[00:32] <mrroth> is it possible to make ubuntu server in to a secondary domain controller in my Active directory forest
[00:32] <mrroth> widnows 2008
[00:39] <TenKTech> Im not aware of anything fully funtional but heres a link to something that could help  https://help.ubuntu.com/10.04/serverguide/C/samba-dc.html
[00:44] <twb> For AD I believe you need Samba 4
[01:14] <fluvvell> twb, have they fully released that yet? Or is it still in testing?
[01:14] <twb> I don't know
[01:15] <twb> They aren't in main/ as at lucid, which was enough for me to dismiss it
[01:37] <fluvvell> quentusrex, http://www.kernel.org/pub/linux/utils/raid/mdadm/ANNOUNCE
[01:38] <quentusrex> fluvvell, do you know of a unified annouce list or rss feed?
[01:38] <fluvvell> druciferre, which acls? Are you talking windows file shares or apache acls ?
[01:39] <druciferre> neither... I'm talking about the service from the package acl in ubuntu
[01:40] <druciferre> (i.e. sudo apt-get install acl)
[01:40] <druciferre> setfacl ...
[01:40] <fluvvell> quentusrex, slashdot has a lot of announcements - do you visit there?
[01:40] <quentusrex> fluvvell, I haven't ina while.
[01:46] <twb> druciferre: I think those are "posix acl"s
[01:46] <twb> Yes, they are
[01:47] <twb> ``Most of the Unix and Unix-like operating systems (e.g. Linux,^[1] BSD, or Solaris) support POSIX.1e ACLs, based on an early POSIX draft that was abandoned.''
[01:47] <twb> https://secure.wikimedia.org/wikipedia/en/wiki/Access_control_list#Filesystem_ACLs
[01:51] <mrroth> okay thanks
[01:51] <druciferre> twb, i don't see anything in the article about whether the acl table is stored on the hard drive (i.e. if I put the hard drive in another system will it still see the acls?)
[01:51] <twb> druciferre: I don't see where else they could be stored.
[01:52] <druciferre> twb, I imagine they could be stored in /etc/ somewhere
[01:52] <twb> druciferre: I very much doubt that is the case, but ICBW
[02:13] <fluvvell> druciferre, I have acl installed, but have yet to find a package or piece of software that requires them or uses them
[04:03] <lickalott> gents, trying to remove dhcp client (establishing static IP) and apt-get remove dhcp-client or dhcp-client3 aren't working
[04:03] <lickalott> any suggestions?
[04:06] <druciferre> lickalott, I eventually setup my router to provide dhcp and assign certain computers certain ips based on their mac's
[04:06] <lickalott> already done
[04:07] <druciferre> lickalott, then why are you removing dhcp ?
[04:07] <lickalott> i'm trying to figure out why i lose my znc every once in a while.
[04:07] <lickalott> sometimes, the whole thing, sometimes just a random user/bot
[04:07] <lickalott> starting at the network and working up
[04:08] <druciferre> I have never used a ZNC, so someone else may have to help in that regards.
[04:49] <qman__> lickalott, dpkg -l | grep dhcp
[04:58] <mrroth> can I rsynch a mac OSX to a linux server
[04:58] <mrroth> anyone done this
[04:59] <twb> rsync(1) works on Linux.  IIRC it historically had problems on OS X, but presumably these have since been fixed.
[05:02] <mrroth> and I would not then need to brother with time machine
[05:04] <twb> Well, time machine is probably smarter than a simple rsync
[05:05] <mrroth> but it does ntot do offsite
[05:05] <mrroth> so if something happen were the machine and the backup is located
[05:06] <greppy> time machine is closer to rsnapshot instead of just rsync.
[05:06] <mrroth> ahh
[05:07] <twb> Yep, although rsnapshot is really just "cp -al cur $(date -I)" + rsync
[05:08] <mrroth> ahh
[05:08] <mrroth> does it compress the files sent, or archive in a tar or just send it over
[05:08] <mrroth> mirror coppied
[05:08] <greppy> rsync should only send stuff that has changed
[05:09] <lickalott> qman__ just saw that.  tks!
[05:11] <mrroth> for ubuntu server rspanshot is like time machine
[05:12] <twb> I thought time machine hooked into various syscalls
[05:14] <twb> https://secure.wikimedia.org/wikipedia/en/wiki/Time_Machine_(software)#How_it_works
[05:14] <twb> I'm wrong, time machine is literally just rsnapshot
[05:15] <twb> Of course, hard links themselves are implemented in a royally fucked up way on OS X -- inside the HFS+ filesystem driver
[05:15] <mrroth> oh that fucked up
[05:15] <mrroth> so time machine is just a nice front end GUI for rsnapshot
[05:16] <twb> mrroth: sure looks that way
[05:16] <twb> Well, and an API for apps to talk to it
[05:16] <mrroth> so I wonder why then I can't send my time machine offsite if rsnpshop can do offsite
[05:17] <twb> Also rsnapshot can use ssh or rsyncd protocols to be remote pull-based, whereas TM looks to be local only
[05:17] <mrroth> yea or to a time capulses backup
[05:17] <mrroth> and the media to backup to must be HFS
[05:17] <twb> mrroth: sorry, when I said "literally" I meant it's using the exact same technique -- it's not running rsnapshot code itself
[05:17] <twb> mrroth: HFS+
[05:17] <twb> mrroth: HFS is not HFS+
[05:17] <mrroth> oh and oh yea hfs+
[05:18] <mrroth> yea hfs was used in mac classic HFS + is used now
[05:18] <mrroth> oh I see
[05:18] <twb> Apparently Time Machine can use AFP to another host on the LAN
[05:19] <twb> Obviously this is inferior to ssh, which doesn't care which network it's on, and further is encrypted over the wire.
[05:21] <mrroth> "Although it is unsupported and potentially corrupting the backups after a while, users and manufacturers have configured Linux servers and network-attached storage systems in a similar fashion and used them to store Time Machine backups. ahh
[05:21] <mrroth> hmm
[05:22] <mrroth> yea I wonder how OSX server gets backup maybe third party tools or rsnch
[05:23] <jmarsden> mrroth: http://8help.osu.edu/1247.html has some ideas...
[05:24] <mrroth> oh wow Backup, by Apple, is a program provided to .Mac users. It is a simple-to-use backup program that allows users to backup to their .Mac account, a second hard drive,
[05:26] <jmarsden> mrroth: Carbon Copy Cloner seems to be an rsync wrapper with some OSX-specific enhancements... I've not used it, but it might work for you if you know and like rsync.
[05:27] <mrroth> sweet
[05:27] <mrroth> I will check it out
[05:28] <jmarsden> If you want to do it the harder way, see http://www.bombich.com/rsync.html
[05:29] <mrroth> ahh I see
[05:29] <mrroth> thanks
[05:35] <twb> jmarsden: why not just ports/fink?
[05:35] <twb> I guess if it's not a dev box he might not already have ports set up...
[05:35] <jmarsden> I'm not sure, I think there is Mac specific metadata that the standard rsync port does not back up.
[05:36] <twb> jmarsden: resource forks, sure, but nobody uses them much anymore
[05:36] <twb> Ah, apparently there are some other things, I see
[05:36] <twb> Suck 80s-style vendor unix :-/
[05:37] <twb> I never understood why people want to go back to the unix wars
[05:38] <jmarsden> I think I'd just say that if you are running an Apple-specific OS on Apple-specific server hardware, you probably should use an Apple-specific backup too.  Why people pick Apple or WIndows etc is a *whole* different topic :)
[05:39] <twb> jmarsden: granted
[06:57] <cjs> I'm trying to declare a bridge without a "bridge_ports ethN" line; is this ok? With it, it worked ok, without it, the bridge is never created
[06:58] <cjs> Nobody has any thoughts on how to deal with that?
[07:01] <cjs> Ah, perhaps "bridge_ports none".
[07:04] <twb> Well, I'll show you what I have...
[07:05] <twb> Hm, apparently I always have bridge_ports
[07:06] <cjs> Yes. There's actually a separate manpage for that stuff, bridge-utils-interfaces(5). You must have a bridge_ports for the bridge to be created, but you can use "none" to have it created without initially being attached to any interfaces.
[07:06] <cjs> Learn something new every day. Particularly when it's a new hidden manpage. :-)
[07:07] <twb> cjs: hum, I attach it but don't assign an IP
[07:07] <twb> IOW you can reach the VMs bridged to it, but not the VM server itself
[07:08] <cjs> That's quite normal. No need for an IP, really, unless you want the host running the bridge to talk on that network.
[07:08] <cjs> If you want to reach the VM server itself, just assign an IP on that network to the bridge device.
[07:08] <cjs> (A common mistake is instead to assign it to one of the interfaces connected to the bridge.)
[07:09] <twb> OK, sorry, I'm used to dealing with stupid people :-)
[07:09] <jussi> twb: aww, you shouldnt call me that :P
[07:09] <cjs> Probably just ignorant. Not everybody has a couple of decades of network administration under his belt. :-)
[08:10] <maxagaz> hi
[09:12] <tiger2wander> hi there
[09:12] <tiger2wander> I've got error when apache2's virtualhost config larger than 1024
[09:13] <tiger2wander> Error: fcgi: socket file descriptor (2975) is larger than FD_SETSIZE (1024), you probably need to rebuild Apache with a larger FD_SETSIZE
[09:13] <tiger2wander> I have re-build apache with all FD_SETSIZE variable set to higher that value but still got that problem :(
[09:14] <tiger2wander> please help me to solve this!
[09:14] <aliverius_> does this apply for ubuntu server too? https://help.ubuntu.com/community/VMware/Server
[09:14] <tiger2wander> anyone?
[09:16] <remix_tj> aliverius_: do you think that ubuntu server is different from ubuntu desktop?
[09:17] <aliverius_> not that much... my worries are more abuot running it headless
[09:17] <aliverius_> anyway i will take that as a yes
[09:18] <remix_tj> is a yes, you can. VMWare Server works as a service, does not need GUI, it exposes a web interface
[09:19] <aliverius_> great :)
[09:19] <aliverius_> i was using kvm so far
[09:19] <aliverius_> all was good
[09:19] <aliverius_> till i changed to an atom board
[09:19] <_ruben> vmware server is eol though
[09:19] <aliverius_> i didnt know it doesnt support h/w virt
[09:20] <_ruben> hope you're not planning on doing serious virtualiation without h/w virt
[09:20] <aliverius_> so now i am constranained to vmware-server
[09:23] <aliverius_> _ruben: a router
[09:26] <aliverius_> and it fails to install too
[09:26] <aliverius_> so far
[09:26]  * aliverius_ misses kvm
[09:36] <aliverius_> i should buy a second mobo to play the router role
[09:39] <arnoud> Hello, I'm trying to install an LDAP client on Ubuntu 11.04 and I can't get it to work. Is this the right place to ask questions about it?
[09:42] <maxagaz> hi, I have incoming ip packets from an ip x.x.x.x that I would like to give a low priority, can someone tell me how to do it ?
[10:05] <_ruben> aliverius_: why do you even need a seperate instance for the router role?
[10:05] <_ruben> maxagaz: define "low priority"
[10:08] <maxagaz> _ruben, I'm sure how it works, I guess all incoming ip packets have a same priority in the queue, I would like packets from a given ip to have a lower priority, so that they are processed after others
[11:31] <Daviey> Anyone fancy tackling the nut merge? bug #811976 .. doesn't look too complex.
[11:31] <uvirtbot`> Launchpad bug 811976 in nut "Please merge nut 2.6.1-2 (main) from Debian unstable (main)" [Wishlist,New] https://launchpad.net/bugs/811976
[11:33] <danners> hey i want to move one installation of a server to another system with all of the data and serverconfiguration should be the same. how would i do it?
[11:33] <Daviey> danners: take the hd's out and swap them over. :)
[11:33] <danners> Daviey: the old one is a vm
[11:34] <Daviey> danners: rsync or !clone (and copy /etc.)
[11:34] <Daviey> !clone | danners
[11:35] <danners> Daviey: rsync shouldn't destroy the package index and also copy the configuration right?
[11:37] <Daviey> danners: correct.
[11:37] <danners> Daviey: thanks will try that
[11:51] <maxagaz> I have ethernet camera with an ip that I don't know, how can I find it if I plug it to a laptop ?
[12:13] <quentusrex> maxagaz, do you have any guesses?
[12:16] <maxagaz> quentusrex, use nmap ?
[12:17] <quentusrex> maxagaz, I mean do you have any guesses what the ip or subnet would be.
[12:17] <quentusrex> yes, using nmap would be probably one of the better options.
[12:17] <maxagaz> quentusrex, no, I have no clue
[12:20] <quentusrex> maxagaz, best I could say would be to try the common ones.
[12:20] <maxagaz> quentusrex, okay, thanks
[12:21] <quentusrex> maxagaz, you might also be able to reset the camera to factory to defaults
[12:21] <quentusrex> this might set the ip back to the default.
[12:22] <maxagaz> quentusrex, you're right, that's what I should do
[12:22] <quentusrex> it might be fun to learn how to locate that needle in the haystack, but the faster route would be to reset it.
[12:38] <CrazyGir> hello! I'm trying to build nginx from source, but am having a difficult time specifying the library/header paths correctly
[12:38] <CrazyGir> nginx wants to see pcre and ssl libraries, so I have both libpcre3-dev and libssl-dev
[12:38] <CrazyGir> but I can't quite determine the correct paths
[13:43]  * CrazyGir is all set
[13:43] <CrazyGir> nginx is smart enough to figure it out itself :)
[14:05] <uvirtbot`> New bug: #812131 in mysql-5.1 (main) "operation="mknod" profile="/usr/sbin/mysqld" denied_mask="c"" [Undecided,New] https://launchpad.net/bugs/812131
[14:07] <RoAkSoAx> jamespage: what were the packages that you wanted me to sponsor?
[14:07] <jamespage> hey RoAkSoAx
[14:07] <jamespage>  lp:~hudson-ubuntu/+junk/maven-stapler-plugin
[14:08] <jamespage> and lp:~hudson-ubuntu/+junk/stapler-adjunct-timeline
[14:08] <RoAkSoAx> jamespage: build agains ppa:hudson-ubuntu/ppa?
[14:08] <jamespage> use ppa:james-page/jenkins-upload-testing
[14:10] <RoAkSoAx> jamespage: ok cool, ;)
[14:11] <jamespage> RoAkSoAx: thanks - home straight now with Jenkins :-)
[14:17] <RoAkSoAx> jamespage: hehe
[14:18] <Ursinha> lynxman: http://9gag.com/gag/170426
[14:24] <kpettit> I have a CIFS share mounted in /etc/fstab.  That share goes down periodically for maintenance so I need to figure out how to auto remount the share when it's dropped.  Any ideas?
[14:25] <pmatulis> kpettit: cron job?
[14:26] <kpettit> that was going to be my plan B.  I was hoping there is something that automatically does it, but I guess not?
[14:28] <pmatulis> kpettit: well, something has to probe that that specific share is available.  don't know how it can be done automagically.
[14:29] <pmatulis> kpettit: cron can be as simple as issueing 'mount -a' maybe
[14:29] <kpettit> worth a try.
[14:30] <kpettit> Thanks for the suggestion
[14:30] <pmatulis> kpettit: just so long as the cifs share is set 'auto' (for mount -a to work)
[14:31] <kpettit> got ya.  Doing some testing now.
[14:38] <lynxman> Ursinha: lol!
[15:51] <uvirtbot`> New bug: #812367 in bind9 (main) "package bind9 1:9.7.0.dfsg.P1-1ubuntu0.3 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/812367
[17:26] <uvirtbot`> New bug: #812423 in mysql-5.1 (main) "mysql upstart job hangs if database directory not mounted" [Undecided,New] https://launchpad.net/bugs/812423
[18:57] <ChmEarl> In  `update-rc.d xend defaults SS KK` where do the SS|KK values come from? somewhere in /var/lib?
[18:58] <ChmEarl> the suggested values
[19:08] <Merrioc> Question: during a pxeboot preseed install, the install is creating the md raids as md/0 and md/1 but grub is trying to install to md0 and is erroring out
[19:19] <RoAkSoAx> jamespage: ping
[20:13] <alaing> how do i chmod 777 all folders + files in a directory?
[20:14] <greentea-> best way to install packages from a newer release of ubuntu server into a older release
[20:14] <greentea-> like say 11.xx onto 10.04LTS
[20:15] <greentea-> just download the packages or is there a trick with apt-get?
[20:15] <jamespage> RoAkSoAx: pong
[20:16] <ikonia> greentea-: do'nt do it
[20:17] <ikonia> greentea-: it is a terrible idea
[20:17] <greentea-> doesn't matter
[20:17] <greentea-> I need a newer release of nut-ups
[20:18] <Pici> I don't see that package in the Ubuntu repositories.
[20:18] <greentea-> might just be called nut
[20:18] <greentea-> anyways 10.04lts has 2.4.3? and I want 2.5 or whatever it's up to now because they fixed some brain damagve
[20:18] <ikonia> you can't mix software from different versions
[20:19] <greentea-> so wtf do I know then is what I am asking
[20:19] <greentea-> compiling it is not really an option
[20:19] <ikonia> you can control your language please
[20:19] <greentea-> never :P
[20:19] <ikonia> greentea-: then leave the channel
[20:19] <greentea-> I was joking, calm down :P
[20:20] <patdk-wk> greentea, what are you? 10?
[20:20] <greentea-> 37
[20:20] <greentea-> I can't even control my lang in real life anymore
[20:20] <greentea-> mind is going, oh well
[20:20] <ikonia> greentea-: if you mix software packages from different distros you'll find it will mess up your dependencies and cause issues
[20:20] <greentea-> ikonia: that's why I am here
[20:20] <greentea-> and why I asked what's the best way
[20:20] <greentea-> I only need the nut package
[20:20] <ikonia> greentea-: the best way it so package the software for your version
[20:21] <greentea-> it's a c3 diskless
[20:21] <ikonia> greentea-: linked and built against the library versions on your system
[20:21] <greentea-> compiling it is NOT going to happen
[20:21] <ikonia> greentea-: you can log a wishlist/update on launchpad.net
[20:21] <greentea-> unless I setup a buildhost vm somewhere
[20:21] <ikonia> someone else may do the update for you
[20:21] <Pici> You might be able to find a PPA for it, but you should be aware of what a PPA is and what it provides.
[20:21] <greentea-> I can setup a buildhost somewhere I guess
[20:21] <patdk-wk> buildhosts are pointless, use launchpad's ppa
[20:22] <greentea-> whoa
[20:22] <greentea-> this is cool
[20:22] <greentea-> I was going to have a buildhost though anyways
[20:22] <greentea-> for compiling coreboot, openwrt, and some other stuff
[20:23] <patdk-wk> my buildhost is just the place I store all my patchs :)
[20:23] <Pici> !ppa
[20:23] <greentea-> this is a vm so I am not wasting any resources other than disk space
[20:23] <Pici> (just for the warning)
[20:23] <greentea-> Pici: I am not that worried about it
[20:24] <greentea-> basically this embedded box monitors a ups
[20:24] <RoAkSoAx> jamespage: packages uploaded
[20:24] <greentea-> then logs into a bunch of vmware hosts and runs vm and host shutdown scripts
[20:24] <greentea-> followed by shutting down the nas's and san's
[20:24] <greentea-> only package going to be non-standard would be nut-ups
[20:24] <greentea-> everything else is staying stock
[20:24] <Pici> greentea-: thats fine, just needed to say it :)
[20:25] <kpettit> What's the best search app for Ubuntu?  I've got about 8TB and slocate isn't doing it for me.
[20:26] <greentea-> how would I find out if someone has already made a nut deb for 10.04lts
[20:26] <greentea-> man I used to know this stuff, damn head
[20:26] <patdk-wk> kpettit, heh, I always use find :)
[20:27] <kpettit> I shudder to think of that on a 8TB system.
[20:27] <patdk-wk> find good on mine
[20:27] <patdk-wk> but the average file is 10gb
[20:27] <patdk-wk> doing that on a 8tb mailstore though :)
[20:28] <kpettit> wow, how long does it take you to do a find on that?
[20:28] <jamespage> RoAkSoAx: thankyou!
[20:28] <kpettit> I'm mainly trying to do stuff like "locate *.xslt" trying to find files I forgot about, etc.
[20:28] <RoyK> patdk-wk: striped mirrors?
[20:29] <patdk-wk> royk, currently, concat :)
[20:29] <RoyK> erm, no redundancy?
[20:29] <patdk-wk> nope
[20:29] <RoyK> playing with matches and petrol...
[20:30] <patdk-wk> it's only 8 drives, what's the chances :)
[20:30] <RoyK> about 100% chance of failure within 6-12 months :P
[20:31] <patdk-wk> damn, I'm on like year 4
[20:31] <patdk-wk> I do change out the drive the first signs of issues
[20:31] <patdk-wk> hopefully I'll get my replacements in sept, and I can start on the raid6
[20:32] <RoyK> it's quite usual the first sign of failure for a drive is a dead drive
[20:32]  * kpettit uses a Drobo
[20:32]  * RoyK uses ZFS
[20:32] <patdk-wk> royk, never had that, normally I notice slowdowns and delays long before it goes dead
[20:32] <patdk-wk> I only scared of the disk to stop spinning
[20:32] <RoyK> patdk-wk: that happens too, yes
[20:33] <patdk-wk> but I don't think I have ever had that issue since my FH ibm 10meg drives
[20:33] <RoyK> patdk-wk: but a drive losing a head is also quite common
[20:33] <patdk-wk> never had that :(
[20:33] <RoyK> just saying it happens
[20:33] <patdk-wk> ya
[20:34] <patdk-wk> my worst is when drives just won't spin at all
[20:34] <RoyK> so rather use raid[56] or striped mirrors, the latter for performance
[20:34] <patdk-wk> only ever had that happen on raid systems so far
[20:34] <RoyK> random i/o performance on raid[56] isn't very good
[20:34] <patdk-wk> well, 90% of this data is replaceable anyways :)
[20:34] <patdk-wk> it's basically just online storage for all the damned dvd's and cd's I have
[20:35] <RoyK> then my only question is: how long will it take to rip all those CDs and DVDs if one drive fails now? ;)
[20:36] <patdk-wk> well, just what was missing on that one drive :)
[20:36] <patdk-wk> not all the drives
[20:36] <patdk-wk> if it was striped, ya it would be hell
[20:36] <greentea-> i'm going to try mixing packages and if it blows up
[20:36] <greentea-> oh well
[20:36] <greentea-> temp box anyways
[20:37] <patdk-wk> greentea-, without rebuilding the package?
[20:37] <patdk-wk> I would be suprised if it installed, due to missing deps
[20:37] <greentea-> i'm seeing if i can meet the dependencies
[20:38] <greentea-> oh wait
[20:38] <greentea-> that's true
[20:38] <greentea-> if I have a source package then I can just do a rebuild right?
[20:38]  * RoyK setup a home server for his brother a couple of months back, two 2TB drives in RAID-5, and then added a new drive a few weeks ago - it took perhaps 12 hours to rebuild the raid, but no downtime :)
[20:38] <greentea-> I don't have to figure out how to make my own tree etc
[20:38] <patdk-wk> yep
[20:38] <patdk-wk> apt-get source nut-ups
[20:38] <greentea-> ok that works :)
[20:38] <patdk-wk> apt-get build-dep nut-ups
[20:39] <patdk-wk> dpkg-buildpackage -us -uc
[20:39] <greentea-> I need to get the .src file first right?
[20:39] <patdk-wk> nope
[20:39] <patdk-wk> atleast not if it's in apt
[20:39] <patdk-wk> I normally just modify my sources.list to get it
[20:40] <greentea-> is there a way to list what version that will pull down
[20:40] <greentea-> or specifiy it
[20:40] <RoyK> patdk-wk: iirc the 'correct' way is to add a new file in /etc/apt/sources.d
[20:40] <patdk-wk> royk still not good
[20:41] <patdk-wk> I wish I could just do dep-src ..., and apt-get source would get it
[20:41] <patdk-wk> but it won't
[20:41] <patdk-wk> it will only get the source to the binary version it finds
[20:41] <greentea-> yeah which is the old one
[20:41] <patdk-wk> so I find it easier to just do a search/replace of sources.list get it, and then change it back
[20:41] <RoyK> patdk-wk: the debian/ directory is the key - just copy that to the new source tree
[20:42] <greentea-> patdk-wk: you mean change the sources to like 11
[20:42] <greentea-> fetch the file
[20:42] <greentea-> then swap it back
[20:42] <patdk-wk> yep
[20:42] <greentea-> I thought about that, wasn't sure how well that would work
[20:42] <uvirtbot`> New bug: #812539 in cloud-init (main) "FQDN does not get set correctly in /etc/hosts" [Undecided,New] https://launchpad.net/bugs/812539
[20:42] <patdk-wk> just make sure you apt-get update after each change
[20:42] <greentea-> yeah I knew that
[20:42] <greentea-> :P
[20:42] <patdk-wk> well, for apt-get source it's fine
[20:43] <patdk-wk> for apt-get install, results will vary :)
[20:43] <greentea-> i'lll setup a buildhost vm tonight then, thx
[21:02] <uvirtbot`> New bug: #812548 in nova (universe) "bridge not set up correctly with LXC and all-in-one system" [Undecided,New] https://launchpad.net/bugs/812548
[21:05] <Kazilla> whats wrong with locate, it works pretty well, faster then windows index dare i say
[21:06] <maxagaz> hi
[21:06] <uvirtbot`> New bug: #812553 in nova (universe) "LXC instance fails to start" [Undecided,New] https://launchpad.net/bugs/812553
[21:09] <smoser> lynxman, bug 812539
[21:09] <uvirtbot`> Launchpad bug 812539 in cloud-init "FQDN does not get set correctly in /etc/hosts" [Undecided,New] https://launchpad.net/bugs/812539
[21:09] <andygraybeal> hey guys, i'm trying to get the nagios-libvirt package to compile on my box. it says that i need the 'libvirt library' -- i have it installed.  here is the package that i'm referring to:  http://people.redhat.com/~rjones/nagios-virt/   has anyone else experienced this?
[21:09] <andygraybeal> or tried to install this nagios plugin?
[21:10] <lynxman> smoser: oh man... :/
[21:10] <lynxman> smoser: let me have a look, I don't setup the FQDN afaik
[21:11] <lynxman> smoser: adam_g already provided a patch looks like, although according to the manual 127.0.1.1 should be setup as I did
[21:11] <lynxman> adam_g: don't you agree? :)
[21:11] <smoser> yeah.
[21:11] <smoser> we need to set things up as debian says
[21:11] <smoser> not how lynxman or adam_g say
[21:12] <lynxman> smoser: +1 on you sir
[21:12] <smoser> and also have to deal with eucalyptus (or other cloud providers) where local-hostname might not be set up
[21:12] <smoser> (in eucalyptus local-hostname is some bogus value)
[21:12] <smoser> (ip address)
[21:12] <adam_g> so what is the conensus?
[21:12] <lynxman> http://www.debian.org/doc/manuals/reference/ch05.en.html in 5.1.2
[21:12] <lynxman> adam_g: I think neither you or me
[21:13] <adam_g> ok
[21:13] <smoser> i'll let you all sort it out
[21:13] <smoser> and please update bug
[21:13] <smoser> i have to go now.
[21:13] <lynxman> adam_g: your patch looks good though, it just needs to have the fallback scenario
[21:13] <lynxman> smoser: have fun ;)
[21:13] <adam_g> lynxman: well if what ive done violates debian spec, lets change that
[21:14] <lynxman> adam_g: it just does in that it doesn't fall back if there's no FQDN, whereas I just hardcoded the fallback
[21:14] <adam_g> oh, one sec. let me actually read the spec :)
[21:14] <lynxman> adam_g: sure
[22:03] <bencc> is there a limit to files I can put under one folder?
[22:03] <bencc> I want to cache 50K avatars, do I need sub folders?
[22:23] <Kazilla> bencc: i'm very certain that you can do that
[22:23] <SpamapS> bencc: not really a limit.. but.. it will slow down for certain operations.
[22:24] <SpamapS> bencc: you may be better served by using hashed directories above the dir you have the files in.
[22:26] <bencc> SpamapS:  hashed directories?
[22:26] <bencc> how can I find out what file system I have now? ext3 or ext4
[22:26] <Kazilla> if you can fit 50k worth of html elements in your web browser everyday, i'm pretty sure you can fit 50k with of files..
[22:27] <utlemming> bencc: blkid
[22:27] <SpamapS> bencc: if you put 50,000 files in a directory, you will cause a lot of random I/O reading them. If you break it up into 10 dirs of 5,000 files each, the I/O will be less random.
[22:27] <SpamapS> bencc: ext3 performs similarly
[22:28] <Kazilla> if you're browsing it sure, might cause i/o
[22:29] <Kazilla> but for hosted files, where they know where to look for, less so.
[22:29] <SpamapS> Kazilla: inodes are a single FS block each, so thats 50,000 * block size, which is usually 4k, so thats 195MB of inodes..
[22:30] <SpamapS> yeah single lookups are pretty fast
[22:30] <SpamapS> thanks to the index support
[22:30] <bencc> how do I use blkid to know the file system?
[22:30] <SpamapS> bencc: just do 'mount'
[22:30] <SpamapS> that tells you the fs
[22:30] <SpamapS> of all mounted filesystems
[22:31] <bencc> ok, ext3
[22:32] <RoyK> SpamapS: with dir_index, 50k files may work well, but then, with 250k files, it's still dead slow
[22:32] <Kazilla> is this why facebook use haystack
[22:32]  * RoyK just discovered a 250k file folder
[22:32] <Kazilla> just chucking that in there
[22:32] <bencc> something like mongodb gridfs is nice
[22:33] <bencc> but there is no good server plugin for serving files
[22:35] <SpamapS> RoyK: I would expect an index that works well w/ 50k to work well with 250k ... whats the trouble?
[22:35] <SpamapS> RoyK: Is it not just a b-tree?
[22:36] <SpamapS> bencc: mapping urls to Key/Value stores is pretty trivial...
[22:37] <Kazilla> whats the optimum way to serve a web application anyway? if you're running a cluster. central storage? distributed storage?
[22:37] <Kazilla> well web applications shall i say
[22:40] <Kazilla> anyone know where is the best place to talk about web infrastructures?
[22:54] <SuperMiguel> if i want to run fluxbox on ubuntu, and dont want all the crap that comes with the desktop edition, is there a reason not to install the server version?
[23:23] <ChmEarl> SuperMiguel, don't install server unless you are OK with typing in terminal
[23:34] <Ethos> is there anything better to view disk io than iotop?
[23:34] <Ethos> my linode is saying it's high, but iotop is minimal
[23:35] <StevenR> Ethos: dstat maybe.
[23:35] <Ethos> thanks