[00:00] <Patrickdk> there is none, that I know of
[00:00] <Patrickdk> this is what pre-up, up, post-up are useful for
[00:25] <twb> kaushal: one moment
[00:26] <twb> kaushal: here is mine http://paste.debian.net/139315/
[00:26] <kaushal> twb: ok
[00:27] <kaushal> twb: if i just set gateway someip in /etc/network/interfaces ?
[00:27] <kaushal> is that sufficient ?
[00:28] <twb> kaushal: that depends what you want.
[00:28] <kaushal> I mean make it permanent
[00:28] <kaushal> am i doing it wrong ?
[00:28] <twb> Probably.
[00:29] <kaushal> twb: Probably ?
[00:29] <kaushal> not sure i understand
[00:29] <twb> kaushal: I mean: I think you're an idiot, and you want someone to do your homework for you.
[00:31] <kaushal> twb: mind your words ?
[00:31] <kaushal> its pretty bad
[00:34] <twb> I have a bunch of LXC jails, which drop SYS_ADMIN, and thus cannot lower niceness levels.  Cron is hard-coded to do so, so I get messages like this: bleed CRON: pam_limits(cron:session): Could not set limit for 'nice' to soft=20, hard=0: Operation not permitted; uid=0,euid=0
[00:35] <twb> Now, this I can understand, but recently I created a new jail, and for some reason it is saying the same with uid=20 instead of uid=0.
[00:35] <twb> How can this be?  That uid is not in use, and UIDs below 100 are reserved for distro-wide hard-coded use.
[01:31] <amero> why in oneiric, nic aliasing only show the main ip of the nic?
[01:31] <twb> amero: pastebin the output of "ip a"
[01:32] <patdk-lap> who exactly are you *showing* this info?
[01:33] <twb> s/who/how/ ?
[01:33] <amero> i could see all the additional ips in 'ip a' but not in ifconfig -a
[01:33] <patdk-lap> heh, it worked both ways :)
[01:33] <twb> amero: that's because ifconfig is crap
[01:33] <patdk-lap> ipconfig has been depressiated a long time ago
[01:33] <twb> amero: it predates the idea that interfaces can have >1 address
[01:33] <patdk-lap> how did you make the alias?
[01:33] <twb> patdk-lap: er, *ifconfig; *deprecated.  Go drink some coffee :-)
[01:34] <patdk-lap> heh
[01:34] <twb> And "depreciated" is something that happens to money, not code
[01:34] <patdk-lap> well, it happens to code :) we just call it bitrot :)
[01:38] <patdk-lap> looks like ubuntu opted to not fix it, but passed it upstream to debian, and waiting for them to fix it
[01:39] <patdk-lap> bug #876829
[01:49] <twb> patdk-lap: wow, someone actually made a nontrivial change to ifupdown?
[01:49] <twb> Like, made it use ip(8) ?
[01:50]  * stgraber notices a familiar bug # :)
[01:50] <stgraber> so yeah, regarding that bug, I'm not familiar enough with ifupdown's code to fix it, especially as the bug is in Debian's 0.7 branch too
[01:51] <twb> ifupdown is fucking awful code
[01:51] <stgraber> the aliasing actually works (as in, the IP gets assigned to the interface), the problem is the missing label
[01:51] <stgraber> so hopefully now that the source of the problem is known, it should be quite easy for the author to fix it. Once it's fixed upstream, I'll try to isolate the fix and push it to Oneiric as an SRU.
[01:52] <patdk-lap> stgraber, should be a simple fix :)
[01:53] <patdk-lap> figured someone would have patched it by now though
[01:53] <patdk-lap> maybe if I have a moment this week I'll make one
[01:53] <patdk-lap> but that big isn't big for me, as I don't use oneiric
[01:57] <patdk-lap> hmm, 7hours before I can verify this patch works :(
[02:30] <stgraber> patdk-lap: I started looking at the diff last week, though the 0.6 -> 0.7 diff is pretty big so figuring out exactly what broke wasn't too trivial. Though now that I know it's definitely label related, pin pointing the change and reverting/fixing it should be a lot easier.
[02:30] <stgraber> I may try to do that next week if I'm not too busy with pre-UDS preparation
[02:40] <patdk-lap> stgraber, can I see the diff?
[02:40] <patdk-lap> it really should be a very simple patch, should be pretty ovious
[02:46] <stgraber> patdk-lap: http://pastebin.com/p2Eht102 is the diff from 0.6.10ubuntu5 to 0.7~alpha5.1ubuntu1 (so basically from the latest 0.6 release to the earliest 0.7 release)
[02:48] <stgraber> patdk-lap: http://pastebin.com/d40HanPQ when restricting to ifupdown.nw
[02:52] <patdk-lap> there you go :)
[02:52] <patdk-lap> that paste pastebin
[02:52] <patdk-lap> line 940
[02:52] <patdk-lap> ip link set dev %iface% up label %iface%
[02:52] <patdk-lap> should do it
[02:53] <patdk-lap> maybe for 985/986 also? or is that part of a comment?
[02:53] <twb> I thought ifupdown 0.7~a was still classed as "here be dragons"
[02:56] <patdk-lap> actially, I wonder if that would work
[02:57] <patdk-lap> wonder what iface is set too there
[02:58] <patdk-lap> I'll have to install an oneiric machine and play
[02:59] <stgraber> twb: yeah, my understanding is that we got it in Oneiric to have it well tested for the LTS and because we wanted IPv6 support
[02:59] <twb> I guess
[02:59] <patdk-lap> ipv6 support for what?
[03:00] <stgraber> patdk-lap: IIRC ifupdown 0.7 brings a few more options to "inet6 static" and the one currently in Debian brings dhcpv6 support too
[03:01] <patdk-lap> hmm, I haven't done dhcpv6 yet, just stick with static and auto v6 so far
[03:01] <twb> The main use is resolv.conf
[03:11] <bao_> how to delete mysqmail
[03:33] <Takyoji> Anyone know of a way to have scripted input for debconf when installing packages that require human input? (such as for the case of deploying the installation of something across a network of systems, silently; without having to put in data manually at each system.
[03:33] <twb> Takyoji: DEBIAN_INTERFACE=noninteractive
[03:34] <twb> But you probably want something like puppet
[03:34] <Takyoji> because I'm trying to deploy a LDAP/NFS setup on fresh installations and so on
[03:44] <Takyoji> ahh; apparently you can even just set it (as a variable) before any apt-get command rather than having to permanently set it, as I thought was the only option
[03:44] <twb> Also preseeding
[03:51] <Takyoji> preseeding being?
[03:51] <twb> Takyoji: answering questions in advance
[03:52] <twb> Like debconf-set-selections <<<'ldap-auth-config ldap-auth-config/override boolean false'
[03:52] <Takyoji> I'm most likely to be going without puppet; I pretty much already have a bash script written
[03:53] <Takyoji> with non-interactive; does that mean you can just specify the values using stdin, or?
[03:54] <twb> Takyoji: no, it means you can't specify anything
[03:54] <Takyoji> So it just ignores the configuration part entirely?
[03:54] <twb> It doesn't prompt you at all
[03:54] <Zanzacar> Hi I am trying to mount a western digital usb hdd. but I am runing into errors
[03:54] <Zanzacar> mount -t vfat /dev/sdc /media/external -o uid-1000,gid=1000,utf8,dmask=027,fmask=137
[03:54] <Zanzacar> that is the command I used.
[03:54] <Takyoji> so it assumes some default value instead then, or?
[03:55] <twb> Takyoji: it'll take the defaults (or preseed), yes
[03:55] <twb> Takyoji: occasionally it'll email you (e.g. debmirror) if there is no useful default
[03:55] <Zanzacar> it said wrong fs type, bad option, bad superblock on /dev/sdc
[03:55] <Takyoji> /dev/sdc is a whole harddrive, not a specific partition
[03:55] <twb> Zanzacar: try sdc1
[03:55] <Zanzacar> tried that got the same error
[03:56] <twb> Zanzacar: pastebin "cat /proc/partitions"; pastebin "file -s /dev/sdc*"
[03:56] <Zanzacar> I was following the ubuntu help guide on it till I got the errors then came here.
[03:56] <twb> Make that "sudo file -s /dev/sdc*"
[03:57] <Zanzacar> http://paste.pocoo.org/show/497153/
[03:58] <Zanzacar> http://paste.pocoo.org/show/497154/
[03:58] <twb> Blergh
[03:59] <twb> Well, /dev/sdc1 is definitely right, so something else is wrong -- like maybe you weren't running mount as root?
[03:59] <twb> Surprised they're shipping with FAT though, you'd think NTFS would be saner...
[04:00] <Zanzacar> I had a typo, uid-1000 needed to be uid=1000
[04:00] <Zanzacar> sorry about that guys
[04:02] <twb> Zanzacar: oh, sorry I didn't spot that
[04:05] <Zanzacar> no I just need to figure out how to copy all .mp4 movies over to the hdd haha.
[04:05] <twb> upsmon to upsd, that's client pull oriented, right?
[04:14] <Zanzacar> yay I figured it all out, now I can copy 125gigs in less then 5 days.
[04:14] <Zanzacar> usb vs. lan...... lan wins
[04:14] <Zanzacar> whoops
[04:14] <Zanzacar> usb vs ftp.... usb wins
[04:14] <Zanzacar> haha
[04:18] <nathwill> zanzacar: sneaker-net always wins :)
[04:18] <twb> station-wagon full of tapes
[04:39] <Zanzacar> what is flush-8 command? I am copying a bunch of movies from my server HDD to a fat32 HDD
[04:39] <Zanzacar> I am looking at my IO read writes and its says its reading and writing a lot
[04:40] <Zanzacar> man pages says something about email, smtp but I didnt think I had a mail service on this server
[04:40] <twb> Zanzacar: it's probably a kernel thread responsible for getting the data onto the actual platters of the disk
[04:40] <twb> Zanzacar: it's 8 because you have 8 CPUs
[04:41] <twb> Zanzacar: the anpage
[04:41] <twb> Zanzacar: the manpage is for postfix; unrelated
[04:41] <Zanzacar> twb: thanks I was wondering why it didnt make any since.
[04:41] <twb> If you're in top and you hit "c", or in ps auxf output, kernel threads have []'s around them
[04:41] <Zanzacar> th0mz: thanks for the info
[04:42] <Zanzacar> twb: O thats pretty cool, ya it says its a kernel thread
[04:42] <twb> Yeah, well, they piss me off
[04:42] <twb> ps auxf | grep -v ]$
[04:43] <Zanzacar> the kernel threads piss you off?
[04:45] <Zanzacar> o thats really cool its like a tree of where the commands are coming from basically?
[04:51] <twb> They piss me off when they fill up the first three pages of top because I have 8 cores instead of 2
[04:55] <Zanzacar> I am still learning all this, I only started using linux about 2 months ago maybe.
[04:55] <Zanzacar> had a server up after about 1 month.
[04:55] <Zanzacar> every problem I run into I try to figure it out research it, etc.
[04:56] <Zanzacar> I did a bunch of bash stuff but got really frustrated with the syntax so I learned python instead.
[04:56] <Zanzacar> All I know is now that I know linux I never want to go back to pc
[04:58] <twb> I don't really care about your personal epiphanies.
[05:00] <Zanzacar> twb: ok
[05:06] <Takyoji> How do you specify a preseed file when installing packages with apt-get? Or do I just append it to some system-wide file, or?
[05:11] <twb> Takyoji: you pass it to debconf-set-selections before installing the package in question
[05:12] <Takyoji> ahh, I think I understand what you were specifying earlier.
[05:12] <twb> Quite
[05:14] <Takyoji> and you generate the preseed config how? Do the installation as normal and get the added info to debconf-get-selections, or?
[05:25] <twb> Takyoji: using your brain
[05:25] <twb> debconf-get-selections helps
[05:31] <josePhoenix> So, I just did do-release-upgrade... it's been stuck on installing a new version of /etc/mysql/debian-start for a few minutes.
[05:31] <josePhoenix> How long until I should be worried?
[05:34] <josePhoenix> Well.. I'm already worried. How long until I should ctrl-c? Or what should I do?
[05:35] <RoyK> josePhoenix: you may want to try to strace it
[05:36] <josePhoenix> I don't know which process invoked by do-release-upgrade is actually running
[05:36] <josePhoenix> ah, pstree..
[05:37] <twb> josePhoenix: if you interrupt d-r-u it'll explode
[05:37] <twb> josePhoenix: it really doesn't cope well with that
[05:37] <josePhoenix> Okay, noted
[05:37] <twb> at least, that's my experience -- I get told off for bad-mouthing d-r-u all the time
[05:38] <josePhoenix> can I kill the hung process (looks like restarting mysql) and will d-r-u move on
[05:38] <josePhoenix> ?
[05:39] <twb> Dunno
[05:39] <josePhoenix> So... what's the least-bad way out of here?
[05:40] <twb> Probably kill it off as cleanly as possible, then pick up the pieces
[05:40] <twb> I would try killing off just the mysql restart at first
[05:40] <josePhoenix> Cool, that seems to have gotten it moving again
[05:41] <josePhoenix> I guess I have to figure out why mysql won't start some other time -.-
[05:41] <twb> Cos it's crap
[05:41] <twb> postgres ftw
[05:41] <josePhoenix> Well yeah
[05:41] <josePhoenix> have to support both though
[05:41] <josePhoenix> not my choice
[05:41] <twb> Stupid users :-/
[05:42] <josePhoenix> you're telling me
[05:42] <twb> Just tell them it's a security risk
[05:53] <josePhoenix> and they'll complain about not being able to use wordpress
[05:53] <josePhoenix> which is also a security risk >.>
[06:00] <twb> hear, hear
[06:09] <josePhoenix> What the hell
[06:09] <josePhoenix> Have a script that spawns some processes on boot (using "@reboot" in a user crontab)
[06:09] <josePhoenix> 60 sec after they're spawned, they get SIGKILL and exit
[06:10] <josePhoenix> Is there some kind of throttling going on?
[06:12] <josePhoenix> no, that's not it.. those lines don't show up until later
[06:31] <twb> josePhoenix: what's the proc named?
[06:31] <twb> A thing I've seen sometimes is the init script does like "killall squid" and kills both the squid process and itself, because it's called /etc/init.d/squid
[06:31] <josePhoenix> I think I figured it out.. apparently php5-fpm doesn't clean up its own pid file, and my startup script isn't handling that nicely
[06:58] <josePhoenix> Hm. How can I find per-user crontabs? I need to update this in a few users' crontabs at once
[06:58] <twb> josePhoenix: /var/spool/cron/ somewhere
[06:59] <twb> Not sure if you should be blatting them directly, though
[06:59] <josePhoenix> Isn't that warning just because cron needs a sighup to know to rescan the crontabs?
[06:59] <josePhoenix> I'm going to be rebooting anyway
[06:59] <twb> something like that, yeah
[07:02] <josePhoenix> might be able to fix this without touching the crontabs though... so that'd be nice
[07:45] <jamespage_> morning all
[07:48] <RoyK> morning
[08:08] <elz89> morning all, what would be the best way of making all of my ubuntu client single sign on against my fresh install of oneiric server?
[08:29] <TeTeT> elz89: probably via LDAP authentication, https://help.ubuntu.com/community/OpenLDAPServer and https://help.ubuntu.com/community/LDAPClientAuthentication
[08:29] <elz89> TeTeT: thank you
[09:11] <lynxman> morning o/
[09:33] <koolhead17> hi all
[09:57] <uksysadmin> hello koolhead17
[10:02] <koolhead17> hey uksysadmin !!! how are you sir? :)
[10:02] <koolhead17> lynxman: hola
[10:02] <koolhead17> hey Daviey
[10:04] <uksysadmin> I'm good thanks. and you?
[10:04] <koolhead17> uksysadmin: am awesome!! so how is hacking coming along?
[10:08] <trapmax> logcheck gives "subprocess installed post-installation script returned error exit status 2" on 10.04. after failed upgrade aptitude shows partially installed. how do i fix this?
[10:08] <trapmax> *partially configured
[10:12] <uksysadmin> I like awesome. I'm getting there. I know the #juju channel will like me saying this - but that's equally as awesome (and all I've done is type juju bootstrap and juju deploy a mysql server!) lol
[10:13] <koolhead17> uksysadmin: yes juju rocks!!  you can try charm for CMS like drupal or wordpress as well :)
[10:15] <uksysadmin> indeed - I'm following kim0 tutorial and its working
[10:16] <uksysadmin> just limited in hardware at the mo which is being alleviated this week and over the next few weeks
[10:16] <koolhead17> uksysadmin: awesome!! :)
[10:17] <uksysadmin> indeed
[10:17] <uksysadmin> ubuntu rocks [just trying to get it back on track for this room] ;-)
[10:17] <koolhead17> uksysadmin: i need to install oneric as well on my laptop to use LXC and stop using AWS for deplyment test
[10:18] <uksysadmin> (that's me standing on the stage and saying "Anyone here use Ubuntu?" followed by cheers... ;-)
[10:18]  * koolhead17 thinks most folks around here are still in weekend hangover!1 :P
[10:19] <uksysadmin> yeah I need to do that - but I'm primarily focussed on OpenStack so providing I can get that up and running then do juju stuff to install to it - that's enough for a couple of demos to the engineers and execs here to show the power of it all
[10:19] <koolhead17> hmm.
[10:20] <uksysadmin> btw - I wouldn't upgrade to oineric on a desktop... its great for the OpenStack stuff, but Unity sucks (sorry for those who like it and have allegiance to it - its not for me) plus its buggy as hell.
[10:21]  * uksysadmin puts flame-retardant underwear on
[10:21] <koolhead17> uksysadmin: we are server folks here, and you always have choice :D
[10:22] <koolhead17> i will install xubuntu/lubuntu on my netbook
[10:23] <uksysadmin> I've had about 10,000 hits over a few days to my blog since putting up a fix that is caused by an upgrade to 11.04 to 11.10 in some circumstances.
[10:23] <uksysadmin> that's not the desktop distro that has previously "just worked" out of the box that people came to love
[10:23] <uksysadmin> though us Linux folk are fickle I guess - and that comes through choice.
[10:24] <koolhead17> :P
[10:24] <koolhead17> uksysadmin: am using LTS on my desktop
[10:24] <koolhead17> currently
[10:24] <uksysadmin> keep it that way ;-)
[10:26] <w00> uksysadmin, link to the post?:)
[10:26] <koolhead17> w00: google for uksysadmin i think :D
[10:26] <w00> oh right
[10:26] <uksysadmin> http://uksysadmin.wordpress.com/2011/10/14/upgrade-to-ubuntu-11-10-problem-waiting-for-network-configuration-then-black-screen-solution/
[10:27] <w00> oh, i read that already cause i got bitten by it :)
[10:27]  * w00 thanks uksysadmin
[10:27] <uksysadmin> w00, its poor for something that IMHO shouldn't have gotten past testing...
[10:28] <uksysadmin> BUT - how many people upgraded 11.04 (or maybe earlier) to one of the betas? I didn't - I tested the betas using a fresh install
[10:29] <w00> uksysadmin, tell me about it, i'm kind of dissapointed as a user who just started using Ubuntu few months ago
[10:29] <uksysadmin> I can only guess that it didn't handle /var/run (a mounted fs) in 11.04 and earlier which has files that possibly are in use because you're doing a live upgrade... it doesn't/can't remove it whilst its in use... so when you reboot /var/run doesn't have the needed structure that /run has which is a new tmpfs area
[10:29] <trapmax> dpkg --configure logcheck also gives same error
[10:30] <uksysadmin> and most/all packages refer to /var/run so 11.10 symlinks to /run... on a clean install this is the case, on an upgrade you're left with a /run that isn't linked to /var/run
[10:31] <uksysadmin> its made worse by the fact it means networking fails - so if its your one and only desktop - you can't even google for a solution
[10:32] <uksysadmin> dhclient eth0 (or whatever) will sort that out on a dhcp network - but desktop users who have ditched Windows for a cleaner experience mightn't know the magic keypresses and commands to get this going again.
[10:32]  * uksysadmin stops ranting now
[10:33] <koolhead17> :P
[10:43] <patdk-lap> heh, I have had issues with tmpfs /var/run before
[10:43] <patdk-lap> mainly programs making subdirs in it, and using them
[10:44] <patdk-lap> that means you either can't have them use /var/run, or you have to recreate that dir each boot
[11:28] <trapmax> http://pastebin.com/2QjTHNZS
[11:30] <Daviey> Anyone want to tackle a bitesize bug?
[11:35] <koolhead17> Daviey: ?
[11:36] <Daviey> koolhead17: bug 875262 is a good bitesize bug to tackle
[11:37] <koolhead17> Daviey: waoo. i saw/faced it last night itself
[11:37] <Daviey> koolhead17: Great!  It should be pretty easy to fix... :)
[11:37] <koolhead17> during installation of owncloud2
[11:38] <Daviey> koolhead17: note, you can't use UDD for it. :/
[11:38] <koolhead17> Daviey: show me the path o master!1 :)
[11:38] <Daviey> koolhead17: $ pull-lp-source php5
[11:38] <Daviey> :)
[11:38] <koolhead17> k
[11:41] <koolhead17> Daviey: the suggestion was to remove /etc/php5/conf.d/sqlite3.ini as the same bug was reported for 8.04 as well
[11:41] <Daviey> koolhead17: *no*, /etc/php5/conf.d/sqlite.ini
[11:42] <koolhead17> Daviey: oops
[11:42] <koolhead17> yes
[11:47] <koolhead17> Daviey: am on it then. :)
[11:49] <koolhead17> i have it downloaded now.
[11:52] <koolhead17> Daviey: https://bugs.launchpad.net/ubuntu/+source/php-sqlite3/+bug/281714 are they both not linked ?
[11:52] <Daviey> Hah, no - they are opposite. :)
[11:53] <Daviey> koolhead17: That was a bug on hardy.
[11:54] <koolhead17> Daviey: yeah. so what about apps which uses sqlite not sqlite3 ?
[11:54] <koolhead17> even now?
[11:54]  * koolhead17 is curious
[11:55] <Daviey> koolhead17: SoL
[11:55] <koolhead17> Daviey: SoL ?
[11:56] <koolhead17> Daviey: sorry am finding too n00b with all the terms!! :D
[11:59] <Ursinha> good morning
[11:59] <Daviey> koolhead17: sorry, Sad Out of Luck
[12:00]  * koolhead17 writes the meaning in his slag list :D
[12:00] <Daviey> koolhead17: as per, bug 747325
[12:11] <lynxman> Daviey: hey small question for you, what's our HA ipvs of choice right now? :)
[12:27] <Daviey> lynxman: simple answer?
[12:27] <lynxman> Daviey: yeah
[12:28] <Daviey> lynxman: We have none recommended.. If you want kernel level, ip_vs_dh
[12:29] <Daviey> I don't believe it is in Lucid.
[12:30] <lynxman> Daviey: was looking for something with a bit more punch, virtual IP and active monitoring of DRBD
[12:31] <soren> Daviey: Not in Lucid? Really? It was there in Hardy.
[12:32] <soren> Daviey: It's there in Lucid, too.
[12:38] <Daviey> soren: Oh, i didn't realise it was a module in hardy or lucid.
[12:39] <Daviey> ISTR having to do something funky when using it with Lucid, but i could be wrong.
[12:39] <Daviey> lynxman: Really need more detail :)
[12:41] <Daviey> lynxman: looked at pacemanker or drbd8 and OCFS2
[12:41] <lynxman> Daviey: pacemaker, that's it :)
[12:41] <lynxman> Daviey: thanks!
[12:48] <koolhead17> Daviey: so is sqlite will be removed from precise :D
[12:52] <soren> Daviey: Every single Ubuntu release ever has had it as a module.
[12:52] <soren> Daviey: All the way back to Warty.
[12:55] <amero> since ifconfig cant recognize alias nic in oneiric, what's the alt way to disable the alias nic?
[13:02] <patdk-wk> damn, ifdown on a alias, downs the whole interface, but doesn't mark the raw interface as down
[13:02] <patdk-wk> that is annoying
[13:06] <amero> holy.. i was gonna try it with ifdown just now. so glad that i haven't done it
[13:07]  * patdk-wk tests if this works for alias
[13:07] <patdk-wk> but hmm, making ifdown not break everything will be harder I think
[13:12] <patdk-wk> yep, quick easy fix to make ifup work :)
[13:14] <patdk-wk> in ifupdown-0.7~alpha5.1ubuntu5/ifupdown.nw
[13:14] <patdk-wk> line 4380, add: label %iface%, to the end
[13:15]  * koolhead17 is bit excited
[13:15] <patdk-wk> I also did it on 4494, but probably not needed, but might be, for tun interfaces
[13:15] <patdk-wk> the testbuild file needs to be edited to add the label onto those checks for all ipv4 tests though, easy enough
[13:15] <patdk-wk> see if I can locate the ifdown issue
[13:15] <patdk-wk> well, locate is esay, fixing it not so much
[13:21] <patdk-wk> hmm, ifdown isn't too hard, depending
[13:21] <patdk-wk> need to add label %iface% to the ip flush dev
[13:21] <patdk-wk> but then, need not all link set down at all, or it brings the whole interface down
[13:34] <Ursinha> Daviey, hello
[13:34] <BrixSAT> Hello
[13:34] <BrixSAT> im trying to make wpa_suplicant connect to a hidden ssid but not successfull :(
[14:14] <jdstrand> Daviey: fyi, while cyrus-imapd-2.2 is in universe, I thought I might point out http://people.canonical.com/~ubuntu-security/d2u/. there are several simple merge candidates for several releases. this could be good practice for core-dev/motu for someone here
[14:15] <jdstrand> Daviey: and hello! :)
[14:30] <MuNk``> Wondering if anyone here has had any success in getting Ubuntu to install on HP rx2620? after booting from the installation CD you enter setup and it then fails to continue to reconnize the CD anymore
[14:47] <Daviey> jdstrand: Ooo, thanks!
[14:47] <Daviey> Ursinha: heya
[14:51] <Ursinha> Daviey, hey
[14:58] <Daviey> hey Ursinha
[15:00] <Daviey> jdstrand: Am i being a banana?  There are not yet LP bugs for those CVE's are therE?
[15:01] <Daviey> (I assume you don't help a helper tool for raising bugs, where you throw in a CVE number? :)
[15:03] <jdstrand> Daviey: I don't see bugs for that, no. we also don't have a tool to add a bug to LP. I won't go into why, but suffice it to say that LP does not meet the security team's needs for CVE triage, and we work out of https://launchpad.net/ubuntu-cve-tracker (add bug references to it when they come in)
[15:04] <jdstrand> Daviey: if you open a bug, ping me and I'll add it to the open CVEs
[15:05] <Daviey> jdstrand: sad, it looked like LP did start to do some good stuff for CVE work.  I guess it didn't get finished.
[15:06] <jdstrand> Daviey: well, we need to track just way to much stuff. to get an idea, see README in uct
[15:07] <Daviey> jdstrand: It's been a few months since i tried to switch to uct, is it that time again to try again? :)
[15:07] <jdstrand> hehe
[15:11] <Daviey> jdstrand: is bug 880909 correct?  (series targets?))
[15:15] <jdstrand> Daviey: yeah. fyi, there is also CVE-2011-1926 and CVE-2011-3208
[15:15] <zul> mornin
[15:15] <Daviey> uvirtbot: Yeah, i was about to open seperate bugs for those.
[15:15] <Daviey> err, jdstrand
[15:15] <jdstrand> ack
[15:50] <hallyn> stgraber, do you see any reason why we need resolvconf in the default list of packages installed by the ubuntu lxc template?
[15:50] <hallyn> (re bug 880020)
[16:01] <stgraber> hallyn: nope, it's actually the first thing I remove in my containers
[16:02] <patdk-wk> stgraber, I updated that ifupdown bug with a mostly working fix :)
[16:05] <hallyn> stgraber, ok, i'm still wondering why that fails to install, but i'll go ahead and remove that
[16:05] <hallyn> i guess we can just remove it upstream and let it trickle down?  or do you think it's more urgent than that?
[16:06] <hallyn> it only fails when inetutils-ping is also in the packagelist, btw
[16:06] <stgraber> cool, I saw some comments in the debian bug too (for the ifupdown one)
[16:07] <patdk-wk> what is the debian one?
[16:07] <patdk-wk> see if I can steal any ideas to totally fix mine :)
[16:07] <stgraber> yeah, I'm fine with removing it upstream
[16:07] <patdk-wk> not that I care, as long as it's fixed for 12.04 :)
[16:08] <dkn> hey all, accidentally mounted my external over a mount point that already had another partition mounted to it, everyworked, except, now the original mount is stuck and i can't stop the process relying on the mount, and i can't unmount cause the process is relying on it.... ? i'm sure a reboot would fix it, but i can't really do that right now
[16:08] <stgraber> its linked from the one in launchpad, sorry I'm on my cell :)
[16:08] <patdk-wk> no problem, just didn't notice it
[16:11] <patdk-wk> heh, the debian people are misdirected on this issue :(
[16:13] <hallyn> stgraber, funky actually i think resolvconf isn't hte problem, but inetutils-ping
[16:17]  * patdk-wk wonders if his email got lost, or if debian just takes forever to update bugs
[16:18] <stgraber> It took alma
[16:18] <stgraber> doh...
[16:19] <stgraber> It took almost an hour last time I sent a mail to the bugtracker
[16:19] <patdk-wk> yuk
[16:26] <patdk-wk> heh, it updated :)
[16:30]  * koolhead17 thinks why is he so dumb
[16:36] <scalability-junk> does someone have a rutorrent installation running on ubuntu 11.10?
[16:43] <elz89> Why is it that when I 'apt-get update' on a fresh oneiric install, the last package fails with Hash Sum Mismatch?
[16:48] <rbasak> Daviey: I've commented on bug 878180, you aren't subscribed. Not sure whether to put certain comments in the bug or the merge request or what.
[16:51] <elz89> I was told to uncomment the CD line in sources.list, but this does not help the matter, and I can't figure out why I have the same CD line commented twice. I really do not understand how a fresh installation can be bug ridden. This is not the first stupid and trivial little problem I have had with this new Ubuntu Server Oneiric either?
[16:53] <Ursinha> elz89, after commenting the lines, did you run a sudo apt-get update
[16:53] <Ursinha> ?
[16:53] <elz89> Ursinha: O it was you! lmao yes, of course I did, I just followed your instructions :-)
[16:55] <Ursinha> elz89, :)
[16:55] <Ursinha> elz89, and you keep getting the same error?
[16:55] <Daviey> rbasak: wow, helluva script
[16:55] <elz89> yeah
[16:56] <rbasak> Daviey: yeah well, it beats doing it by hand again and making mistakes :)
[16:56] <rbasak> Daviey: I don't expect that the script covers all cases but I'm pretty sure it covers this one OK
[16:57] <rbasak> Daviey: (since the results are almost identical to what I got by hand last time)
[16:57] <rbasak> Daviey: (and I've checked the differences by hand)
[16:57] <elz89> Ursinha: yes
[16:58] <Ursinha> elz89, hmm.. can you paste the whole thing for me on a pastebin? pastebinit is a really useful tool to do that :)
[16:59] <Daviey> rbasak: well it looks good, but i am getting confused myself..
[16:59] <Daviey> :)
[16:59] <rbasak> Lots of things changed between those three versions :)
[17:05] <elz89> Ursinha: ok
[17:05] <koolhead11> grrrrrrrrrr
[17:07] <koolhead11> Daviey: am still at same :(
[17:20] <hallyn> stgraber, are you still on cell phone?
[17:21] <hallyn> stgraber, if not, could you review and sponsor http://people.canonical.com/~serge/lxc-resolvconf.debdiff ?
[17:44] <stgraber> hallyn: I'm back from lunch. Having a look now
[17:45] <stgraber> hallyn: oh nice, I meant to write that /run/lock fix as I had to do it for some of my containers already :)
[17:46] <stgraber> hallyn: patch looks good, uploading now
[17:47]  * koolhead11 needs some help with pbuilder
[17:48] <hallyn> stgraber, thanks
[17:48] <koolhead17> http://paste.ubuntu.com/718047/
[18:01] <hallyn> guess i'ts time to sync lxc
[19:06] <BuenGenio> guys, I will try to stay calm - it's 3am and I've been in the DC installing this new server for the last 6 hours
[19:06] <BuenGenio> I have one question
[19:06] <BuenGenio> # apache2ctl start
[19:06] <BuenGenio> Action 'start' failed.
[19:06] <BuenGenio> The Apache error log may have more information.
[19:06] <BuenGenio> what is the flippin' problem?
[19:07] <BuenGenio> the log file is empty
[19:07] <BuenGenio> and for some reason the apache binary can't tell me what the error is on the command line
[19:07] <SpamapS> BuenGenio: nothing in /var/log/apache2/error.log ?
[19:07] <BuenGenio> no
[19:07] <SpamapS> BuenGenio: did you try just 'apache2' ?
[19:08] <BuenGenio>  apache2
[19:08] <BuenGenio> apache2: bad user name ${APACHE_RUN_USER}
[19:08] <SpamapS> sudo -u www-data apache2
[19:09] <BuenGenio> sudo -u www-data apache2
[19:09] <BuenGenio> apache2: bad user name ${APACHE_RUN_USER}
[19:09] <SpamapS> oh haha right
[19:09] <SpamapS> sudo sh -c 'APACHE_RUN_USER=www-data APACHE_RUN_GROUP=www-data apache2'
[19:10] <BuenGenio> (2)No such file or directory: apache2: could not open error log file /etc/apache2/${APACHE_LOG_DIR}/error.log.
[19:10] <BuenGenio> Unable to open logs
[19:10] <BuenGenio> finally something..
[19:10] <RoyK> BuenGenio: what's the owner/mode of that dir?
[19:10] <SpamapS> BuenGenio: that may be a red herring ;)
[19:10] <SpamapS> apachectl sets that variable IIRC
[19:11] <BuenGenio> where?
[19:11] <RoyK> SpamapS: shouldn't that be apaceh2ctl?
[19:12] <SpamapS> yeah
[19:12] <BuenGenio> now the error log says
[19:12] <BuenGenio> 2)No such file or directory: apache2: could not open error log file /etc/apache2/${APACHE_LOG_DIR}/error.log.
[19:12] <BuenGenio> Unable to open logs
[19:12] <BuenGenio> root@cyclon:/etc/apache2#
[19:13] <RoyK> BuenGenio: which distro version is this?
[19:13] <BuenGenio> 11.10
[19:14] <SpamapS> Heh.. 11.10 and you're up at 3am to fix it?
[19:14] <RoyK> and, excuse me for asking, but why would you want to install something bleeding edge as oneiric on a server?
[19:14] <SpamapS> RoyK: there are plenty of reasons.. if your website is only going to live for 6 months.. ;)
[19:14] <RoyK> lol
[19:15] <BuenGenio> ffs
[19:15]  * SpamapS has flashbacks to consulting for "marketing companies"
[19:15] <BuenGenio> bleeding-edge?
[19:15] <RoyK> BuenGenio: it was released a week ago
[19:15] <RoyK> BuenGenio: use LTS for servers
[19:15] <SpamapS> BuenGenio: yes, 11.10 would be considered bleeding edge, having been released 2 weeks ago, and only having 18 months of updates
[19:15] <BuenGenio> yes, well
[19:15] <SpamapS> BuenGenio: still, it should work. :)
[19:16] <SpamapS> I switched to nginx on my test servers back in 11.04 or I'd test it out for you there... ;)
[19:16] <RoyK> SpamapS: famous last words....
[19:16] <koolhead17> and only having 18 months of updates
[19:16] <RoyK> BuenGenio: really, stick to LTS for servers unless you have a very good reason not to
[19:16] <SpamapS> BuenGenio: anyway, back to your issue....
[19:17]  * RoyK hands BuenGenio a lucid cd
[19:17] <SpamapS> BuenGenio: how about 'sh -x /etc/init.d/apache2 start'
[19:17] <SpamapS> BuenGenio: thats going to print *a lot*
[19:17] <SpamapS> BuenGenio: but the thing I'm most interested in is the part where it execs apache2
[19:18] <BuenGenio> http://pastebin.com/zsQS1wC9
[19:18]  * SpamapS loves a good debugging partner who knows how to pastebin fast. :)
[19:18] <genii-around> apache2 is not yet converted to upstart?
[19:19] <SpamapS> no
[19:19] <SpamapS> its probably never going to be
[19:20] <RoyK> SpamapS: why??
[19:20] <SpamapS> I honestly see no point in converting everything.
[19:20] <SpamapS> all the desktop stuff, the boot-supporting stuff, yes
[19:21] <SpamapS> but once you get to runlevel 2 .. just start normally
[19:21] <SpamapS> RoyK: maybe when cgroups is in effect so we don't have to deal with expect fork.
[19:21] <SpamapS> BuenGenio: ls -la /var/log/apache2
[19:22]  * RoyK sometimes wonders why lucid was moved to upstart in the first place
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data  563077 Oct 25 02:41 access.log
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data   66324 Oct 23 16:54 access.log.1
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data     346 Oct 12 11:18 access.log.2.gz
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data   28853 Oct 25 03:22 error.log
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data  350245 Oct 23 06:25 error.log.1
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data    4675 Oct 16 06:25 error.log.2.gz
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data 1161523 Oct 25 02:41 other_vhosts_access.log
[19:22] <BuenGenio> -rw-r-----  1 www-data www-data 1498700 Oct 22 12:39 other_vhosts_access.log.1
[19:22] <SpamapS> BuenGenio: ack!
[19:22] <BuenGenio> -rw-r--r--  1 www-data www-data    8939 Oct 15 14:46 other_vhosts_access.log.2.gz
[19:22] <RoyK> !pastebin
[19:22] <BuenGenio> sorry
[19:23] <SpamapS> BuenGenio: looks like there's stuff in error.log
[19:23] <BuenGenio> it's way past my bed time
[19:23] <BuenGenio> ;)
[19:23] <BuenGenio> SpamapS, I set LogLevel to debug
[19:23] <BuenGenio> that's why
[19:23] <SpamapS> RoyK: upstart is necessary and useful for many, many things
[19:23] <BuenGenio> http://pastebin.com/SX9qfYjK
[19:23] <RoyK> SpamapS: sure, but sysv scripts work too.....
[19:23] <BuenGenio> nothing unusual there
[19:24] <SpamapS> RoyK: but for just the run of the mill network services.. I don't really think its a place to focus much attention.
[19:24] <SpamapS> like, samba is integrated into auth and boot stuff.. its all intertwined, and needs the event based approach to start when needed
[19:24] <SpamapS> but apache.. apache just needs to work.
[19:25] <zul> apache and upstart? bah...
[19:25] <SpamapS> BuenGenio: do you by any chance have any *encrypted* SSL keys?
[19:26] <RoyK> BuenGenio: I'm just saying it'll be a lot easier to maintain the system if you chose lucid - it's LTS, supported until 2015...
[19:26] <BuenGenio> it just seemed natural to go for the latest version....
[19:26] <SpamapS> BuenGenio: think about how long that server is expected to live
[19:26] <RoyK> BuenGenio: I know, but on ubuntu, that's not the case for servers
[19:27] <SpamapS> BuenGenio: if its more than 18 months.. you *must* upgrade or be left without patches.
[19:27] <RoyK> BuenGenio: that's why there are two tracks, LTS and "normal"
[19:28] <RoyK> imho the "normal" track should be labelled "frequent"
[19:32] <pmatulis> BuenGenio: maybe 'sudo strace -o start.txt apache2ctl start' and pastebin start.txt
[19:32] <BuenGenio> I think it's done
[19:32] <BuenGenio> it was an SSL website
[19:32] <BuenGenio> I just toook it out ouf the config
[19:33] <BuenGenio> well, can I not upgrade from 11.10 to an LTS release
[19:33] <BuenGenio> when it's out?
[19:33] <pmatulis> BuenGenio: yes, of course
[19:34] <pmatulis> BuenGenio: which is a very reasonable approach, don't let these guys scare you  :D
[19:35] <BuenGenio> i just wish it apache told me what the problem was right away =)
[19:35] <pmatulis> BuenGenio: you had an encryted ssl private key then?
[19:36] <BuenGenio> tayes
[19:36] <BuenGenio> yes
[19:38] <alaing> I'm not sure if this is the right place but here goes. I'm using ubuntu server edition 11.04 and apache. I've just installed coldfusion which by default places its admin pages under /var/www now I've moved those out inthe the filing system for security reasons. I'm getting the basic page layout but the css/images are not working because they looking for the files under /var/www is there away...
[19:38] <alaing> ...I can get apache to redirect those requests to the new folder structure?
[19:39] <cjs226> where's the proper place to put an init script if not using upstart?  I've tried /etc/rc0.d/K01…  but it's not working
[19:41] <SpamapS> BuenGenio: actually thats most likely a bug in code that I wrote.
[19:41] <BuenGenio> well, I hope I helped you fix it ;-)
[19:41] <BuenGenio> I'm really looking foward to bed now :P
[19:42] <RoyK> BuenGenio: 12.04 will be released in half a year or so
[19:42] <SpamapS> BuenGenio: SSLPassPhraseDialog  exec:/usr/share/apache2/ask-for-passphrase
[19:42] <SpamapS> BuenGenio: don't you still need that one SSL site to be up? ;)
[19:42] <RoyK> BuenGenio: and will change the update manager config to default to upgrading only LTS
[19:43] <BuenGenio> thanks guys, wiill parse the chat logs tomorrow ;)
[19:43] <SpamapS> cjs226: /etc/init.d
[19:43] <BuenGenio> thanks
[19:43] <SpamapS> cjs226: for all that I said about not using upstart.. I do think its easier to write an upstart job for simple services than it is to write a sysvinit script ;)
[19:44] <SpamapS> BuenGenio: anyway, when you're ready to try and fix that SSl site, try commenting that line out of /etc/apache2/mods-enabled/ssl.conf ... if that fixes it.. please file a bug. :)
[19:44] <cjs226> SpamapS: easier even for shutdown?  i need to include some logic and am thinking I'd rather just use init.d.  to my question, i have the script in /etc/init.d and link it to /etc/rc0.d/K01…  is that the correct procedure?
[19:45] <SpamapS> cjs226: yes its easy for shutdown too
[19:46] <SpamapS> cjs226: no, the correct thing to do is put the script in /etc/init.d , then use 'update-rc.d' to enable /disable it.
[19:46] <cjs226> SpamapS: true. i did that and it appeared to just link it like the ole days
[19:47] <cjs226> SpamapS: but didn't run upon shutdown
[19:47] <SpamapS> cjs226: are you sure it didn't run? maybe it had an error
[19:49] <cjs226> SpamapS: i don't believe it did.  in fact i appended everything with a echo out to a log which was never updated.  ok, maybe i should start fresh.  if i wanted to use upstart, the normal logic, etc. I'd use in an init script would go in a pre-stop stanza?
[19:57] <SpamapS> cjs226: depends on a few factors
[19:58] <SpamapS> cjs226: if you need to do more than just send SIGTERM .. then yes, do that in pre-stop
[20:06] <cjs226> SpamapS: thx
[20:08] <SpamapS> cjs226: also if you are doing anything that will take longer than 10 seconds, its advisable to 'stop on starting rc RUNLEVEL=[016]' and 'start on runlevel [2345]' ..
[20:08] <SpamapS> cjs226: the shutdown only waits about 10 seconds for stuff to stop cleanly if it does 'stop on runlevel [016]'
[20:12]  * RoAkSoAx will be back in an hour
[20:41] <lauris> hi, what's the best choice of IMAP server on ubuntu?
[20:42] <lauris> in terms of security/performance
[20:42] <jMCg> lauris: Last time I managed one, it was dovecot
[20:42] <andol> lauris: Well, dovecot is at least not the wrong choice.
[20:42] <jMCg> These days I use Zimbra.
[20:43] <lauris> i tried Zimbra once, the feel was like it is very heavy & slow
[20:43]  * RoyK too
[20:43] <RoyK> that is, I use zimbra
[20:43] <RoyK> it's on the heavy side, truely, probably because of the java bits, but it works well and is easy to administer
[20:44] <lauris> how about performance?
[20:44] <RoyK> but then, you need a dedicated machine or VM with 2GB RAM or so
[20:44] <RoyK> lauris: sufficient
[20:44] <jMCg> I give mine about 3.5G, IIRC.
[20:44] <lauris> broadly speaking, zimbra is an alternative to exchange?
[20:44] <RoyK> lauris: yes
[20:45] <lauris> RAM is cheap these days
[20:45] <jMCg> Java eats RAM, and lots of it. The more RAM you can spare the more performance it'll deliver.
[20:45] <lauris> so this is not an issue
[20:46] <lauris> is there a mobile app for zimbra?
[20:46] <lauris> iOS/Android
[20:46] <RoyK> lauris: depends how long you strech it ... MS concludes that memory overcommit isn't needed in hyper-v for that reason :P
[20:48] <jMCg> memory overcommit isn't needed. [Period]
[20:48] <lauris> :))
[20:48] <lauris> i'll try it
[20:49] <RoyK> jMCg: it's very useful with xen and vmware .....
[20:50] <jMCg> http://lwn.net/Articles/104179/
[20:51] <jMCg> The red part is where they explain how memory overcommit + OOM killer work nicely together in an analogy understandable to everyone.
[20:52] <jMCg> RoyK: what works well for Xen and VMware is COW. And we've know and used that kind of stuff since ~1974. Even Linux has it.
[20:53] <RoyK> jMCg: memory overcommit in kvm sucks rather badly, it's better in xen and it works even better in vmware - I have that info from people who have used all of those platforms
[20:53] <RoyK> memory overcommit != CoW
[20:53] <lauris> RoyK, does open source edition have all the features from commercial version?
[20:53] <RoyK> lauris: zimbra?
[20:54] <lauris> yes
[20:54] <RoyK> lauris: IIRC the OSS zimbra lacks some smartphone things, outlook connector and a few others - ask on #zimbra
[20:54] <lauris> oh, sorry :)
[20:55] <RoyK> np, just pointing you to a better place ;)
[20:55] <jMCg> RoyK: my experience so far is limited to Xen and KVM, but the point I'm arguing against here I really experienced *in* the guests. I give them 4G of memory, they pretend to have unlimited amount, give it freely on malloc() and then start OOMkilling processes, because, low and behold, we're running out of *real* memory.
[20:56] <jMCg> That's something I really like about Solaris: You want to spawn a process? You want a malloc()? Well sorry, no more memory. Let's start swapping like crazy. If you don't want to see this, buy more RAM.
[20:57] <lauris> RoyK, i found a page of product comparison
[20:57] <RoyK> k
[20:58] <jMCg> lauris, RoyK - I use icssync for the calendar on my Android and the mail works just fine with the mail app.
[20:58] <jMCg> But you don't get push.
[20:59] <lauris> for my personal needs i use gmail, it syncs perfectly on my android phone :)
[20:59] <jMCg> While forwarding all my data to google for ad-analysis.
[20:59] <RoyK> that's not really relevant to zimbra, is it? ;)
[21:03] <Wietienkie> hi, I'm having a problem with RAM detection on a 64bit installation
[21:04] <Wietienkie> there's 6 GB in the server, but only 3 GB are recognised...
[21:04] <Wietienkie> I'm using 10.04
[21:04] <Wietienkie> anyone know what might be the cause?
[21:04] <Wietienkie> been googling all evening, couldn't really find something
[21:06] <RoyK> Wietienkie: what does uname -a have to say?
[21:10] <Wietienkie> RoyK: already checked that, but here goes:
[21:10] <Wietienkie> Linux hkserv 2.6.32-34-server #77-Ubuntu SMP Tue Sep 13 20:54:38 UTC 2011 x86_64 GNU/Linux
[21:10] <Wietienkie> so yeah, it's x86_64
[21:11] <RoyK> Wietienkie: does linux see less memory than what's reported by BIOS?
[21:12] <Wietienkie> I have no access to the bios now, sadly
[21:12] <RoyK> I guess you'll have to check that first
[21:12] <RoyK> also, pastebin lshw output
[21:12] <Wietienkie> I checked with cat /proc/meminfo though, and it reports 3GB only, so I do fear the bios sees 3GB only
[21:13] <Wietienkie> the mobo is supposed to support 8GB
[21:13] <Patrickdk> if they installed high desity ram, instead of low density, then yes, it would do that
[21:15] <Wietienkie> http://pastebin.com/19sPbmDL
[21:16] <Patrickdk> maybe they where put in the wrong banks?
[21:16] <Patrickdk> so that the 1g stick is paired with the 2gig stick? instead of 1g to 1g
[21:16] <Wietienkie> well that's what I'm thinking of too right now, but I'm not sure though
[21:16] <Wietienkie> yeah, the 1 GB is paired with the 2 GB stick
[21:16] <RoyK> wierd - the DIMMs are therë... I've seen similar stuff on linux machines, though
[21:17] <Wietienkie> fffff, shouldn't have been so lazy and have looked up what's the best
[21:17] <Patrickdk> royk, that isn't too odd :)
[21:17] <Wietienkie> damnit, google is revealing something more to me
[21:17] <Patrickdk> that just reports what the stick says it is, 2gigs
[21:17] <Patrickdk> doesn't mean the computer can see the whole 2gig stick :)
[21:17] <Wietienkie> Due to Intel 965 chipset limitation, max addressable memory is 4GB
[21:17] <Wietienkie> chipset is P965...
[21:17] <RoyK> Patrickdk: but 3GB? not 4?
[21:17] <Patrickdk> most likely
[21:18] <Patrickdk> it limits each ram slot to 1gig
[21:18] <Patrickdk> actually, that was 64bit
[21:18] <Patrickdk> depends
[21:18] <Wietienkie> yeah well, don't really get it
[21:18] <Patrickdk> not sure how that chipset works, it might use up that space for pci stuff
[21:18] <Patrickdk> I have had systems limited to 2.4gigs ram cause the rest was pci/pcie/agp crap
[21:18] <RoyK> Patrickdk: 1+2+1+2 should be either 6GB or 4GB IMHO
[21:18] <Wietienkie> I know the P945 was a bitch on supporting max 4 GB while declaring to be 64 bit
[21:19] <Wietienkie> strange thing
[21:19]  * RoyK is off to bed - night all
[21:19] <Patrickdk> royk, yes, but you have reserved memory space
[21:19] <RoyK> Patrickdk: not 1GB!
[21:19] <Patrickdk> if the computer can't map that extra ram to >4g, you loose it
[21:19] <Wietienkie> also on the details of the motherboard it states it supports up to 8 GB
[21:19] <Patrickdk> roy, I have lost up to 1.5gigs cause of it
[21:20] <RoyK> sounds very unlikely
[21:20] <RoyK> but then - it's late - night ladies
[21:20] <Wietienkie> night
[22:46] <webPragmatist> anyone have an idea how to get sftp to obey a umask
[22:47] <webPragmatist> i tried pretty much everything on http://jeff.robbins.ws/articles/setting-the-umask-for-sftp-transactions
[22:47] <webPragmatist> currently i have Subsystem sftp /bin/sh -c 'umask 0002; /usr/lib/openssh/sftp-server'
[22:52] <webPragmatist> also tried modifying pam.d
[22:52] <webPragmatist> :(
[22:52] <webPragmatist> what the heck
[22:54] <Takyoji[laptop]> Is it possible to have apt-cache (or whatever works best) on a 10.04 LTS server and passively cache packages for 11.10 desktops?
[22:55] <Takyoji[laptop]> I'm working on a school setup; and the download speed is only ~1Mbps (T1)
[22:57] <webPragmatist> it doesn't run the .sh if i modify the Subsystem
[22:58] <qman__> Takyoji, yes
[22:58] <qman__> how you do it depends on which software you choose, but proxies will cache anything you want
[22:58] <qman__> and mirrors you have to set up each distro you want to mirror for
[22:58] <Takyoji[laptop]> So what would be a reasonable minimalistic setup for such>
[22:59] <qman__> caching proxy
[22:59] <qman__> probably apt-proxy or squid
[22:59] <qman__> mirrors are on the very costly side, proxies only keep what you download
[23:00] <qman__> but if your internet is really slow or unreliable a mirror may be the way to go
[23:00] <Takyoji[laptop]> and I'm also thinking what precautions would be needed for the sake of limiting it from being a way around the firewall
[23:00] <qman__> you run it through the firewall just like everything else
[23:01] <Takyoji[laptop]> But the server is exempt from all the rules
[23:01] <qman__> all it needs to do is download HTTP from the mirror sites
[23:01] <qman__> why, exactly?
[23:02] <Takyoji[laptop]> but anyway, doesn't apt-cache-ng pretty much passively cache?
[23:02] <webPragmatist> rather… the sftp-server.sh script works but umask 0002; does nothing
[23:02] <qman__> don't know
[23:02] <qman__> but the best way to do it IMO is a transparent inline proxy
[23:02] <qman__> no client setup needed
[23:02] <Takyoji[laptop]> I think the reason is: there's no internet access allowed at all, unless if you HTTP auth with the proxy
[23:03] <qman__> well, if you're already running a proxy, you'll have to configure that proxy to cache apt
[23:03] <qman__> you can't use two different proxies at once
[23:03] <webPragmatist> dir is drwsrwsr-x
[23:03] <qman__> or chain them with firewall trickery
[23:04] <qman__> but I can see that causing some issues
[23:06] <qman__> actually, you probably could use two different proxies if you configure apt to one and browsers to another
[23:06] <Takyoji[laptop]> but as stated, the server itself is exempt, thus doesn't even need to configure with a proxy. It's some awkward trickery of an Astaro server (or at least of how they configured it)
[23:06] <qman__> but it's still a mess you really don't need, just configure the one proxy
[23:06] <Takyoji[laptop]> my point is that I don't want the apt package cache as a way around the firewall for client desktops with malintent
[23:07] <qman__> then set the firewall up properly
[23:07] <qman__> the server shouldn't be completely exempt from the firewall, it should have exceptions for what it needs to do
[23:07] <qman__> but if your clients are already using a proxy
[23:07] <qman__> just configure that proxy to cache apt
[23:08] <Takyoji[laptop]> The reason for the firewall is so that you have to authenticate to get internet access entirely. I don't think there's any domain-specific restrictions or anything. Also, the firewall isn't a Windows/Linux system, it's some proprietary server rack, not even sure how it's configured
[23:10] <qman__> black boxes aren't good security practice
[23:10] <qman__> but even so, if you go the squid route, you can configure squid to only allow outgoing connections to the mirrors you want to cache
[23:11] <qman__> a site whitelist
[23:12] <Takyoji[laptop]> and yes, if I was the decision maker, I wouldn't have something that's pretty much a black box for security at all. xP
[23:12] <qman__> it's not that it's a proprietary system, it's that you don't know how it's set up and what it does
[23:13] <webPragmatist> daskjfdsfjsdfsadiojf\\\
[23:13] <webPragmatist> what the heck
[23:26] <SpamapS> Takyoji[laptop]: so you can get the admins to poke a hole out for a restricted proxy, right?
[23:28] <Takyoji[laptop]> It's how they originally set it up; I haven't had any intervention
[23:29] <Takyoji[laptop]> Anyone know the different between apt-proxy and apt-cacher?
[23:29] <Takyoji[laptop]> difference*
[23:29] <twb> Takyoji[laptop]: yeah, they do the same job but were written by different people
[23:30] <SpamapS> Takyoji[laptop]: apt-cacher-ng is "the next gen" cacher.. I guess. ;)
[23:30] <Patrickdk> they work differently though
[23:30] <Patrickdk> apt-proxy I thought worked before the fact, and apt-cacher was as needed
[23:30] <twb> I have used both, and they have lots of annoying heisenbugs
[23:30] <Patrickdk> apt-cacher is slower than dirt for me, apt-cacher-ng works good
[23:30] <twb> Either use a proper (if partial) mirror with debmirror or apt-mirror, or use a "dumb" cacher like squid
[23:31] <Patrickdk> twb, I have had nothing but issues with squid and .deb
[23:31] <twb> I use debmirror now and I have no problems with it, and the end result is MUCH fast
[23:31] <twb> Patrickdk: it's mainly an issue if you rice up your squid.conf, like "don't cache small files"
[23:31] <Patrickdk> apt-mirror is nice, just a huge hog if you don't need everything
[23:31] <twb> I don't know about a-m, but debmirror allows you to exclude stuff
[23:31] <Patrickdk> twb, na, I had it cache everything up to 20gigs
[23:32] <Patrickdk> I'm happy with apt-cacher-ng though
[23:32] <Patrickdk> I use one hostname for it
[23:32] <twb> Shrug
[23:32] <Patrickdk> and use local overrides at different locations, so that it goes to my home apt-cacher-ng, or colo apt-cacher-ng if I'm roaming
[23:33] <Patrickdk> so I pretty much always have something relatively fast
[23:33] <twb> http://paste.debian.net/139591/ is what I do, it uses 78GiB today, and it pulls from an unmetered ISP mirror so I don't care.
[23:33]  * Patrickdk is annoyed do-release-upgrade isn't using apt-get proxy settings :(
[23:33] <twb> And that means all my jails have a bind-mounted or NFS-mounted file:// mirror which means it's super fast
[23:34] <Takyoji[laptop]> also can a person resize a (md) software RAID 1? :P
[23:35] <Patrickdk> sure
[23:35] <Patrickdk> just give md a larger drive
[23:35] <Takyoji[laptop]> I mean smaller
[23:35] <Patrickdk> sure, just reformat
[23:36] <SpamapS> Actually I believe there is a known issue with squid and the archive having to do with Release and Packages files coming from slightly out of sync mirrors .. but it only should affect dev releases and occasionally security/updates when they change.
[23:36] <Patrickdk> dunno md supports that really
[23:36] <Takyoji[laptop]> When I was installing the OS through the alternate CD, it didn't really give me the option of more than one partition for the RAID
[23:36] <SpamapS> Something to do with the mtimes and Expires headers that the mirrors give back
[23:36] <Takyoji[laptop]> and I hope I don't have to fully reformat and install again. xP
[23:36] <Patrickdk> spamaps, I seem to hit that issue a few times a week
[23:36] <Patrickdk> I finally gave up on squid
[23:39] <SpamapS> Patrickdk: yeah I think there may actually be investigation underway to handle this better
[23:40] <Patrickdk> I seemed to have pushed the ifupdown issue along on debian some
[23:40] <Patrickdk> but someone with enough knowledge about it is going have to finish fixing it
[23:41] <Patrickdk> that part doesn't sound promising
[23:41] <SpamapS> Patrickdk: working at using iproute2 for the static method?
[23:42] <Patrickdk> well for aliases
[23:42] <Patrickdk> would defently annoy me in 12.04 :)
[23:42] <Patrickdk> and my ipvs config
[23:42] <Patrickdk> that is the whole point I'm here :)
[23:42] <Patrickdk> to push along things, and resolve issues that affect me :)
[23:44] <Patrickdk> I totally don't get whatever language ifupdown.nw is written in though
[23:44] <SpamapS> nowebm
[23:44] <SpamapS> Its C+postscript .. totally nuts
[23:44] <SpamapS> "literate" programming
[23:44] <Patrickdk> oh, a cobal offtake? :)
[23:44] <SpamapS> the theory is you print out the program with its documentation or something like that
[23:44] <SpamapS> I find it to be *maddening*
[23:45] <Patrickdk> ya, I had a huge issue attempting to tell code from documentation
[23:45] <SpamapS> because its not just nowebm that makes it hard to understand, it also does some weird, weird code generation
[23:45] <Patrickdk> then I seemed to notice it was kind of like a badly made makefile
[23:46] <Takyoji[laptop]> So awkward that postfix is hinted to be installed when installing mdadm
[23:46] <Patrickdk> why?
[23:46] <Patrickdk> how else would mdadm warn you of a bad drive?
[23:47] <Takyoji[laptop]> ahh, true
[23:47] <SpamapS> the defn2c.pl script is my favorite
[23:48] <SpamapS> perl to translate some made up thing into C
[23:48] <Nicolas> Hi! I have a problem with setting up 2 network cards with 2 static ip adresses. I set up the /etc/network/interfaces file but no way I can bring eth1 to life... or eth0:1 just to give a try to the virtual network card.... Can somebody help? Thanks! (ubuntu 11.10 server 32 bit)
[23:48] <matrix3000> I need assistance with DNS, for some reason I am unable to establish a lookup from another system. When pointing to the bind dns server, on ubuntu, and i updated the /etc/resolv.conf on the client system to point to the ubuntu bind dns server I cannot browse the internet. Was there anything I am missing in terms of allowing a lookup against the root zone?
[23:49] <Patrickdk> nicolas, what exactly is a virtual network card?
[23:49] <omni_> Is there a command to log out of a GUI back into terminal?
[23:49] <Patrickdk> omni_, ubuntu-server has no gui
[23:50] <omni_> Patrickdk, i know, i installed one and realized it was not what i wanted. Now when i re-boot i log into the Gnome and want to get out of it.
[23:50] <Nicolas> Preickdk: I have just seen in many how to that you can set up virtual interfaces with this, but I have to phisical NIC, eth0 and eth1
[23:51] <Patrickdk> totally not sure what exacly eth0:1 has to do with a virtual interface, *:x are aliases, not virtual
[23:51] <Nicolas> yes, alias, my mistake sorry
[23:51] <Patrickdk> what does ifconfig eth1 show?
[23:51] <Nicolas> the thing is that a separate IP address is needed for one domain but the others are on a different and that is why i would like to setup eth1
[23:52] <Patrickdk> nothing in that last sentense made any sense at all, but it all sounds completely non-relevent
[23:52] <Nicolas> auto lo
[23:52] <Nicolas> iface lo inet loopback
[23:52] <Nicolas> # The primary network interface
[23:53] <Nicolas> auto eth0
[23:53] <Nicolas> iface eth0 inet static
[23:53] <Nicolas>         address 94.247.88.164
[23:53] <Nicolas>         netmask 255.255.255.0
[23:53] <Nicolas>         network 94.247.88.0
[23:53] <Nicolas>         broadcast 94.247.88.255
[23:53] <twb> Urk
[23:53] <Nicolas>         gateway 94.247.88.1
[23:53] <Nicolas>         #dns-* options are implemented by the resolvconf package, if installed
[23:53] <Patrickdk> nicolas, you need lots of help
[23:53] <Nicolas>         #dns-search binga.hu
[23:53] <Nicolas> auto eth1
[23:53] <Nicolas> iface eth1 inet static
[23:53] <Nicolas>         address 94.247.88.245
[23:53] <Nicolas>         netmask 255.255.255.0
[23:53] <Patrickdk> use a pastebin, and no one asked for that
[23:53] <Nicolas>         network 94.247.88.0
[23:53] <Nicolas>         broadcast 94.247.88.255
[23:53] <Nicolas>         gateway 94.247.88.1
[23:53] <Nicolas>         #dns-* options are implemented by the resolvconf package, if installed
[23:53] <twb> He needs to have his network access administratively revoked
[23:53] <Nicolas>         #dns-search barnabees.hu
[23:53] <Nicolas> oops sorry :$
[23:53] <Patrickdk> that config though, is totally messsed up
[23:54] <Nicolas> yes, I know :(
[23:54] <Patrickdk> linux really doesn't like to have multible nics on the same network
[23:54] <Patrickdk> normally you get lots of, packet received on wrong nic, errors
[23:54] <twb> Patrickdk: short of bridging
[23:55] <twb> Nicolas: just put both addresses on one interface
[23:55] <Patrickdk> twb, wouldn't bridging just cause stp to make one not work? therefor fixing the issue?
[23:55] <matrix3000> anyone want to help me with a dns issue? basically i can resolve domains on teh dns, but it isn't using the root servers fro anything else
[23:55] <twb> Patrickdk: er, sorry, trunking
[23:55] <Patrickdk> ya, trunking is totally different :)
[23:55] <Patrickdk> trunking/bonding
[23:57] <Nicolas> and how can I do it?
[23:57] <Nicolas> sorry I am very noob to this....
[23:57] <Patrickdk> Nicolas, what is your goal?
[23:57] <Patrickdk> you haven't told us what your attempting
[23:57] <Patrickdk> other then show us what you did, and we said, wouldn't work
[23:58] <Nicolas> my goal is to make both IP addresses work, one is for only one domain and the other is for many
[23:58] <methods1> where are init jobs configured now ?
[23:58] <methods1> there is no more inittab right ?
[23:58] <Patrickdk> Nicolas, what is *a domain*
[23:58] <twb> http://paste.debian.net/139595/
[23:58] <twb> Nicolas: ^^ that is an interfaces stanza that puts multiple addresses on one interface
[23:59] <Nicolas> there is a domain name that needed a separate ip addreess because of an ssl certificate or something like this and I got a new IP from the ISP and I would like to make it work in the sever
[23:59] <Patrickdk> nicolas, ya, using eth1 is not going be easy for that
[23:59] <Patrickdk> use eth0:1 instead
[23:59] <Patrickdk> and when you do that, use a netmask of 255.255.255.255