[00:10] <SpamapS> deadsmith: re your 32TB RAID 50 ... the io schedule would be more dependent on the workload and the physical hardware.
[00:11] <SpamapS> deadsmith: the size/RAID type are only a smart part of the decision
[00:11] <deadsmith> SpamapS:  Thanks for the reply... The work will be a lot of high-memory processes running on large files... but there will be some large files generated too...
[00:12] <deadsmith> SpamapS:  what else should I be considering?
[00:12] <SpamapS> deadsmith: if the hardware has a smart enough raid controller, then you want the noop scheduler (just throws io at the disk controller blindly)...
[00:14] <deadsmith> SpamapS:  okay, I think the big array has a smart controller... there's another machine that's more like 10TB; more active file copying, less longterm storage... do you feel the same way about that one?
[00:14] <deadsmith> SpamapS:  and, I'm happy to read, I just had a hard time finding relevant info...
[00:15] <SpamapS> deadsmith: this one is actually pretty good, despite its critical tone, it spells them out quite well http://blogs.sybase.com/database/2010/03/io-schedulers-is-linux-really-an-enterprise-os/
[00:16] <deadsmith> thanks a bunch!
[00:16] <SpamapS> a bit dated at kernel 2.6.18 and RHEL 5.4
[00:22] <Thirtysixway> how do i get matlab
[00:22] <Thirtysixway> ...oops wrong channel >.<
[02:42] <Error404NotFound> I just configured chrooted SFTP on a lucid box, problem is that it takes way long to connect, even on shell it takes like 5 minutes
[02:46] <pmatulis> strange
[02:46] <pmatulis> local accounts?
[02:55] <patdk-lap> dns lookup timeouts?
[03:08] <kirkland> SpamapS: hmm, i haven't tried that, actually
[03:09] <kirkland> SpamapS: i switched to squid-deb-proxy yesterday...  still working around its, um, differences
[05:58] <jongbergs> hi, i just setup an ubuntu server, now i am planning to setup a mail server usin postfix..domain is working fine now..my question is: do i still need to set the hostname in order for this to work?
[06:21] <SpamapS> kirkland: I haven't found squid-deb-proxy to be the right solution for chroots/vms ..
[06:23] <amiralul> I'm kinda newbie in the server-land. Can I install a virtual Windows Server guest on my ubuntu box without any GUI? I'm thinking about KVM, but other options are welcomed as well.
[06:25] <SpamapS> amiralul: https://help.ubuntu.com/community/WindowsXPUnderQemuHowTo
[06:26] <SpamapS> a bit out of date.. hrm.. 8.04
[06:26] <SpamapS> https://help.ubuntu.com/community/SeamlessVirtualization
[06:27] <SpamapS> amiralul: that might be better :)
[06:28] <amiralul> SpamapS: thanks, I'll have a look
[06:30] <amiralul> I have to choose between virtualising Windows Server inside Ubuntu or make a goddam apsx website run on apache
[06:31] <SpamapS> amiralul: http://www.mono-project.com/ASP.NET
[06:31] <amiralul> thanks again, I'm still struggling with that :)
[08:22] <ttx> Daviey: typo on "peices" in http://ubuntuserver.wordpress.com/ new look
[08:37] <MTecknology> time to bump my servers up to Ubuntu 11.04
[08:40] <popey> the ubuntuserver.wordpress.com meeting minutes are unreadable because < nicnkname> is stripped out when posted, so you can't see who said what.
[08:40] <popey> http://ubuntuserver.wordpress.com/2011/01/21/ubuntu-server-meeting-minutes-from-2011-01-18/ for example
[09:33] <Daviey> ttx, ooo, well caught!
[09:33] <Daviey> popey, and you.
[09:34] <ttx> Daviey: I have the curse of ultimate noticing
[09:35] <jpds> Daviey: "bits and p*ei*ces".
[09:37] <Daviey> "sack of nuts and bolts"
[09:39] <jpds> Daviey: Blood, sweat and toil.
[09:39] <Daviey> and cloud.
[09:39] <Daviey> or is that equally split between those three, jpds ?
[09:40] <jpds> I think the preferred term nowadays is 'claude'.
[09:51] <JamesPage> Daviey: would that Apache 2 POST segfault bug be impacting my cobbler installation by any chance?
[10:04] <Daviey> JamesPage, hmm, perhaps
[10:05] <Daviey> JamesPage, are you seeing Segfaults? :)
[10:05] <Daviey> JamesPage, zul said he found a fix for it yesterday, i image he uploaded it.
[10:06] <amiralul> brb
[10:06] <amiralul> brb
[10:08] <JamesPage> Daviey: I'll update and see if it still fails
[10:09] <JamesPage> Daviey: it was failing on cobbler check - the POST to apache hung and with a segfault in the Apache error log
[10:13] <JamesPage> Daviey: new version of apache2 -> cobbler working fine :-)
[10:20] <Daviey> JamesPage, \o/
[10:55] <aliverius> can upstart start stop daemons?
[10:56] <aliverius> if yes how? is it true that starting them with /etc/init.d/blah start makes them start at boot without taking advantage of upstart's speeedup features
[10:56] <aliverius> ?
[10:57] <jpds> aliverius: sudo service <service> stop|start
[11:34] <aliverius> i have two nics and i want to swap their if names, so that eth0 becomes eth1 and vice versa
[11:34] <aliverius> how do i do that?
[11:35] <aliverius> i remember that i could remove the current eth0 nic, erase some file and then eth 1 becomes eth0
[11:35] <jpds> aliverius: Edit: /etc/udev/rules.d/70-persistent-net.rules
[11:36] <aliverius> oooooooook
[11:38] <aliverius> i did udevadm trigger
[11:38] <aliverius> what else do i have to do, restart some network service?
[11:39] <aliverius> i could reboot but i wanted to do it the cool way :p
[11:40] <jpds> I'm not sure - I usually reboot after editing that file.
[11:40] <aliverius> ok i will do so too
[11:40] <aliverius> thanks for your help
[11:40] <jpds> Well, I only edit it when I want to change the MAC address of an interface.
[11:41] <phaidros> is there a know stable xen setup with recent ubuntu server version? ($kernel, $xen) ?
[11:43] <g0bl1n> is there a channel for AWS Ubuntu AMI's ?
[11:44] <g0bl1n> I'm trying to find ami-7e5c690a AMI but can't find it in AWS. This is the Ubuntu 10.10 Server 32Bit AMI
[11:58] <soren> g0bl1n: http://uec-images.ubuntu.com/server/releases/maverick/release/
[11:59] <g0bl1n> soren, yes, but the ID for 32b europe AMI can't be found in Amazon "Request Instances Wizard" in Management Console
[12:00] <g0bl1n> soren, that is  ami-7e5c690a
[12:00] <Devo-Kun> soren: I don't see 7e5c690a in the list on http://uec-images.ubuntu.com/releases/10.10/release/, I think you want ami-339ca947
[12:01] <g0bl1n> Devo-Kun, if you press the shown AMI, you'll see the AMI's for severall zones
[12:02] <g0bl1n> Devo-Kun,  ami-339ca947 can't seem to be found in AWS too :-(
[12:04] <Devo-Kun> g0bl1n: I've had problems with the AWS web manager before. What happens if you try to start the instance from the command line?
[12:04] <soren> g0bl1n: $ ec2-describe-images ami-339ca947 --region eu-west-1
[12:04] <soren> IMAGE	ami-339ca947	099720109477/ubuntu-images/ubuntu-maverick-10.10-i386-server-20101225	099720109477	available	public		i386	machine	aki-4deec439			instance-store	paravirtual
[12:05] <g0bl1n> ah ok, let me try Devo-Kun soren
[12:21] <g0bl1n> got it
[12:32] <yann2> hello
[12:32] <yann2> how would you characterize the state of btrfs in 11.4? usable? or "wow be careful"?
[12:32] <yann2> am interested in its snapshot features
[13:20] <hXm> hello
[13:21] <hXm> i have configured a webdav directory on /var/www that works fine from mac and linux, but windows is not able to connect them
[13:21] <hXm> it asks for an user that is never validated
[13:22] <hXm> where can i read information aboout this issue?
[13:31] <caution> how do I check disk performance and how heavy the disk activity is?
[13:34] <pmatulis> caution: bonnie++ and iostat
[13:34] <caution> thanks
[13:42] <plm> Hi all
[13:42] <RoyK> http://wimp.com/twocellos/ ftw!
[13:42] <plm> people, anyone know how to raise the open file limit in Ubuntu?
[13:44] <shauno> plm: sysctl -w fs.file-max=integer.  'sysctl fs.file-max' alone will give you the current value
[13:45] <pmatulis> plm: http://tinyurl.com/6ym9sb3
[13:47] <shauno> hm, they're all using ulimit.  I tend to change things system-wide instead.  not sure what the tradeoffs are there
[14:32] <RoAkSoAx> morning all
[14:36] <RoAkSoAx> ping Daviey:
[14:36] <smoser> good morning mr RoAkSoAx
[14:37] <RoAkSoAx> smoser: hi there mr smoser ;)
[14:37] <RoAkSoAx> how;s it going today
[14:39] <zul_> *grumble* mysql testsuite *grumble*
[14:40] <Daviey> RoAkSoAx, o/
[14:49] <Roasted> hey guys. anybody here set up freeradius before?
[15:02] <Krhome> Hi
[15:03] <Krhome> I've got a problem with my network card. I've put my HD with ubuntu-server in another PC, and the new Network card is not recognized. However, when I make a lspci, the netword card is present
[15:04] <Krhome> Any idea?
[15:07] <pmatulis> Krhome: pci id?
[15:09] <sergevn> Roasted: yes me
[15:09] <sergevn> Roasted: while ago though
[15:09] <Roasted> sergevn, did you set it up on ubuntu by cahnce?
[15:09] <zul> Daviey: http://en.wikipedia.org/wiki/Products_produced_from_The_Simpsons#Tomacco
[15:09] <zul> Daviey: doh! https://code.launchpad.net/~devcamcar/openstack-dashboard/trunk
[15:09] <sergevn> yes
[15:10] <Roasted> sergevn, what exactly do you do after you install it?
[15:10] <Daviey> zul, oooo
[15:10] <Roasted> sergevn, I was told not to change anything because its default config is supposed to handle almost everything.
[15:10] <zul> Daviey: git er done
[15:10] <Roasted> sergevn, but I'm just kind of lost on how to get things started.
[15:10] <Krhome> thanks pmautils for your help
[15:11] <Krhome> but i've format everything
[15:11] <sergevn> Roasted: by default freeradius authenticates againts local system users
[15:11] <Krhome> next time ;-)
[15:11] <Daviey> zul, I am so pleased it's Django :)
[15:11] <Krhome> Have a good day
[15:11] <Roasted> sergevn, so therefore, local users cannot get on wireless. you need to be a domain user. correct?
[15:11] <zul> Daviey: tickled pink?
[15:12] <sergevn> Roasted: have you configured your AP to autheticate via freeradius, you configured that correctly already?
[15:12] <Daviey> zul, aye
[15:12] <Roasted> sergevn, no, I'm not that far yet. I'm remoted into the server (it's running on VMWare) trying to finish the server setup.
[15:12] <zul> Daviey: so what do you want to do about it?
[15:13] <Roasted> sergevn, I found a guide on the official site with some "tests" to run.
[15:13] <Roasted> sergevn, it unfortunately failed when I tried to launch the server in debug mode. It just said command not found.
[15:13] <sergevn> Roasted: could you give me the url?
[15:13] <Roasted> sergevn, surely.
[15:13] <Roasted> http://freeradius.org/doc/?
[15:13] <zul> Daviey: i was thinking putting it in the WI tracker for alpha-3
[15:13] <sergevn> Roasted: doesnt freeradius start?
[15:13] <Roasted> They noted to change as little as possible, and if you do change things, make 1 change, save, test, to make sure its okay.
[15:14] <Roasted> sergevn, because of that I wanted to run it in debug mode sicne that was the first step-test to do.
[15:14] <Roasted> sergevn, I have no idea how to check. It is an active process that should be running>
[15:14] <Roasted> ?
[15:14] <sergevn> first stop freeradius
[15:14] <sergevn> service freeradius stop
[15:14] <sergevn> as root ofcourse
[15:14] <Roasted> done
[15:15] <Roasted> completed fine
[15:15] <sergevn> now do:  radiusd -X
[15:15] <sergevn> and keep that terminal open.
[15:15] <Roasted> okay, snag.
[15:15] <Roasted> if I run that as root, command not found
[15:15] <sergevn> now it's in debug mode
[15:15] <Roasted> if I run without root, it says...
[15:16] <sergevn> did you install via apt-get?
[15:16] <Roasted> The program 'radiusd' can be found in the following packages. and it lists 3 packages.
[15:16] <Roasted> am I to install them?
[15:16] <sergevn> in your terminal
[15:16] <RoAkSoAx> kirkland: howdy!! Sorry about last night!! anyways, when free... let me know
[15:17] <sergevn> sec
[15:17] <sergevn> did you install via apt-get or from source?
[15:17] <Roasted> sergevn, what, freeradius?
[15:17] <sergevn> yes
[15:17] <Roasted> apt-get
[15:18] <sergevn> i think radiusd is called different in ubuntu/debian
[15:19] <sergevn> try freeradiusd or something
[15:19] <Roasted> nada
[15:19] <Roasted> :(
[15:19] <sergevn> meh
[15:19] <sergevn> sec m8
[15:19] <Roasted> thanks bro. your help is mucho appreciated.
[15:20] <Roasted> in fact Im going to step away from the keyboard for a minute, brb.
[15:20] <sergevn> Roasted: you have luced?
[15:20] <sergevn> *lucid
[15:23] <Roasted> yes. 10.04.1
[15:23] <Roasted> sergevn
[15:23] <sergevn> the prodigy is perfect debug music
[15:24] <sergevn> it's freeradius -X
[15:24] <Roasted> bingo
[15:24] <sergevn> on CentOS it's radiusd....
[15:24] <Roasted> accepting requests now
[15:25] <Roasted> sergevn, yeah I couldnt find any distro specific guides. just these general guidelines.
[15:25] <Roasted> that's one thing I hate about linux. one can be so different from the other.
[15:25] <sergevn> yeah, freeradius lacked that also 2 years back when I configured it also for 802.1x wireless network
[15:25] <Roasted> lacked what?
[15:25] <sergevn> debian docs
[15:26] <Roasted> ah I gotcha
[15:26] <sergevn> but that was only for testing back then, production went to CentOS
[15:26] <Roasted> I see.
[15:27] <Roasted> so if I can go into debug mode, I'm good, right? Do I need to tinker with anything in here?
[15:27] <sergevn> in debug mode you can watch some text scroll by ;)
[15:27] <sergevn> and check your config file for errors
[15:27] <Roasted> my config file should be default. I haven't touched it.
[15:27] <sergevn> also if someone authenticates, in debug mode you see that
[15:27] <Roasted> I know I haven't touched it beeeecause I don't even know where it is...
[15:27] <Roasted> :P
[15:28] <sergevn> so now you have to add a user, and then configure your AP
[15:28] <Roasted> so debug mode is handy for me to hang out in when I get another client laptop and AP that's hooked up to this server.
[15:28] <sergevn> yes
[15:28] <sergevn> watch out for plaintext password going by..
[15:28] <Roasted> unfortunately I don't have an AP, and I won't be able to get one today because of bad weather. But I'll take notes here to do that.
[15:28] <Roasted> How do I configure the AP to look at this radius server? Is it just IP based?
[15:28] <sergevn> yes, ip based
[15:29] <andho_busy> hi
[15:29] <andho_busy> is it necessary to give a class C ip to the cluster
[15:29] <sergevn> BUT, for encrypted network traffic you will need certificates and such
[15:29] <plm> Ho change this limit: ulimit -n
[15:29] <plm> 1024
[15:29] <Roasted> sergevn, so wait... if I get into the config, put in the IP of this ubuntu server in the radius section of the AP config, and then try to connect, I'll see these entries passing through debug mode?
[15:29] <andho_busy> regarding eucalyptus that is
[15:29] <plm> how change that?
[15:30] <sergevn> Roasted: yes
[15:30] <RoyK> plm: man ulimit? google?
[15:31] <Roasted> sergevn, do I need to set my wireless adapter on the laptop to have specific settings in order to connect? Or do I simply just need to be logged into a domain account?
[15:31] <hallyn> soren: awesome, about qemu-nbd.  I was pretty sure the lock_kernel removal was to blame, but I was thinking that BKL was a mutex, not semaphore, so was looking for a more blatant error
[15:31] <hallyn> soren: thanks for finding it
[15:31] <Roasted> sergevn, speaking of whcih, how does the linux radius server even know to authenticate domain accounts by now?
[15:31] <sergevn> it doesnt
[15:31] <sergevn> it only knows the users you configure in the configfile
[15:32] <sergevn> check the docs you posted to me, in the Initial Tests section
[15:32] <Roasted> sergevn, so I would just have to set up a local test user for testing purposes on this.
[15:32] <Roasted> sergevn, then when completed, THEN domain users come after.
[15:32] <Roasted> am I right?
[15:33] <sergevn> yes
[15:33] <Roasted> okay, okay. this is making sense now.
[15:33] <sergevn> then you have to install the ldap module etc..
[15:33] <sergevn> that's the hard part
[15:33] <sergevn> are you using Windows Active Directory?
[15:34] <Roasted> yes
[15:35] <Roasted> is there a way to get out of debug mode, or do I just close terminal?
[15:36] <sergevn> ctrl c
[15:36] <sergevn> to quit command in unix/linux use ctrl+c
[15:36] <Roasted> ah yeah, I know that. I just wasnt sure if there was a command to disable it or if ctrl c was okay
[15:48] <zul> robbiew: i know you probably dont know this but is the desktop team suppose to be still handling likewise-open
[15:49] <zul> robbiew: or am i going to be stuck with these two merge requests with my name beside them
[15:50]  * patdk-wk likes you name better
[15:54] <robbiew> zul: I have no idea...sorry...try asking pitti
[15:56] <kirkland> RoAkSoAx: hey
[15:56] <RoAkSoAx> kirkland: yo
[15:57] <kirkland> RoAkSoAx: hey man -- powernap?
[15:57] <RoAkSoAx> kirkland: sure!
[15:58] <RoAkSoAx> kirkland: so anyways, saw the load monitor!! pretty cool
[15:58] <kirkland> RoAkSoAx: ;-)  did you try it?
[15:58] <kirkland> RoAkSoAx: i like it
[15:59] <zul> robbiew: k
[15:59] <RoAkSoAx> kirkland: it's pretty cool
[16:00] <RoAkSoAx> kirkland: anyways, I refactored the powernapd_loop
[16:01] <RoAkSoAx> kirkland: do you wanna chat to make it faster?
[16:01] <kirkland> RoAkSoAx: sure, let me finish my current email
[16:01] <RoAkSoAx> kirkland: ok ;)
[16:16] <kirkland> RoAkSoAx: okay
[16:16] <kirkland> RoAkSoAx: skype?
[16:16] <Bipul`> well i have a domain and i have installed apache on ubuntu
[16:16] <Bipul`> now what shud i do /
[16:17] <RoAkSoAx> kirkland: usre
[16:21] <zul> Bipul`: umm...create a webpage maybe?
[16:21] <Bipul`> zul,  check this http://113.19.130.140/
[16:22] <Bipul`> and now i have register a domain on.tk but dont know the next step
[16:24] <compdoc> you have a static ip address?>
[16:24] <Bipul`> no
[16:24] <compdoc> you are going to host the site at your location?
[16:24] <Bipul`> dyanmic
[16:25] <Bipul`> yes
[16:25] <compdoc> you'll need a service like dyndns.com
[16:25] <Bipul`> well
[16:25] <Bipul`> i know
[16:26] <Bipul`> dyndns but i dont found any option
[16:26] <Bipul`> there t
[16:26] <compdoc> they have a free version, but you have to use their domain name. Or, a pay for version, where you can use your own domain name
[16:26] <compdoc> they can also register your domain name
[16:27] <Bipul`> well i have my domain
[16:27] <Bipul`> but whear to fix my domain i dont have any idea
[16:28] <compdoc> a dns provider will point it to you
[16:28] <compdoc> who did you register with?
[16:29] <Bipul`> l33t-life.tk
[16:29] <Bipul`> is mine domain
[16:29] <compdoc> heh
[16:29] <Bipul`> and i have made an account in dyndns
[16:29] <Bipul`> now whear to fix this domain name
[16:29] <patdk-wk> what is dyndns?
[16:30] <patdk-wk> you need to find their ns servers, and update your domain with them
[16:30] <patdk-wk> looks like you registered it via dot.tk
[16:30] <compdoc> DynDNS Custom: Managed DNS Hosting Solution
[16:31] <patdk-wk> there are hundreds of dyndns companies
[16:31] <compdoc> if dot.tk has a dns hosting service, that would be the way to start
[16:34] <skrite> hey all
[16:59] <Daviey> zul, I noticed you added the Dashboard work item... Do you especially want to do that?
[17:05] <SpamapS> zul: I haven't looked yet. When you merged mysql cluster 7.1, did you drop the libmysqlclient packages?
[17:33] <zul> SpamapS: yeah
[17:33] <zul> Daviey: i think so
[17:34] <zul> SpamapS: mysql-cluster was never part of debian though
[17:35] <SpamapS> zul: oh right
[17:35] <SpamapS> "imported the latest"
[17:36] <zul> SpamapS: yeah
[17:36] <zul> 5.1.54 has been uploaded as well
[17:36] <SpamapS> zul: sweet. I plan to spend a couple hours figuring out how to make pic builds of 5.5 go
[17:37] <SpamapS> zul: I wonder if we couldn't just disable them initially to get something uploaded
[17:37] <zul> SpamapS: okies...i created a PPA to get people testing for SRU fixes as well
[17:37] <zul> SpamapS: i still have to sit down and look at your packaging
[17:37] <SpamapS> hmm?
[17:37] <SpamapS> zul: the cmake transition is maddening
[17:38] <zul> SpamapS: i spent some time this morning closing mysql bugs asking people to test a backported 5.1.54 and we can go from there
[17:40] <zul> kirkland: where are we in the mcollective stuff?
[17:41] <kirkland> zul: uploading today
[17:41] <kirkland> zul: right after lunch ;-)
[17:41] <zul> kirkland: k cool
[17:41] <kirkland> zul: i have one in my ppa, i just wanted to do a little sniff testing of the binaries
[17:41] <zul> url?
[17:41] <kirkland> zul: i was waiting on elmo or someone from Canonical IS to give me some testing feedback
[17:42] <kirkland> zul: i haven't heard from them yet, so i figured i'd just upload a clean source package that builds and installs well
[17:42] <kirkland> zul: and deal with everything else as bugs later
[17:42] <kirkland> zul: that okay by you?
[17:42] <zul> kk
[17:42] <kirkland> zul: rock
[17:42] <kirkland> zul: url coming ...
[17:42] <zul> you might want to talk to lynxman as well
[17:43] <kirkland> zul: my first cut is in https://launchpad.net/~kirkland/+archive/ppa
[17:44] <kirkland> zul: i have a few minor changes since that
[17:44] <kirkland> zul: anyway, it'll be in the NEW queue today
[17:44] <zul> kirkland: do you want me to review it after you upload it?
[17:44] <kirkland> zul: you looking for release team highlights?
[17:44] <zul> kirkland: nope not today
[17:44] <kirkland> zul: ah
[17:45] <lynxman> hey kirkland o/
[17:45] <kirkland> zul: also, i'll upload orchestra today too
[17:45] <kirkland> lynxman: howdy
[17:45] <kirkland> lynxman: you're familiar with mcollective?
[17:45] <lynxman> kirkland: yes, playing with it right now :)
[17:45] <kirkland> lynxman: cool, let me get you a fresh binary package built, and if you can test it and provide some feedback, that would be great ;-)
[17:45] <lynxman> I actually backported your mcollective natty packages to maverick
[17:45] <lynxman> kirkland: sure!
[17:46] <lynxman> I already have some dependency thingies you might want to talk about
[17:46] <kirkland> lynxman: oh?  cool ... what are they?
[17:46] <lynxman> just gimme 5 mins since I'm in the middle of a mumble session
[17:46] <kirkland> zul: oh, one thing ....
[17:47] <kirkland> zul: i talked to the puppet guys and they want me to package 1.0.0 instead of 1.1.0
[17:47] <kirkland> lynxman: i don't know if ^ matters to you
[17:47] <kirkland> > 1.1.x is our development branch so I'd want to avoid getting that included.
[17:47] <kirkland> >
[17:47] <kirkland> > My commitment to the community is to only do bugs and packaging changes in
[17:47] <kirkland> > 1.0.x which should be in line with your needs.
[17:47] <kirkland> zul: ^ is what I got from PuppetLabs
[17:48] <zul> kirkland: ok well lemme look and i think it makes more sense doing 1.1 if that is upstream recommends
[17:48] <kirkland> zul: upstream recommends 1.0
[17:48] <kirkland> zul: not 1.1
[17:48] <zul> kirkland: yeah i got that mixed up
[17:50] <kirkland> zul: cool
[17:53] <Ubuntufanatic> join #ubuntu-offtopic
[17:53] <lynxman> kirkland: the thing I was missing mostly is that the mcollective package (per se) doesn't have any plugins so it's mostly useless except for low level RPC calls
[17:53] <lynxman> kirkland: I did some PPA's in launchpad with what I needed at https://launchpad.net/~lynxman/+archive/mcollective-maverick
[17:55] <lynxman> kirkland: it's work in progress so I'm completely ignoring the deb packaging standards so far
[18:01] <JamesPage> SpamapS: remind me again how I get stats on PPA downloads?
[18:01] <SpamapS> JamesPage: there's a tool.. I'll push it up to a bzr branch
[18:02] <JamesPage> SpamapS: ta
[18:02] <lynxman> SpamapS: howdy, you're also SpamapS on serverfault right?
[18:03] <SpamapS> lynxman: yes!
[18:04] <lynxman> SpamapS: so you were giving me grief about encryption bloatiness :P
[18:06] <SpamapS> lynxman: Indeed I was. 40% is really, really a lot more than I've ever seen.
[18:06] <kirkland> lynxman: zul: okay, so i think there should be a dependency on rabbitmq and rabbit-stomp, right?
[18:06] <lynxman> SpamapS: yeah, I was talking about IPSEC 4k vs 1k and stuff, I had a quite long discussion with my boss 1+ years ago about it
[18:06] <kirkland> lynxman: zul: which of the binary packages need to depend on that?
[18:07] <lynxman> kirkland: hmm kinda since mcollective is also the one that runs on the clients so making it depend on rabbitmq (which is just in one machine) would be confusing, I've thought about that though
[18:07] <RoAkSoAx> kirkland: so what does the LoadMonitor do when you set the threshold to "n" again?
[18:07] <lynxman> SpamapS: the problem is that I read OpenVPN like 5 posts too late :D
[18:07] <kirkland> RoAkSoAx: it scales it to however many CPUs you have online
[18:07] <SpamapS> lynxman: IPSec's encryption re-keys at around the same rate as openvpn IIRC .. public keys are just too big and too unwieldy to be using for more than initial startup and forward secrecy.
[18:07] <kirkland> RoAkSoAx:         if config['threshold'] == "n":
[18:07] <kirkland>             self._threshold = commands.getoutput("getconf _NPROCESSORS_ONLN")
[18:08] <patdk-wk> using public/private keys to encrypt data?
[18:08] <lynxman> SpamapS: yeah but my data was showing that
[18:08] <kirkland> RoAkSoAx: if there's a way to do that getconf without forking a shell, that would be great ;-)
[18:08] <patdk-wk> that has to be cpu taxing
[18:08] <lynxman> SpamapS: in special, bandwidth usage on scp over an IPSEC link vs same bandwidth usage out of the tunnel
[18:08] <kirkland> RoAkSoAx: i couldn't find a getconf() builtin
[18:09] <lynxman> patdk-wk: could be, but in this case the CPU wasn't topping up
[18:09] <RoAkSoAx> kirkland: so if there's 2 processors, then if the loadavg is more than 2, then it will detect activity, otherwise it won;t?
[18:09] <SpamapS> lynxman: I will say that the enc algorithm and the packet profile matters a lot. For his tiny packets of key presses, there is at least one block of overhead .. so for AES256 .. that would at least be enough to make the IP packet take a little longer to build/transmit.
[18:09] <SpamapS> lynxman: right, so enc+enc means double overhead
[18:09] <lynxman> SpamapS: yeah fair enough, was just trying to explain my case here :)
[18:09] <kirkland> lynxman: hmm, i'm confused about your statement on rabbit/stomp?
[18:09] <SpamapS> lynxman: so in a way you're right .. but 40% still seems way too high. :)
[18:10] <lynxman> kirkland: mcollective needs to be installed on every single node whereas rabbitmq just needs to sit in one machine
[18:10] <lynxman> SpamapS: I was shocked as well believe me
[18:10] <SpamapS> lynxman: This is why its always important to compress first, then encrypt. :)
[18:10] <lynxman> SpamapS: hear hear, completely agreed
[18:11] <lynxman> kirkland: so in this case there's no dependency between mcollective and any stomp server
[18:11] <lynxman> kirkland: maybe a Suggests?
[18:12] <lynxman> kirkland: or do a metapackage mcollective-server in which installs mcollective and adds the dependency for a stomp server (like rabbitmq or activemq)
[18:14] <SpamapS> JamesPage: lp:~clint-fewbar/+junk/lptools
[18:15] <kirkland> lynxman: yeah, i like the latter
[18:16] <lynxman> kirkland: me too :)
[18:16] <kirkland> lynxman: mcollective-server meta package it is!
[18:16] <lynxman> wohoo \o/
[18:16] <SpamapS> kirkland: btw, thank you for errno. I've always used mysql's 'perror' .. but this is much better. :)
[18:16] <kirkland> SpamapS: \o/
[18:17] <kirkland> SpamapS: it's been moved from ubuntu-dev-tools to its own package, errno, FYI
[18:17] <SpamapS> kirkland: yeah I just saw that fly by and I'm waiting for it to hit my mirror. :)
[18:17] <kirkland> SpamapS: you'll need to apt-get install errno next time you upgrade your ubuntu-dev-tools ;-)
[18:17] <kirkland> SpamapS: :-)  anyway, thanks for your 'thanks';  i wrote that in 2003, when I was doing some systems level development
[18:18] <kirkland> SpamapS: and found myself digging through manpages and *.h files
[18:18] <SpamapS> kirkland: IIRC, Replaces: can be used to get apt to do that during transition.
[18:18] <kirkland> SpamapS: understood
[18:19] <RoAkSoAx> kirkland: and yeah was the eval() thingy so I'm gonna remove the quotes from the config
[18:19] <kirkland> SpamapS: well, actually, ubuntu-dev-tools would have to recommend errno
[18:19] <RoAkSoAx> kirkland: and the eval for the LoadMonitor
[18:19] <kirkland> SpamapS: which ain't gonna happen, given my talks with the maintainer of that package
[18:19] <kirkland> RoAkSoAx: \./
[18:19]  * kirkland grew *really* long arms
[18:19]  * RoAkSoAx loves kirkland's LoadMonitor for PowerNap
[18:20]  * kirkland highfives RoAkSoAx 
[18:20] <SpamapS> kirkland: yeah you can actually doo  Replaces: ubuntu-dev-tools (<< xxxx) Breaks: ubuntu-dev-tools (<< xxxx) ..
[18:20] <kirkland> SpamapS: bzr branch lp:errno
[18:20] <kirkland> SpamapS: and you'll that
[18:20] <kirkland> SpamapS: and you'll see that
[18:21] <SpamapS> kirkland: ahh ok, then yeah, as long as ubuntu-dev-tools recommends it.. should work. Or is there some suggestion that we don't want to do that?
[18:22] <RoAkSoAx> Fedora Servers compromised http://is.gd/qr1rlr
[18:22] <lynxman> RoAkSoAx: again?!
[18:23] <RoAkSoAx> lynxman: that's yesterdays news, so yeah I guess that again
[18:25] <kirkland> lynxman: http://paste.ubuntu.com/558652/ ?
[18:26] <lynxman> kirkland: oh libstomp-ruby adds the rabbitmq server stomp functionality? cool
[18:26] <lynxman> kirkland: looks beautiful
[18:27] <kirkland> lynxman: SpamapS showed me that :-)
[18:27] <lynxman> SpamapS kudos for that one
[18:27] <lynxman> and here I was making a new package for that
[18:27] <kirkland> lynxman: i didn't know about it either;  i had asked SpamapS to look at packaging it
[18:27] <kirkland> lynxman: hah, i started the same thing
[18:27] <lynxman> kirkland: heh :)
[18:28] <lynxman> RoAkSoAx: yeah they were compromised as well like a year and scratch ago
[18:28] <kirkland> zul: okay, last thing now ... init -> upstart
[18:29] <kirkland> zul: were you working on that?  or should i?
[18:29] <zul> go ahead
[18:29] <kirkland> zul: cool, looks easy
[18:30] <zul> kirkland: the makefile patch looks a bit overkill to me couldnt you just make the directories and copy the approiate files in the debian/rules
[18:33] <SpamapS> zul: how does upstream do installs?
[18:33] <SpamapS> kirkland: ^^ ?
[18:33] <kirkland> SpamapS: dunno
[18:33] <kirkland> zul: only one of those patches is getting applied
[18:34] <kirkland> zul: see the series file
[18:34] <zul> kirkland: then get rid of it ;)
[18:34] <kirkland> zul: well, i haven't looked at what they do and don't do yet :-)
[18:35] <zul> kirkland: ok
[18:38] <kirkland> zul: okay, upstart done
[18:41] <kirkland> zul: initlsb.dpatch  and makefile.dpatch should definitely go away
[18:41] <kirkland> zul: not sure about conffile.dpatch
[18:49] <kirkland> RoAkSoAx: i think i see why powernapd is not starting ....
[18:49] <kirkland> RoAkSoAx:         dh_installinit --noscripts --error-handler=true
[18:49] <kirkland> RoAkSoAx: we need to drop --noscripts
[19:00] <kirkland> zul: lynxman: okay http://people.canonical.com/~kirkland/mcollective
[19:00] <kirkland> zul: upstart script now
[19:00] <lynxman> great, this one is for natty right?
[19:00] <kirkland> zul: lynxman: let me know if there's anything else
[19:00] <kirkland> lynxman: yes
[19:00] <kirkland> lynxman: should (?) install on maverick
[19:01] <zul> kirkland: thanks also why is it a native package?
[19:01] <kirkland> mcollective (1.0.0-0ubuntu1) unreleased; urgency=low
[19:01] <kirkland> zul: ?
[19:01] <lynxman> kirkland: yeah, it works right away, just need to build my plugin packages on top of that then
[19:01] <kirkland> zul: it's a -0ubuntu1
[19:01] <zul> kirkland: there is not *.orig.tar.gz
[19:02] <kirkland> zul: [ ]	mcollective_1.0.0.orig.tar.gz	11-Dec-2010 17:07 	158K	
[19:02] <zul> kk
[19:02] <kirkland> zul: were those trick questions?
[19:02] <kirkland> :-)
[19:02] <kirkland> "just checking"
[19:03]  * kirkland goes for lunch finally
[19:03] <kirkland> lynxman: zul: let me know if there's anything else
[19:04] <lynxman> kirkland: can't think of anything right now, have a good one, I'll afk for the evening in a few mins as well
[19:11] <Pici> Ah, natty.
[19:12] <gobbe> ah, natty <3
[19:12] <genii-around> Interesting how there is the same report in like 4-5 languages
[19:19] <RoAkSoAx> kirkland: Just pushed my changes!! It works beatifully for me with 1 monitor of each type!! It actually doesn't seem to add much load
[19:21] <RoAkSoAx> kirkland: and ConsoleMonitor and WoLMonitor (ports 7 and 9) are enabled by default
[19:21] <RoAkSoAx> kirkland: the only thing left is to update the manpages
[19:22] <kirkland> RoAkSoAx: rock
[19:31] <patdk-wk> fix the wol detection?
[19:53] <kirkland> RoAkSoAx: let me know when you have the manpages updated
[19:53] <kirkland> RoAkSoAx: i'm testing now
[19:58] <Mip5> Hi - what is the preferred way to have a script run at boot in ubuntu server 10.04? The script will issue "ip rule add .." and "ip route add..."
[19:58] <RoyK> Mip5: /etc/network/interfaces
[19:58] <RoyK> Mip5: just add 'up ip rule ...' in the interface block
[19:59] <Mip5> RoyK: Really? I thought about that.
[19:59] <Mip5> RoyK: So I need it to be indented (and below the rest of my interface description).
[19:59] <RoyK> the 'up' keyword in front will run that commmand when the interfaces is upped
[19:59] <RoyK> 'down' works the same way
[20:00] <RoyK> Mip5: I don't know if indentation is needed, but it looks better that way :P
[20:00] <Mip5> RoyK: Excellent - thanks very much!
[20:00] <RoyK> np :)
[20:04] <SpamapS> JamesPage: were you able to use that get-ppa-stats script btw?
[20:12] <hggdh> anyone knows what happened to the AMD64 ISO image for today?
[20:20] <genii-around> Maybe because apache2 failed to build from source
[20:25] <RoyK> For those interested in disk lifetimes etc, this is good reading  http://www.cs.cmu.edu/~bianca/fast07.pdf
[20:25] <RoAkSoAx> kirkland: will have them finished by the end of the day... I'm on my way out for lunch right now
[20:25] <kirkland> RoAkSoAx: cool
[20:25] <RoAkSoAx> kirkland: if your testing goes well I think we can release right after getting the manpages finished
[20:25] <kirkland> RoAkSoAx: ack
[20:25]  * RoAkSoAx lunch finally!! (starving)
[20:33] <JamesPage1> SpamapS: with a couple of local tweaks - works great - whats the update interval on stats? or is it live....
[20:35] <SpamapS> JamesPage: no it lags a lot
[20:35] <JamesPage> SpamapS: thought so :-)
[20:35] <SpamapS> JamesPage: I'd guess.. by about a day
[20:35] <SpamapS> maybe 2
[20:37] <kirkland> SpamapS: hmm, lintian/upstart question for you
[20:37] <kirkland> W: powernap: script-in-etc-init.d-not-registered-via-update-rc.d /etc/init.d/powernap
[20:37] <kirkland> SpamapS: looks like a bullshit warning to me
[20:37] <kirkland> SpamapS: its seeing the symlink installed by upstart
[20:37] <kirkland> actually .... i thought i fixed this in lintian itself
[20:37]  * kirkland checks
[20:38] <SpamapS> kirkland: yeah thats b.s. if its the symlink to upstart-job
[20:39] <kirkland> SpamapS: http://paste.ubuntu.com/558704/
[20:40] <SpamapS> kirkland: lintian should be patched already to not complain in that case
[20:40] <SpamapS> kirkland: if not.. definitely needs a bug report
[20:40] <kirkland> SpamapS: yeah, i sponsored a related fix from jiboumans
[20:41] <kirkland> http://pastebin.ubuntu.com/558705/
[20:41] <SpamapS> kirkland: maybe it got lost in a recent merge
[20:42] <kirkland> SpamapS: hmm, last merge was  -- Benjamin Drung <bdrung@ubuntu.com>  Mon, 26 Jul 2010 10:26:45 +0200
[20:43] <kirkland> SpamapS: looks like the maintainer modified jiboumans' *working* patch
[20:43] <kirkland> SpamapS: with something that don't worky
[20:43] <kirkland>     + [RA] Exclude symlinks to upstart-job from init script syntax checks.
[20:43] <kirkland>       Based on a patch by Jos Boumans.  (Closes: #569492)
[20:44] <SpamapS> DOh!
[20:45] <SpamapS> kirkland: and really, that has a whole other approach in debian.. I'm surprised there was even an attempt made to handle it there.
[20:45] <kirkland> SpamapS: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=569492
[20:46] <kirkland> SpamapS: aggravating, frustrating, disappointing
[20:47] <SpamapS> kirkland: well either way, the merge should not have used the debian patch. Their policy on upstart jobs is *completely* different.
[20:47] <kirkland> SpamapS: yeah, i'm going to re-introduce our patch
[20:48] <kirkland> SpamapS: or part of it, at least
[20:49] <SpamapS> kirkland: indeed, Russ Allbery's description is spot on.. the policies are different and ubuntu will have to maintain delta. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=569492#36
[20:49] <SpamapS> kirkland: note that Steve Langasek is working on changing that... and I've promised to help him with that where I can.
[20:50] <SpamapS> I believe the policy now has been changed to allow packages to carry a .init and a .upstart file ..and install both.
[20:50] <kirkland> SpamapS: agreed;  the person who did the merge should have carried or ported the ubuntu specific change and he did not
[21:43] <RoAkSoAx> kirkland: anything that you'd like to see in the manpage?
[21:46] <kirkland> SpamapS: http://paste.ubuntu.com/558735/
[21:46] <kirkland> SpamapS: diff works for me
[21:47] <kirkland> SpamapS: wanna eyeball that for me?
[21:47] <SockPants> hi, i need to test my server for heat problems. i've downloaded cpuburn but it only runs on one of the processors, even if i run 2 instances.
[21:48] <SpamapS> kirkland: I see nothing wrong off hand, but I have to run to an appointment unfortunately.. so I can't give it my full attention right at the moment. Will look at it in a couple hours.
[21:48] <kirkland> SpamapS: cool;  i'm uploading now
[21:48] <kirkland> SpamapS: i'm happy with it
[22:05] <markit> hi, I've a problem with ACL setup. I need, for a school, a dir RW for (group) teachers and (group) students. I've setup ACL so if a teacher creates a file there, everything is ok. But if teacher creates a file in his home and then copies it there (dolphin), the file uses the original permission and it ends up with "effective r--". How can I do?
[22:05] <markit> I want teachers to be free to create content on their home and then copy/paste in the shared dir
[22:12] <RoAkSoAx> kirkland: is the threshold monitor gonna be shipped by default?
[22:12] <kirkland> RoAkSoAx: you mean the LoadMonitor?
[22:12] <RoAkSoAx> s/threshold/Load
[22:13] <RoAkSoAx> kirkland: yep :)
[22:13] <kirkland> RoAkSoAx: yes, and was going to make threshold = "n" in the default config
[22:13] <kirkland> RoAkSoAx: also, i want to disable the process monitor by default
[22:13] <RoAkSoAx> kirkland: now is threshold = n
[22:13] <kirkland> RoAkSoAx: currently, it always matches init
[22:13] <kirkland> RoAkSoAx: cool
[22:14] <RoAkSoAx> kirkland: ok, anything else that comes into your mind?
[22:14] <kirkland> RoAkSoAx: hmm
[22:14] <kirkland> RoAkSoAx: well, i kinda wanted to revisit besteffort = 0
[22:14] <kirkland> RoAkSoAx: i kinda thinking powersave = 0 makes the most sense
[22:15] <RoAkSoAx> kirkland: sure, and leave best effort as 4 or change to 2, then pm-suspend 3, etc
[22:16] <kirkland> RoAkSoAx: do we need a "best effort" ?
[22:16] <kirkland> RoAkSoAx: default to 0 -> powersave
[22:16] <kirkland> RoAkSoAx: if someone wants to be more aggressive then they can
[22:16] <markit> (btw, if someone can suggest me a better place where ask, is welcomed. I've been told to ask here but could be wrong)
[22:16] <StrangeCharm> how can i, as a user, disable ssh password or keyboard-interactive logins via ssh (for my account, obviously)?
[22:17] <RoAkSoAx> kirkland: we don't really need best-effor, though it first test if /etc/powernap/action is executable, if not it evaluates pm-suspend, pm-hibernate, poweroff
[22:18] <kirkland> RoAkSoAx: right
[22:19] <mathiaz> kirkland: hey!
[22:19] <mathiaz> kirkland: who is looking after hadoop and (java) testing this cycle?
[22:20] <kirkland> mathiaz: very likely one mister JamesPage
[22:20] <RoAkSoAx> kirkland: i'll just swap and put powersave -> 0, besteffort -> 4
[22:20] <mathiaz> kirkland: great - thanks!
[22:20] <kirkland> mathiaz: you bet ;-)
[22:21] <kirkland> RoAkSoAx: sounds good
[22:23] <Brdavs> Hello! I have a freshly updated install of maverick and missing /dev/disk/by-uuid... Anyone know the problem?
[22:23] <RoAkSoAx> kirkland: btw.. how were your tests?
[22:23] <kirkland> RoAkSoAx: haven't gotten to it yet;  got pulled over to mcollective stuff
[22:24] <RoAkSoAx> kirkland: ok ;)
[22:26] <MTecknology> rsyslog is killing my CPU; any ideas what I can do to figure out why. It seems to only do it now that I'm on 11.04 and it's not eating up the Disk IO any..
[22:31] <MTecknology> none of my logs are growing abnormally either
[22:44] <Brdavs> 11.04 is very beta.
[22:46] <Brdavs> Missing /dev/disk-by-uuid in 10.10, latest update. Anyone seen that?
[22:48] <RoAkSoAx> Brdavs: try "/dev/disk/by-uuid/"
[22:50] <RoAkSoAx> Brdavs: /win 4
[22:50] <RoAkSoAx> lol
[22:50] <RoAkSoAx> sry
[22:50] <StrangeCharm> how can i, as a user, disable ssh password or keyboard-interactive logins via ssh (for my account only)?
[22:52] <MTecknology> StrangeCharm: I don't know that you can do it per user..
[22:53] <RoAkSoAx> StrangeCharm: "man sshd_config" will tell you
[22:54] <StrangeCharm> RoAkSoAx, doe that give settings for the ssh server administrator, or settings that an individual user can put in ~/.ssh ?
[22:55] <MTecknology> DANGIT!
[22:55] <MTecknology> I'm getting sick of spam!
[22:56] <RoAkSoAx> StrangeCharm: http://ubuntuforums.org/showthread.php?t=1487425
[22:56] <yann2> btw if anyone could review bug #705648 it'd be much appreciated
[22:57] <yann2> reported it a week ago, hasn't seen much attention, though it is quite important (dealing with fs corruption)
[22:57] <yann2> happy to provide more infos if needed
[22:58] <StrangeCharm> RoAkSoAx, that doesn't resove the problem, because i'm a user, not an administrator
[23:02] <RoAkSoAx> StrangeCharm: uhmm not sure then, sorry
[23:11] <realmatt> having difficulty getting an nfs share to mount at boot time.
[23:24] <Brdavs> disk-by-uuid disappearing is because of LXC containers screwing up badly
[23:24] <Brdavs> this LXC will be the death of me it seems.
[23:29] <bobobo> hello everyone
[23:30] <bobobo> is someone able to tell me how to set a different background color for the tty1? like white background and black fonts?
[23:32] <bobobo> it's just to have a better readibility
[23:35] <Brdavs> use ascii color codes in .bashrc?
[23:35] <Brdavs> i mean ansi
[23:35] <bobobo> I've read and tried a lot.....but I don't know HOW
[23:36] <bobobo> like I can change the color of the command line, but not the background
[23:37] <Brdavs> how about running "xterm -fg red -bg green"
[23:37] <geekbri> is there an official ubuntu doc for creating custom AMI's with 10.04 LTS ?
[23:37] <bobobo> xterm Xt error: Can't open display:
[23:37] <bobobo> xterm:  DISPLAY is not set
[23:37] <Brdavs> nope xterm is in X, not tty1
[23:37] <bobobo> ....I'm in text-only mode
[23:37] <Brdavs> yup I gather that now
[23:38] <bobobo> hehe, you know, when you want to read something in the night i feel white background is better
[23:40] <Brdavs> http://www.linuxselfhelp.com/howtos/Bash-Prompt/Bash-Prompt-HOWTO-6.html
[23:40] <Brdavs> try something alon this line.
[23:43] <bobobo> thx
[23:44] <bobobo> waw I've been googling it for 1 hour
[23:44] <bobobo> :D that's nice
[23:51] <RoAkSoAx> kirkland: just pushed manpage updates. So if your testing goes well, and you will what I added to powernapd.8, feel free to upload
[23:52] <bobobo> not working :(
[23:52] <bobobo> time to code now
[23:52] <bobobo> byebye