[01:24] <tshauck> Hi, I'm trying to use an ubuntu ec2 instance and cloudinit to copy files from s3 to the instance, but it isn't working… what's the best way to diagnose the issue
[01:24] <sarnold> "isn't working" is vague -- what specifically doesn't work?
[01:26] <tshauck> sorry - so the files aren't getting copied over from s3 to the instance
[01:26] <sarnold> do you get any errors or warnings anywhere along the process?
[01:27] <tshauck> No, I tried checking the cfn-init and cloud-init logs, but didn't notice anything
[01:28] <tshauck> although I'm a neophyte at this
[01:28] <tshauck> I tried it with the standard amazon ami and it worked
[01:29] <sarnold> tshauck: does it matter if the instance is up and running or spun down or ..? (I'm newer at it than you, hehe)
[01:32] <tshauck> ya… don't know, I mean the whole idea of cloud formation is to start up the instance with the config of the cloud formation
[01:33] <sarnold> tshauck: so, both amis were spun down when doing the cloudinit?
[01:35] <tshauck> there was only one instance, cloud formation takes a template then using cloudinit to startup the instance
[01:36] <neunon> hi everyone. who owns uec-images.ubuntu.com content? for an internal tool, my company's been using an automatically generated JSON file from there (containing a big list of EC2 AMIs and some metadata), and the file vanished sometime over the past few days.
[02:20] <Tohuw> When isc-dhcp-server is configured to update bind9, to where are the A records for new clients written?
[02:33] <patdk-lap> into bind
[02:33] <patdk-lap> it directly modifies the zone file
[04:06] <gerbilcabbagehai> U.S.12.04 , How do I check if the server can view the internet via CLI
[04:08] <gerbilcabbagehai> for that matter, what do i type to see what packages are installed on the server?
[04:09] <Guest98952> this room is dead bro
[04:09] <Tohuw> gerbilcabbagehai: Define "view the the Internet". If you mean testing to see if it can reach networks across the WAN, a ping test never hurts: ping 8.8.8.8 or ping google.com (assuming a DNS client is configured).
[04:09] <gerbilcabbagehai> true
[04:09] <Tohuw> gerbilcabbagehai: To view installed packages, try dpkg --get-selections
[04:10] <Guest98952> tohuw how do you get the address lik www.yoursite.com
[04:10] <Tohuw> Guest98952: Register a domain with a registrar, then point the domain to something. A registrar's site often has helpful guides on this process. Try www.namecheap.com.
[04:12] <gerbilcabbagehai> what is the unix form of dir /p I've tried |more but it only does one list item more, rather than a whole page full
[04:13] <gerbilcabbagehai> looking up man more now
[04:14] <Tohuw> gerbilcabbagehai: If you want to list files a page at a time, try ls | more
[04:14] <Tohuw> Good on you for surfing the manpages. That will get you far.
[04:15] <gerbilcabbagehai> oh lol. i was overthinking. page up page down will probably do it
[04:15] <gerbilcabbagehai> i pushed enter
[04:15] <gerbilcabbagehai> space is the equivalent. lol i feel silly. thanks for your help so far Tohuw
[04:18] <gerbilcabbagehai> alright, ill take a step back, how can i see if my server box is connecting to the router?
[04:19] <Tohuw> gerbilcabbagehai: The best first step is probably to see the current active configuration of the network interfaces. Check man ifconfig
[04:20] <gerbilcabbagehai> network not accessible. reading man now
[04:21] <gerbilcabbagehai> im not scared of cli anymore
[04:22] <gerbilcabbagehai> i used to be. This is the fourth time I've installed ubuntu (since 5.04) and have slowly been learning to use it
[04:22] <gerbilcabbagehai> a couple of times i uninstalled after getting frustrated and went back to win. but i haven't given up, and now things are starting to make sense...(have to unlearn win/dos bad habits)
[04:26] <gerbilcabbagehai> does ifconfig stand for InterFace Configuration?
[04:32] <joeb3_> gerbilcabbagehai, yes it does.
[04:32] <gerbilcabbagehai> thanks joeb3_
[04:33] <gerbilcabbagehai> im starting to think my server is not installed correctly. would i need to restart the server after having disconnected then reconnected the router, or will the server sense the reconnection?
[04:35] <joeb3_> it should sense the connection
[04:40] <gerbilcabbagehai> i rebooted just in case and now it wont boot at all. im going to check this out. ill be awhile. thanks for your help
[04:57] <gerbilcabbagehai> i think... the server might have been running off a dvd... when i removed it, and rebooted it could no longer work...
[04:58] <gerbilcabbagehai> problem was, i removed the dvd a week ago. :)
[06:33] <gerbilcabbagehai> is it common to need to reinstall US1204 often?
[06:49] <gerbilcabbagehai> ok, my server is up and running again. how do i see the public www page from another computer on the same network? typing in my address bar the name of the server isn't working so im typing something wrong
[06:49] <gerbilcabbagehai> is it //servername ?
[06:51] <gerbilcabbagehai> never mind. i found the ip address of teh server and its working
[07:21] <jibel> cloud-init update to 0.7.0~bzr659-0ubuntu1 is broken. I filed bug 1053239
[08:18] <railsraider> hi i need to automate the aptitude safe upgrade and keep the local version of grub installed how to i set it via debconf or something?
[08:25] <xnox> railsraider: google for apt pinning, then use unattended-upgrade
[08:26] <railsraider> xnox: thanks
[08:58] <linocisco> How could I get free LAMP cloud server with webconsole without having credit card of my own?? I looked at Amazon EC2 but that registration needs credit card info.
[08:59] <linocisco> I dont want to use one month cloud offier by ubuntu cloud. which is short term
[08:59] <linocisco> I dont want to use one month cloud offer by ubuntu cloud. which is short term
[09:04] <Japje> linocisco: build your own cloud, then you can have a lamp server for free
[09:04] <linocisco> Japje, I have no domain registered . How can I?
[09:05] <Japje> linocisco: http://bit.ly/S9r9yi
[09:05] <Japje> there you can find how to build a cloud
[09:06] <linocisco> Japje, thanks
[09:06] <linocisco> Japje, I will look at it
[09:07] <Japje> np
[09:08] <linocisco> Japje, it is nothing. I searched similiar like that. I want to know the site with free cloud based LAMP server ready for me to upload only web application with free registration
[09:10] <Japje> linocisco: i dont think there are such providers
[09:10] <Japje> free is never free
[09:11] <eagles0513875> hey guys I ran a few more tests to see why i cannot connect to my dovecot imaps + postfix setup everythign is configured correctly yet squirrelmail gives me an error 111 connection refused and nothing is showing up in the mail logs or apache logs
[09:11] <eagles0513875> any help would be greatly appreciated
[09:15] <linocisco> Japje, someone said http://java.dzone.com/articles/free-ec2-cloud-server-based-0 . but it asked for Credit card when I click on EC2 tab
[09:16] <Japje> why dont you just pay for something you need?
[09:16] <Japje> like the rest of the world
[09:17] <Japje> buy some quality, and you know what you want, so you can shop for it
[09:17] <Japje> if the product is free, you are the product
[09:17] <linocisco> Japje,  As I am user of ubuntu, I like the spirit of ever Free
[09:17] <Japje> Free as in free speech, not free beer
[09:32] <linocisco> Japje, free in imagination not in reality?
[09:39] <CluelessPerson> Hello all.
[09:40] <CluelessPerson> Can someone help me with chroot with a fstab mounted shared folder on virtualbox?
[09:41] <Japje> linocisco: then goodluck with finding a free cloud server with lamp ;)
[09:41] <CluelessPerson> When I try to SFTP in as group:user I get the error  "fatal: bad ownership or modes for chroot or component 'mounted drive' "
[09:42] <CluelessPerson> My fstab is setup: SFTP /media/sftp vboxsf rw,auto,umak=0755,gid=sftp,uid=root,async 0 0
[09:43] <CluelessPerson> My understanding is that the direcotry that's shared is required to be locked down with chmod 755 or 750.  Has to be root, but group accessible to the sftp user.
[12:38] <Korsakof> Hi everyone! Anyone is using LXC?
[12:39] <Korsakof> I want to migrate my old shared web server with traditionnal chroot to lxc virtual server.
[12:39] <Korsakof> Since each lxc server receives a private ip, I don't know how to route request from the host to the container...
[12:41] <Korsakof> ex: in the dns www.mysite.com -> (1.2.3.4 thehost.com) but all request to www.mysite.com should be route to 192.168.52.4 lxc
[12:43] <Korsakof> so a client that "ssh user@www.mysite.com" connect automatically in the container forwarded from the host :)
[12:51] <Redoubt> Hello all. I'm need some preseeding help in Ubuntu Server
[12:52] <Redoubt> Is there a reason that, in the my custom preseed file, I cannot access /cdrom ?
[12:55] <Redoubt> As a simple example, take the following late_command: preseed/late_command string df > /target/home/df.txt; ls -Rl /cdrom > /target/home/cdrom.txt; ls -Rl /target/media/cdrom > /target/home/target_cdrom.txt\
[12:55] <Redoubt> Oops-- missed the d-i
[12:55] <Redoubt> d-i   preseed/late_command string df > /target/home/df.txt; ls -Rl /cdrom > /target/home/cdrom.txt; ls -Rl /target/media/cdrom > /target/home/target_cdrom.txt
[12:56] <Redoubt> The only part of that command that works is the df. And the result of the df do indeed show /cdrom! The Ubuntu Server install tells me that command fails, and neither of the other two files contain any text
[12:58] <Redoubt> All I want to do is add a repository key offline. So I included the key with the ISO. But I can't seem to get at the key!
[13:03] <smoser> hallyn, ping
[13:04] <smoser> hallyn, i'd like your thoguhts on https://bugs.launchpad.net/maas/+bug/1051626 (i tried to summarise in comment 14)
[13:08] <gerbilcabbagehai> on my server is it safe to just type sudo reboot now?
[13:08] <gerbilcabbagehai> or are there some unmounting i need to do first?
[13:09] <hallyn> smoser: looking
[13:10] <Redoubt> gerbilcabbagehai: Should be safe
[13:12] <smoser> hallyn, i can set you up access to a recreate if you're interested
[13:12] <smoser> this does seem to me like "should work"
[13:15] <smoser> hallyn, random other somewhat related question, would you know how to make libvirt manage a network but not run dnsmasq on it at all ? ie, i dont want it to run a dns server *or*  a dchp server.
[13:15] <smoser> maybe jdstrand knows that
[13:18] <hallyn> smoser: i think you're supposed to be able to do that - not specify dhcp?
[13:19] <hallyn> yeah i think if you don't specify <dhcp> ... </dhcp> in the network.xml
[13:19] <smoser> http://bazaar.launchpad.net/~smoser/maas/maas-pkg-test/view/head:/maas-ephemeral-test-quantal.txt
[13:19] <smoser> see line 32 there.
[13:19] <smoser> thats what i'm wrigint.
[13:19] <smoser> but it still runs a dnsmasq (and a DNS server)
[13:20] <smoser> s/wrigint/writing/
[13:23] <smoser> jibel, thank you for opening 1053239. fix just uploaded.
[13:24] <hallyn> smoser: near as i can tell that's a bug.  but what release is that?
[13:24] <smoser> that was precise pxelinux, but quantal everywhere else.
[13:24] <smoser> you would think bug in pxelinux, right?
[13:25] <smoser> or are you saying the dnsmasq is a bug
[13:25] <hallyn> dnsmasq
[13:25] <smoser> ah.
[13:25] <smoser> that was quantal.
[13:25] <smoser> i haven't tried anywhere else, but i can trivally test.
[13:26] <hallyn> smoser: hm, but the comment in code says:     /* start dnsmasq if there are any IP addresses (v4 or v6) */
[13:37] <hallyn> smoser: heh.  ok.  yeah, so i guess pxelinux bug for not picking its own ip addr on the interface from which the request came?
[13:40] <smoser> hallyn, well, yeah. thats what i thought.
[13:40] <hallyn> xkvm?
[13:40] <doko> zul, I assume there was a FFe for nova?
[13:41] <zul> doko: afaik there is a standing FFE for nova
[13:41] <doko> ahh, ok
[13:41] <hallyn> stgraber: all sorts of acks coming in on eric's most recent userns patchset to lkml.  \o/
[13:42] <stgraber> hallyn: yay!
[13:42] <stgraber> hallyn: so we might be getting something working for 3.7 then? or do we need even more patches than what he submitted?
[13:43] <smoser> hallyn, https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1053408
[13:43] <jdstrand> smoser: remove the default network, or have it autostart
[13:44] <smoser> hallyn, xkvm is just some crappy thing i wrote.
[13:44] <jdstrand> smoser: rather, not have it autostart
[13:44] <smoser> http://smoser.brickies.net/git/?p=tildabin.git;a=blob;f=xkvm;h=18da66440b23ad1100e24c0e85d1357ad17d12eb;hb=HEAD
[13:44] <smoser> it just wraps kvm but does tap networking and teardown more easily.
[13:44] <jdstrand> smoser: actually, that is more than you want
[13:44] <jibel> smoser, yw. thanks for the fix.
[13:45] <smoser> jdstrand, i'm confused.
[13:45] <smoser> see bug. i was wanting to run my own dns (bind) on that network
[13:45] <jdstrand> smoser: you want a libvirt managed network with no dns or dhcp. what I suggested gets rid of the network (and by extension, dbs and dhcp), but that is too much
[13:46] <smoser> and my own dchp. i was just wanting to have libvirt create it so it would bring it up on startup and do the NAT stuff.
[13:46] <smoser> jdstrand, right. its nto the 'default' network that was a problem.
[13:46] <smoser> it runs one on my crafted network explicitly.
[13:47] <hallyn> smoser: of course, i fail to see how writing libvirt network xml is easier than just creating your own bridge by hand
[13:47] <smoser> its easy to create the bridge, yes.
[13:47] <ninjix> smoser: hah. I ran into the same dnsmasq/user problem with libvirt the other day
[13:47] <smoser> but then i have to:
[13:47] <smoser> a.) have it come up on boot
[13:47] <smoser> b.) do the NAT
[13:47] <jdstrand> smoser: so, people have often done this by using a bridged network, but that gets rid of nat. I've not done what you are requesting, but you can look at http://libvirt.org/formatnetwork.html
[13:47] <hallyn> trivial, see lxcbr0
[13:48] <smoser> so i have to write my own upstart job
[13:48] <hallyn> yup
[13:48] <hallyn> well no
[13:48] <hallyn> i've done it with /etc/network/interfaces before
[13:48] <smoser> you're suggesting that is easier and preferred over using an existing documented tested supported solution.
[13:48] <smoser> network/interfaces doesn't support NAT
[13:48] <smoser> that i know of
[13:48] <smoser> and that means i have to edit that file
[13:48] <hallyn> no, but it supprots running scripts after it come up
[13:49] <smoser> so anyway... i decided to use libvirt because it sure seems easier to me.
[13:49] <hallyn> and no, i'm not saying you should have to do that.  just saying you and i would have done it differently :)
[13:49] <hallyn> not excusing the libvirt bug
[13:49] <smoser> additionally, i actually *want* libvirt running.
[13:49] <smoser> as i'm going to use it to power on and off nodes
[13:50] <smoser> but i still think just the creation of the network and nat is more straight forward by declaring it and saying "bring that up on boot"
[13:50] <smoser> than writing upstart jobs and fiddling with /etc/network/
  that's all in your biased descriptions :)
[13:51] <hallyn> I say i'ts "just defining the bridge and its startup", you say it's "writing upstart jobs and fighting the minotaur'
[13:51] <smoser> hallyn, tell you what. you figure out how you can do the setup in that bug in a shorter bit of text for copy and pasting and easily checking for errors (set -e) and i'll take your solution.
[13:52] <hallyn> nah.  i've got other bugs i need to track down.  thanks for the offer :)
[13:52] <hallyn> like bug 1053408
[13:52] <smoser> :)
[13:52] <smoser> hallyn, thanks for your work.
[13:52] <smoser> so where should i go with my pxelinux ug?
[13:53] <smoser> it seems serious, and strange to exist dont you think?
[13:53] <pmatulis> hallyn: i punted your lxc branch back to you
[13:55] <hallyn> pmatulis: saw that - thanks, i'll take a look in deatil later today
[13:55] <hallyn> smoser: well it's isc-dhcp-server responsible, or what?
[13:55] <smoser> hallyn, i dont think so.
[13:55] <hallyn> doesn't it handle next-server?
[13:55] <smoser> why would you think isc-dhcp-server is responsible.
[13:55] <smoser> oh. sorry.
[13:56] <smoser> isc-dhcp-server is *told* to use that address by maas (which i consider a bug in maas)
[13:56] <smoser> and its easily fixable by telling it 'next-server' of the bridge.
[13:56] <smoser> but it should work anyway
[13:56] <smoser> and clearly pxeboot downloads the pxelinux.0 content from that tftp server, so it *can* get there.
[13:57] <hallyn> is there an option to tell it 'use whatever address belogns to the interface'?
[13:57] <smoser> probably not.
[13:57] <smoser> but that wouldn't be necessarily desireable anyway
[13:57] <hallyn> i dunno, if you tell isc-dhcp-server to use the other addr, i'm not sure asking it to second-guess is right
[13:58] <smoser> i very well might want to run a tftp server on a separate host (and network) from my dhcp server.
[13:58] <hallyn> of course
[13:58] <smoser> and thats essentially what i'm doing.
[13:58] <smoser> (when misconfigured)
[13:58] <hallyn> but you segragated networks are really nothing out of the ordinary
[13:58] <smoser> right.
[13:58] <hallyn> so there must be a standard way to handle this
[13:58] <smoser> i think it "should work"
[13:59] <hallyn> and you can't say "on eth0, use ip1, on eth1, use ip2" ?
[13:59] <smoser> i'm confused.
[13:59] <smoser> yes.
[13:59] <smoser> i can make this work
[13:59] <smoser> but i think it should work without that.
[14:00] <smoser> it seems to me that pxelinux is freaking out.
[14:00] <hallyn> when you say 'pxelinux' you mean tftpd-hpa?
[14:00] <smoser> no. i mean pxelinux.
[14:01] <hallyn> i'm lost in the stack
[14:01] <smoser> https://bugs.launchpad.net/maas/+bug/1051626/comments/7
[14:01] <smoser> look at that comment, hallyn
[14:02] <smoser> in the "working" case:
[14:02] <smoser>  a.) pxeboot gets an ip address, sees next server, downloads pxelinux.0
[14:02] <hallyn> yes i was looking at that comment
[14:02] <smoser>  b.) requests pxelinux.cfg/01-52-54-00-12-34-01
[14:02] <smoser> in the broken case:
[14:02] <smoser>  a.) same as above
[14:02] <smoser>  b.) requests pxelinux.cfg/00000000-0000-0000-0000-000000000000
[14:03] <smoser> why did it decide to do something different?
[14:03] <smoser> it clearly has access to the tftp server.
[14:03] <smoser> (it got the pxelinux.0 payload)
[14:03] <lordievader> Good afternoon
[14:03] <hallyn> is 01-52-54-00-12-34-01 the macaddr of the guest?
[14:04] <smoser> yes.
[14:05] <smoser> well, '01' is not part of the MAC, but that is part of pxe
[14:06] <rbasak> smoser: looks like a syslinux bug to me. Either that or a BIOS bug, but that seems less likely.
[14:06] <smoser> rbasak, i agree, but what i cant figure out is how such a bug would exist
[14:06] <smoser> as this seems not at all an uncommon setup
[14:07] <hallyn> smoser: set one up physically and see what it does
[14:07] <rbasak> smoser: can you check the exact number of zeroes in the request that it does make?
[14:07] <smoser> its the SYSUID
[14:07] <smoser> rbasak,
[14:08] <rbasak> yeah, but is it, or is it a corrupt pointer pointing to the middle of memory that holds the SYSUID?
[14:08] <smoser> right. thats also possible.
[14:08] <rbasak> eg. off-by-one going off the end of the IP address string and into the next one that happens to be the SYSUID
[14:09] <rbasak> hence I wanted you to count the zeroes :)
[14:10] <smoser> rbasak, 32 zeros
[14:11] <hallyn> rbasak: before he yells at you, quick, open https://bugs.launchpad.net/maas/+bug/1051626/comments/7
[14:11] <hallyn> jdstrand: i was trying to bisect the cause of that qemu bug, but i was thwarted by random segfaults at startup.  i HATE that
[14:13] <smoser> anyone know of a "diffpastebin" ?
[14:13] <smoser> http://www.diffchecker.com/diff http://www.textdiff.com/ http://www.comparemyfiles.com/
[14:13] <smoser> are some. the first is actually nice.
[14:13] <smoser> but i can't give a link.
[14:14] <smoser> oh wait. i can!
[14:14] <smoser> http://diffchecker.com/dasJyJ16
[14:14] <rbasak> smoser: how reproducible is it? Consistently?
[14:14] <rbasak> smoser: how about on a different VM?
[14:14] <smoser> always.
[14:15] <smoser> there are different vms ?
[14:15]  * rbasak finds this intriguing
[14:15] <smoser> i'm not going to try it in virtualbox if that is what you mean
[14:15] <smoser> :)
[14:15] <rbasak> I mean another one eg. with a different MAC
[14:15] <smoser> well, i can try different macs. i've tried same one then :01, :02, :03 ...
[14:15] <smoser> note, the mac that 'xkvm' chooses is actually the same one that kvm would choose if not provided.
[14:16] <smoser> it uses that range.
[14:17] <smoser> rbasak, the othere thing i really *should* try... before i got hallen and you dragged in is a newer pxelinux.0
[14:17] <smoser> or older
[14:17] <smoser> :)
[14:18] <rbasak> That's an idea :)
[14:18] <rbasak> I was going to ask you what version you were using
[14:18]  * rbasak is pulling the Source
[14:18] <smoser> its what comes down from maas-import-pxe
[14:18] <smoser> (which is precise)
[14:18] <rbasak> OK
[14:23] <hallyn> jdstrand: hm, looking at the fix for the CVE, (505597e4476a6bc219d0ec1362b760d71cb4fdca), it does look like it was undone in later changes!
[14:23] <hallyn> gotta run for a bit, bbl and will test that
[14:23] <jdstrand> hallyn: oh! well, thank goodness for QRT :)
[14:24] <jdstrand> hallyn: if you confirm it, let me know, it'll need to be addressed on oss-security (or upstream may announce it-- keep me posted)
[14:31] <undave> hi all, new to ubuntu servers and servers in general. have three ubuntu 12.04 servers running as file backup servers using ssh on a non-standard port. is there any "security" tools i need to be using, please?
[14:31] <Daviey> jamespage: we seem wedged, http://pb.daviey.com/nYx7/
[14:32] <Daviey>  (/job/precise_folsom_deploy/332/console)
[14:32] <rbasak> undave: make sure you install security updates promptly, and run regular backups. That's all I do.
[14:32] <lordievader> undave: I recommend reading this: http://ubuntuforums.org/showthread.php?t=510812
[14:33] <undave> thanks to you both, i'll go read that now
[14:33] <rbasak> I disagree with some of that. "In general, do not write your passwords down, and if you must, keep them in a secure place (Do not put them on a sticky note attached to your monitor for example)." for example
[14:34] <Daviey> jamespage: odd, it became unwedged.. was that you?
[14:34] <jamespage> yep
[14:34] <Daviey> ah
[14:34] <rbasak> Bruce Schneier, who has just about the best security expert credentials possible: http://www.schneier.com/blog/archives/2005/06/write_down_your.html
[14:35] <roaksoax> zul: do you have an example of a package that replaces a config file, and avoids the dpkg question?
[14:35] <zul> which dpkg question?
[14:36] <jamespage> Daviey, it all looks a bit crud to me
[14:36] <Daviey> jamespage: Yeah, but that is what we have to deal with :)
[14:36] <roaksoax> zul: when you upgrade, and the previous config file differs from the new one (due to user modifiying it). it asks whether you want to replace it, or not do anything, etc
[14:37] <zul> roaksoax: not off the top of my head samba might be one
[14:37] <zul> roaksoax: but i think you want to look at ucf
[14:37] <zul> the configuration management tool
[14:37] <zul> not the package :)
[14:37] <Daviey> jamespage: I assume you didn't mean create, read, update and delete ?
[14:37] <roaksoax> zul: cool thanks
[14:37] <undave> rbasak: yeah i disagree with that as well. odd that it'd be on a security thread.
[14:38] <rbasak> undave: I think the first rule of security is to be careful about who you take your advice from
[14:38] <roaksoax> zul: so yeah, I want to avoid ucf from asking questions
[14:39] <zul> roaksoax: ssh might be another one
[14:39] <roaksoax> cool thanks
[14:40] <jamespage> Daviey, its looking a bit better now
[14:41] <jamespage> I've done nothing....
[14:41] <jamespage> I really need to get adam_g to give me a 'catchup' on the CI stuff
[14:41] <jamespage> I've been lagging badly - and I want to plug in ceph at some point in time as well.
[14:43] <Daviey> jamespage: no, it's still poorly
[14:43] <Daviey> it's the damn chmod issue
[15:04] <WormDrink> hi
[15:05] <WormDrink> someone installed ubuntu desktop on a headless server here - I cleaned up most things thats not needed - like X, and networkmanager and dbus - but should I have acpid running ?
[15:06] <WormDrink> also - something is setting cpu cpufreq/scaling_governor scaling to ondemand
[15:06] <WormDrink> what is this and how do I stop it ?
[15:07] <ogra_> why would you want to stop it
[15:07] <WormDrink> cos I suspect it set cpufreq/scaling_governor to ondemand ...
[15:08] <WormDrink> and maybe did some other things which is highly innappropriate on a server
[15:08] <ogra_> yes, thats a default in every modern distro nowadays
[15:08]  * patdk-wk loves using cpu scaling on his servers
[15:08] <ogra_> its not inappropriate at all
[15:08] <patdk-wk> I defently don't want to be burning/heating at 130w's per cpu, all night long
[15:08] <ogra_> yeah
[15:09] <ogra_> or waste the energy it takes to keep it hot when idling
[15:09] <patdk-wk> if the server has to respond in realtime to something, sure
[15:09] <patdk-wk> but that normally isn't the case
[15:09] <ogra_> and that would likely require even deeper changes
[15:09] <ogra_> like a specially built kernel
[15:11] <WormDrink> ogra_, not in gentoo ;)
[15:11] <ogra_> heh, gentoo
[15:12] <WormDrink> patdk-lap, ogra_ - there is marked performance drops when using ondemand
[15:12] <ogra_> what are you serving ? stock data in a bank ?
[15:12] <WormDrink> building a project on my laptop takes 2x as long with ondemand
[15:12] <ogra_> then there is something wrong with your laptop
[15:12] <WormDrink> nope
[15:12] <WormDrink> you can do some tests yourself - there is something wrong with ondemand
[15:12] <ogra_> anyway, up to you, just disable ondemand, but never upgrade again
[15:13] <WormDrink> wow ubuntu is quality stuff
[15:13] <ogra_> (since the defaults will be restored on upgrades for tha, given it is utter nonsense and extremely energy wasting (and CPU killing) to use preformance)
[15:13] <patdk-wk> you could work on uninstalling, and keeping it removed and all that
[15:14] <patdk-wk> but it's so much easier to just disable it in the bios
[15:14] <ogra_> if your bios offers that, yeah
[15:14] <patdk-wk> I haven't seen one that doesn't
[15:15] <ogra_> and i doubt it would be fun uninstalling the initrscripts package or keeping it uninstalled (if you can even get rid of it without bricking)
[15:15] <patdk-wk> also odd
[15:15] <patdk-wk> none of my ubuntu server have any cpu scaling installed
[15:15] <patdk-wk> using normal instal
[15:15] <patdk-wk> but then, they are also all virtual machines
[15:16] <patdk-wk> so probably not used for them
[15:16] <ogra_> dpkg -l |grep initscripts
[15:16] <patdk-wk> orga_: cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors
[15:17] <ogra_> what about it ?
[15:17] <patdk-wk> well, if it doesn't exist, oviously it can't be used :)
[15:18] <ogra_> ah, yeah, but thats a kernel thing
[15:18] <patdk-wk> and the kernel controls cpu scaling
[15:18] <ogra_> might be that the vm kernels have no scaling
[15:24] <ogra_> WormDrink, but since you know so much about governors, why dont you just add the runes needed to /etc/rc.local  to force performance
[15:25] <epinky>  Hello, I'm getting an odd message [ldap] bind to localhost:389 failed: Can't contact LDAP server, but in my /etc/freeradius/modules/ldap I have another IP(not localhost) configured.
[15:25] <ogra_> (its not like it is secret how to do that)
[15:25] <WormDrink> ogra_ I just did update-rc.d ondemand disable
[15:25] <WormDrink> seems to have worked
[15:25] <ogra_> until next upgrade
[15:26] <WormDrink> yes yes - next upgrade wont be gentoo ;)
[15:26] <ogra_> update-rc.d isnt for admin use
[15:26] <WormDrink> I mean will be gentoo
[15:26] <WormDrink> ubuntu isnt for use
[15:26] <ogra_> rather mv around the K links
[15:27] <ogra_> WormDrink, well, for people that know what they are do ubuntu is fine, for people wanting to turn ubuntu into gentoo its probably not
[15:27] <ogra_> *doing
[15:29] <WormDrink> ogra_, yeah - I'm just being nasty - but I have to say ubuntu always freaks me out
[15:30] <ogra_> WormDrink, i bet gentoo would freak me out the same ;)
[15:32] <ogra_> i lost my urge to compile the world somewhere in the early slackware days
[15:35] <RoyK> heh
[15:36] <smoser> rbasak, fyi, hardy through quantal pxelinux behave identically
[15:40] <rbasak> smoser: interesting. I wish I had time to dig into it!
[15:40] <rbasak> Just changing latest quantal maas-enlist, cloud-init and your ephemeral image generator now
[15:40] <rbasak> changing?
[15:40] <rbasak> testing
[16:05] <smoser> rbasak, how'd that go?
[16:06] <rbasak> smoser: still on it
[16:06] <rbasak> it takes ages for the generation thing to run. probably because of qemu
[16:13] <hallyn> jdstrand: actually i think this is different - the 1, function 3 has no_hotplug set an dis not being deleted.  it's 1, function 2 (which is the usb bridge) which is being unplugged, and causing the assert failure.
[16:31] <rbasak> smoser: it's got through to d-i now, so the ephemeral image must be good. Thanks! Looks like it's using the latest cloud-init and maas-enlist, and the only hack is the mountall from the PPA that I built for armhf by hand
[16:32] <rbasak> smoser: so the next step: can we get a precise ephemeral image with this fix?
[16:32] <smoser> rbasak, well.. yeah, thats a bit more of a PITA
[16:32] <smoser> as a lot of stuff has to go back.
[16:33] <smoser> slangasek was handling the moutnall fix, but he's out all this week on holiday
[16:34] <smoser> so i'm tempted at the moment to build quantal daily ephemeral images with that ppa as it is.
[16:34] <smoser> rbasak, how do i get arm ppa build?
[16:34] <smoser> if i do that, then i can build you dailies with the only non-archive thing being that ppa and the mountall.
[16:35] <jdstrand> hallyn: ah, so maybe a different issue. still 'just a bug' which we should try to get fixed before release. should adjust qrt then for now
[16:35] <zul> jamespage:  so i was thinking
[16:36] <hallyn> jdstrand: yeah, trying to ping mstsirkin on #qemu.  hopefully will get it resolved soon.
[16:52] <zul> smb:  ping
[16:53] <doko> zul, Daviey: nova pulls in python-babel (MIR missing), and there are other missing
[16:53] <smb> zul, hm?
[16:54] <zul> smb: so we are seeing a kernel bug on precise with lvm and dd
[16:54] <plars> jamespage: for the 1G space problem with ubuntu-server installs, any chance to look into it?  Collin is looking at the 128M install failure but commented that the OOMs installing with 256M+ were probably due to not having swap space (because I was manually partitioning to make everything fit)
[16:54] <smb> zul, gotta bug report?
[16:54] <zul> doko: like what?
[16:55] <zul> smb:  yeah https://bugs.launchpad.net/cinder/+bug/1023755/
[16:55] <jgriffith> zul: Ping
[16:55] <zul> hey jgriffith
[16:55] <jgriffith> zul: Howdy
[16:55] <doko> zul, see http://people.canonical.com/~ubuntu-archive/component-mismatches.svg
[16:55] <zul> smb:  its actually jgriffith bug im just the conduit
[16:56] <jgriffith> smb: zul is just lucky that way :)
[16:56] <zul> doko:  ok ill fix it for the next upload
[16:57] <jgriffith> zul: Who should I talk to here?
[16:57] <zul> jgriffith:  smb
[16:57] <zul> smb: meet jgriffith
[16:58] <jgriffith> :)
[16:58] <jgriffith> smb: Are you familiar at all with this issue?
[16:58] <jgriffith> smb: Or would a recap be helpful?
[16:58] <smb> heh
[16:58] <smb> jgriffith, Well I am reading the report right now
[16:59] <jgriffith> smb: K... in  a nut shell
[16:59] <jgriffith> smb: Performing a dd if/dev/zero to an LVM device sometimes hangs
[16:59] <jgriffith> smb: In the case of an LVM snapshot device it almost always hangs
[16:59] <jgriffith> smb: Usually when this happens the entire system is locked up and I can't do anything
[17:00] <jgriffith> smb: I just so happen to have one now where the system is responsive, so I thought it might be useful to debug
[17:00] <smb> jgriffith, So you say you would sometimes see it on a plain lv, too, just not that often
[17:00] <jgriffith> smb: Initially I thought I would *never* see it on plain LVM
[17:00] <jgriffith> smb: But this morning it looks like that may have changed
[17:01] <xnox> jgriffith: deleting lvm volume does not zero it out.
[17:01] <jgriffith> smb: It's possible that this is slightly different, but it seems it must be related
[17:01] <jgriffith> xnox: correct
[17:01] <xnox> jgriffith: I use lvm volumes as hard-drives for libvirt-manager
[17:01] <jgriffith> xnox: We're sending a dd command before lvmremove
[17:01] <jgriffith> xnox: the dd command is what hangs
[17:02] <xnox> and boy I was surprised when *fresh* volume had old partitions "resurrected"
[17:02] <jgriffith> xnox: :)
[17:02] <smb> jgriffith, Hm, ok. So next question is a) is this a bare metal host or a vm and b) is the vg on local disks or something like iscsitarget
[17:02] <jgriffith> xnox: Thus the dd before we remove
[17:02] <xnox> I fixed bugs in debian-installer now to work around that =)
[17:02] <jgriffith> smb: This happens to be on a VM, I've reproduced on bare metal as well
[17:02] <xnox> jgriffith: I would set oflags=direct or whatever it is, to set directio
[17:02] <jgriffith> smb: It's a loopback device to a local file
[17:02] <jgriffith> xnox: done
[17:02] <xnox> because lvm device is not really a block device.
[17:02] <xnox> hmmmm
[17:03] <xnox> jgriffith: strace & file a bug report.
[17:03] <smb> xnox, Not true lvm lv's are block devices. not physical ones though
[17:04] <xnox> smb: ok. But I wish I new the i/o difference between lvm, md, SSD, spinny drives, scsi...
[17:05] <jgriffith> smb: http://paste.openstack.org/show/21027/
[17:05] <xnox> because kernel does handle it differently and mdadm had a lot of work on trim support and i/o optimisations....
[17:05] <patdk-wk> heh? trim in mdadm?
[17:05] <jgriffith> smb: and a portion of kern.log http://paste.openstack.org/show/21028/
[17:06] <patdk-wk> or do you mean trim support for device-mapper?
[17:06] <xnox> patdk-wk: i think the later but I'm not sure =) i'm just lurking on the linux-raid mailing list ;-)
[17:06] <jgriffith> smb: So I tried running apport, but got honked a bit
[17:06] <jgriffith> smb: Can try again if it would be helpful?
[17:07] <smb> jgriffith, not sure those io error look good... later its definitely waiting for some request to complete...
[17:07] <jgriffith> smb: Yeah, and I'm pretty sure it's the dd which hangs forever
[17:07] <jgriffith> smb: It's interesting because usually when I'd see this the entire kernel would hang and the system was useless
[17:08] <jgriffith> smb: This is the first time I've been able to do anything on the system when this happens
[17:08] <jgriffith> smb: I made the mistake of killing the dd commands that were in progress...
[17:08] <jgriffith> smb: They did in fact die, but lvm commands still hang
[17:09] <jgriffith> smb: So it seems that whatever it is that's going on it just throws LVM into a complete tizzy
[17:09] <smb> jgriffith, Ah well, so here the lvdisplay tries to read from one lv (meta-data) which hangs. Can be a result from the dd breaking the device rq
[17:10] <jgriffith> smb: makes sense
[17:10] <jgriffith> smb: So How the heck do I use this apport in console mode?
[17:10] <jgriffith> smb: I'm at the login/register screen
[17:11] <jgriffith> smb: I made the mistake of using vi commands... that was *no bueno*
[17:11] <smb> Hm, there was a way to write the info to file... though that might run just into the same problem of accessing the bad device
[17:15] <jgriffith> smb: Would it be helpful if I got you access to the system?
[17:16] <smb> jgriffith, Partially. Though I might just try to reproduce it
[17:16] <Daviey> smoser: Hey, on the next publisher run, ubuntu-cloud-keyring  should be NEW'd into precise-proposed.  Can you independently test it works ok?
[17:16] <jgriffith> smb: cool... you're call, just lemme know
[17:17] <smb> jgriffith, For that it would help to get a as detailed description of the setup in the bug report (if I did not miss it)
[17:17] <smb> Especially how the VG is constructed
[17:17] <smoser> Daviey, is there a bug ?
[17:17] <smoser> that i'd post comment on ?
[17:18] <Daviey> smoser: bug 1053452
[17:18] <jgriffith> smb: I can break things apart and document it, OR might be easier to download devstack and do it that way
[17:19] <jgriffith> smb: After that it's just simply cinder create 1, cinder snapshot-create <volume-uuid>, cinder snapshot-delete
[17:19] <jgriffith> smb: Up to you on what's best there
[17:22] <zul> jgriffith/smb: is there a way to see if its a bad device?
[17:22] <smb> jgriffith, Best to reduce the testcase to as basic things as possible. If I have to install and configure a load of things, I may get bored. ;) So I think the really interesting parts are what the device-mapper stack looks like. That loop mount on a file and make a pv of that kind of things
[17:23] <jgriffith> smb: Sounds reasonable, I'll disect and update the bug
[17:24] <jgriffith> smb: Is there any reason to keep this machine in it's current state hanging around?
[17:24] <jgriffith> zul: I don't know... but good question
[17:24] <smb> jgriffith, Don't think so. It sounds like not too hard to reproduce
[17:27] <smb> zul, jgriffith Hm, wonder if cat /sys/block/<dev>/inflight would be >0 and not going down...
[17:28] <zul> i can find out
[17:37] <zul> smb/jgriffith: ok maybe not easily
[17:38] <jgriffith> zul: :(
[17:39] <zul> jgriffith: devstack is taking forever
[17:39] <smb> zul, jgriffith Depending on ways to cause an unhandled NMI or sending sysrq triggers https://wiki.ubuntu.com/Kernel/CrashdumpRecipe may be helpful on gathering info. At least I would try that when reproducing tomorrow (as my today reaches its end)
[17:40] <zul> smb:  ok ill see what i can do
[17:41] <jgriffith> zul: smb Here's an psuedo script of what we do: http://paste.openstack.org/show/21029/
[17:41] <jgriffith> zul: smb Some things need filled in with the count  and size blah blah blah
[17:41] <jgriffith> zul: smb But this is the basic sequence at any rate
[17:42] <jgriffith> zul: smb One thing to note, it is more readily producible IF your volume size is 2 Gig or greater
[17:42] <jgriffith> zul: smb less than 2G it's a bit harder to hit
[17:42] <smb> jgriffith, Sure. Looks like something to work on. Will that go into the bug report?
[17:43] <jgriffith> zul: smb One thing this leaves out is the iscsi target creation but I don't think that should matter
[17:43] <zul> jgriffith: can you add the script to the bug report so we can document it
[17:43] <jgriffith> smb: Yes, I can add a link to the paste in the bug
[17:43] <smb> jgriffith, probably better do paste it as a comment. I do not trust paste links to remain...
[17:44] <smb> Hm so iscsitarget (of some kind) _is_ involved here
[17:45] <jgriffith> smb: Yes, on the create, we then create an iscsi target for the lvm device
[17:46] <jgriffith> smb: Suppose that could be related, but it seems I found soem bug reports on google about devmapper snapshot devs and dd kernel hangs
[17:48] <smb> jgriffith, Certainly could be some factor. It sounds more plausible to loose a request that way. I would certainly try a setup backed by a local disk first. Since that gets rid of complexity if it also shows the problem
[17:48] <zul> smb:  this is my /sys/block/dm-1/inflight when deleting the snapshot: http://pastebin.ubuntu.com/1217211/
[17:48] <jgriffith> smb: Yeah, and the thing is the iscsi target is deleted before this dd command is sent
[17:48] <smb> But if that does not and only the iscsitarget setup has it, then it may allow to get rid of device-mapper iin that
[17:49] <smb> jgriffith, o_O What is then backing the volume?
[17:49] <jgriffith> smb: ?
[17:50] <smb> zul, Definitely a lot of inflight
[17:50] <jgriffith> smb: The backing file is still present, we just no longer have it mapped ot an iscsi tgt
[17:50] <jgriffith> smb: Unless I'm mistaken on the sequence...
[17:51] <smb> jgriffith, If the setup is <lv>+<snapshot> --> <vg> --> <iscsitarget> then you could not remove iscsitarget and expect writes to anything of the vg to work.
[17:57] <jgriffith> smb: I believe it's <vg>---> <lv> ---> <iscsitarget> ---->snapshot
[17:57] <jgriffith> smb: No iscsi target is created for the snap BTW
[17:58] <jgriffith> smb: So the snapshot_delete is  <dd if=/dev/zero> ---> lvremove snapshot
[17:59] <smb> jgriffith, Ah ok, that makes more sense than what I was thinking. So the iscsi device is exported from the snapshot (iow top of the stack)
[17:59] <jgriffith> smb: yup
[17:59] <smoser> Daviey, it is still not there? (ubuntu-cloud-keyring) ?
[17:59] <jgriffith> smb: errr...
[17:59] <jgriffith> smb: Actually no, we don't do any iscsi target work with the snapshot at all
[18:00] <jgriffith> smb: we require you do a create from snapshot to go that way
[18:00] <jgriffith> smb: The snap is never exported... we would do a create from snap <lvm-snap> ---> <new lvm> ---> <export iscsi>
[18:02] <jgriffith> smb: Sorry.. have to run, catch up with you later, or tomorrow if you're around
[18:02] <adam_g> isn't the original LV that backs a snapshot required to exist and be ACTIVE?
[18:02] <smb> jgriffith, Probably mixing up the names. But I guess what is done is to export something via iscsi that is the part one can modify and then gets discarded after usage.
[18:03] <smb> jgriffith, Ok, yeah. I try to have tried looking into it by then
[18:09] <Redoubt> Can I preseed selections in the "install additional software" window?
[18:10] <Redoubt> I can obviously pkgsel/include whatever I want, but I would just like to preselect openssh server
[18:31] <njin> hallo, is /etc/init.d/networking restart deprecated ?
[18:32] <njin> if yes is right to passs sudo service networking restart ?
[18:33] <njin> or is sudo servoce networking stop
[18:33] <njin> sudo service networking start ?
[18:46] <RoyK> njin: both should work
[18:46] <njin> RoyK, thanks
[18:49] <RoyK> "service" something just runs the script
[18:50] <RoyK> no magick involced
[18:55] <jgriffith> smb: Still around?
[18:58] <escott> njin, service is preferred
[19:00] <njin> escott, thanks
[19:11] <pmatulis> hallyn: re lxc branch, does the validate script give errors for you?
[19:16] <hallyn> pmatulis: nope
[19:16] <hallyn> pmatulis: did you get that after a merge, or straight from my branch?
[19:17] <hallyn> pmatulis: sha1sum gives me 37b04b97fc82659d5d1d8fa9e7d79e3a604428f5  virtualization.xml
[19:19] <pmatulis> hallyn: strange.  i branched the latest trunk and merged your branch (started over fresh)
[19:20] <skaet> Woot!   way to go ubuntu-server team!   5 of the 5 bugs release noted with Beta 1 have been fixed for Beta 2.   :D
[19:21] <skaet> https://wiki.ubuntu.com/QuantalQuetzal/TechnicalOverview
[19:21] <highvoltage> Nice.
[19:22] <highvoltage> eek, beta2 is next week already!? it never ends! :p
[19:23] <pmatulis> hallyn: i actually got 10 times the error i posted in my review comment (pointing to different lines)
[19:30] <hallyn> pmatulis: can you pastebin the virtualazation.xml you have, or push to a tmp branch
[19:37] <pmatulis> hallyn: https://code.launchpad.net/~petermatulis/serverguide/temporary_branch_merge_123642
[19:46] <hallyn> pmatulis: ok thanks, that makes more sense - that at least is on a listitem :)  still don't see the problem, but looking...
[19:49] <epinky> could anyone give a hand with RADIUSproxy?
[19:53] <hallyn> pmatulis: so it's complaining that the <listitem> which immediately follows <itemizedlist> is not following itemizedlist.  Bug in the validator?
[19:54] <hallyn> oh or does it need a para
[19:56] <hallyn> that did it
[19:57] <hallyn> pmatulis: re-pushed.  or do you prefer i push a versoin based on your temp version?
[19:59] <pmatulis> hallyn: well, i don't understand why you didn't get the same errors
[20:00] <hallyn> pmatulis: i did once i ran the validator by hand.  i expected 'make serverguide-html' to run it
[20:00] <pmatulis> hallyn: ah, no
[20:00] <LazyTown> please help!  installed apache mpm_worker module and now my server isn't parsing php,  what did I miss?
[20:01] <pmatulis> hallyn: let me try again to get your stuff
[20:01] <sarnold> LazyTown: what error messages do you get in your logs?
[20:01] <hallyn> pmatulis: thanks
[20:03] <mcb_> Using Ubuntu 10.10, I made a disk image from a server to an external driver. Reinstaled grub in the external drive (chroot used), changed the UUIDs in /etc/fstab  and in /boot/grub/grub.cfg. I more then double checked this and they are all right. I use this same method with Debian 6 and works to perfection. In Ubuntu, Grub loads OK, but it hangs just after, with no erro message, just the blinking cursor in the left upper conner of the screen. I can 
[20:05] <LazyTown> sarnold: I got the server to serve php again, which is a plus, but now I get a 500 and this is the error from my log: [Thu Sep 20 20:04:08 2012] [error] [client 199.96.116.26] Request exceeded the limit of 10 internal redirects due to probable configuration error. Use 'LimitInternalRecursion' to increase the limit if necessary. Use 'LogLevel debug' to get a backtrace., referer: http://mysubdomain.mycompany.com/
[20:06] <sarnold> LazyTown: I wonder if that is from excessive (or broken) mod_rewrite ...
[20:06] <sarnold> LazyTown: does a simple phpinfo(); page load?
[20:06] <LazyTown> let me test
[20:07] <pmatulis> hallyn: validated, making html...
[20:07] <LazyTown> sarnold, no
[20:08] <sarnold> LazyTown: how about static pages?
[20:08] <LazyTown> sarnold: plus, when I browse to my home domain I get a 404: The requested URL /cgi-bin/php5/index.php was not found on this server.
[20:08] <LazyTown> and yes, I can get static pages to work
[20:15] <epinky> I want to proxy EAP-TTLS (PAP credentials) to some other server, what should I configure?
[20:15] <epinky> I'm using freeradius
[20:19] <mcb_> Using Ubuntu 10.10, I made a disk image from a server to an external driver. Reinstaled grub in the external drive (chroot used), changed the UUIDs in /etc/fstab  and in /boot/grub/grub.cfg. I more then double checked this and they are all right. I use this same method with Debian 6 and works to perfection. In Ubuntu, Grub loads OK, but it hangs just after, with no erro message, just the blinking cursor in the left upper conner of the screen. I can 
[20:23] <sarnold> mcb_: you're cut off at "screen. I can"
[20:24] <mcb_> sarnold: Just after the Grub screen, all that i get is the blinking cursor in the left upper corner of the screen, nothing more happens...
[20:25] <sarnold> mcb_: does grub need to load USB, SCSI, or filesystem modules to find the kernel on the external drive?
[20:27] <Daviey> smoser: SpamapS didn't accept the binary :), now done.  Will NOW show on next publisher run
[20:27] <Daviey> SpamapS: Hope you didn't mind me accepting it.
[20:27] <mcb_>  sarnold:  No. No adicional module needed. And if it was an module issue i should expect a "Module not found" erro message...
[20:28] <mcb_>  sarnold: This lack of erro message, or anything is the worst part...
[20:29] <sarnold> mcb_: indeed. Even lilo would print L I L O after each successful portion of loading to give you some hint what was going no...
[20:31] <mcb_>  sarnold: You made me think. It does not show that GRUB is entering level2 or level 1.5. In Debian system it show a message. Does in Ubuntu grub show in each stage it is?
[20:32] <sarnold> mcb_:  good question, I pay remarkably little attention when things work well. :(
[20:33] <mcb_> sarnold: LOL, but does not help me.... :) I think i will make a virtual server and look in a clean install....
[20:34] <mcb_> sarnold: At least i will know if GRUB in ubuntu talks some more or less....
[21:04] <SpamapS> Daviey: don't mind at all no. ;)
[21:04] <SpamapS> Daviey: as I said, I'm not an AA, so I wouldn't know to accept binaries ;)
[21:05] <Daviey> heh
[21:06] <Daviey> OK.. So.  should fortunes-ubuntu-server be kicked out of the archive.
[21:06] <Daviey> It was a package i created, and not been updated since 2010
[21:06] <Daviey> (kicked out for Quantal)
[21:10] <SpamapS> Daviey: what does it contain? old quotes from mathiaz and dendrobates ? ;)
[21:19] <Daviey> SpamapS: Helpful tips i put together.. pretty crappy
[21:19] <Daviey> There is a bug that many of the URL's no longer exist
[21:20] <SpamapS> Daviey: well yeah if its not helpful just drop it. Perhaps put out a call to the users who like it to keep it up to date or risk it fading into memory.
[21:24] <NcA> Any reccommendations for a parallel ssh client on Ubuntu? I'm needing to run commands on 5+ boxes simultaneously
[21:29] <SpamapS> NcA: heh... parallel-ssh comes to mind ;)
[21:31] <NcA> I checked it out, to elaborate a bit further... I'm trying to run sudo commands, so something interactive is preferred, as I need to auth with ssh, then when commands run
[21:31] <NcA> pssh just takes args and goes
[21:31] <SpamapS> NcA: this sounds like a really bad idea...
[21:31] <NcA> uh oh...
[21:32] <SpamapS> NcA: sudo? authing ? you sure you want to do this?
[21:33] <NcA> SpamapS: I fail to see the issue? If I'm connected to multiple servers via SSH, passing a password via the encrypted stream...
[21:34] <SpamapS> NcA: this is sort of why configuration management was invented..
[21:34] <NcA> SpamapS: We're currently working on Puppet...
[21:34] <SpamapS> NcA: "tell me what you want, now what you want me to do"
[21:34] <SpamapS> s/now/not/
[21:34] <SpamapS> NcA: no better time than right now to roll it out :)
[21:35] <NcA> SpamapS: So, on Mac, iTerm has a feature which you can initiate ssh sessions to multiple remote hosts, and run commands interactively, and simultaneuosly
[21:36] <sarnold> NcA: how does it handle if the authentication fails on one of the five hosts?
[21:36] <sarnold> NcA: how does it handle if one of the five hosts has a full disk and drives itself off a cliff with stupid/silly/incomprehensible errors?
[21:38] <SpamapS> NcA: there are ways to do that with the terminals on Ubuntu as well...
[21:38] <SpamapS> NcA: I'm suggesting that this is a great time to start using tools made for this job, rather than abusing tools not made for this job. :)
[21:44] <hallyn> ahs3: hey, are you around?  i've been asked to build a libnl3 version of netcf for experimental, but i can't find libnl-3-dev and libnl-route-3-dev in exp archives...
[21:44] <hallyn> do you see them?  what am i missing?
[21:45] <ahs3> hallyn: hrm.  haven't looked.  libnl-3-dev or libnl3-dev libnl-dev would be what i'd search for, tho
[21:46] <ahs3> whups -- not libnl-3-dev; that's an unlikely name; are you sure they've been uploaded to experimental?
[21:49] <hallyn> ahs3: well, not absolutely.  they've been in ubuntu for quite some time
[21:50] <hallyn> searching for libnl just gives me 1 and 2
[21:50] <hallyn> cyphermox: ^ do you know?
[21:50] <NcA> sarnold: SpamapS: sorry, got dragged away into firewall land
[21:50] <ahs3> hallyn: nod; you may need to poke the maintainer to get a libnl3 uploaded/available
[21:51] <NcA> the servers I intend to use this on are identical members of a cluster (torquebox to be specific)
[21:51] <SpamapS> NcA: identical.. until you run these commands, and then one fails.. then you have 4 identical, and one special snowflake. ;)
[21:51] <cyphermox> hallyn: just a second
[21:52] <sarnold> SpamapS: "special snowflake" :)
[21:52] <cyphermox> ah, libnl-3-dev is indeed the right name
[21:52] <hallyn> hm, so maintainer would be Heiko
[21:52] <sarnold> NcA: if the machines are really intended to be part of a cluster, could you just get ssh keys that allow you to log in directly as the user account that needs to be authenticated?
[21:52] <sarnold> NcA: or configure sudo to allow passwordless access?
[21:53] <cyphermox> hallyn: mbiebl asked me about libnl3 in exp/debian/jessie this morning
[21:53] <cyphermox> it's not in experimental but 3.2.7-4 is in sid
[21:53] <hallyn> waitaminute
[21:53] <hallyn> d'oh
[21:54] <hallyn> my ami is on squeeze+exp, not sid+exp
[21:54] <hallyn> sorry!
[21:54] <cyphermox> yeah
[21:54] <cyphermox> alright, good that this is sorted ;)
[21:54] <hallyn> ahs3: mind if i poke you in a bit (or likely tomorrow) about uploading netcf 0.2.2-1 to experimental?
[21:55] <ahs3> hallyn: nope.  holler when ready
[21:55] <hallyn> ahs3: thanks!  good night
[21:55] <cyphermox> hallyn: oh, you're the netcf maintainer in Debian. now I understand
[21:55] <hallyn> cyphermox: yeah, but it looks like i need to write a reliable user-data file for setting up my test instances :)
[21:57] <hallyn> there.  that shouldn't be happening again.
[22:25] <hallyn> ahs3: i need to do some testing, but http://people.canonical.com/~serge/netcf022/netcf_0.2.2-1.dsc is what i'm basing off of
[22:25] <ahs3> hallyn: ack