 along?
[21:23] <kirkland> (Daviey)?
[21:59] <adam_g> RoAkSoAx: conflicts with which? lp:orchestra ?
[22:01] <RoAkSoAx> adam_g: yes
[22:01] <adam_g> RoAkSoAx: oh jeez, i was based on the ubuntu branch. one min
[22:01] <RoAkSoAx> adam_g: lol no worries, gonna go for lunch will merge when I get back
[22:01] <adam_g> cool
[22:08] <elz89> I am having problems when setting up slapd. I would like to know how to remove it properly, plus all configuration, so that I can start over?
[22:12] <Daviey> hallyn: Soft-freeze for a1 is now live... Makes sense to wait until end of week to remove etherboot?
[22:13] <Daviey> Doesn't seem to be something we need to acquire a freeze request for, right?
[22:18] <adam_g> RoAkSoAx: rebased on trunk, sent another proposal
[22:39] <RoAkSoAx> adam_g: cool will merge it in a few
[22:53] <elz89> sudo apt-get purge slapd doesn't seem to remove everything?
[22:56] <SpamapS> elz89: what did it leave behind?
[23:11] <jdstrand> hallyn: I seem to be having a problem with qemu and -usb. see http://paste.ubuntu.com/753100/
[23:11] <jdstrand> hallyn: basically, on oneiric if I specify -usb, I get a UHCI Host Controller without having to do anything else. on precise, I do not
[23:12] <jdstrand> hallyn: this might be a 0.14.1 vs 0.15 thing, but http://wiki.qemu.org/ChangeLog/0.15 wasn't enlightening. have you seen this? am I doing something wrong?
[23:13] <jdstrand> hallyn: hello btw :)
[23:14] <jdstrand> I'm going to file a bug
[23:20] <jdstrand> hallyn: fyi, bug #897466
[23:23] <RoAkSoAx> smoser: orchestra released
[23:31] <kirkland> looks like you guys are kicking butt on Orchestra today :-)
[23:31] <RoAkSoAx> kirkland: hehe :)
[23:41] <hallyn> jdstrand ok thx will look tonight
[23:58] <Xelmep> HELP About PPTP Server on Ubuntu 10.04 !!!
[23:59] <Xelmep> HELP About PPTP Server on Ubuntu 10.04 !!!