[01:54] <jameswf-home> anyone know why I have to put my seed file in the initrd for it to work?
[05:36] <chmac> How do I grep for a /? grep -E / -F and any combination of \/home/sites\\\/ doesn't seem to work.
[05:36] <chmac> I'm trying to find the string "/home/sites/"
[05:36] <ScottK> \/
[05:37] <ScottK> \/home\/sites I think will work.
[05:39] <hads> Just put it in quotes.
[05:40] <chmac> ScottK: `grep -rHn "\/home\/sites\/" *` doesn't work as expected, it produces all sorts of spurious results which don't appear to contain home or site
[05:41] <chmac> hads: You mean juse use `grep -rHn "/home/sites/" *` ? That also produces spurious results for me
[05:41] <ScottK> What hads said should work.
[05:41] <jameswf-home> whats wrong with ls
[05:43] <hads> Yeah, `grep "/home/sites/" *` should be fine.
[05:46] <chmac> Ok, so it does work, the error was elsewhere, sorry
[05:46] <chmac> I was running `grep -rHn "/home/sites/" * | less`
[05:47] <chmac> Then in less, /home didn't return any results, but the text was there, less just didn't find it properly
[05:47] <jameswf-home> chmac: what are you tring to acomplish
[05:51] <chmac> jameswf-home: I've moved a site from one server to another, the old path was /home/sites/ the new one is different
[05:51] <chmac> I wanted to find all instances of /home/sites/ in the new htdir
[05:52] <chmac> I was piping it to less because it was returning a line from an SQL cache file which was extremely long, so it was cluttering my screen
[05:52] <chmac> I thought that grep wasn't working because when I tried searching in less, for "home", it wasn't found, so I thought grep was returning nonsense
[05:53] <chmac> Turns out that less wasn't finding the text home, probably because it was too far along the line or something
[05:53] <chmac> As in, the line was too long
[06:04] <hads> If there's something you don't want in your results just pipe a grep -v on the end
[06:09] <jameswf-home> maybe perl or sed would be better suited
[06:09] <jameswf-home> grep ia a bit of a brut
[06:18] <A|ysum> hello - how can I install libcurlftpfs on the old Ubuntu dapper ?
[06:19] <A|ysum> i cannot compile it from source - too many errors
[06:19] <hads> jameswf-home: grep is brut? grep is fantastic.
[06:19] <jameswf-home> hads for globs and rough searches yes for specifics not so much
[06:20] <hads> Then there's egrep :)
[06:22] <jameswf-home> for i in `find /path/to/dir/` do sed -n '/\/home\/sites\//p'; done
[06:23] <jameswf-home> doj
[06:23] <jameswf-home> for i in `find /path/to/dir/` do sed -n '/\/home\/sites\//p' $i; done
[06:23] <hads> Umm, ok.
[06:26] <chmac> Thanks for all the grep help guys, got it sorted now and the site is running well :)
[06:26] <chmac> Now to buy some furniture...
[06:55] <A|ysum> hello can I pls get some help installed libcurl3-dev ? so many dependencies....
[07:24] <A|ysum> how how do you pipe in ftp commands after ftp host command?
[07:59] <jameswf-home> nc
[08:33] <kraut> moin
[08:49] <pschulz01> Howdy.. anyone sucessfully pushing 'search-domains' to clients using DHCP?
[08:59] <osmosis> im stuck on this apache vhost issue
[09:01] <soren> osmosis: "this apache vhost issue"?
[09:07] <osmosis> soren: Im setting a ServerName for my VirtualHost, but its still using the default config rather then the new config I am specifying.
[09:11] <osmosis> here is the config I am using, http://dpaste.com/75192/
[09:12] <osmosis> the second virtual host is not working. I dont know why.
[09:12] <soren> The first one doesn't list a servername?
[09:12] <soren> Is that intentional?
[09:13] <soren> And please elaborate on "is not working".
[09:29] <pschulz01> Anyone use procmail ?
[09:31] <soren> pschulz01: Yes.
[09:31] <pschulz01> Are you using any rules on  the "Delivered-To:"?
[09:31] <pschulz01> header?
[09:32] <pschulz01> soren: I'm trying to use the procmail to keep a globol archive of incoming email...
[09:33] <soren> pschulz01: No, no rules on Delivered-To
[09:33] <pschulz01> soren: .. but .. it looks as though some email doesn't get a 'Delivered-To:' header (or the match doesn't work) which causes procmail to barf and bounce the message back through postfix.
[09:34] <soren> Er... What?
[09:34] <soren> If a procmail rule doesn't match, it just falls through and lands in the default mailbox.
[09:34] <pschulz01> Ok.. (email)->Postfix->Procmail
[09:35] <pschulz01> Procmail fails with "procmail: Exceeded LINEBUF".. which is a a little stupid, as I've made the LINEBUF huge (32k).
[09:35] <soren> How have you configured postfix to pass things to procmail?
[09:35] <pschulz01> .. and the log indicates that procmail: No match on "^Delivered-To: \/[^@]+"
[09:36] <soren> What do your delivered-to headers look like?
[09:38] <pschulz01> soren: Default ubuntu-server install (I believe).. but..  (looking)
[09:38] <pschulz01> soren: Ahha! in main.cf
[09:38] <pschulz01> main.cf:mailbox_command = /usr/bin/procmail
[09:41] <AnRkey> i have an Ubuntu 7.10 postfix and procmail box with dovecot for imapd and pop3, what files do i need to move to my new 8.04 server to migrate this setup?
[09:51] <AnRkey> bbl
[09:52] <soren> pschulz: Ok. Yeah, that looks just right. Do you have an example e-mail that fails like that? (with the "Exceeded LINEBUF" thing)
[10:08] <broonie> Is there any chance someone could have a look at the NIS package?
[10:57] <Bambi_BOFH> broonie, what sort of look?
[10:58] <broonie> Bambi_BOFH: The major thing would be to disable Network Manager support by default (since it causes no end of bug reports)
[10:59] <broonie> Bambi_BOFH: But there's also some long-standing issues with the ordering of NIS in the init process which appear to be getting no love at all.
[11:00] <Bambi_BOFH> not sure where to follow that up. -dev, i guess, but you could try -motu or -bugs
[11:01] <broonie> It's in main so motu isn't appropriate.
[11:01] <broonie> -devel suggested I try here for someone who might want to look at it.
[11:02] <Bambi_BOFH> hm.
[11:02]  * broonie is the Debian manintainer; I could provide patches but my experience hasn't been all that positive.
[11:03] <Bambi_BOFH> nod
[11:03]  * Bambi_BOFH wont be the one to step forward, not a lot of time atm :(
[11:25]  * NCommander is away: This creature sleeps beyond the reaches of time itself
[11:25] <ghaleb__> hello, how can I use ls command to show subdirectories
[11:41] <thefish> ghaleb__: -R is recursive, it will show all subdirs and contents recursively
[11:43] <ghaleb__> thefish: thank you, I want it to tape data using cpio
[11:44] <thefish> ghaleb__: you want to back up just the listing?
[11:46] <ghaleb__> no
[11:46] <ghaleb__> ls | cpio -oVc > /dev/rmt/0
[11:46] <ghaleb__> ls -R would backup subdirectories as well
[12:52] <KoS> hi all
[12:53] <KoS> i have one problem with ubuntu server installation
[12:53] <KoS> how can i get root pwd ?
[12:53] <Bambi_BOFH> !root
[12:56] <KoS> when I installing it dont ask to type root pwd.
[12:57] <Bambi_BOFH> no, it wouldnt have
[12:57] <Bambi_BOFH> look at the wiki page meantioned
[12:57] <KoS> ok thanks
[15:02] <confidential> I think something went wrong during the install when it was configuring apt, where can I find out what sources.list is supposed to contain in ubuntu-server 7.10 ?
[15:13]  * delcoyote hi
[15:23] <soren> broonie: I remember some talk between you and someone else a while back... Was it Michael Vogt perhaps?
[15:28] <broonie> Not me.
[15:29] <broonie> I did ask about this on -devel a couple of days ago. I also spoke via msg with someone who was patching a FTBFS but they didn't seem terribly confident in doing anything substantial.
[15:32] <broonie> Oh, I probably spoke to mvo about n-m causing constant problems.
[15:33] <soren> broonie: Yeah, that sounds about right.
[15:33] <soren> broonie: What are the nature of your problems, exactly? In Ubuntu Server we don't have network-manager by default.
[15:34] <broonie> As far as network manager goes it appears that a reasonable number of people enable NIS on their clients then have n-m break it.
[15:35] <broonie> So it'd be nice if someone would disable n-m support by default for release at least.
[15:36] <broonie> That'd stop the overwhelming majority of bug reports about NIS in Ubuntu.
[15:37] <broonie> I don't think it's desparately related to server itself either but -devel bounced me here for someone who might care (which I suppose makes some sense since servers tend to have clients).
[15:39] <ivoks> woho!
[15:40] <ivoks> dell reseler in croatia now offers ubuntu :)
[15:50] <soren> broonie: Disable n-m- for release? I don't think I follow?
[15:50] <soren> broonie: You want us to have the development releases have n-m enabled, but then disable it at release time (kind of like apport)?
[15:50] <broonie> soren: Or just disable it all the time, I don't care either way.
[15:51] <broonie> (the n-m integration in NIS, this is)
[15:52] <soren> broonie: I'm not sure what n-m integration NIS does?
[15:52] <broonie> ypbind can use n-m to detect when there is a worknig network connection.
[15:52] <soren> I've not used NIS in a non-server environment, so no network-manager.
[15:52] <soren> broonie: Ok. Sounds reasonable enough?
[15:52] <broonie> Unfortunately on some systems n-m constantly lies about this.
[15:53] <soren> ivoks: Neat! Maybe Denmark will follow soon!
[15:53] <ivoks> soren: heh, laptop for 300euros :)
[15:54] <ivoks> otherwise they are starting at 500€ here
[15:55] <soren> broonie: Ok.. I realise this is a problem for NIS, but it doesn't sound like anything you can solve from the NIS side.
[15:55] <soren> broonie: Let's try it this way: What do you propose as an alternative?
[15:58] <broonie> ypbind can be told to ignore n-m by default
[15:58] <broonie> Which sidesteps the issue.
[15:59] <soren> But in case n-m isn't lying and there's actually no network available, I don't suppose that's handled particularly gracefully?
[15:59] <broonie> This is what Debian shipped for etch and will ship for lenny.
[16:00] <broonie> It handles it as gracefully as it ever did which isn't too bad - there's a long timeout at startup but other than that it works.
[16:00] <soren> broonie: I suppose there's a bug open about this?
[16:01] <broonie> I strongly expect a very small proprotion of users actually make use of NIS in configurations where this is a likely occurrence and their system is still usable without it.
[16:01] <broonie> There's ones like #224828
[16:02] <soren> bug 224828
[16:05] <soren> broonie: From the looks of that bug, it seems that nis only checks for network during startup. Is that accurate?
[16:06] <soren> broonie: I would have expected it to spring to life when the network connection came around.
[16:06] <broonie> soren: no, that's not the case at all.
[16:06] <soren> Well, not so much "checks for network", but "checks whether HAL thinks we have network".
[16:06] <soren> broonie: Ok.
[16:07] <broonie> It checks whenever n-m claims the status changes.
[16:07] <soren> Ok.
[16:07]  * soren rereads
[16:09] <soren> broonie: Ok, so the problem is that n-m says "connected" too soon, so nis checks and finds that there's actually no network, and since n-m's state never changes, nis doesn't recheck?
[16:09] <soren> If that's not it either, can you please explain it again?
[16:09] <broonie> When n-m goes to connected state it may not be connected.
[16:10] <broonie> The n-m status may change again but since it would be leaving the connected state nis will beleive that the network has been disconnected
[16:11] <broonie> So when connected is reported it is too soon and since it only generates events on status changes there's no kick to re-check.
[16:11] <broonie> NIS should be coping better with this but disabling the network manager integration is a simple, non-invasive workaround.
[16:12] <soren> broonie: Have you talked to network-manager people about this?
[16:12] <broonie> soren: They don't appear to care.
[16:12] <soren> broonie: No response at all?
[16:12] <broonie> Basically, yes.
[16:13] <broonie> AFAICT NIS is only marginally supported in Ubuntu anyway.
[16:14] <soren> broonie: Have you checked n-m 0.7's behaviour?
[16:15] <soren> broonie: I just checked, and nm-tool doesn't seem to say "connected" until I'm actually connected.
[16:15] <broonie> It's always worked on a lot of systems.
[16:15] <soren> ...
[16:15] <broonie> Appears to be related to the NIC driver.
[16:15] <broonie> Or possibly the local network.
[16:16] <soren> I wonder... Why does nis go and recheck whether the network is there, if n-m says it's there?
[16:16] <broonie> It doesn't.
[16:17] <broonie> What it does is try to use the IP address of the interface which falls over because none has been assigned.
[16:17] <soren> Ok, is there a bug that actually has all this information in it, so that I don't have to sit here and guess all the right questions to ask?
[16:18] <broonie> Probably not a single one.
[16:18] <broonie> They mostly get assigned over to n-m and then triaged/closed away.
[16:19] <soren> Example?
[16:19] <soren> Also, have you checked it with n-m 0.7?
[16:19] <broonie> Sorry, I don't archive the relevant info.
[16:19] <broonie> No, I've never been able to reproduce this on my own systems.
[16:20] <soren> You said something in the beginning about patches..
[16:21] <broonie> Someone (not me) patched a FTBFS in nis in ubuntu recently.
[16:21] <broonie> They spoke to me about it for some reason.
[16:22] <soren> 10:02:43  * broonie is the Debian manintainer; I could provide patches but my experience hasn't been all that positive.
[16:25] <broonie> Oh, that. Essentially that's a case of figuring out how to put a patch into lp and get it applied; most of the problems are process/lp usage ones.
[16:28] <soren> So you have patches? Or not?
[16:29] <broonie> It's a one line patch, I can generate it trivally when I get home.
[16:29] <broonie> Or now, for that matter.
[16:30] <soren> does it just pass --no-dbus to nis?
[16:30] <soren> Or does it fix network-manager?
[16:30] <broonie> The former.
[16:30] <soren> It's really not very appealing. Asking network-manager about the status of the network seems like a reasonable thing to do.
[16:31] <broonie> I agree; ideally there would be a better fix.
[16:35] <soren> I would very much like to know if the problem persists in Intrepid (i.e. with n-m 0.7)
[17:05] <broonie> soren: AFAICT the best thing for that is to wait for release and see if anyone complains.
[17:06] <broonie> nis users tend not to be early adopters :/
[17:12] <jmedina> I dont know any admin/friend who uses nis, instead all they use ldap/nss
[18:16] <siretart> broonie: perhaps there is some way to disable nm integration at runtime rather than compile time?
[18:26] <broonie> siretart: That's what I'm suggesting people enable.
[18:27] <broonie> siretart: There's a default configuration file in the package, I'm suggesting changing that default to disable the n-m integration.
[18:32] <frith> an ubuntu directory server is a great idea
[20:29] <acemo> how can i make a sort of symlink to /home/torrentflux/acemo in a proftpd folder so that the symlinked folder is actually accessable through ftp?
[22:14] <siretart> broonie: that sounds reasonable to me. do you happen to know if there is already a bug about this that I could reference in the changelog?
[23:56] <pschulz01> Morning..
[23:56] <pschulz01> soren: ping?
[23:57] <pschulz01> I'm still having procmail issues, which I'd like to get to the bottom of..
[23:58] <pschulz01> but.. I would rather write up the problem I am having somewhere as a 'blueprint' or 'answer' as I'm sure that there is a better way of doing what I am trying to do.