[00:08] <owh> How do I tell LP that it's a bug report about hardy?
[00:11] <kgoetz> owh - https://bugs.launchpad.net/ubuntu/hardy/+bugs
[00:12] <owh> kgoetz: Yes, but how do I add my bug report, Bug #201237 to that list - do I just add a "hardy" tag?
[00:13] <kgoetz> owh: if you click 'report bug' it should reporta  bug on hardy, aiui
[00:14] <owh> kgoetz: Hmm, that seems to indicate that I need to delete the bug report I just made and do it again. That doesn't seem right.
[00:14] <ScottK2> owh: Let me look at the bug.
[00:15] <owh> https://bugs.launchpad.net/ubuntu/+source/tasksel/+bug/201237
[00:15] <owh> ubotu seems asleep at the wheel again :)
[00:16] <ScottK2> Yeah
[00:16] <ScottK2> The only thing I see it nominate for a release.
[00:16] <ScottK2> Dunno if that's the right way though
[00:16] <ScottK2> owh: I'd suggest ask in #ubuntu-bugs
[00:16] <owh> Cool, will do.
[00:22] <ubotu> Launchpad bug 201237 in tasksel "tasksel does not work if installed on a machine with no network" [Undecided,New] https://launchpad.net/bugs/201237
[00:24] <owh> ROTFL
[00:25] <owh> FYI the suggestion in #ubuntu-bugs was to indeed nominate it for hardy, then it may or may not actually be approved by the maintainer for such inclusion.
[00:26] <ScottK2> OK
[00:59] <owh> I need some help. I'm still trying to track down why a cron.daily job isn't running. The script runs fine from a command-line. I'm wondering how I can "simulate" the cron environment so I can see what it's doing. There was a suggestion to install mail on this server to attempt to catch any output that might be generated, but I'm reluctant to do that on this production server. Any suggestions / comments / reading material?
[01:00] <andguent> I would second the email notify setup, you never know what else isn't working and is trying to notify you via mail
[01:00] <kgoetz> i'll third it - why wouldnt you want mail on a production sytem?
[01:00] <infinity> There will be some minimal logging to syslog.
[01:01] <kgoetz> the only other thing i can think of is redirect syslog to another system and bump up crons reporting
[01:01] <andguent> another option is to cat /etc/crontab, look for the daily line, and run that
[01:01] <infinity> You could also wrap your cronjob in something that logs to a logfile.
[01:01] <andguent> it might rotate all of your logs though :)
[01:01] <infinity> owh: What does the script do?
[01:01] <kgoetz> first thing i'd check with not-runnign cronjobs is crons $PATH
[01:02] <infinity> And, yes, I was going there. ;)
[01:02] <owh> infinity: It backs up vmware images.
[01:02] <infinity> cron's PATH is, by default, rather lacking.
[01:02] <infinity> You should set PATH either in your cronjob, or the script being run by said cronjob.
[01:03] <owh> That's a point that I had not considered.
[01:03] <owh> Also, the mail side of things, under Debian I can setup local mail only, is that available under Feisty-server, I'm unsure.
[01:03] <infinity> Yup.
[01:03] <infinity> Both postfix and exim4 should give an option to be local-only.
[01:04] <owh> Well, that simplifies things.
[01:04] <infinity> postfix might even default to it these days.  Not sure, as I'm a disciple of Exim.
[01:04] <owh> Ok, given the "overwhelming" suggestion to install mail, I might do that first.
[01:05] <infinity> (Exim-lover though I am, though, I'd recommend postfix from the "it's the Ubuntu default, and you're more likely to get IRC support for it" perspective)
[01:05] <owh> Yeah, exim here too, but I'll go with the Ubuntu default.
[01:05] <kgoetz> i'm not familar with postfix's debconf, but exims is good for that sort of 'just need basic mail' stuff
[01:05] <infinity> If you know exim well enough, do Exim.  We support both (ie: both are in main)
[01:05] <infinity> exim4-daemon-light, select "local only", profit.
[01:05] <andguent> i personally love ssmtp, its so much less of a headache to configure
[01:05] <anthony> Any particular reason you like exim btw?  (/me doesn't know about either)
[01:05] <owh> I'll do a quick test before I do it for real :)
[01:05] <andguent> don't know about any local only settings though
[01:06] <infinity> ssmtp has the distinct disadvantage of not delivering local mail. :)
[01:06] <infinity> Nice for getting mail OFF the host, though.
[01:06] <owh> I'm expecting that tasksel will do the right thing.
[01:06] <owh> Ok, thanks for knocking me into place, I'll investigate some more :)
[01:06] <kgoetz> infinity: which is really all you need in a server :)
[01:06] <infinity> anthony: I've been using exim for a decade, I know the spec inside out and backwards, and I've contributed hudreds, if not thousands, of lines of code to the project.
[01:06] <infinity> anthony: In other words, I'm terribly biased.
[01:06] <kgoetz> hehe
[01:07] <andguent> me on the other hand, have never contributed code to either :) I just know that 5-6 lines of conf files and i'm done :)
[01:07] <anthony> infinity: That's a pretty good reason.  :)
[01:08] <owh> infinity: You would have been a handy person to know some years ago when two instances of exim refused to send email between each other for some unfathomable reason. Network sniffing indicated that one instance would just give up. Never did find out why, but upgrading one side "fixed" it :)
[01:08] <faulkes-> bah, you love sendmail.cf, admit it
[01:08] <owh> ROTFL
[01:08] <infinity> I think postfix likely wins in the "ease of use" category, but I still feel like I can do a lot more with exim's config files, when I want to do something "weird".
[01:08] <owh> faulkes-: My introduction to sendmail.cf was on an Apple AU/X server :)
[01:09] <owh> For those wondering, Apple made Unix before OS/X :)
[01:09] <infinity> faulkes-: sendmail is why I love exim, to be honest.  Exim is nowhere near as confusing, but manages to be just as powerfully configurable.
[01:09] <owh> s/Unix/Unix servers/
[01:09] <infinity> owh: s/made/licensed/
[01:09]  * faulkes- lalalalala m4 I can't hear you
[01:09] <owh> infinity: Yeah, a bit of that too :)
[01:09] <faulkes-> heh, you had to love the bat book
[01:10] <infinity> owh: It's okay, your secret's safe with me.  My first at-home UNIC box was an Amiga.
[01:10]  * faulkes- remembers doing some terribly awful things to sendmail.cf
[01:10] <infinity> (First UNIX machine I played with that I didn't own was OSF)
[01:10] <owh> infinity: Now that's personal torture :)
[01:10]  * infinity misses DEC.
[01:11]  * kgoetz gives infinity an Alphaserver 5/300
[01:11]  * owh misses the Alpha on the desktop and the VAX heard in the back office :)
[01:11] <kgoetz> *3000
[01:11]  * kgoetz missed the whole era :@
[01:11] <infinity> I have a DEC3k/900 lying around here.
[01:11]  * owh started with 3.5kb RAM on a 6502 based Vic20 :)
[01:12]  * kgoetz goes to sulk
[01:12] <infinity> Never got around to porting the turbchannel drivers from Linux/MIPS to Linux/Alpha, so the thing still boots NetBSD/Tru64.
[01:12] <owh> infinity: It seems that you have waaaay too much time on your hands :)
[01:14] <infinity> If only that were true.
[01:15]  * owh misses a serial link between the Vic20 and the Apple ][ joystick port to transfer images from floppy to the Vic20 tape drive :)
[01:15] <infinity> If I had more spare time, the HPPA port would be in better shape, we'd have an Ubuntu/Alpha port "just cause", and I'd still be working on m68k glibc/kernel stuff.
[01:15] <owh> infinity: This morning I was lamenting all the things I want to do, but just don't have enough hours in the day for, so I know what you're talking about.
[01:16] <kgoetz> infinity: are you just a serial porter?
[01:17] <infinity> kgoetz: Well, no.  I work for Canonical on the buildd infrastructure.  But, yeah, I used to do a lot of porting before this.
[01:17] <infinity> kgoetz: No time anymore, really.
[01:17] <owh> The irony is that while I was lamenting that, I remembered a TV show last night about the Royal Family and the strange things that people get up to. While I might not get to do all that I want, at least what I'm doing is improving the community around me, so it made me feel a tad more useful than the accountant who plays police officer for Liz when she visits Scotland one week in the year.
[01:18] <owh> infinity: So, think of all the good you're doing :)
[01:19] <owh> How do I stop an installer from apt-getting stuff during installation?
[01:19] <kgoetz> remove the network cable
[01:20] <owh> Hmm, yes, I suppose I can disable the interface in VMware.
[01:20] <owh> Sigh, I thought there might be a more elegant way :)
[01:20] <kgoetz> log into the installer console and try to ifdown the nics?
[01:21] <owh> That might work after it's configured them.
[01:21] <owh> Excellent suggestion.
[01:23] <andguent> i'm a few minutes late, but you could also repoint its dns server to something bogus....
[01:24] <owh> Fair idea, bit of extra work beyond Alt-F2, Enter, ifconfig down eth0
[01:24] <owh> But valid :)
[01:25] <kgoetz> owh: or even shorter: ifdown eth0 (if the installer supports it)
[01:25] <owh> kgoetz: Nope :)
[01:26] <kgoetz> aww
[01:26] <owh> Tried that :)
[01:26] <kgoetz> :)
[01:30] <kgoetz> would i be correct in guessing these dirs are misnamed? http://archive.ubuntu.com/ubuntu/dists/hardy/main/installer-sparc/
[01:37] <owh> Can someone please explain why if I've just logged into a machine, either via a login prompt or by unlocking the screensaver, I still am required to enter the same password in sudo? I'm unsure why this might be considered "safer". I'm not saying, never prompt for a sudo password, having it expire in the same way as it does between sudo calls is fine, but why?
[01:38] <anthony> Anyone have opinions on ebox?
[01:38] <owh> anthony: It will be the server gui of choice RSN.
[01:40] <anthony> owh: And does that translate to "it's decent" or "it's the best of server GUIs, but all server GUIs are terrible"?
[01:40] <owh> :)
[01:41] <owh> anthony: Personally, I'm leaning toward the latter, but that is based on looking at it, not using it. I don't feel that I'm a fair observer.
[01:42] <owh> anthony: I'm sure that there are other opinions in this room.
[01:42] <anthony> owh: Fair enough.  I _can_ handle cli configuration, but might give it a look anyway, since if it works that would be very convenient at times.
[01:54] <owh> According to the debian manual aptitude can install tasks, but the man page doesn't divulge how that might be achieved. With apt-get, you can specify with a trailing caret (^), this doesn't work with aptitude. How do I do this?
[01:55]  * owh cannot use tasksel because it hangs at 100% due to a bug.
[02:02] <andguent> owh: what are you trying to install and how?
[02:02] <owh> The task mail-server
[02:02] <owh> :)
[02:04] <owh> I wanted to just run tasksel, but the box is running Feisty and it stops at 100% - this is on my VM instance, not production. So, then the work-around is to install the task from the commandline: apt-get install mail-server^, but my preference is for using aptitude.
[02:04] <owh> I'm not gaim to try this on the production server just yet :)
[02:04] <owh> s/gaim/game/ :)
[02:05] <andguent> ok.. what happens when you do a 'apt-cache search mail-server'
[02:05] <owh> I get procmail and gup back.
[02:05] <owh> There are no packages with that name, the '^' in the apt-get command takes care of that.
[02:06] <andguent> i'm not sure what documentation is throwing an extra ^ in the line, but apt-get doesn't use them in any special way
[02:06] <owh> No, the way to install a task with apt-get is to supply the task name with a '^' at the end.
[02:07] <andguent> if you are still looking to get tasksel installed, you can try 'apt-get install tasksel' -- or maybe i'm just lost and not helping much :)
[02:07] <owh> No, tasksel is installed, but a bug in debconf stops it from working properly. There's a bug about it.
[02:07] <owh> This is leaning towards, suck it up and use apt-get rather than aptitude :)
[02:09] <andguent> i've only used apt-get, ignorant and happy about it :)
[02:09] <owh> ROTFL :)
[02:09] <owh> I'm just going to emulate a different install and see what gives.
[02:11] <owh> Well, well, well. aptitude install postfix worked just fine and dandy, no dovecot, and other stuff :)
[02:14] <andguent> mmmmmkay, glad you figured it out, i think i need to figure out if this screen is causing my headache or not :)
[02:15] <owh> andguent: What is the refresh rate set to?
[02:17] <andguent> 60hz :) I usually can handle that fine.... its TFT too
[02:17] <infinity> owh: Tasks in apititude are something like ~t^taskname, I think.  It's been a while.
[02:17] <infinity> owh: apt-get's interface works for me, and it's what I use on the buildds.
[02:18] <owh> infinity: Tah, I ended up just installing postfix with no mail-server task :)
[02:20] <lamont> infinity: there's another interface besides apt-get ??? :-)
[02:20] <lamont> well, I mean other than dselect
[02:20] <owh> lamont: Now that there is autoremove, not really :)
[02:21] <owh> kirkland: ping
[02:21]  * lamont was introduced to the aptitude version of easter eggs the other day... most cute.
[02:23] <owh> lamont: I suppose it beats the flight simulator in Excel :)
[02:24] <andguent> apt-get moo?
[02:24] <owh> andguent: And others.
[02:24] <lamont> andguent: that's the apt-get version... aptitude has a similar one
[02:26] <andguent> somehow, that just doesn't surprise me
[03:51] <owh> infinity: On the buildds, can you run "extra" scripts and get reports?
[03:53] <owh> infinity: kirkland and I are trying to get init.d scripts to output a meaningful status. I've created some scripts to figure out what has an init.d, and I'm grepping those for "status)", but I wondered if the packages all get installed during build testing and if they are subjected to any standard tests.
[03:54] <owh> infinity: Likely it isn't possible, but I just wondered.
[04:00] <kgoetz> owh: packages are checked by lintian/linda for validity, or do you mean something past that?
[04:02] <owh> kgoetz: We're trying to determine first of all which packages have an init.d script, then which of those supports "status", then which of those is LSB compliant. Fix bugs/Add code accordingly.
[04:02] <owh> I just wondered if the buildds would accommodate that, rather than installing each package and testing.
[04:03] <owh> At present I have all the steps untill "installing the package" and running the init.d script with "status".
[04:03] <kgoetz> owh: you dont have to install - you could download the packages and dpkg -x them (i did lots of looking into those same sorts of questions)
[04:03] <kgoetz> brb though, say 5 min
[04:03] <owh> kgoetz: Cool.
[04:08] <kgoetz> back
[04:09] <kgoetz> determining if something has an init script should be relatively easy. not sure about status (not sure what this means) or lsb compliance, because i dont know whats involved
[04:09] <owh> kgoetz: Yes, I've written the scripts and have all that detail, now I need to determine what it might output.
[04:09] <owh> kgoetz: https://wiki.ubuntu.com/OnnoBenschop/ubuntu-server/init.d-status
[04:11] <kgoetz> what is status?
[04:11] <kgoetz> is it an upstart thing?
[04:11] <owh> kgoetz: It tells you if the package is running etc.
[04:13] <kgoetz> owh: any reason to `find -type f` instead of `find -name \*.deb`?
[04:13] <owh> kgoetz: Yes, there are udebs in there too.
[04:13] <owh> kgoetz: Besides the package dirs should only contain packages :)
[04:14] <kgoetz> they shouldnt contain spaces either :) (which i think are fair assumptions, fwiw)
[04:16] <owh> :)
[04:16]  * owh wrote a nice big disclaimer at the top :)
[04:16]  * kgoetz saw, and was a bit confused by it
[04:17] <owh> kgoetz: It really wasn't intended for anything other than initially getting a handle on the status. If we're going to do this for real, I suspect I'll have a little think about it. Mind you, dpkg -x should handle it, the spaces are simple enough.
[04:17] <owh> faulkes-: Does your graphing magic handle IRC logs?
[04:18] <kgoetz> owh: quote the variables and it shouldnt be a problem
[04:18] <owh> kgoetz: Yup.
[04:19] <owh> kgoetz: I didn't because I got beat up by echo -e and trying to output a \t.
[04:19] <ScottK2> Why do we care about status in init scripts?
[04:20] <owh> Well, when we finally get ebox to do its thing, it might be useful :)
[04:20] <owh> I'm sure that there is a high-brow LSB answer too somewhere :)
[04:21] <owh> s/high-brow/official/
[04:21] <ScottK2> I'll try it a different way then ...
[04:21] <ScottK2> owh: What caused you to care enough to write the script?
[04:21] <owh> Ah, there was some discussion about it. Lemmie have a gander in my logs. One mo.
[04:23]  * kgoetz thought upstart meant no init scripts
[04:23] <kgoetz> *init.d
[04:23] <owh> ScottK2: Yesterday: (08:34:04) kirkland: owh: mathiaz and I were talking earlier today about getting more of the scripts in /etc/init.d LSB-compliant
[04:24] <owh> ScottK2: So blame mathiaz and kirkland :)
[04:25] <owh> Really, getting a status back is not a bad idea. You can then use that status in other places. If there is a standard way of doing it, makes life easier.
[04:25] <owh> Does that answer the question, or did I just head off on a wild goose chase?
[04:26] <owh> kirkland started volunteering and I had a few hours to put some scripts together.
[04:32] <ScottK2> Makes me wonder what the heck status is supposed to provide, but not so much I'm going to go hunt it down.
[04:32] <owh> kirkland told me he had a link, I'll get him to dig it up for us for tomorrow's meeting :)
[04:36] <owh> ScottK2: http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html
[04:39] <owh> Current version is 3.2.0, but the specs for this appear to be the same.
[04:39] <ScottK2> That makes sense I guess.
[04:40] <owh> The nice thing about it is that you can then use the status in other places, a web gui, a cron job, etc.
[04:53] <owh> kgoetz: Is it a reasonable assumption in my script that: a package with init.d script has it stored inside the /etc/init.d directory, not created by post installation script
[04:54] <kgoetz> owh: yes. as a rule
[04:54] <owh> That is, if I unpack it with dpkg -x, I'll have an etc/init.d directory.
[04:54] <owh> Cool
[04:54] <owh> So far so good then :)
[04:56] <kgoetz> what i'm not sure of is if it will have a directory structure, or if you should grep the install scripts for a mv blah /etc/init.d
[04:57] <owh> kgoetz: Well thus far I've had etc/init.d directories, but a grep for etc/init.d might be useful.
[05:18] <owh> kgoetz: Lots of false positives, example scripts, help documents, procps (which has 'cp' in the name :), but nothing that indicates any other means of creating an init.d script.
[05:18] <owh> s/Lots/Found lots/
[05:19] <kgoetz> owh: what did you grep for?
[05:19] <owh> grep -ra '/etc/init.d' ${temp_dir}/* | grep -E 'mv|cp'
[05:20] <owh> The only one that jumped out at me in any way was python-twisted-core, but that appeared to be sample code only.
[05:22] <owh> I suppose we could install *all* applications and see what falls out the bottom :)
[05:22]  * owh is kidding.
[05:25] <owh> kgoetz: Hmm, I suppose for a package to be compliant it needs to register all its files. The ones installed are plain, but the ones we're looking for would perhaps be classed as config files.
[05:27] <kgoetz> owh: what do you mean by 'registered'?
[05:28] <owh> In the dpkg list, there's a file for each package. I cannot recall what its called.
[05:28] <owh> /var/lib/dpkg/info/*.list
[05:30] <owh> That directory appears to be full of init.d references :)
[05:31] <kgoetz> yep.
[05:31] <kgoetz> have you thought about grabbing source packagefs instead of binary?
[05:31] <nxvl> how do i write new blueprints?
[05:31] <owh> kgoetz: Well, at present I'm just using the .iso
[05:31] <nxvl> i don't find the option
[05:31] <kgoetz> owh: ah right
[05:31] <owh> kgoetz: I figured I'd just start at one end :)
[05:32] <kgoetz> nxvl: https://blueprints.launchpad.net/ubuntu/ 'register a blueprint'
[05:32] <kgoetz> ?
[05:32] <kgoetz> owh: yep
[05:32] <nxvl> kgoetz: thnx
[05:33] <owh> The comments we're making here aren't going to waste, so keep poking me in the eye! (I'm adding them to a list to add to the wiki page.)
[05:37] <kgoetz> i'm stopping- i'm heading off :)
[05:38] <owh> Thanks for your comments!
[05:38] <kgoetz> afk. heading home
[05:38] <kgoetz> no worries :)
[05:42] <nxvl> kgoetz: thnx
[07:55] <Jeeves_> Morning!
[08:00] <_ruben> mornin Jeeves_ .. having fun with your sun servers? ;)
[08:03] <Jeeves_> _ruben: Not yet :)
[08:03] <Jeeves_> Connected them with alom
[08:04] <Jeeves_> But need to fix some stuff first
[08:04]  * Jeeves_ is having arguments with the Cisco they're connected to :)
[08:05] <_ruben> hehe
[08:05] <_ruben> looked at the pics the other day .. lookin sweet ;)
[08:06] <_ruben> quite a shitload of disks ;)
[08:07] <Jeeves_> Indeed :)
[08:08] <_ruben> was kinda curious about the cooling though .. they look rather tightly stacked together
[08:08] <Jeeves_> I just requested an seperate vlan and IP-space, so we should be able to play with them this afternoon
[08:08] <Jeeves_> _ruben: What about it?
[08:10] <_ruben> well .. there's not much room for airflow i think?
[08:22] <Jeeves_> _ruben: There's a lot of heat coming from the rear. So the airflow manages to get there :)
[08:41] <amer> hi , how can i list the runlevel processes?, please any help
[08:44] <Jeeves_> 09:37 < bitrot> Config on router jun1.kelvin changed by user 'daniel' with comment add Ubuntu vlan
[08:44] <Jeeves_> Noc is doing their thing ;)
[08:47] <_ruben> heh .. irc feedback of noc operations? how sweet :)
[08:47] <nijaba> morning
[08:48] <Jeeves_> _ruben: snmptraps
[08:50] <_ruben> Jeeves_: ah .. something ive been meaning to investigate as well .. am in the progress of replacing our crappy d-link switches with 3com ones at one of our hosting environments
[08:50] <_ruben> there's only very little info to be found on the details of the various traps
[08:51] <Jeeves_> _ruben: Replacing 'crappy' switches with 3com?
[08:51] <Jeeves_> Those are crappy as well! :)
[08:54] <_ruben> Jeeves_: cisco had my preference, but would've cost 2-3 times more .. and compared to the d-link ones, the 3com ones are rather spectacular .. and who knows, maybe they'll be replaced with cisco's eventually .. if these turn out to not be adequate either ;)
[10:13] <henkjan> jdstrand: stateless ipv6 autoconfiguration is something common used in an desktop environment. in any case its more common used than dhcpv6
[11:04] <jdstrand> henkjan: I read up on it and committed it.  it should hit sometime today
[11:10] <henkjan> jdstrand: okay
[11:10] <henkjan> thnx for quick fixing :)
[11:15] <jdstrand> np ;)
[11:17] <henkjan> jdstrand: would you like to have access to an ipv6 enabled host for testing?
[11:18] <henkjan> i can create a xen domU for testing purposes
[11:18]  * jdstrand ponders that
[11:19] <thewoland> Hi, I would like to install a server version of Ubuntu on a tiny PC
[11:19] <thewoland> The PC is built on VIA Eden running at 200MHz and booting from a CF card.
[11:19] <thewoland> Any suggestion on where to find a good manual for this task?
[11:41] <zul> morning
[12:02] <jdstrand> henkjan: I appreciate the offer, but I think at this time it is not required
[12:03] <jdstrand> henkjan: not a lot are using ipv6 right now, and I think we are ok.  Perhaps we can revisit this in the future?
[12:14] <henkjan> jdstrand: okay
[12:22] <tzn> hello everyone
[12:23] <tzn> got a question about rdac module
[12:23] <tzn> im currently using some DSxxxx hw with centos systems
[12:23] <tzn> what is current support for rdac in ubuntu?
[12:33] <Jeeves_> http://ipv6.google.com/
[12:33] <Jeeves_> 13:03 < jdstrand> henkjan: not a lot are using ipv6 right now, and I think we are ok.  Perhaps we can revisit this in the future?
[12:34] <Jeeves_> jdstrand: Too bad you feel that way
[12:34] <jdstrand> Jeeves_: please explain.
[12:34] <jdstrand> ipv6 is fully supported in ufw
[12:35] <Kamping_Kaiser> Jeeves_, hey mate
[12:35] <jdstrand> I just didn't want henkjan to have to go through all the administrative overhead of giving me access to one of his machines
[12:35] <jdstrand> when I can test ipv6 here locally
[12:35] <Kamping_Kaiser> i wont be able to installfest tomorrow :( one of the test team decided he wanted to play with solaris
[12:36] <jdstrand> Jeeves_: ^^
[12:37] <Jeeves_> jdstrand: Cool that it's supported! Did I misinterpreted the 'not a lot are using ipv6 right now, and I think we are ok'?
[12:37] <jdstrand> Jeeves_: I think so
[12:37] <jdstrand> Jeeves_: I was responding to henkjan asking if I wanted access to an ipv6 enabled machine of his
[12:38] <Jeeves_> I thought you made the often made chicken-egg mistake. My bad.
[12:38] <jdstrand> np
[12:38] <Jeeves_> Kamping_Kaiser: I'm about to configure alom here and there
[12:39] <Jeeves_> And I've got two of them :)
[13:23] <Jeeves_> It pings!
[13:23] <Jeeves_> Anyone knows the default T1000 ALOM pass?
[14:25] <Jeeves_> Booting T1000 ...
[14:26] <sommer> noooooo... he'll go after John Conner!
[14:27] <_ruben> haha
[14:27] <sommer> here all week
[14:27] <faulkes-> sommer? sounds a little close to conner
[14:27] <faulkes-> and as we all know, john does use computers
[14:27] <faulkes-> something you aren't telling us?
[14:28] <sommer> it's not me it's skynet...
[14:28]  * _ruben hides
[14:33] <Jeeves_> Hmm
[14:33] <Jeeves_> I don't have keyboard on the console
[14:35] <zul> hasta la vista baby?
[14:53] <shirish> guys, there was some discussion about a Ubuntu Small Business Server edition, is that still somewhere or no?
[14:56] <Jeeves_> shirish: ebox? :)
[15:00] <shirish> Jeeves_: dunno what ebox is
[15:00] <dexem> shirish: http://www.ebox-platform.com
[15:00]  * shirish thinks it might be something to do with server & enlightment wm
[15:01] <shirish> dexem: thanx, will check it out.
[15:01] <Jeeves_> shirish: Webinterface for SBS-ish stuff
[15:01] <shirish> btw Ubuntu Server doesn't have its own homepage? Everything on the wiki only?
[15:04] <shirish> also does somebody have info. about the specific kernel which ships with the server image, any mail/link explaining the difference between the desktop kernel & server kernels would be great :)
[15:07] <shirish> found it, just sharing it with u guys as well, http://www.serverwatch.com/tutorials/article.php/3715071 seems pretty nice :)
[15:22] <zul> mathiaz: for 3.0.28a debian dropped a bunch of patches so we dropped them as well
[15:24] <mathiaz> zul: cool !
[15:25] <zul> http://pastebin.ca/939747
[15:26] <kirkland> mathiaz: hey, https://bugs.launchpad.net/ubuntu/+source/apache2/+bug/201287
[15:26] <ubotu> Launchpad bug 201287 in apache2 "apache2 init script support for 'status'" [Undecided,New]
[15:27] <mathiaz> kirkland: exzzellent...
[15:27] <kirkland> mathiaz: so I'm working on cron now
[15:28] <mathiaz> kirkland: now could you also check if all the other bits required for lsb compliant scripts are available ?
[15:28] <kirkland> mathiaz: apache was nice, in that it already had a built in "pidof" function and call and dependencies
[15:28] <kirkland> mathiaz: cron does not
[15:28] <kirkland> mathiaz: pidof is part of sysvutils
[15:28] <mathiaz> kirkland: I think that pidof is actually a function provided by the lsb init functions
[15:29] <kirkland> mathiaz: well, the command line utility is in sysvutils
[15:29] <kirkland> perhaps there's a function....
[15:29] <kirkland> mathiaz: so i didn't want to add sysvutils as a Depends
[15:29] <kirkland> mathiaz: but I was thinking of adding it to Recommends
[15:30] <kirkland> and making `etc/init.d cron status` fail gracefully if pidof not found
[15:30] <kirkland> mathiaz: http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html
[15:30] <kirkland> that's what's required by lsb
[15:31] <kirkland> mathiaz: http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptfunc.html
[15:31] <kirkland> lsb also requires a pidofproc function
[15:31] <mathiaz> kirkland: there is a pidofproc function
[15:31] <kirkland> k
[15:31] <mathiaz> kirkland: available in /lib/lsb/init-functions
[15:31] <kirkland> let me try that
[15:32] <kirkland> mathiaz: LSB has some tests
[15:36] <kirkland> mathiaz: pidofproc works like a charm!
[15:36]  * kirkland high fives mathiaz
[15:54] <Jeeves_> zpool1                 8.4T    39K   8.4T     1%    /zpool1
[15:54] <Jeeves_> Hmmm
[15:56] <soren> kirkland: "pidofproc"?
[15:56] <kirkland> soren: http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptfunc.html
[15:56] <kirkland> soren: /lib/lsb/init-functions
[15:57] <soren> kirkland: That's a rather unfortunate description.. "Only process identifiers of running processes should be returned"
[15:58] <kirkland> soren: unfortunate why?
[15:58] <soren> kirkland: So only if they're in actual running state in which case it'd be a no-op on a non-smp system.
[15:58] <soren> or..
[15:58] <soren> on processes which are in the process table, which is redundant since that's the only way they could have a pid.
[15:59] <soren> s/^on/only/
[16:00] <zul> yay: https://bugs.edge.launchpad.net/ubuntu/+source/samba/+bug/201442
[16:00] <ubotu> Launchpad bug 201442 in samba "ffe for samba 3.0.28 to 3.0.28a" [Undecided,New]
[16:00] <kirkland> soren: seems okay for me for daemons, which is the target here
[16:01] <kirkland> soren: are you quabling with the wording of the spec, or the actual use of the function?
[16:01]  * kirkland still can't tell when soren is joking ;-)
[16:02] <soren> kirkland: The wording of the spec.
[16:02] <kirkland> soren_not_joking: thanks, that will keep things clear... just keep /nick up to date at all times :-)
[16:02] <soren_joking> I will.
[16:03] <_ruben> Jeeves_: that could hold a lot of pr0n^H^H^Hrepository files
[16:03] <Jeeves_> :P
[16:03] <kirkland> soren: right, so running processes means to me processes in the process table
[16:03] <kirkland> soren: which may or may not be active/sleeping
[16:04] <soren> Right. Which other kinds of processes have pid's?
[16:04] <soren> None.
[16:04] <soren> So it's redundant (and thus confusing)
[16:04] <kirkland> soren: ps -ef | grep defunct
[16:04] <soren> Specifications should be exact and to the point.
[16:04] <_ruben> Jeeves_: what kind of raid setup is that btw?
[16:05] <Jeeves_> _ruben: No clue.
[16:05] <Jeeves_> Trying to rebuild the whole stuff using zfs
[16:05] <soren> kirkland: So all /runnable/ processes?
[16:05] <_ruben> Jeeves_: hehe
[16:06] <Jeeves_> I'm new to ZFS
[16:06] <_ruben> Jeeves_: but i do hope there's hardware raid involved? :)
[16:06] <soren> kirkland: "runable", perhaps.
[16:06]  * soren is not sure
[16:06] <Jeeves_> _ruben: zfs is software-raid, AFAIK
[16:06] <Jeeves_> _ruben: But it's built for being a fileserver
[16:07] <soren> kirkland: Where do I send patches for that spec? :)
[16:07] <kirkland> soren: i know a couple of the IBM guys who work on it
[16:07] <kirkland> let me check
[16:09] <kirkland> soren: it's Linux Foundation, now, so perhaps in a month in Austin
[16:09] <_ruben> Jeeves_: seems zfs has raid0/raid1 and 'custom versions' of raid5 and raid6 .. looks nice on first sight
[16:09] <soren> kirkland: I'll print out 500 copies of a patch and distribute it all over the place and see what happens :)
[16:10] <kirkland> :-)
[16:11] <kirkland> okay, so the apache2 package doesn't have a -1ubuntu in its name
[16:11] <kirkland> which means that it's just a straight up debian package?
[16:11] <soren> Yup.
[16:11] <mathiaz> kirkland: yes
[16:12]  * mathiaz thinks about all the trees in Denmark 
[16:12] <kirkland> mathiaz: looking at the changelog, i don't even see a -1ubuntu version in its history... does that mean ubuntu patching it is unprecedented (and therefore highly unlikely)?
[16:12] <mathiaz> kirkland: it was - just that when a package is synced, the changelog is dropped
[16:13] <mathiaz> kirkland: otherwise we'd have to keep a merge
[16:13] <kirkland> mathiaz: i see....
[16:13] <kirkland> mathiaz: how, then, do I get around this error: "dpkg-source: error: Version number suggests Ubuntu changes, but Maintainer: does not have Ubuntu address
[16:13] <kirkland> "
[16:14] <mathiaz> kirkland: did you set the Maintainer field to core-dev ?
[16:14] <nxvl> mathiaz: did you see the blueprint i started yesterdays
[16:14] <nxvl> mathiaz: i subscribed yourself to it
[16:14] <kirkland> mathiaz: nope, I suppose that has to be done, since we've modified the package, then
[16:16] <mathiaz> nxvl: I've noticed - still have to read thourgh your proposal
[16:16] <mathiaz> nxvl: did you look at the previous specs and the pages in wiki.ubuntu.com about previous attempts ?
[16:17]  * zul thinks about all the bacon in denmark
[16:17] <mathiaz> kirkland: yes - https://wiki.ubuntu.com/DebianMaintainerField
[16:17] <nxvl> mathiaz: i based the wiki page on ufw one
[16:17] <mathiaz> nxvl: I meant about the content
[16:17] <nxvl> mathiaz: i'm not still finished with the spects, but the base idea is already descriebed
[16:17] <mathiaz> nxvl: there are a couple of spec written up for simplifing the administration of an ubuntu server
[16:18] <nxvl> mathiaz: yes, i take a look at some of them
[16:18] <nxvl> mathiaz: but nothing so similar as i want to
[16:19] <nxvl> mathiaz: also, the tool i want can be used in the future for complete desktop and/or serve administrator, as it is modular
[16:33] <Jeeves_> Who is my date for tomorrow morning, for playing with the T1000's ?
[16:38] <Jeeves_> henkjan: Says it is pschulz01
[16:39] <kraut> are you right or left handed?
[16:41] <kirkland> mathiaz: see http://launchpadlibrarian.net/12607821/apache2_init_status.diff
[16:42] <kirkland> mathiaz: that look better?
[16:44] <mathiaz> kirkland: yes - what about the comment on pidof apache2 in pidof_apache2 ?
[16:44] <mathiaz> kirkland: # if pidof is null for some reasons the script exits automagically
[16:44] <mathiaz> kirkland: # classified as good/unknown feature
[16:44] <kirkland> mathiaz: :-)  yeah, funny, huh?  i did hit that bug
[16:44] <kirkland> mathiaz: and I found that putting the || true inside the shell call fixed it for me
[16:44] <kirkland> mathiaz: hence my comment
[16:45]  * kirkland notes that those comments were unchanged by kirkland
[16:45] <mathiaz> kirkland: yes - but the apache maintainer added a "# classified as good/unknown feature"
[16:46] <mathiaz> kirkland: if you've fixed a good feature, it ain't no good ;)
[16:46] <kirkland> mathiaz: i took that as a poor attempt at comedy
[16:46] <kirkland> mathiaz: calling a "bug" a "feature"
[16:46] <mathiaz> kirkland: under which condition did you hit the bug ?
[16:46] <kirkland> mathiaz: in the case where there are no apache2's running
[16:46] <kirkland> that pidof command just bombs
[16:46] <kirkland> and the script promptly exits
[16:47] <kirkland> silently
[16:47] <kirkland> no information
[16:47] <mathiaz> kirkland: right - and now it would just say that apache2 is not running
[16:47] <kirkland> it is definitely NOT a good feature
[16:47] <mathiaz> kirkland: agreed
[16:47] <Theo__> hello
[16:47] <kirkland> mathiaz: i think the outermost || true could be removed
[16:47] <kirkland> mathiaz: as well as the smart-alec comment
[16:48] <kirkland> mathiaz: but I went with the bare minimum needed for my bug fix
[16:48] <kirkland> mathiaz: alternatively, i could have used pidofproc
[16:48] <kirkland> mathiaz: but it seemed silly, since this script had one specifically tailored toward apache2 built in
[16:48] <kirkland> figured i better use that
[16:50] <mathiaz> kirkland: yes - so we could get rid of the second truee
[16:50] <mathiaz> kirkland: and just use PIDS=`pidof apache2 || true`
[16:50] <kirkland> mathiaz: i agree
[16:50] <kirkland> mathiaz: it's superfluous
[16:51] <mathiaz> kirkland: could you also fix the echo statement ?
[16:52] <mathiaz> kirkland: and use the lsb function instead
[16:52] <kirkland> sure
[16:54] <kirkland> mathiaz: what echo do you speak of?
[16:55] <kirkland> mathiaz: I did change it to this:                 if pidof_apache >/dev/null; then
[16:55] <kirkland> that takes care of it spitting process id's at you
[16:55] <mathiaz> kirkland: there is an echo "ERROR:..." in pidof_apache
[16:56] <kirkland> mathiaz: ah, I see
[16:56] <kirkland> mathiaz: log_failure_msg ?
[16:57] <mathiaz> kirkland: seems like the best option
[17:00] <kirkland> mathiaz: http://pastebin.com/d2311aa2c
[17:03] <kirkland> mathiaz: look better?
[17:05] <mathiaz> kirkland: meh - I don't think we should do that
[17:05] <mathiaz> kirkland: log_failure_msg echoes to stdout
[17:05] <kirkland> mathiaz: crap, i just posted to launchpad
[17:05] <mathiaz> kirkland: so we may end up with PIDS with the error message
[17:06] <mathiaz> kirkland: -> PID=$(pidof_apache)
[17:06] <mathiaz> kirkland: actuall - no
[17:06] <kirkland> mathiaz:         exit 2
[17:06] <mathiaz> kirkland: yes :)
[17:06] <kirkland> mathiaz: it exits immediately
[17:07] <kirkland> mathiaz: so you're happier with it now?
[17:08] <mathiaz> kirkland: one more comment - could you also fix the echo "MPM ..." at the begining of the script
[17:08] <mathiaz> kirkland: and make it use an lsb function
[17:09] <kirkland> mathiaz: log_daemon_msg ?
[17:09] <kirkland> log_end_msg
[17:10] <dfullo> I am trying to use ldap authentication on 7.10 server
[17:10] <dfullo> when I make the changes in nsswitch.conf and reboot the server it hangs
[17:10] <dfullo> passwd:  files ldap
[17:11] <dfullo> group: files ldap
[17:11] <kirkland> mathiaz: I'd also need to move the . /lib/lsb/init-functions
[17:11] <kirkland>  up
[17:11] <dfullo> shadow: files ldap
[17:12] <mathiaz> kirkland: log_failure_msg seems to be more appropriate
[17:12] <dantalizing> mathiaz: have you had a chance to check on status of the -openvz flavor for hardy?
[17:12] <mathiaz> kirkland: log_daemon_msg are debian/ubuntu specific IIRC
[17:12] <mathiaz> kirkland: there are not part of the LSB standard
[17:12] <kirkland> mathiaz: oky
[17:12] <mathiaz> dantalizing: nope
[17:13] <dantalizing> mathiaz: k.  i asked in #ubuntu-kernel and got no response, just fyi
[17:13] <kirkland> mathiaz: http://pastebin.com/dac660c4
[17:14] <mathiaz> dantalizing: gitweb doesn't show anything new on that front
[17:14] <dantalizing> mathiaz: yah thats why i'm asking...i check during my lunchtime ... :(
[17:14] <mathiaz> dantalizing: you may wanna ping BenC directly as he is the one that said openvz should land in hardy
[17:15] <dantalizing> mathiaz: didnt wanna breach any existing protocol and hurt the chances by being a pest... wasnt sure the best way to follow up
[17:16] <dantalizing> mathiaz:i'll try though
[17:17] <mathiaz> kirkland: ok - much better.
[17:18] <mathiaz> kirkland: so to keep improving things, the exit code for the MPM not here should not be 0
[17:18] <mathiaz> kirkland: according to http://refspecs.freestandards.org/LSB_3.2.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html
[17:18] <mathiaz> kirkland: it should be 6
[17:19] <kirkland> mathiaz: done.
[17:20] <mathiaz> kirkland: do you have a debdiff somewhere ?
[17:20] <kirkland> with the exit 6?
[17:21] <mathiaz> kirkland: yes
[17:21] <mathiaz> kirkland: or the whole diff
[17:21] <kirkland> mathiaz: one second...
[17:21] <mathiaz> dantalizing: it's okay to ping someone every once in a while
[17:21] <mathiaz> dantalizing: as long as you don't harress them on a hourly basis
[17:22] <kirkland> mathiaz: http://pastebin.com/d5bdc341f
[17:23] <mathiaz> kirkland: could you also document the exit 6 in the changelog
[17:23] <mathiaz> kirkland: something like "exit code compliant with the LSB policy."
[17:23]  * dantalizing deletes his new ircpingbot
[17:25] <kirkland> mathiaz: http://pastebin.com/d73b8e639
[17:30] <mathiaz> kirkland: so to keep on LSB compliant exit codes, there are two exit 1 that are wrong
[17:30] <mathiaz> kirkland: for reload) and *)
[17:31] <mathiaz> kirkland: reload should exit 6 when there is a configuration problem with apache
[17:31] <mathiaz> kirkland: and *) should exit 3 with an action is not implemented
[17:32] <kirkland> mathiaz: okay, i'll make a sweep of all exit codes
[17:32] <kirkland> mathiaz: this patch was originally just supposed to add a status) section...  now we're well into full LSB compliance work
[17:32] <mathiaz> kirkland: I wouldn't touch exit 2 in pidof_apache
[17:32] <kirkland> mathiaz: is that the new goal?
[17:32] <mathiaz> kirkland: hum... :D
[17:33] <mathiaz> kirkland: not full LSB compliance
[17:33] <mathiaz> kirkland: just fixing stuff here and there
[17:33] <mathiaz> kirkland: after that I'll be happy :)
[17:33] <kirkland> mathiaz: sure you will .... :-)
[17:33] <kirkland> *mathiaz like the iterative process
[17:35] <kirkland> mathiaz: http://pastebin.com/d4caee1f5
[17:40] <mathiaz> kirkland: great - I'm happy now
[17:40] <mathiaz> kirkland: can you attache the debdiff to the bug and I'll sponsor your upload
[17:41] <kirkland> sure thing, will do now
[17:43] <kirkland> mathiaz: debdiff uploaded to https://bugs.launchpad.net/ubuntu/+source/apache2/+bug/201287
[17:43] <ubotu> Launchpad bug 201287 in apache2 "apache2 init script support for 'status'" [Undecided,New]
[17:44] <kirkland> mathiaz: here are the one I was going to tackle immediately...
[17:44] <kirkland> mathiaz: http://pastebin.com/d5c09ebc2
[17:45] <kirkland> i think I can do most of those today
[17:45] <kirkland> owh has a better list, and I'm running his script on my mirror for all of main and universe
[17:45] <mathiaz> kirkland: seems great - I'd suggest that you check in the Debian bug tracker to see if there are patches for the status action
[17:46] <mathiaz> kirkland: and also in LP
[17:46] <kirkland> mathiaz: yep, i've been checking LP
[17:46] <kirkland> mathiaz: haven't checked Debian
[17:46] <kirkland> mathiaz: what do I need to do to push this apache patch to Debian?
[17:46] <mathiaz> kirkland: So as a first step I'd start to file bugs in LP to implement the status action
[17:47] <mathiaz> kirkland: then check if there is a bug in debian
[17:47] <mathiaz> kirkland: if so - link it to LP
[17:47] <kirkland> mathiaz: it to LP, or LP to it?
[17:47] <mathiaz> kirkland: you link it in LP
[17:48] <kirkland> mathiaz: gotcha
[17:48] <kirkland> mathiaz: what about pushing the apache patch to debian?
[17:49] <mathiaz> kirkland: yes - always a good idea to do it.
[17:49] <mathiaz> kirkland: usually I do that once I've uploaded the package to ubuntu
[17:50] <mathiaz> kirkland: with the submittodebian script
[17:51] <kirkland> mathiaz: okay, so you'll take care of this one then, since you're the uploader?
[17:57] <mathiaz> kirkland: I'd rather you do it - so that you can use to the process
[17:58] <kirkland> mathiaz: fair enough
[17:58]  * kirkland is happy to learn
[17:58] <mathiaz> kirkland: actually - submittodebian is more for merges as it requires the ubuntu and the debian dsc files
[17:58] <mathiaz> kirkland: in our current case we only have the ubuntu change.
[18:02] <mathiaz> kirkland: ok - so could you use the sponsorship process for your apache2 diff ?
[18:02] <mathiaz> kirkland: https://wiki.ubuntu.com/MOTU/Sponsorship/SponsorsQueue
[18:03] <kirkland> mathiaz: sure, let me go read that
[18:11] <kirkland> mathiaz: ubuntu-main-sponsors right?
[18:11] <mathiaz> kirkland: yes - apache2 is in main
[18:12] <kirkland> mathiaz: okay, i renamed the patch to .debdiff, set the status to confirmed, and subscribed ubuntu-main-sponsors
[18:12] <mathiaz> kirkland: excellent.
[18:12] <kirkland> mathiaz: i think that's all i was missing from the SponsorQueue page
[18:12] <mathiaz> kirkland: it was just to introcude you to the sponsorship program
[18:12] <mathiaz> kirkland: now I can get your debdiff from there
[18:13] <kirkland> mathiaz: sure, no problem
[18:13] <kirkland> that's good
[18:14] <mathiaz> kirkland: ok - so I've uploaded your cyrus-sals2 merge
[18:14] <kirkland> mathiaz: 'your' cyrus-sals2 merge :-P
[18:14] <kirkland> mathiaz: i'm not to be blamed for that mess :-)
[18:14] <kirkland> j/k
[18:15] <mathiaz> kirkland: you should have received an email from LP
[18:15] <mathiaz> kirkland: now the next thing to do is to forward patches to debian
[18:15] <kirkland> mathiaz: okay....
[18:16] <kirkland> mathiaz: wiki page?
[18:17] <spiekey> hi
[18:17] <spiekey> does someone know a firewall testing tool to test udp and tcp ports?
[18:17] <spiekey> like ftester
[18:18] <mathiaz> kirkland: https://wiki.ubuntu.com/Bugs/ReportingToDebian
[18:21] <kirkland> mathiaz: cool, querybts shows no relevant bugs
[18:23] <kirkland> mathiaz: okay, so i run submittodebian
[18:23] <kirkland> mathiaz: take out the maintainer field changes
[18:23] <kirkland> mathiaz: leave everything else
[18:24] <mathiaz> kirkland: well - I tend to actually break the diff into functional patches
[18:24] <mathiaz> kirkland: so that the debian maintainer doesn't have to go through a giant patch with multiple bugs fixed in it
[18:25] <kirkland> mathiaz: this isn't exactly a "giant patch" though
[18:25] <mathiaz> kirkland: right - it was more a general comment
[18:25] <kirkland> mathiaz: k
[18:25] <mathiaz> kirkland: let me check what I'd do
[18:26] <kirkland> mathiaz: normal severity (6)
[18:27] <mathiaz> kirkland: so I wouldn't include the postinst diff
[18:27] <mathiaz> kirkland: it's ubuntu specific
[18:27] <mathiaz> kirkland: the change to rules make sense for debian
[18:28] <mathiaz> kirkland: for the control file I'd open a bug mentioning the reason to do this
[18:28] <kirkland> mathiaz: there is no postinst in my diff
[18:28] <mathiaz> kirkland: so I'd file two bug reports - one with the clean target diff
[18:28] <mathiaz> kirkland: and one for the control file
[18:29] <mathiaz> kirkland: ok - you're probably not using the one I've uploaded
[18:29] <kirkland> mathiaz: are we talking about apache2 here?
[18:29] <mathiaz> kirkland: no - cyrus-sasl2 !
[18:29] <kirkland> mathiaz: jebus
[18:30] <mathiaz> kirkland: I was refering to cyrus-sasl2 because I've just uploaded the merge
[18:30] <kirkland> mathiaz: i think you should submit the cyrus-sasl2 ones, since you have the updated sources.  i'll do the apache2 one, for the experience, and because I have all the info
[18:30] <mathiaz> kirkland: ok.
[18:34] <kirkland> mathiaz: apache2 patches sent to debian
[19:16] <chris____> hello, does anyone here have experience with BIND and DNS?
[19:16] <chris____> I've set myself up a bind server and it works fine locally, its just when I do an nslookup on the domain from my local computer, it doesn't return anything
[19:18] <mathiaz> soren: how can I setup a bridged network for libvirt/kvm ?
[19:21] <sommer> chris____: have you put the dns server into /etc/resolv.conf?
[19:22] <chris____> sommer: yes it is the first entry in my /etc/resolv.conf
[19:22] <sommer> is port 53 open on the server?
[19:25] <chris____> sommer: yes, nmap locahost shows port 53 is open
[19:26] <sommer> chris____: you might try restarting bind9 and checking /var/log/syslog for errors
[19:26] <chris____> ill check it out right now, thanks
[19:27] <sommer> does dig -x domain.name work on the dns server?  another check anyway
[19:27] <sommer> dig is part of the dnsutils package... if you don't have it installed
[19:28] <chris____> here is the output from syslog, btw my domain name is my325i.com and the IP is 69.93.237.13
[19:28] <chris____> Mar 12 15:27:01 chris named[19418]: zone 237.93.69.in-addr.arpa/IN: loaded serial 2008031000
[19:28] <chris____> Mar 12 15:27:01 chris named[19418]: zone my325i.com/IN: NS 'ns1.my325i.com' has no address records (A or AAAA)
[19:28] <chris____> Mar 12 15:27:01 chris named[19418]: zone my325i.com/IN: loaded serial 2008031200
[19:28] <chris____> Mar 12 15:27:01 chris named[19418]: zone localhost/IN: loaded serial 1
[19:28] <chris____> Mar 12 15:27:01 chris named[19418]: running
[19:28] <chris____> sorry for pasting in here, I should have pasted it in a text file
[19:28] <sommer> chris____: try creating an A record for ns1
[19:29] <sommer> "zone my325i.com/IN: NS 'ns1.my325i.com' has no address records (A or AAAA)"
[19:30] <chris____> okay, in the db.my325i.com file ? Here's my dig output http://69.93.237.13/dig.txt
[19:36] <chris____> I added it in the db.my325i.com file and I got a new error, heres the output from /var/log/syslog http://69.93.237.13/syslog.txt, the entry I tried was "@ IN A ns1.my325i.com"
[19:37] <nijaba> :( ubotu left us
[19:37] <chris____> haha hes back
[19:38] <chris____> I have to go, but I will be back on later probably asking the same question, thanks for helping me troubleshoot
[19:39] <sommer> doh
[19:53] <soren> mathiaz: You need to set up your bridge first manually.
[19:53] <soren> mathiaz: After that, it'll show up in virt-manager.
[19:54] <mathiaz> soren: ah - so I should not try to get an xml file for libvirt
[19:55] <soren> Nah, don't bother.
[19:59] <kirkland> mathiaz: maintainer field change added to changelog
[20:08] <mathiaz> soren: I'm following this wiki page for setting a bridge to use with kvm: https://wiki.ubuntu.com/KvmWithBridge
[20:08] <mathiaz> soren: there is a reference to a tap0 device, but I don't have it on my hardy server
[20:15] <chris____> hello, I'm back, I was having trouble with DNS on my server. I can't seem to get it to work properly. I posted up my "dig -x my325i.com" (my325i.com is my domain) at http://69.93.237.13/dig.txt, I also posted up a snippet from my /var/log/syslog file at http://69.93.237.13/syslog.txt
[20:15] <nxvl> mathiaz: server meeting starts in 45 minutes, didn't it?
[20:15] <chris____> if anyone could help, it would be greatly appreciated
[20:16] <chris____> by the way, the syslog snippet I posted was after adding "@ IN A ns1.my325i.com." to my db file
[20:16] <mathiaz> nxvl: the server meeting will start in 45 minutes - at 21:00 UTC
[20:17] <twb> Is there an equivalent to snapshot.debian.net for Ubuntu?
[20:18] <kirkland> mathiaz: see if this one is better on the first time through....  https://bugs.launchpad.net/ubuntu/+source/cron/+bug/201504
[20:19] <ubotu> Launchpad bug 201504 in cron "cron init script support for 'status'" [Undecided,Confirmed]
[20:19] <nxvl> mathiaz: thnx, i still have problem with UTC time
[20:19] <nxvl> and i can't add it to my gnome clock
[20:20] <mathiaz> nxvl: date -u is your friend
[20:22] <ScottK2> kirkland: Is there really a point in filing these bugs right now?  This is really a Hardy +1 issue.
[20:24] <kirkland> ScottK2: in almost every case, they are trivial to fix
[20:24] <kirkland> ScottK2: it is an LSB-compliance issue
[20:25] <kirkland> ScottK2: and it's something that linux admins rely upon in other linux distros--i think it's important for the ubuntu server to provide such informaiton
[20:27] <kirkland> ScottK2: we're concentrating (for now) on the core services in ubuntu-server which have the bug
[20:27] <kirkland> ScottK2: you can see my patches so far for apache, and cron
[20:28] <kirkland> ScottK2: for the many, many other init scripts (non ubuntu-server), sure, Hardy+1 is great
[20:32] <kirkland> owh: howdy
[20:32] <kirkland> owh: sorry i missed your pings last night ;-)
[20:32] <owh> It happens when you're asleep :)
[20:33] <owh> Did you see the updated stuff on the wiki page?
[20:33] <kirkland> owh: i did
[20:33] <owh> Did it help?
[20:33] <kirkland> owh: and I have made a few changes to your script
[20:33] <owh> Bring it on!
[20:33] <owh> :)
[20:33] <kirkland> owh: yeah, but there's something else we're going to have to work around to run it against a full mirror
[20:34] <owh> ?
[20:34] <kirkland> owh: so I'm running it against a local copy of something like http://archive.ubuntu.com/ubuntu/pool/main/
[20:34] <owh> kirkland: Wasn't that the intent? Or are you talking about something else?
[20:35] <kirkland> owh: now have a look at final subdirectory
[20:35] <kirkland> http://archive.ubuntu.com/ubuntu/pool/main/a/abiword/
[20:35] <kirkland> so I can filter that down some with find -type f -name *_i386.deb
[20:35] <kirkland> to get rid of the dupliacate arches
[20:35] <owh> kirkland: Yes, that will need to be filtered :)
[20:35] <owh> What about udeb?
[20:36] <owh> The .iso was full of them.
[20:36] <kirkland> owh: but the versions present a problem...  we only want the latest
[20:36] <owh> I didn't see that, one mo.
[20:37] <owh> kirkland: Yes, that's right. I can deal with that with a sort and a tail.
[20:37] <kirkland> owh: almost; maybe a fancy sort
[20:37] <owh> kirkland: Based on that the naming of the latest version will be alphabetically consistent though.
[20:37] <kirkland> owh: because those dirs have multiple packages
[20:38] <kirkland> owh: abiword-common, abiword-gnome, etc.
[20:38] <kirkland> i think sort can split up parts of the string
[20:38] <owh> kirkland: So, perhaps we'd need to start at the other end of the problem, that is, start from a package list.
[20:39] <kirkland> owh: right, so I have some perl that might help
[20:40] <owh> kirkland: It depends what we're doing this for. Is it a one-off, or is it going to be an ongoing compliance check?
[20:40] <owh> kirkland: If it's one off, we can construct a list, if not we need to be able to handle what comes our way.
[20:40] <Erick> i need some help getting apache to work on ubuntu
[20:40] <kirkland> owh: a one-off, I hope
[20:41] <kirkland> owh: LSB has compliance tests
[20:41] <kirkland> owh: http://www.linux-foundation.org/en/Testing
[20:41] <owh> kirkland: How big is a full listing of a mirror?
[20:41] <kirkland> owh: one minute, i'll pastebin it for you
[20:42] <kirkland> owh: this is for main and universe
[20:42]  * owh gets ready to copy/paste :)
[20:42] <kirkland> owh: a listing of all /etc/init.d/* scripts, sorted by the package that owns it
[20:42] <Erick> is anyone able to help me?
[20:43] <owh> Erick: You need to tell us what is wrong.
[20:43] <Erick> owh:  http://www.pastebin.ca/940179
[20:43] <nijaba> Erick: ask your question, don't ask to ask
[20:45] <kirkland> Erick: try starting it more like "sudo /etc/init.d/apache2 start"
[20:46] <owh> I suspect that Erick has another copy of a web-server running.
[20:46] <Erick> owh, no
[20:46] <csguest> could somebody take a look at: https://bugs.launchpad.net/ubuntu/+source/xen-3.2/+bug/199533
[20:46] <ubotu> Launchpad bug 199533 in xen-3.2 "Device 0 (vif) could not be connected. Hotplug scripts not working." [Undecided,Confirmed]
[20:46] <csguest> it is confirmed and we have proposed fixes
[20:47] <Erick> kirkland, when i start it like that it tells me it is already running
[20:47] <kirkland> Erick: "sudo /etc/init.d/apache2 restart"
[20:47] <owh> kirkland: So, this listing is output, or is it a dir listing of the mirror?
[20:48] <Erick> kirkland, still cant connect to my computer
[20:48] <kirkland> owh: perl script is here http://pastebin.com/d39fab86e
[20:48] <kirkland> owh: uses a find under the covers
[20:49] <kirkland> owh: uses a dpkg -c to list the contents of the package (such that I don't have to explode it like you do)
[20:49] <kirkland> owh: currently, only looks for ./etc/init.d/
[20:49] <kirkland> owh: needs your code to bust it open, and search for "status)"
[20:49] <nijaba> Erick: can you ping your server?
[20:49] <owh> kirkland: Did you see the discussion that kgoetz and I had about that yesterday? It may well be that there are other ways to create init.d scripts :(
[20:49] <kirkland> Erick: have you customized the apache2 configuration files already?  or are you using the stock config
[20:50] <Erick> nijaba, yes i can ping it
[20:50] <kirkland> owh: hmm, i don't remember seeing that
[20:50] <Erick> kirkland, ive configured it a different site, but have not done anything with apachee.conf
[20:51] <owh> kirkland: I summarised it on the wiki: https://wiki.ubuntu.com/OnnoBenschop/ubuntu-server/init%2ed-status#head-0325a23d7857a153c12a28aedfe10ccf6c6fa10c
[20:51] <kirkland> owh: here's the first part http://pastebin.com/d6ba4dded
[20:51] <nijaba> Erick: can you paste your apache.conf on pastebin.ca?
[20:51] <Erick> sure
[20:51] <kirkland> nijaba: i doubt its his apache.conf...  it's probably one of the sites-enabled
[20:52] <nijaba> Erick: sorry I meant your confing for the other site
[20:52] <kirkland> Erick: can you restore the default configurations and see if apache2 will start?
[20:52] <Erick> whoa, you lost me nijaba
[20:52] <Erick> kirkland, i dont know how
[20:52] <nijaba> Erick: you said that you did not touch apache.conf
[20:52] <Erick> yeah
[20:52] <Erick> i didnt
[20:52] <Erick> well, a little bit with the servername
[20:53] <Erick> but thats all
[20:53] <nijaba> Erick: so I ma wondering how you configured the "other site"
[20:53] <Erick> i told apache to use that one instead of the old one
[20:53] <kirkland> Erick: list the contents of "ls /etc/apache2/sites-enabled/"
[20:54] <Erick> kirkland, no such file or directory
[20:55] <kirkland> Erick: what version of Ubuntu are we talking about?
[20:55] <Erick> gusty
[20:56] <kirkland> Erick: ls /etc/apache2/sites-available/
[20:56] <Erick> ok, got it
[20:56] <Erick> the site is: ericksite
[20:56] <Erick> i configured it to use that one
[20:57] <Erick> when i go to http://localhost/ i get that page, too
[20:57] <kirkland> Erick: using a2ensite?
[20:57] <Erick> kirkland, yes
[20:57] <Erick> then i restarted apache
[20:58] <kirkland> hmm, i have doubts about your installation if you're lacking an /etc/apache2/sites-enabled/ directory, but whatever... if http://localhost/ is working, it's a problem within your config file
[20:58] <soren> twb: What are you looking for specifically?
[20:58] <kirkland> Erick: post your ericksite config to a pastebin
[20:58] <kirkland> stripping anything sensitive, of course
[20:58] <Erick> kirkland, i did get the sites enabled
[20:58] <Erick> that was it
[20:58] <Erick> ericksite
[20:58] <kirkland> Erick: cat ericksite
[20:58] <Erick> cat?
[20:59] <kirkland> Erick: man cat
[20:59] <Erick> what do you mean?
[21:00] <kirkland> Erick: I'm giving you commands to run
[21:00] <Erick> ok
[21:00] <Erick> lol
[21:00] <nealmcb> server team meeting now in #ubuntu-meeting
[21:00]  * owh heads over to the meeting.
[21:00] <kirkland> Erick: man is how you read documentation from a command line in linux
[21:00] <kirkland> Erick: "man cat" will tell you that cat is the command used to print a file to the screen
[21:00] <Erick> ok
[21:01] <Erick> i printed it
[21:01] <kirkland> Erick: copy and paste that to a pastebin if you want help with your configuration
[21:01] <deshantm> is there someone that could pass on a bug to the ubuntu Xen team or Chuck?  #199533
[21:01] <deshantm> it is confirmed and we have proposed solutions
[21:01] <Erick> kirkland:  http://www.pastebin.ca/940241
[21:02] <Erick> whoa, wait
[21:02] <kirkland> Erick: I don't need to see the documentation to "cat" ...
[21:02] <Erick> cat ericksite doesnt work
[21:02] <twb> soren: the ability to grab debs that were in an Ubuntu repo, but are no longer (because they've been superseded.)
[21:03] <kirkland> Erick: probably because you aren't in the right directory?
[21:03] <soren> twb: They're all on launchpad. Why do you need them, though?
[21:03] <twb> soren: for example, right now I wanted the newest version of coreutils that didn't cause LP 191884.
[21:03] <ubotu> Launchpad bug 191884 in coreutils "wrong id behaviour on a system with LDAP" [Undecided,New] https://launchpad.net/bugs/191884
[21:04] <soren> twb: Ok, launchpad it is, then.
[21:04] <twb> I don't understand what you mean by `they're all on launchpad'
[21:05] <kirkland> Erick: cat /etc/apache2/sites-available/ericksite
[21:05] <kirkland> Erick: or, cd /etc/apache2/sites-available; cat ericksite
[21:06] <Erick> kirkland:  http://www.pastebin.ca/940254
[21:11] <Erick> kirkland, i need to go
[21:12] <kirkland> Erick: the problem is almost certainly that you have multiple web servers running
[21:12] <kirkland> Erick: look hard for another server running
[21:12] <Erick> i dont though
[21:12] <Erick> never messed wither servers before
[21:12] <Erick> the system was only installed 2 weeks ago
[21:12] <kirkland> and try the /etc/init.d/apache2 restart thing again
[21:12] <owh> Erick: That looks a lot like the default site. I think you have both default and ericsite enabled on the same port.
[21:12] <Erick> but i need to go, thanks for the help
[21:13] <kirkland> Erick: sorry, good luck
[21:14] <kirkland> Erick: owh: could be... you should set each site to listen on a unique Listen port or ServerName
[21:14] <owh> kirkland: I think Erick copied default to ericsite, edited it a bit and enabled it.
[21:15] <kirkland> owh: right, he doesn't have a proper sites-enabled directory
[21:15] <kirkland> owh: and he should be using a2ensite
[21:15] <owh> kirkland: I'm putting that down to a typo on entering the ls command.
[21:16] <kirkland> owh: http://pastebin.com/d4fca412e
[21:16] <kirkland> owh: complete list
[21:17] <kirkland> owh: that's main only, actually
[21:17] <owh> kirkland: That's not looking too scary.
[21:18] <kirkland> owh: i'm running universe now
[21:18] <kirkland> gonna take 5...
[21:18] <owh> Sure.
[21:19] <owh> kirkland: You going to be around for a bit after the meeting?
[21:19] <kirkland> owh: yep
[21:19] <owh> k
[21:28] <soren> twb: Which package do you want, I'll show you how to find them.
[21:30] <twb> OK, for the sake of the exercise let's say coreutils=5.97-5.6ubuntu1
[21:31] <soren> https://launchpad.net/distros/ubuntu/+source/coreutils/
[21:31] <soren> twb: Find the release you want and click the version number of it.
[21:32] <soren> That brings you to https://edge.launchpad.net/ubuntu/+source/coreutils/5.97-5.6ubuntu1
[21:32] <soren> Go down to the builds sections. You want amd64 or i386?
[21:32] <twb> Yes.
[21:32] <soren> Which one? :)
[21:32] <twb> Preferably as in an apt repo, like snapshots.d.n
[21:33] <twb> s/as//
[21:33] <soren> snapshots.d.n is not an apt repo.
[21:33] <twb> It's a whole bunch of apt repos
[21:33] <twb> One for each package
[21:33] <soren> Ah.
[21:33] <twb> Well really it's one repo, but with /package instead of /main
[21:33] <soren> WEll, do you want the package or not? :)
[21:34] <twb> One moment, I'm still taking notes
[21:34] <soren> Ok, for the sake of the excercise, let's choose amd64. Click hardy amd64, which brings you to https://edge.launchpad.net/ubuntu/+source/coreutils/5.97-5.6ubuntu1/+build/472902.
[21:35] <soren> On the left, there's a box called "Resulting binaries". Click the binary package you want..
[21:35] <soren> ..and there it is ready for your downloading pleasure.
[21:35] <soren> The url for that last page is:
[21:35] <twb> OK.
[21:35] <soren> https://edge.launchpad.net/ubuntu/hardy/amd64/coreutils/5.97-5.6ubuntu1
[21:35] <twb> That's better than nothing, but it's very tedious.
[21:36] <soren> Not if you just construct the url if you know the binary, arch, and version.
[21:36] <twb> Often I don't know exactly which version is the one that introduced a regression
[21:36] <soren> If you often need older binaries, that's a symptom that something is very wrong somewhere.
[21:36] <twb> So I want to add all of them to my sources.list and downgrade them one version at a time
[21:37] <soren> "all of them2?
[21:37] <soren> "?
[21:37] <twb> soren: yes, I would never downgrade on a normal system
[21:37] <twb> "all of them" as in all versions
[21:37] <twb> This is the snapdhots.d.n workflow
[21:37] <soren> What would you need it for?
[21:37] <twb> Finding what release introduced a regression
[21:38] <twb> Or finding what is the latest version of a library that works with your stupid in-house code.
[21:39] <soren> *shrug* File a bug against soyuz.
[21:39] <soren> Again:22:36:57 < soren> If you often need older binaries, that's a symptom that something is very wrong somewhere.
[21:39] <twb> Granted.
[21:40] <twb> If hardware and software worked reliably, I'd not have a job
[21:40] <twb> :-)
[21:40] <soren> I'm sorry to hear that. Do you work in a support centre?
[21:40] <soren> :p
[21:40] <twb> I'm a sysadmin
[21:47] <kraut> i'm your worst nightmare
[21:48] <soren> Hah!
[21:49]  * zul is an EX-sysadmin..wohoo
[22:04] <nxvl> sorry, i needed to go out for a while, what are we talking about now?
[22:04] <nxvl> wrong channel
[22:04] <nxvl> :P
[22:10] <ivoks> ScottK: this is something we could do for hardy+1
[22:10] <nxvl> mathiaz: did you take a look at the blueprint?
[22:11] <nxvl> i still need to work harder on the wiki page, but the base idea is there
[22:11] <ScottK2> ivoks: Yes
[22:12] <ScottK2> ivoks: You still need you new tasksel revision uploaded, right?
[22:12] <ivoks> ScottK2: right... patch is uploaded
[22:13] <mathiaz> nxvl: not yet
[22:13] <ScottK2> ivoks: So you do not need me to do it?
[22:14] <ivoks> ScottK2: umm...? i'm not core dev, so i can't apply it
[22:14] <ScottK2> ivoks: So you meant you'd uploaded it to LP.  Got it.
[22:14] <ivoks> ScottK2: i've just added new version to lp
[22:14] <ivoks> right
[22:25] <ScottK2> mathiaz: Why do you think the tasksel needs an FFe?  It looks like bug fixing a previous broken upload to me?
[22:35] <kirkland> mathiaz: so we now have a better list for you
[22:35]  * owh runs for cover.
[22:35] <kirkland> mathiaz: http://ubuntu.pastebin.com/m627c07fb
[22:35] <kirkland> mathiaz: owh installed a vm with the latest hardy-server snapshot
[22:36] <kirkland> ticking all the servers offered by the installer
[22:36] <mathiaz> ScottK2: are you refering to https://bugs.launchpad.net/ubuntu/+source/dovecot/+bug/164837 ?
[22:36] <ubotu> Launchpad bug 164837 in dovecot "Dovecot SASL for postfix" [Low,In progress]
[22:36] <ScottK2> mathiaz: Yes.
[22:36] <owh> mathiaz: Then I installed all 7 server tasks.
[22:36] <mathiaz> ScottK2: the debdiff adds a new postinstall script
[22:36] <ScottK2> Yeah.  I'm asking slangesek about it now.
[22:37] <kirkland> mathiaz: of which 4 (apparmor, postfix, postgresql, cupsys) have status sections
[22:37] <kirkland> mathiaz: plus we already have patches for 2 more (apache, cron)
[22:37] <mathiaz> ScottK2: ok.
[22:37] <kirkland> mathiaz: waaay to many for he and i to hack out overnight, as ScottK2 suspected
[22:38] <mathiaz> kirkland: hum - mysql has a status action IIRC
[22:39] <kirkland> mathiaz: ooh, good point, bad grep on our part
[22:39] <kirkland>   'status')
[22:39] <kirkland> quoted, tricky.....
[22:40] <ScottK2> So no ivoks has run off
[22:42] <ScottK2> mathiaz: Did you see slangasek's comments on ubuntu-devel?  I think we need to think about Dovecot again
[22:44] <kirkland> mathiaz: so mysql was the only additional one on that list
[22:44] <ScottK2> mathiaz: At a glance, I don't see anything in the Dovecot part of the tasksel change that's clearly wrong for putting straight in the Dovecot package.
[22:44] <ScottK2> That'd make this whole thing much simpler
[22:45] <kirkland> mathiaz: at this point, should we just pick a few of the most important?  (eg openssh-server) and propse those?
[22:45] <ScottK2> I see you discussing it now ...
[22:48] <Erickj92> kirkland, would having the apache server going on another computer on the same network effect it on the machine im trying to get things working on?
[22:49] <kirkland> Erickj92: nope
[22:49] <Erickj92> darn
[22:49] <Erickj92> is there any other possible thing?
[22:50] <Erickj92> for the record, i cant ping that machine
[22:50] <kirkland> Erickj92: well that would be a problem
[22:54] <faulkes-> if you can't ping the machine, then a more fundamental problem exists than apache
[22:54] <faulkes-> check that the interface is up and configured
[22:54] <Kamping_Kaiser> unless you drop icmp
[22:54] <Erickj92> faulkes-, how?
[22:55] <kirkland> Erickj92: can that machine ping out?
[22:55] <Erickj92> ping out?
[22:55] <Erickj92> oh
[22:55] <faulkes-> well, default install shouldn't affect it (if this is a default install)
[22:55] <Erickj92> yeah
[22:55] <Erickj92> it can ping out
[22:55] <faulkes-> then it has connectivity
[22:55] <faulkes-> so I would say it looks like an iptables issue
[22:56] <Erickj92> i just bought this new static IP from my ISP
[22:56] <Erickj92> is there somthing in my router settings that could be doing it?
[22:57] <faulkes-> if you are dealing solely on the local lan, probably not but that isn't a garrauntee, every network is different
[22:57] <Erickj92> im so confused
[22:57] <faulkes-> but if you can't ping your machine but it can ping out, it sounds like there is something firewalling it (iptables)
[22:58] <Erickj92> from where/
[22:58] <Erickj92> ??
[22:58] <Erickj92> like a software firewall, or hardware
[22:59] <Erickj92> faulkes-, when i go to www.whatismyip.org, it tells me somthing different then the ISP told me
[22:59] <Erickj92> that i had
[23:01] <Erickj92> what is the command to stop apache
[23:01] <Erickj92> ?
[23:15] <iclebyte> is there a package for BASE?.. i cant seem to find one.
[23:16] <soren> "BASE"?
[23:20] <iclebyte> nope.. there is a package called 'acidbase' but it wants to use postgresql and i'm trying to use it in conjunction with the snort-mysql package
[23:28] <faulkes-> hrmmm, I'm going to have to go over todays meeting log
[23:28]  * faulkes- was running back and forth
[23:28] <faulkes-> laptop install, server install, meeting
[23:28] <faulkes-> not enough eyes and hands
[23:41] <soren> ScottK2: dovecot.conf is a conffile? Sure?
[23:41] <ScottK2> soren: Everything in /etc is a conffile since debhelper 3 or 4.
[23:41] <soren> Er.. No.
[23:42] <ScottK2> No?
[23:42] <soren> Anything shipped in /etc is a conffile.
[23:42] <soren> "shipped" being the keyword.
[23:42] <ScottK2> Ah.
[23:42] <soren> dovecot.conf is built during postinst from a template, iirc.
[23:43] <ScottK2> I'd say that's parsing things pretty finely.
[23:43]  * ScottK2 goes and reads the policy
[23:43] <soren> No.
[23:43] <soren> It's really not.
[23:43] <soren> It's done that way to avoid it being a conffile.
[23:43] <soren> Policy forbids you from doing automated editing of conffiles.
[23:44] <soren> ..so if you want to edit a file in /etc e.g. during postinst, you need to make sure it's not a conffile.
[23:44] <soren> "conffile" is not just a clever name for files shipped in /etc.
[23:44] <soren> It's a status that has very specific meaning to dpkg.
[23:45] <soren> smb.conf is not a conffile either.
[23:47] <mathiaz> soren: yes - however dovecot.conf is managed via ucf.
[23:47] <mathiaz> soren: would this matter ?
[23:48] <soren> To what? I havent' followed the discussion. I just saw ScottK's comment in a bug claiming dovecot.conf was a conffile and it's not.
[23:49] <soren> The short version is:
[23:49] <mathiaz> soren: whether a maintainer script can modify a file managed by ucf, but not a conffile
[23:49] <soren> You can choose to mark your configuration files as conffiles, never edit them programmatically, and let dpkg handle upgrades and all that jazz..
[23:49] <soren> or..
[23:49] <soren> handle upgrades gracefully yourself.
[23:50] <mathiaz> soren: let me rephrase
[23:50] <soren> Gracefully means: Don't clobber users' changes, don't ask the user to confirm changes that *you* made from a script, etc.
[23:50] <mathiaz> soren: whether a maintainer script *from another package* can modify a file managed by ucf, but not a conffile
[23:50] <soren> mathiaz: It's not strictly forbidden, afair, but frowned upon.
[23:51] <soren> If you need that sort of thing, the package that "owns" the configuration file should provide a script or something to do the relevant changes.
[23:51] <mathiaz> soren: well - in the case of dovecot, the proposal is to enable sasl by default
[23:52] <soren> http://www.debian.org/doc/debian-policy/ch-files.html#s10.7.4 for reference.
[23:53] <soren> mathiaz: Which package will need to edit which package's configuration?
[23:54] <mathiaz> soren: looking at https://bugs.launchpad.net/ubuntu/+source/dovecot/+bug/164837
[23:54] <ubotu> Launchpad bug 164837 in tasksel "Dovecot SASL for postfix" [Undecided,Invalid]
[23:55] <mathiaz> soren: and the last patch from ivoks :http://launchpadlibrarian.net/12606104/tasksel.diff
[23:55] <mathiaz> soren: there are two things that needs to be done: configure postfix to use sasl (via postconf) and modify dovecot to run a sasl server
[23:56] <mathiaz> soren: the latter requires a modification of dovecot configuration file
[23:56] <soren> That, in itself, is not a problem.
[23:56] <mathiaz> soren: the former call postconf to setup postfix correctly, which is allowed in the policy
[23:57] <mathiaz> soren: right - but it's frown upon.
[23:57] <soren> To maintain sanity, the code to actually enable sasl in dovecot should be in the dovecot package.
[23:57] <soren> ...and the tasksel postinst can call that (or whatever).
[23:57] <mathiaz> soren: well it seems that we could actually ship dovecot with sasl enabled by default
[23:57] <mathiaz> soren: which would solve our problems
[23:58] <soren> mathiaz: Well, no, it's fine to edit another package's configuration. You just need to have the mechanics straight.
[23:58] <soren> mathiaz: That would be lovely.
[23:59] <soren> I've never liked the idea of the tasksel postinst.