[00:01] <keithclark> maybe scrap?
[00:08] <keithclark> I unplugged, revbooted and plugged back in.  No access to the drive though
[00:08] <keithclark> This is messed up
[00:08] <escott> :-/
[00:08] <keithclark> Windows is much easier!
[00:12] <escott> keithclark, whenever anyone says that i always ask "have you ever actually tried to install windows"
[00:12] <keithclark> Yup
[00:14] <keithclark> I've installed Windows from 3.1 and up.  OS/2 as well
[00:15] <escott> i guess its what you are used to then
[00:15] <keithclark> yup but doesn't help me here
[00:16] <escott> keithclark, if you can be more specific about the errors when you try to boot we can try to offer suggestions
[00:16] <keithclark> I told you the error on bootup
[00:16] <escott> keithclark, "no accesss to the drive" that one?
[00:16] <escott> a little more context would help
[00:16] <escott> when do you see that during the boot
[00:17] <keithclark> It is stuck at 'mountall:  fsck /boot .....'
[00:17] <escott> keithclark, yes, and then you tried the install cd and what did it say
[00:18] <NomadJim> Already asked this in #ubuntu, but just found this channel.  I'm coming from debian. Is there much of a difference administrating an ubuntu server versus a debian one?
[00:19] <escott> NomadJim, its more or less the same people in this channel as others
[00:19] <keithclark> rebooting now....
[00:19] <escott> NomadJim, so crossposting is discouraged
[00:19] <NomadJim> heh
[00:20] <escott> also this channel is super quiet tonight
[00:20] <keithclark> escott, you are doing fine here@!
[00:21] <keithclark> escott, fsck /boot [390] terminated with status 1
[00:22] <escott> keithclark, so that indicates that /boot had errors but they were fixed
[00:22] <escott> keithclark, the important question is "why"
[00:22] <keithclark> escott, never gets past that
[00:22] <escott> keithclark, but that is you booting the installed system
[00:23] <escott> if we cant trust that /boot isn't corrupted then there isn't much point booting the system installed under /boot. so get your install media and boot that
[00:23] <keithclark> yes.  with the usb in.  If I take it out, it boots
[00:24] <escott> im really confused about what is going on now
[00:25] <escott> i thought it wasnt booting
[00:25] <keithclark> So am I
[00:25] <escott> booting the hard disk you get what
[00:25] <keithclark> If I boot no usb drive, it works\
[00:25] <escott> booting the usb you get what
[00:25] <keithclark> If I put the usb drive in no boot
[00:25] <escott> so why are you trying to boot the usb drive?
[00:25] <keithclark> I'm not
[00:26] <keithclark> Just trying to simply add a usb drive.  Should be easy
[00:28] <keithclark> This is a disaster
[00:28] <escott> keithclark, ok. so just boot the system with the usb out for now. then we can fix your fstab entry so that the usb boots
[00:28] <escott> rather so that the usb mounts
[00:28] <escott> because you probably typoed the fstab
[00:28] <keithclark> I copy/
[00:28] <escott> then when that is working we can make sure that you can cleanly reboot
[00:29] <keithclark> pasted the fstab
[00:30] <keithclark> machine rebooted with no usb  drive in
[00:30] <keithclark> drive plugged in
[00:30] <keithclark> (windows is much easier!)
[00:30] <escott> i thought you werent able to boot at all. but it boots without the usb so that means it fails to see the desired drive and passes on it but for some reason with the drive in there it is not passing
[00:31] <escott> basically mountall is trying to mount everything in fstab and for some reason this external is causing that to hang or slow down. perhaps the external is really large and hasnt been fscked?
[00:31] <keithclark> escott, I have no idea
[00:33] <escott> keithclark, lets just start with your fstab
[00:33] <keithclark> ok
[00:33] <escott> why dont you put that on paste.ubuntu.com
[00:35] <keithclark> http://pastebin.ca/2248397
[00:35] <escott> keithclark, the last line is all kinds of wrong
[00:36] <escott> see that line two above it that begins UUID thats what your line needs to look like
[00:36] <escott> UUID=the_uuid_goes_here_without
[00:36] <escott> UUID=the_uuid_goes_here_without_quotes_and_CAPITALIZATION_shouldnt_MATTER /media/external ext4 defaults 0 2
[00:37] <escott> keithclark, does that make sense?
[00:38] <escott> after making that change try to "mount /media/external" and see if it works without sudo. if it does then your fstab is good
[00:41] <keithclark> ok,boots
[00:41] <keithclark> >No access to the drive though
[00:41] <escott> where are you seeing that error message
[00:41] <keithclark> no error
[00:42] <keithclark> no driber
[00:42] <keithclark> no drtivber
[00:42] <escott> you mean once it is booted the /media/external is not mounted?
[00:42] <keithclark> yup
[00:42] <keithclark> No drive
[00:42] <escott> two possibilities
[00:43] <escott> (a) you need to "sudo update-initramfs -uk all"
[00:43] <escott> and if that doesnt work then (b) add "mount /media/external" to /etc/rc.local
[00:45] <keithclark> reboot on a?
[00:46] <keithclark> (windows was plug and play)
[00:46] <escott> yes
[00:48] <escott> keithclark, generally hotplugging a usb would cause it to mount, but that depends on your installing all the gui tools. ive been assuming that since you are in the server channel you did a server install
[00:48] <keithclark> yup, server install
[00:49] <escott> keithclark, so if you do "sudo apt-get install ubuntu-desktop" and login to a normal desktop you would have hotplug usb mounts
[00:49] <escott> but thats only for the logged in user
[00:49] <escott> problem is a server doesn't have a logged in user, so it doesn't know who to mount usb disks for
[00:49] <keithclark> should not need that
[00:49] <escott> which wouldn't matter for an ext4 but most usb disks are ntfs in which case it is important
[00:50] <escott> so rather than having it automount only if its a POSIX filesystem they just say
[00:50] <keithclark> Nope, usb not mounted
[00:50] <escott> you installed a server OS you presumably know how to configure it
[00:50] <keithclark> ok, so nobody is new at servers?
[00:50] <escott> keithclark, then its probably a race between mountall and the usb bus coming up or something
[00:51] <escott> add "mount /media/external" to /etc/rc.local
[00:51] <escott> most people who are new to ubuntu don't jump to the server install
[00:51] <keithclark> I did
[00:51] <escott> they would usually select the desktop os, and once they are familiar they would install the server
[00:51] <keithclark> I've been running Ubuntu in
[00:52] <escott> its fine to do what you did, but you did jump into the deep end
[00:52] <keithclark> since 2007u
[00:52] <escott> the server is meant to be lightweight
[00:52] <keithclark> yup
[00:53] <escott> optimized for the webservers/sql servers etc you run on it
[00:53] <escott> so lots of nice user friendly stuff is not setup by default
[00:53] <keithclark> yup
[00:53] <escott> if you are less performance sensitive you can install ubuntu-desktop
[00:53] <keithclark> Windows server is different
[00:54] <keithclark> Don't like it
[00:54] <escott> in which case you have a desktop that also has server applications
[00:54] <keithclark> Never the less, does not help me
[00:55] <keithclark> I guess I'm screwed?
[00:56] <escott> did you add mount /media/external to rc.local?
[00:57] <keithclark> there is nothing in meida
[00:59] <escott> keithclark, so "mount /media/external" is listed in rc.local above the exit 0 and you rebooted and its still not mounting?
[01:01] <keithclark> yup
[01:01] <minorix> hello all
[01:02] <escott> keithclark, that doesn't make much sense. if "mount /media/external" works then putting that in rc.local should also work. you could check the syslogs in /var/log
[01:04] <keithclark> well this is a stock ubuntu server install
[01:04] <keithclark> maybe a rething
[01:04] <keithclark> k?
[01:05] <escott> keithclark, can you paste your /etc/rc.local just in case there is something wrong there?
[01:07] <keithclark>   GNU nano 2.2.6                            File: /etc/rc.local
[01:07] <keithclark> #
[01:07] <keithclark> # By default this script does nothing.
[01:07] <keithclark> exit 0
[01:08] <keithclark> http://pastebin.ca/2248402
[01:09] <escott> keithclark, you need to put "mount /media/external" in the line just before "exit 0"
[01:09] <escott> and it should NOT have a # at the beginning
[01:10] <keithclark> I don't understand......windows is plug and play
[01:11] <escott> this isnt F**** windows
[01:11] <keithclark> now the mo
[01:11] <escott> if you want plug and play install ubuntu-desktop
[01:12] <keithclark> ubuntu-desktop=overhead
[01:12] <holstein> yeah, you just plug a USB stick in using nautilus, or most other GUI file managers keithclark
[01:12] <holstein> keithclark: that "overhead" is what auto mounts USB drives... and "plus and plays"
[01:13] <holstein> plugs*
[01:13] <escott> the whole "windows does it better" when i've told you three times to put this one line in rc.local and you haven't done it is kinda pissing me off
[01:15] <keithclark> escott to be honest,you've told me lots of one line fixes that have not rung true./
[01:16] <holstein> keithclark: and to be fair, you are asking for desktop functionality from a server OS
[01:16] <escott> keithclark, the first one you typed in wrong
[01:16] <keithclark> holstein, true
[01:16] <holstein> keithclark: you can use a desktop version of ubuntu and run all the server services.. might be easier to "get your feet wet"
[01:17] <escott> keithclark, when you install a server OS there is an expectation that you will be able to read and understand manpages
[01:17] <holstein> thats the way i did it.. i used linux for years before jumping into a headless commandline only rig.. and then i knew how to search google for how to mount a USB stick... or to ask the bots
[01:17] <escott> so if you would prefer i can just !fstab you and leave you on your own
[01:17] <holstein> !mount
[01:17] <escott> but its really rude when im trying to help you to constantly be complaining that windows is better
[01:18] <keithclark> escott, point taken
[01:18] <keithclark> And rightly so
[01:19] <escott> keithclark, if the last two lines of your rc.local are "mount /media/external" followed by "exit 0" then it should work the way you want
[01:20] <escott> having to put this in rc.local is a little unusual, but its possible with a USB disk that maybe the USB stack isn't up when mountall is run so it just doesn't mount the USB disk because it isn't there
[01:23] <keithclark> hmm didn't work....not mounted
[01:24] <escott> keithclark, but if you type "mount /media/external" at the console right now it does work
[01:25] <holstein> keithclark: what are you mounting? i typically just run "sudo fdisk -l" and i have a mount point created and sudo mount the partition i want to the location i want.. as in !mount
[01:25] <keithclark> there is nothing in /media
[01:26] <holstein> keithclark: i added one drive to !fstab for auto mounting on a server...
[01:26] <holstein> keithclark: i usually just specify
[01:26] <keithclark> there is nothing  in /media
[01:27] <escott> keithclark, this is what is driving me nuts. earlier you said when i asked you the same question, that "mount /media/external" was working and now its not?
[01:27] <keithclark> ok never mind
[01:27] <keithclark> i give up
[01:28] <holstein> keithclark: on mounting? manual mounting is not trivial.. have you read...
[01:28] <holstein> !mount
[01:29] <holstein> sudo fdisk -l lists the drives and partitions... then you can just follow the commands to mount to where ever you lke
[01:29] <holstein> lkike*
[01:29] <holstein> like**
[01:30] <keithclark> holstein, yup, windows just does it
[01:30] <escott> alright im leaving this channel before i start throwing stuff
[01:30] <holstein> keithclark: sure.. so does ubuntu.. and other desktop os's
[01:31] <holstein> keithclark: you are not using an environment which supports auto mounting
[01:31] <keithclark> holstein, automouting is enabloed\
[01:31] <holstein> keithclark: ive never used automounting in a command line system
[01:32] <keithclark> holstein, there you go
[01:32] <holstein> keithclark: i have just mounted.. on the command line... as i am suggesting you do... i have also added to the fstab
[01:32] <holstein> keithclark: i have never wanted the overhead of automounting on a command line system
[01:33] <holstein> keithclark: i cant imagine the work flow where you would want a usb stick auto mounting with a server os.. though, everything is open and anything is possible
[01:33] <holstein> keithclark: i usually suggest using the desktop os, and getting used to the ins and outs of linux.. you are dealing with a lot of unknowns
[01:35] <holstein> keithclark: if i wanted something to auto mount with the server,, i would just add it to fstab.. though, since it is implied that the machine is more than likely going to stay on, its not so much of a concern
[01:35] <keithclark> Ok, Linux is not for me then
[01:35] <holstein> keithclark: you mean, ubuntu server is not for you then
[01:36] <keithclark> yes, that is the easiest
[01:36] <holstein> keithclark: if you want usb sticks to mount when you plug them i.. ubuntu can do that.. lots of os's, including windows can
[01:36] <holstein> keithclark: what is the easiest? ubuntu server?
[01:36] <holstein> keithclark: ubuntu server is *not* the easiest.. its the most "minimal" in some ways..
[01:41] <holstein> https://help.ubuntu.com/community/Mount/USB is what i use keithclark
[01:42] <holstein> for a typical ntfs drive sudo mount -t ntfs-3g /dev/sdb1 /media/external
[01:43] <holstein> OR, from ubuntu, xubuntu, ubuntustudio, kubuntu.. any of the desktop variants, as well as windows you can just plug it in the machine, open a file manager and start browsing[C
[01:57] <chilicuil> hi, good morning, has anyone tried maas in ubuntu 12.10?, I've made it performe in ubuntu 12.04, but right now in ubuntu 12.10 isn't working.., I've already downloaded the iso files.., however it don't provide ip's, the web interface show nothing about dhcp/dns management, and I installed maas-dns & maas-dhcp
[04:51] <James> hi
[04:51] <Guest96387> ok
[04:51] <Guest96387> help
[04:51] <Guest96387> how to reduce power when i hosting server
[04:51] <Guest96387> ?
[04:52] <holstein> reduce power consumption?
[07:04] <cornfeed> hello, i am trying to setup an irc server and have hit an odd wall. I have got inspircd running, and I can use telnet & ircII to test it locally, but i cant connect to it remotely for my life. I have no firewall, but just to be safe, iptables -L shows all ACCEPT
[07:05] <cornfeed> any ideas about how I can get remote connections to work? netstat -tap shows *:ircd
[09:16] <Iceman_B> lo
[09:20] <Iceman_B> how can I configure dhclient to keep requesting a lease until it gets one, upon server (re)boot?
[09:32] <webwurst> In Quantal you can create snapshots of runnig virtualized qemu-kvm-geuest with "$ sudo virsh snapshot-create-as". To freeze the guests filesystem you can add the parameter "--quiesce". In the guest you need to have a running qemu-guest-agent for this feature to work. But in Quantal this agent "qemu-ga" seems missing?
[09:32] <webwurst> I can find it in Precise in the package "qemu-kvm".
[09:33] <webwurst> Why is it removed from Quantal?
[10:10] <webwurst> Has anyone used "QEMU Guest Agent (qemu-ga)" on Ubuntu?
[10:12] <jamespage> webwurst, I feel I should have done now
[10:12] <jamespage> webwurst, its in the qemu-kvm package
[10:13] <jamespage> ah - I see your point now
[10:13] <jamespage> lemme take a look
[10:16] <koolhead17> jamespage: I hope we have https://bugs.launchpad.net/quantum/+bug/1069966 backported :)
[10:16] <koolhead17> for cloud archive
[10:17] <koolhead17> also hello jamespage :)
[10:17] <jamespage> koolhead17, as its already landed in stable/folsom it will make it
[10:17] <jamespage> koolhead17, can't say exactly when just yet tho
[10:17] <jamespage> koolhead17, morning!
[10:18] <koolhead17> cool. hopefully by this week? :)
[10:52] <taalas> Hi. I have started using Ubuntu Server on our application server a while ago, and so far I am very happy with it. One thing I would need advice with: Is there a way to (periodically) get info about my software raid status? Currently I am not sure if I would be notified if any errors occured? What's the best way to solve this?
[10:53] <xnox> taalas: install monitoring. There are lighweight options like: logwatch / logcheck
[10:53] <xnox> taalas: or you can setup nagios+check_mk. Check_mk is a utility that generates nagios configs for you and it has support to auto-detect raid.
[10:54] <xnox> taalas: that means you will get a nice history/web interface to check status as well as receiving notifications if it goes down.
[10:54] <taalas> monitoring is a package?
[10:54] <xnox> no. it's a statement. As in "You should invest time and create monitoring for your systems"
[10:55] <taalas> Ah ok, i wasn't sure if I would be able to succesfully configure a full Nagios install, check_mk seems like a good solution then.
[10:55] <taalas> I see (monitoring)...sorry for the confusion ^^
[10:56] <taalas> Will give Nagios a try then and possibly also monitor other things as well
[10:56] <taalas> Many thanks
[11:07] <ak5> hello, anyone around?
[11:11] <vezq> sure
[11:21] <ivoks> taalas: if somethign happens with mdadm raid, mdadm-monitor will send you an email
[11:21] <ivoks> you don't need to install anything
[11:21] <ivoks> of course, you do have to have MTA set up correctly
[11:26] <taalas> ivoks: yes, that was what I was hoping, thanks. The MTA should be working correctly (though I did have some confusion as to what the right way is to forward emails to root to another user, me)
[11:27] <ivoks> taalas: edit /etc/aliases and run newaliases
[11:28] <taalas> check_mk is great tip nonetheless, since I was looking for a simple monitoring solution for other components as well. Nagios3 seems to run fine from the package, I am not sure yet if check_mk needs to be started somehow, currently it doesn't seem to be running after the package installation
[11:29] <taalas> ivoks: yes, that's what I did. I did notice rare occasions though, where mails to root bounced. Quite possibly another mistake on my part though
[11:29] <ivoks> if you misconfigured mail server...
[11:30] <taalas> Other processes send mail just fine (although not to root)
[12:15] <xnox> Daviey: about servercloud-r-vmbuilder - I totally agree with "vmbuilder" should die in favor of launching cloud-images.
[12:16] <xnox> Daviey: I actually want a static download that can be quickly relaunched locally on the ssh+libvirt/kvm, instead of running slow debootstrap / creating my own images.
[12:16] <xnox> rbasak: ^^
[12:16] <xnox> =)
[12:16] <Daviey> xnox: \o/
[12:17] <Daviey> xnox: seen cloud-localds?
[12:17]  * xnox is talking as a person who had to run "internal cloud" which was a single beefy KVM host with no openstack magic.
[12:18] <rbasak> xnox: I have a work item to wrap cloud-localds into a friendly one-step tool for "I want a virtual machine NOW"
[12:18] <xnox> Daviey: was that around in 2011? First time I see it =) looks nice.
[12:18] <xnox> rbasak: + add blogging =)
[12:18] <rbasak> xnox: juju deploy blog :-P
[12:19] <xnox> rbasak: in soviet russia, cloud deploys YOU =)))))
[12:19]  * xnox is on Wednesday of my post-UDS processing
[12:20] <Daviey> eek
[12:23] <xnox> Daviey: what does maas/juju use python-oauth for?
[12:23] <Daviey> xnox: maas uses it to create an authenticated metadata service.
[12:24] <xnox> Daviey: between master and it's nodes?
[12:24] <Daviey> right
[12:24] <xnox> ack.
[12:26] <xnox> Daviey: that basically means maas/juju doesn't need porting to python-oauthlib \0/ less work =)
[12:26] <Daviey> sounds good... i thiink
[12:28] <xnox> Daviey: yeah. We were in the session of s/python-oauth/python-oauthlib/ but were not sure about maas & friends. Turns out openstack is python2 and not moving any time soon & you use server bits from oauth (not just client bits).
[12:28] <xnox> Daviey: so the python3-oauthlib will be pushed on the desktop only, e.g. for client apps that need python3 porting.
[12:29] <Daviey> xnox: well...
[12:29] <Daviey> xnox: it's not captured... but something i wanted us to try and achieve this cycle is porting openstack depends to py3.
[12:29] <xnox> Daviey: you want python3-only on the server cd, while still using python2 code? =)))))
[12:29] <Daviey> So we can support openstack moving to py3 quicker.
[12:29] <xnox> I see, you do.... =)
[12:30] <Daviey> xnox: No... no hurry for py3 this cycle or next
[12:30] <Daviey> but if we can get as many deps to py3 as possible, we are getting near :L)
[12:31] <xnox> Daviey: the best way to do that, is to add an extra tab on the python3-only blueprint named something like "openstack" or "server". Barry is coordinating both core-dev & community/motu efford around porting packages there.
[12:31]  * Daviey feels co-ordinated already!
[12:31] <xnox> It is mostly tracking upstream (dead, doing the port, welcomes the port, patches available), status in debian and status in ubuntu.
[12:32] <xnox> Some of the rdeps were as easy as simply package needs patching.... others were more than that.
[12:32] <xnox> Daviey: qa team are very interested in python3-libvirt which is not there currently.
[12:33] <Daviey> rbasak: I suspect you are just triaging, but do you want to review the debdiff on bug 1074357 ?
[12:33] <Daviey> xnox: python3-libvirt is C bindings?
[12:34] <xnox> Daviey: I don't know =) qa is asking for it and I didn't have time to look at it last cycle.
[12:34] <Daviey> it's something that would scare me, i suspect.
[12:36] <Sander^work> Is it safe to do an apt-get upgrade compeard to dist-upgrade in a production enviroment?
[12:37] <xnox> Daviey: that's easy bit $ python3 -c "from medication import anxiety; anxiety.feed('Daviey')"
[12:37] <Daviey> ImportError: No module named medication
[12:38] <xnox> Sander^work: depends. apt-get upgrade is ok for -updates & -security. Between releases you want do-release-upgrade or apt-get dist-upgrade.
[12:38] <xnox> Sander^work: otherwise you will be stuck in a partial upgrade with packages from both releases.
[12:38] <Daviey> erm, dist-upgrade doesn't bounce between releases
[12:39] <xnox> Sander^work: sorry, I meant if you changed your sources to point to next release.
[12:39] <xnox> If you didn't change your source to point to new release, either upgrade or dist-upgrade should correctly apply updates/security.
[12:40] <Sander^work> xnox, Ok. What's the diffrence between upgrade and dist-upgrade when I dont change the source?
[12:40] <henkjan> Sander^work: smarter dependency handling in dist-upgrade. See 'man apt-get'
[12:41] <Daviey> Sander^work: it is ofc' calculated risk aswell.. Whilst there is a good level of regression testing between updates, it's possible an issue was introduced that causes you pain.  If the machine controlled my life support machine, i'd be slightly more careful.. But for most scenarios, sure
[12:41] <Daviey> You sould keep up with security updates regardless.
[12:41] <Daviey> If i was being more paranoid, i'd have a staging archive for local testing.
[12:43] <xnox> Sander^work: in essence dist-upgrade will consider removing packages to complete upgrades of larger number of packages. This may have unpleasant side-effects if you have non-dpkg managed software that depends on system packages that dist-upgrade decides to remove.
[12:45] <xnox> Sander^work: use dist-upgrade, but read the output it prints, especially the list of packages it tries to remove/additionally install.
[12:50] <Sander^work> xnox, the only diffrence is that it wants to install a new kernel.. from 2.6.32-28-server to 2.6.32-44-server
[12:51] <xnox> Sander^work: that makes sence, because the metapackage gained a dependency on a new package, which previously was not installed at all. Such change is considered "major" by apt-get upgrade.
[12:52] <Sander^work> xnox, question is if there is any security vulnerabilities in my old kernel.
[12:53] <xnox> Sander^work: for example why it is considered "major": it may break by hand compiled kernel modules that are not managed with dkms, e.g. system may fail to reboot into new default kernel.
[12:53] <xnox> Sander^work: well, the decision is yours if you want it or not =) "Power to the sysadmins"!
[12:54] <Sander^work> xnox, Yeah.. I've had a system not boot once, because virtual integration components wasnt compatible with a newer kernel.
[12:55] <xnox> Sander^work: and "security" vulnerabilities in that kernel may not apply to your use-cases & existing security measures. An offline server in a locked bunker doesn't really need security updates.
[12:57] <_ruben> and the other possibility: the sec vulns in the old kernel that *do* pose a risk for you, might still be present in the new one ;)
[12:57] <xnox> _ruben: I like that one =)
[12:59] <Sander^work> xnox, a local root vulnerability is enough for me to upgrade.
[13:01] <xnox> ack.
[13:16] <rbasak> Daviey: elinks debdiffs look perfect
[13:18] <ironm> Hello. When I try to install a solaris 11.1 VM on KVM host (last ubuntu-server 12.04) I am getting the following boot error: http://rsync.it-infrastrukturen.org/.solaris111-kvm-boot-error/solaris11.1-kvm-boot-error.png
[13:19] <ironm> config: http://rsync.it-infrastrukturen.org/.solaris111-kvm-boot-error/sol111vm.xml
[13:19] <ironm> Do you have any idea how to trace it? (kvm debug level or whatever ...). Thank you in advance for any hints.
[13:26] <Daviey> rbasak: thanks
[13:35] <zul> yo
[13:36] <koolhead17> zul: Y0 YO :)
[13:38] <ivoks> hi all
[13:49] <koolhead17> hi ivoks
[14:05] <zul> hallyn: ping when you are awake/conscious/non-jetlag/whatever
[14:07] <hallyn> next week?
[14:07] <hallyn> breakfast, will ping you in a bit
[14:08] <patdk-wk> I hope my breakfast isn't able to move, let alone ping
[14:09] <hallyn> bah!   makes a boring life
[14:09] <hallyn> i'll ping my lunch later
[14:13] <zul> hallyn: k
[14:51] <dimitrig_> hi, what does this line from top mean, when it comes to cpu useage on all my cpu cores: Cpu(s): 65.3%us, 32.6%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  2.1%si,  0.0%st
[14:52] <dimitrig_> i.e. the 'us' , 'ni' , etc
[14:54] <dimitrig_> ok, i found the answer in man pages, sry
[14:59] <jdstrand> Sander^work: for more information on vulnerabilities that were fixed in the latest kernel, you can look at Ubuntu Security Notices (http://www.ubuntu.com/usn/)
[15:05] <hallyn> zul: oh, sorry, i got carried away on email :)  what's up?
[15:05] <zul> hallyn: so i have libvirt 1.0.0. but im getting a weird error when starting domains
[15:05] <Iceman_B> something is seriously hogging my line, so my ssh connection is slooo-o-o-o-o-o-o-w too, how can I get real-time info about my network stats?
[15:06] <hallyn> zul: qemu, lxc, or both types of domains?
[15:06] <zul> hallyn: qemu mainly havent gotten to lxc yet http://pastebin.ubuntu.com/1334899/
[15:07] <hallyn> ugh
[15:07] <hallyn> jdstrand: do you know if there have been a set of LSM for libvirt patches lately from upstream?
[15:07] <zul> hallyn: i can throw it up if you want to play with it
[15:08] <hallyn> zul: i need to do some bookeeping first to make sure etherpads don't go down before i save all their contents, but yeah
[15:08] <zul> hallyn: k
[15:09] <hallyn> zul: if you jsut wanna push to a ppa i'll grab the source as well
[15:09] <rbasak> Iceman_B: try iftop and/or mtr
[15:09] <zul> hallyn: i was just going to upload the source
[15:09] <jdstrand> hallyn: they actually have done quite a bit in there lately
[15:10] <jdstrand> that error is quite generic and I haven't looked at the upstream changes closely
[15:10] <hallyn> zul: ok
[15:11] <hallyn> jdstrand: drat - thanks.
[15:11] <zul> hallyn: have fun with that one :)
[15:11] <jdstrand> hallyn: if I had to guess, I would look at this series: "[libvirt] [PATCH 0/3] apparmor: bug and typo fix and add tapFD relabeling"
[15:12] <hallyn> jdstrand: thanks
[15:23] <zul> shazbut
[15:25] <Iceman_B> rbasak: tnx
[15:33] <Iceman_B> rbasak: neither of the program,s are present
[15:41] <rbasak> !apt | Iceman_B
[16:14] <SpamapS> rbasak: hey, btw, are you going to be able to fix apt this cycle?
[16:14] <SpamapS> Failed to fetch bzip2:/var/lib/apt/lists/partial/us-west-2.ec2.archive.ubuntu.com_ubuntu_dist
[16:14] <SpamapS> s_precise-updates_universe_source_Sources  Hash Sum mismatch
[16:14] <SpamapS> never fails to hit when I need things to work most
[16:14] <SpamapS> :-
[16:15] <rbasak> SpamapS: I'm hoping to do it this cycle, yes. Just sorting out exactly what my schedule will be this week
[16:15] <rbasak> Interestingi that you got it in precise-updates
[16:15] <SpamapS> rbasak: kind of laughable to think you could sort 6 months into the future, isn't it? ;)
[16:15] <rbasak> SpamapS: I need to get it done by FF :-/
[16:16] <SpamapS> rbasak: yeah, updates and security are still vulnerable.. just less churn than dev release
[16:16] <rbasak> SpamapS: we knew it was theoretically vulnerable, but wasn't aware of anyone hitting it in the wild after the cache header tuning on archive.u.c
[16:17] <ironm> Hello. Can anyone confirm the following issue, please? When I try to install a solaris 11.1 VM on KVM host (last ubuntu-server 12.04) I am getting the following boot error: http://rsync.it-infrastrukturen.org/.solaris111-kvm-boot-error/solaris11.1-kvm-boot-error.png
[16:17] <SpamapS> rbasak: since this is the S3 mirrors, its entirely possible its only a problem with those
[16:17] <ironm> config: http://rsync.it-infrastrukturen.org/.solaris111-kvm-boot-error/sol111vm.xml
[16:17] <ironm> Do you have any idea how to trace it? (kvm debug level or whatever ...). Thank you in advance for any hints.
[16:17] <SpamapS> rbasak: I hit it quite often when there are updates actually
[16:17] <rbasak> Hmm
[16:18] <rbasak> Anyway all the more reason to fix it :)
[16:18] <SpamapS> some of the charms ignore and move on.. others stop dead
[16:18] <SpamapS> rbasak: seems so straight forward, but I suppose the politics are as tricky as the solution
[16:19] <rbasak> SpamapS: I think the politics are sorted now. I just have to implement a production-quality apt with a fallback. The fallback is a bit awkward to do in the spaghetti that is the apt pseudo-event-driven download code
[16:20] <rbasak> Interesting that the juju user story only needs apt and not debootstrap.
[16:23] <SpamapS> rbasak: debootstrap was removed from the equation with juju 0.6 using cloud images for containers
[16:23]  * rbasak nods
[16:47] <hallyn> Daviey: oh, you misspoke at the uds summary :)  we are NOT going to not autostart virbr0 in general.  Only if 192.168.122.x is already in use.
[16:49] <Daviey> hallyn: wait, i thought i said.. We are ceasing to autostart?
[16:50] <Daviey> hallyn: if i did get it wrong, it looks like only you noticed... So don't tell anyone.. we'll keep it our secret. :)
[16:53] <hallyn> Daviey: yes, you said we are ceasing to autstart.  that is not the case
[16:53] <hallyn> somebody cheered when you said it, so i got a bit worried :)
[16:53] <hallyn> "we'll keep it our secret" - i can just hit 'backspace' in the irc logs right?  :)
[17:05] <Daviey> hallyn: i thought your irc 'client' allowed /bin/rm?
[17:08]  * jpds puts more apparmor on Daviey's client.
[17:16] <hallyn> heh
[18:08] <hallyn> zul: (phew) done with bookkeeping for a bit (until Daviey yells at me :)  have you pushed the libvirt src you were working with?
[18:08] <zul> hallyn: yep http://people.canonical.com/~chucks/libvirt/
[18:08] <hallyn> might duck out for lunch first.  hey lunch - I'm PINGING You!
[18:08] <hallyn> thanks
[18:19] <RoyK> any news on sanlock for ubuntu_
[18:19] <RoyK> ?
[18:19] <RoyK> setting up a KVM cluster on some shared filesystem is a PITA without it...
[18:22] <hallyn> RoyK: DIUI it's packaged in debian and not in ubuntu?
[18:24] <hallyn> ah i see the needs-packaging bug
[18:24] <elijahchancey> hey everybody. i think there's an issue with the apt repos hosted on AWS. who can i talk to about this?
[18:25] <hallyn> ivoks: is bug 882485 interesting to you?  Interesting enough to help push it?
[18:29] <sarnold> elijahchancey: It's probably a known issue, a resync has been started
[18:29] <elijahchancey> sarnold: will the resync fix the issue described by: http://askubuntu.com/questions/209844/latest-ec2-ubuntu-instance-seems-broken ?
[18:30] <elijahchancey> sarnold: also, how long does that process typically take?
[18:31] <sarnold> elijahchancey: that hash-mismatch was exactly the error I saw reported thta kicked off the resync
[18:31] <sarnold> elijahchancey: they suggested an hour is reasonable, and four hours is the "it should not take longer than this"
[18:32] <med_> stupid question time: why is there a manifest generated for ubuntu-####-desktop but not ubuntu-####-server in releases.ubuntu.com
[18:32] <med_> ?
[18:32] <med_> is that because u-server is more likely to download packages at install time?
[18:34] <elijahchancey> sarnold: great. thank you so much!
[18:36] <sarnold> elijahchancey: you're welcome :)
[18:42] <RoyK> hallyn: DIUI?
[18:44] <hallyn> jdstrand: oh i see, the patces you cited might *fix* the issue :)
[18:48] <hallyn> RoyK: missed a word there :)  do i understand it *right*
[18:50] <hallyn> RoyK: i'll follow up on that, that needs to get packaged, thanks.
[18:50] <RoyK> :)
[18:50] <hallyn> (s/get packaged/pushed into the archive)
[18:52] <RoyK> looks like the package existing is ok
[18:52] <RoyK> but I had some issues with it and moved to centos for this kvm test-cluster
[19:00] <hallyn> stgraber: http://lists.linuxfoundation.org/pipermail/containers/2012-November/030813.html  proposed deprecation of clone-children
[19:17] <_KaszpiR_> http://i49.tinypic.com/rhl74k.png
[19:17] <_KaszpiR_> so I've decided to make an experiment and found out that my server strange locksups
[19:17] <_KaszpiR_> *lockups
[19:18] <_KaszpiR_> as we can see i'm connected via putty over ssh and can see the htop running
[19:18] <_KaszpiR_> but closing app will end in freezeof the console
[19:23] <acidflash> x
[19:32] <RoyK> _KaszpiR_: setup netconsole or an old-fashioned network console, or perhaps syslogging to another server. you may get some results from that, like an OOPS or perhaps a Panic
[19:32] <_KaszpiR_> ah right, i didn; have time for that/forgot
[19:32] <_KaszpiR_> now time to do it
[19:32] <RoyK> syslog may die during such an event, so better use a console thing
[19:43] <sarnold> heh, apt-get purge nano is the second or third step I take on my new installs..
[19:51] <_KaszpiR_> RoyK can I send netconsole to a couple of addressess?
[19:52] <sarnold> yes
[20:21] <smoser> jamespage, still around?
[20:24] <_KaszpiR_> RoyK ok, netconsole set up
[20:25] <_KaszpiR_> time to squeeze the shit out of that box
[20:25] <_KaszpiR_> ;D
[20:50] <Daviey> hallyn: Hey, are you planning to get libvirt 1.0 into ~ubuntu-virt PPA?
[20:50] <NomadJim> ext3 is the standard file system for all linux servers pretty much right?
[20:50] <_KaszpiR_> NomadJim yes
[20:51] <Daviey> erm, probably ext4 now.
[20:51] <NomadJim> cool wanted to double check before diving into learning one
[20:51] <Daviey> there isn't exactly much to learn
[20:51] <NomadJim> that's good to know
[20:52] <Daviey> it's like learning how a lightbulb works :).. Sure, if you are interested... otherwise just use it. :)
[20:52] <Daviey> btrfs has stuff to learn.
[20:52] <NomadJim> i want to know if a filesystem impacts how much data is sent over a network
[20:52] <NomadJim> is the main reason i'm looking into it
[20:52] <_KaszpiR_> hm
[20:52] <Daviey> erm
[20:52] <NomadJim> given that the same file is stored but on different file systems
[20:53] <_KaszpiR_> it's rather more dependand ont he services
[20:53] <patdk-wk> in reality? unlikely
[20:53] <_KaszpiR_> like access the same file via rsync,samba, nfs
[20:53] <patdk-wk> how quickly a filesystem can locate data, via seq, random, ... depends
[20:53] <patdk-wk> but normally the disks themselfs slow it down more
[20:53] <Daviey> in the case of rsync, fat32 tends to transfer more than a linux filesystem :)
[20:54] <NomadJim> interesting hmm
[20:55] <_KaszpiR_> cause fat32 has specific 'modified' times
[20:55] <NomadJim> i was thinking of a border case where you have many small files
[20:55] <NomadJim> and there might be a potential for a lot of empty space per file
[20:56] <_KaszpiR_> there's reiserfs
[20:56] <Daviey> NomadJim: I bet my hand you will struggle to measure a difference.
[20:56] <_KaszpiR_> or xfs
[20:56] <Daviey> especially with the kernel (& readahead) doing caching
[20:57] <NomadJim> love me some caching
[20:58] <NomadJim> On most file systems you can specify block size right?
[20:58] <NomadJim> so my border case isn't really file system dependent i guess?
[21:01] <hallyn> Daviey: well zul is packaging it, but i'm tossing him a debdiff right now to fix a snafu in qemu+apparmor behavior
[21:01] <hallyn> Daviey: i think it was just going to go into the archive...
[21:01] <hallyn> or did you mean for quantal?
[21:05] <Daviey> hallyn: precise/quantal
[21:06] <Daviey> hallyn: seems to be interest, http://t.co/uvptv48w
[21:06] <hallyn> Daviey: perhaps, but (a) i'd have to make sure the versions in there now aren't serving a unique need, and (2) they'd require backproted newer qemus as well
[21:06] <Daviey> gah
[21:06] <hallyn> iirc 1.0 requires qemu 1.2 or somesuch
[21:07] <hallyn> bc they use a newer api instead of cvhecking 'qemu --help' output :)
[21:07] <hallyn> according to someone at the uds session.  i haven't checked
[21:07] <_KaszpiR_> hm maybe the server was locking due to the dust? XD
[21:08] <hallyn> zul: your libvirt-1.0.0 plus http://people.canonical.com/~serge/libvirt-1.0.0-aa.debdiff  launches qemu vms for me
[21:08] <hallyn> (haven't run a full qrt)
[21:09] <hallyn> jdstrand: thanks!  just used two of the patches in the thread you pointed to, al lthat was needed  \o/
[21:09] <jdstrand> hallyn: awesome! :)
[21:09] <hallyn> guess i may as well run qrt on this instance now :)
[21:11] <RoyK> NomadJim: most systems use a filesystem for stuff on drive and then some networked filesystem like nfs or smb/cifs or afs or whatever for transferring data between filesystems. unix filesystems have datestamps too, mtime, ctime and atime
[21:11] <RoyK> atime isn't too much used these days on large setups, since inducing a write on every access is somewhat bad to performance
[21:15] <sarnold> .. and drive lifetimes, with limited write cycles
[21:16] <NomadJim> RoyK:  thanks didn't know about the networked filesystem for transferring
[21:18] <NomadJim> RoyK:  this is something that you have to go out of your way to setup though right? nfs? It's not like the command line tools like scp, rsync, etc use it behind the scenes
[21:24] <RoyK> NomadJim: nfs and smb and the rest are used to mount filesystems like they were local
[21:25] <RoyK> NomadJim: just like in windows where you mont \\somehost\somesher to G:
[21:25] <RoyK> NomadJim: just like in windows where you mont \\somehost\someshere to G:
[21:26] <RoyK> NomadJim: NFS was introduced in 1989, and has improved rather a bit since then...
[21:28] <stgraber> hallyn: might be worth passing --dhcp-leasefile to lxc's dnsmasq with a value other than /var/lib/misc/dnsmasq.leases (the current default)
[21:29] <stgraber> hallyn: otherwise all the dnsmsq on the system end up using the same lease file, making dhcp pretty random :)
[21:29] <stgraber> (mostly a problem when also using libvirt I guess)
[21:29] <hallyn> stgraber: sounds worthwhile, yes
[21:30] <hallyn> stgraber: so that's the default, and so rebooting the host in general will provide persistent mappings as one would expect?
[21:30] <stgraber> ah, actually libvirt is setting --dhcp-leasefile already, so it's really just lxc that'd need fixing
[21:30] <stgraber> yep, if it's set to a path that's not wiped on reboot, we'll have persistent IPs across reboots
[21:31] <stgraber> I suppose we could use /var/lib/misc/dnsmasq.$INTERFACE.leases => /var/lib/misc/dnsmasq.lxcbr0.leases
[21:31] <stgraber> that'd make it easy to find (same path as dnsmasq) and avoid any potential clash
[21:35] <hallyn> stgraber: worth opening a bug for, or will we remmeber next time we're pushing?
[21:38] <stgraber> hallyn: I'll just push the change to bzr
[21:38] <stgraber> hallyn: I'll also make dnsmasq start with -K (authoritative) so that if you move containers from a bridge to another you don't need to wait 15s or whathever the DHCP timeout is
[21:38] <hallyn> do you do that often?
[21:39] <stgraber> not terribly often but there's no good reason not to run the DHCP server as authoritative when we're sure it's the only DHCP server on the network

[21:40]  * RoyK works too much with microsoft systems
[21:42] <stgraber> hallyn: pushed to bzr
[21:42] <RoyK> it's something like drinking too much methanol, making you drunk, irrational and blind all that the same with little to gain
[22:22] <hallyn> jdstrand: d'oh, qrt fails bc 'virsh capabilities' now lists a secmodel entry for 'none' instead of not listing it
[22:40] <blkperl> Hi, I'm having trouble getting a Dell PowerEdge R610 to load ubuntu, have install it boots into a black screen
[22:40] <blkperl> s/have/after
[23:04] <blkperl> figured it out needed to set a rootdelay for grub in our kickstart config
[23:06]  * TheLordOfTime yawns.
[23:06] <TheLordOfTime> evening everyone.  anyone here bugcontrol want to review whether i'm correct on an old php5 bug?
[23:06]  * TheLordOfTime could self-review, but would rather someone else check his work :P
[23:06] <TheLordOfTime> s/work/research./
[23:09] <halvors> How do i setup isc-dhcp-client to request a IPv6 prefix via DHCPv6 Prefix Delegation?
[23:17] <halvors> How do i setup isc-dhcp-client to request a IPv6 prefix via DHCPv6 Prefix Delegation?
[23:19] <blkperl> halvors: you don't have to ask multiple times, its just a matter of waiting for someone with the knowledge to look at the window
[23:19] <blkperl> s/the window/this channel
[23:41] <halvors> blkperl: Ah ok, sorry :) Just that this problem is driving me crazy, i have to do this. :-S
[23:50] <sarnold> halvors: I can't volunteer him to help you, but at UDS dtaht mentioned that exact problem...
[23:57] <TheLordOfTime> halvors, you realize patience is a virtue with everything right?