[00:00] <Diegonat> no infact im using it at home
[00:00] <Diegonat> but i dont have any hipervisor listed? do you know why?
[00:01] <Diegonat> i cannot run any istance
[00:03] <Diegonat> pleia2
[00:03] <Diegonat> any idea?
[00:03] <pleia2> Diegonat: it always works fine for me, you'll really want to ask in #openstack for specific problems :)
[00:06] <zul> hallyn_:  still around?
[00:07] <zul> hallyn_:  im getting some weirdness http://paste.ubuntu.com/6121749/
[00:27] <sarthor> How to check that, my VGA Card Memory is 256, 512, 1GB or How much?
[00:49] <Guest75335> samba server on ubuntu 12.04, moving 309.9GBs to Mint desktop = 5 hours to do, over GB network, does that sound right ???
[00:51] <sarnold> Guest75335: I think that works out to 17 megabytes per second; 100 mbit ought to do around 9 megabytes per second, so you're going at perhaps twice the speed of a 100 mbps network. better than nothing but not full speed of gigabit.
[00:52] <sarnold> Guest75335: depending upon the protocol you're using, the source and destination drives, and makeup of files involved, that might be alright, or it might be slow...
[00:53] <sarnold> for example, I'd be bloody thrilled to get those speeds off my pandaboard :) that's a slow-ass little thing. but I'd be a bit sad about that speed between two high-end i7 machines with SSD drives..
[01:15] <Guest75335> ok thanks man
[02:39] <dre101> hey guys
[02:39] <dre101> dr-willis  i would like to map my pause key to run a shell script when it is press
[03:01] <rafi878> hello
[03:01] <rafi878> anyone here use windows powershell?
[03:35] <subman> is there an easy way to setup an ubuntu server thin client server?   Wouldn't this be a great project for a bunch of netbooks and a high performance server?
[03:37] <subman> Maybe a current how to?
[03:38] <RobbyF> what do you mean a thin client server?
[03:38] <RobbyF> VNC?
[03:49] <subman> RobbyF,  I want my minimal netbooks to be able to boot off the server and run gui apps there.  Instead of trying to run them on the limited resource machines.
[03:52] <RobbyF> I think in 13.04 + desktop editions they support remote logins under ubuntu one accounts or something along those lines. I'm not sure if server has it by default or not. Otherwise something like TightVNC might be helpful.
[03:54] <subman> I thought the ltsp project (or something like that) did thin clients
[03:56] <qman___> yes, the project is LTSP
[03:56] <qman___> and IIRC it's an installer option on the alternate CD, rather than the server CD, since alternate has all the GUI stuff you need for it
[04:22] <subman> anyone else with a good tutorial an a virtual machine server?
[08:42] <jamespage> rbasak, hey
[08:42] <jamespage> rbasak, are you taking care of bug 1162139
[08:42] <jamespage> ?
[08:42] <rbasak> jamespage: o/
[08:43] <rbasak> jamespage: just doing it now actually
[08:43] <jamespage> rbasak, marvellous!
[08:43] <jamespage> rbasak, hows things?
[08:43] <rbasak> (assuming it builds OK)
[08:43] <rbasak> Good thanks
[08:43] <rbasak> How's the family?
[08:43] <jamespage> rbasak, doing well!
[08:43] <rbasak> Excellent!
[08:49] <vedic> Hey guys, not ubuntu server question but based on your experience: can you suggest tool for drawing client-server communication diagram like this: http://fresherstep.blogspot.in/2011/08/tcp-server-communication-udp-client.html
[08:52] <rbasak> vedic: dot for the programmatic way. dia, xfig or libreoffice for a GUI.
[08:55] <vedic> rbasak: Thank you
[09:11] <vedic1> rbasak: From dia, lifeline can used for this. Thanks
[10:35] <brendand> hey, i'm trying to run qemu in -nographic mode on a server. anyone know how to see what's going on while it's running?
[10:44] <lerra> Hi,
[10:44] <lerra> Where can i find a debug kernel for 2.6.32-42-server so I can analyse a kerneldump ?
[10:47] <rbasak> lerra: take a look at https://wiki.ubuntu.com/Kernel/KernelDebuggingTricks#Using_GDB_to_find_the_location_where_your_kernel_panicked_or_oopsed.
[10:47] <rbasak> lerra: also try #ubuntu-kernel
[10:48] <lerra> Aah, thanks :)
[11:22] <Sling> I just updated my Ubuntu 12.04 PXE boot environment to 12.04.3, by using all files in the install/netboot/ubuntu-installer/amd64 directory, yet when booting that kernel and entering the setup menu, it complains about the CD-ROM not being found..  I can't find a reference to a media location anywhere in the install/ folder, where would the old CD contents be on the PXE host?
[11:22] <Sling> the PXE host is a 12.04.2 installation
[12:02] <rbasak> Sling: the PXE host doesn't contain the CD contents. For PXE installation, the installer will fetch them directly from a mirror.
[13:19] <Pici> Is there a good way of deleting messages deferred postfix messages older than a certain date?
[13:19] <Pici> I have some regular status emails that go out, and its been broken since yesterday and once I fix it I don't want my users getting 24 status emails all for the same thing.
[13:23] <rbasak> Pici: mailq and some text manipulation should be able to do that, right? IIRC, postfix does have a way to remove a message by ID, and mailq reports those IDs I think.
[13:23] <rbasak> Pici: or alternatively grep the logs for message IDs?
[13:24] <Pici> rbasak: I figured it would be something like that.  A point in the right direction is all I need, thanks.
[13:46] <patdk-wk_> I think it's pretty simple to google
[13:46] <patdk-wk_> search for someting like pfdel
[13:47] <patdk-wk_> I know I based mine off ones I found on the net
[13:55] <Pici> I just grepped through mailq and piped what I needed to postuser
[13:56] <rbasak> That sounds reasonable
[13:56] <Pici> er, postsuper*
[14:21] <Guest93790> anyone can help me, I have configured a mail server two days ago was working great, now I have in the queue of many emails from yahoo mail that are coming, and I can not block out those emails, delete them but return all my emails to send
[14:41] <hallyn_> zul: what is the weirdness?  (and are you still getting it)
[14:41] <zul> hallyn_:  yeah im still getting it
[14:41] <zul> the connection refused in the strace
[14:41] <hallyn_> are the containers running?
[14:41] <zul> yeah
[14:41] <zul> they were
[14:42] <zul> give me a sec and ill get you an updated one
[14:42] <hallyn_> hm
[14:46] <hallyn_> zul: this is on saucy I assume?  uptodate?
[14:46] <zul> as of this morning
[14:46] <hallyn_> after i run this test i'll do a fresh run and see what hapepns for me
[14:46] <zul> im putting this on an instance and get you access
[14:47] <hallyn_> k
[14:53] <resno> my server has run out of space on "/". where can i look to drop some files quickly?
[14:54] <hallyn_> /var/cache
[14:54] <resno> eh, its only 36M
[14:55] <hallyn_> /var/log?
[14:55] <resno> i havent installed some needed updates would this help or only hurt
[14:56] <rbasak> Extra old kernels that you no longer need, perhaps? Check /boot.
[14:57] <hallyn_> and /lib/modules
[14:57] <rbasak> (but don't remove the files directly - remove the old kernel packages)
[14:57] <hallyn_> but no, don't install updates while out of disk space!
[14:57] <hallyn_> that recently happened to me on two laptops, recovery was not fun :)
[14:58] <resno> hallyn_: how to remove things out of /lib/modules?
[15:01] <hallyn_> resno: well as rbasak said remove the kernel packages themsleves and those should remove the /lib/modules dirs you no longer need
[15:01] <hallyn_> (i do rm -rf them, but that's not recommended, and i pay for it later)
[15:02] <resno> heh
[15:02] <resno> i dont want to pay later
[15:02] <resno> a happy server is a good server
[15:42] <med_> any squid or ceph experts -- http://www.sciencefriday.com/video/08/05/2011/where-s-the-octopus.html  awesome (but totally unrelated to Ubuntu Server)
[16:07] <xibalba> is there a way i can, as a non root user, make it so an app such as ZNC starts up when the server boots up?
[16:07] <patdk-wk_> cron :)
[16:07] <xibalba> ok that makes sense, didn't think of that
[16:07] <xibalba> cron to call a script to check if the app is running, if not, start it
[16:07] <xibalba> any other ways?
[16:07] <patdk-wk_> maybe
[16:07] <patdk-wk_> or just use @reboot
[16:08] <xibalba> oh dang didnt know about that parameter
[16:08] <patdk-wk_> it's only in some versions of cron
[16:08] <RoyK> it's in gnu cron from since like 5 years back (or 10?)
[16:08] <xibalba> great i'm going to give it a try now
[16:09] <patdk-wk_> thought it was just in vixie cron
[16:09] <RoyK> sorry
[16:09] <RoyK> I meant vixie cron
[16:10] <patdk-wk_> well, looks like gnu cron (will, or does) support it, but it's not even beta yet
[16:10] <xibalba> brb trying it now
[16:13] <xibalba> excellent it worked
[16:13] <xibalba> thoguh my resolv.conf is wiped on reboot
[16:14] <RoyK> xibalba: it's generated by your settings in /etc/network/interfaces
[16:15] <xibalba> why did they modify that functionality?
[16:15] <xibalba> i get it for a laptop or desktop
[16:15] <xibalba> but for a server, resolv.conf should be it
[16:15] <xibalba> i have in my interface under 'auto eth0' dns-nameserver 192.168.1.2
[16:16] <RoyK> xibalba: please pastebin the interfaces file
[16:16] <patdk-wk_> xibalba, heh? read the *release notes*
[16:16] <RoyK> ah
[16:16] <RoyK> dns-nameservers
[16:16] <RoyK> plural
[16:16] <xibalba> http://pastebin.ubuntu.com/6124557/
[16:16] <xibalba> ah ok
[16:16] <xibalba> let me try again
[16:17] <xibalba> god damn ubuntu boots fast on vmware
[16:17] <xibalba> thank you royk
[16:17] <RoyK> np :)
[16:18] <xibalba> im still so friggin amazed at the boot time
[16:18] <xibalba> centos is fast, but not this fast
[16:18] <xibalba> like 2 seconds for me
[16:18] <RoyK> no idea what CentOS/RHEL is doing during bootup
[16:20] <patdk-wk_> systemd
[16:21] <RoyK> wasn't systemd also considered for ubuntu?
[16:23] <patdk-wk_> not really, it was later to the game
[16:23] <patdk-wk_> and some thought went into it, but was too heavy into upstart by then
[16:23] <RoyK> nested raids still don't work with upstart, though
[16:23] <RoyK> no idea why
[16:24] <patdk-wk_> lots of things have issues, have to tune the startup of many things
[16:24] <patdk-wk_> like postfix has to start after dovecot (to use dovecot auth/lmtp)
[16:24]  * RoyK considers using debian on his next server install
[16:25] <patdk-wk_> what I really really wish someone would fix, is the iso images
[16:25] <patdk-wk_> have you heard how much damned abuse booting an iso causes?
[16:26] <patdk-wk_> think it would be simple to reorder the files using the ureadahead file, to make the iso not cause so many seeks
[16:26] <patdk-wk_> and boot the iso in <60seconds, instead of 5min
[16:29] <xibalba> weird i haven't used a cd iso in so long
[16:29] <xibalba> im all virtual baby
[16:29] <patdk-wk_> I do it to fix machines
[16:29] <patdk-wk_> non virtual machines
[16:30] <xibalba> i take it back,  i've only used esxi iso cds in the last year
[16:30] <xibalba> until i get some autoDeploy action going
[17:31] <zul> hallyn_:  fixed
[17:35] <smoser> rbasak, http://pad.ubuntu.com/T9qXx9IYHK
[17:35] <smoser> thoughts ?
[17:38] <patdk-wk_> heh? ubuntuone account required for pastebin now?
[17:39] <rbasak> Reading
[17:39] <rbasak> patdk-wk_: the SSO thing. It has been like that for years. AIUI we once had a spam/troll problem.
[17:39] <Pici> pad != pastebin
[17:39] <patdk-wk_> hmm, I have logged in using launchpad before, but haven't hit a ubuntuone one
[17:41] <rbasak> It's the same thing
[17:41] <rbasak> (launchpad/ubuntuone sign on)
[17:41] <rbasak> I didn't realise he'd said pastebin; thanks
[17:41] <Pici> And I'm pretty sure that pastebin only requires SSO if you want to see the raw version.
[17:43] <patdk-wk_> hmm, must of changed on me :)
[17:44] <patdk-wk_> it's just really complaining to me, since I hadn't use the ubuntu one way to login yet
[17:45] <rbasak> At least you don't have to use a second factor. I accept that a second factor is important for some cases. Requiring a second factor for access to the pad that expires daily is overkill.
[17:45] <smoser> paddymahoney, pad.ubuntu.com has needed it for quite some time.
[17:45] <patdk-wk_> :)
[17:45] <smoser> pici is right about pastebin.
[17:45] <smoser> only necessary for download of raw.
[17:46] <paddymahoney> smoser:  what?
[17:46] <smoser> paddymahoney, sorry. bad name complte.
[17:46] <rbasak> smoser: do you know if subarch/$subarch in ephemerals will work if it's a symlink?
[17:46] <paddymahoney> smoser np
[17:46] <smoser> http://paste.ubuntu.com/6124889/ is my 'upget' that scrapes out the raw data from ubuntu pastebin without needing auth.
[17:47] <smoser> so you can 'echo foo | pastebinit' . then elsewhere 'upget http://paste.ubuntu.com/abcdefg > out'
[17:47] <smoser> rbasak, i dont know. i dont particularly care
[17:47] <smoser> :)
[17:47] <rbasak> smoser: then you wouldn't have to overload "subarch"
[17:48] <smoser> ?
[17:51] <Diegonat> hi guys? I have installed devstack on a single machine and it was working. However, I restarted the machine and now openstack is not running. What do I need to do?
[17:56] <rbasak> smoser: would anything break if we shipped ephemeral images with no kernel installed inside at all?
[17:57] <rbasak> (eg. generate initramfs and then purge kernel)
[17:57] <SpamapS> hallyn_: any word on the iscsi question?
[17:57] <xibalba> any of you guys running vmware?
[17:58] <rbasak> smoser: say we did that, once for each kernel. Then ship the kernel and initrd images entirely separately via simplestreams.
[17:58] <smoser> i dont want to have multiple images "assembled" on the other end.
[17:58] <rbasak> They'd all be the same image then.
[17:59] <smoser> i'm fine with base-image-with-no-kernel and kernel/initramfs separate. that is just an optimization of 'kernel_packs'.
[17:59] <rbasak> No need for assembly. Do it dynamically, and only if there's a need for modules.
[17:59] <smoser> existing uses of ephemeral.tar.gz would break though i think. if we didn't put magic in the initramfs to get the right kernel.
[17:59] <smoser> well the "dynamically" is "set up a tgtd"
[18:00] <rbasak> What would break?
[18:00] <smoser> i suspect that something would.
[18:00] <smoser> i cna't think of an explicit example.
[18:01] <smoser> but i agree, if we had a big fat initramfs and that sufficiently prepped the target so that once /sbin/init ran, it all looked good
[18:01] <smoser> then we'd probably be ok.
[18:01] <smoser> one problem is that there is no place for that "big initramfs" to write to
[18:01] <smoser> other than memory (overlafs)
[18:02] <rbasak> Yeah we are relying on overlayfs not to break
[18:02] <smoser> well, that would work fine.
[18:02] <rbasak> I don't think we'll be short of memory in our use case though
[18:02] <smoser> but my problem is that if i copy 120M of kernels to /lib/$(uname -r)
[18:02] <smoser> then i just wasted 120M of memory
[18:02] <smoser> that would have been not used.
[18:02] <smoser> VMs are a valid use case.
[18:03] <smoser> i'd really rather not require 1G memory to boot.
[18:03] <smoser> i guess the initramfs could try to be smart on that even though.
[18:04] <smoser> rbasak, we'd still need the "base" linux and initrd inside
[18:04] <smoser> inside the tarball. not inside the filesystem.
[18:04] <rbasak> For backwards compatibility? Right.
[18:04] <smoser> right.
[18:04] <rbasak> I'm not too worried about that. It can go away in LTS+2 or whatever.
[18:05] <RoyK> smoser: I have ubuntu VMs running on 256MB - works well
[18:05] <RoyK> perhaps down to 96MB
[18:05] <RoyK> below that, initrd gets too large
[18:05] <rbasak> The way I see this, either we ship all kernels installed in the ephemeral image filesystem, or none.
[18:05] <smoser> RoyK, right. but you dont' do overlayfs to RAM in those VMs and copy all your kernel modules to the overlayfs before / is mounted
[18:05] <smoser> :)
[18:05] <rbasak> With none, you get your RAM use if you want to install inside overlayfs
[18:06] <patdk-wk_> heh, I've been fine on 64megs
[18:06] <rbasak> Unless you want to go down the path of having a separate network mounted filesystem (iSCSI/NFS/whatever) for those.
[18:06] <patdk-wk_> but normally smallest I do these days is 512megs ram
[18:06] <rbasak> (which you could I guess)
[18:07] <smoser> rbasak, so you're kind of on board with "kernel packs" ?
[18:07] <rbasak> I don't like the all option. That leads to many kernels. Although I suppose you'll generally download them all anyway, with adding HWE you might want to only download a subset in the future, and bundling them in the fs image stops you from doing that
[18:07] <rbasak> Yes, I think so
[18:07] <smoser> "all option" ?
[18:07] <smoser> oh.
[18:07] <smoser> yeah, id ont like that either. you meanh all installed inside.
[18:07] <rbasak> Right
[18:08] <rbasak> So none, except for backwards compatibility.
[18:08] <xibalba> does ubuntu's sshd support chroot? can't find  it int he man page
[18:08] <rbasak> And access to modules if required through either installing into RAM overlay, or some kind of network filesystem mount
[18:09] <sarnold> xibalba: check the sshd_config manpage
[18:09] <xibalba> what did i just say?
[18:09] <xibalba> " can't find  it int he man page"
[18:09] <sarnold> xibalba: "in the manpage" , you never said which one you read.
[18:09] <RoyK> xibalba: for scp/sftp? try rssh
[18:09] <xibalba> ha, i thoguht it was implied. i'll be more specific
[18:09] <xibalba> in sshd_config.5.html
[18:10] <RoyK> xibalba: chroot for logins or file transfers?
[18:10] <rbasak> smoser: every option seems to have an issue. I can't think of a good answer.
[18:10] <smoser> rbasak, so how will saucy-maas figure out the word 'highbank'
[18:10] <smoser> as right now that doesn't exist anywhere. and i'm not even sure its valid...
[18:10] <smoser> what should id do for arm for short term.
[18:10] <rbasak> smoser: assume that maas knows what it needs
[18:10] <smoser> rbasak,we could in the future set up maas to nfs share the kernel module directories also.
[18:11] <sarnold> xibalba: indeed, there's an sshd(8) manpage which only mentions chroot twice; sshd_config(5) mentions chroot nine times :)
[18:11] <rbasak> smoser: in quantal, detection worked via the tftp requests
[18:11] <smoser> rbasak, i dont follow.
[18:11] <smoser> in maas right now, the user says 'ARCHES=armhf/highbank'
[18:11] <sarnold> xibalba: since it is easy for people to overlook the sshd_config manpage if they don't know it's there, I felt it important to point out the specific one..
[18:11] <smoser> which tells maas that it should add a "subarch' named highbank
[18:12] <smoser> but we want to remove that.
[18:12] <smoser> and let it get everything in the stream (possibly filtering stuff... i surely hoep we can filter)
[18:12] <rbasak> smoser: maas would default to downloading all options, or some specific list (someone else can decide that). Or perhaps a default list specified inside simplestreams.
[18:13] <rbasak> smoser: the user would then customise in config somewhere
[18:13] <smoser> rbasak, thats fine.
[18:13] <smoser> but the simplestream the ephemeral images do not know anything about the word 'highbank'
[18:14] <smoser> they know about the word 'armhf'
[18:14] <rbasak> It might be an idea to sync that with available arches in maas ui's dropdowns
[18:14] <smoser> in the past, maas came up with the world 'highbank' on its own.
[18:14] <smoser> s/world/word/
[18:14] <rbasak> I'm not sure I follow your question
[18:15] <smoser> in order to not regress, somehow saucy maas has to call import-pxe-files with 'highbank
[18:15] <smoser> '
[18:15] <smoser> for arm.
[18:15] <smoser> and at the moment in tych0's work, there is no place where it would ever know the word 'highbank'
[18:15] <rbasak> To further complicate things, the saucy kernel is 'generic' for a highbank machine now
[18:16] <rbasak> There are different methods for determining what a netbooting arm machine is
[18:16] <smoser> but there is no different method for determining what to call 'import-pxe-files' with as 'subarch' arguemnt.
[18:17] <smoser> thats what i'm after. i think.
[18:17] <smoser> ie, look at https://maas.ubuntu.com/images/ephemeral/releases/streams/v1/com.ubuntu.maas:download.json
[18:17] <rbasak> WHy is import-pxe-files being called with an argument?
[18:17] <smoser> because you made it do that.
[18:17] <smoser> because YOU poluted this !
[18:17] <rbasak> A maas install should be able to support multiple arches and subarches at the same time, right?
[18:17] <smoser> :)
[18:18] <xibalba> thanks sarnold
[18:18] <smoser> as it is right now, maas is told via configuration (default configuration) about 'subarch' values of 'generic' and 'armhf'
[18:18] <xibalba> chroot for logins RoyK
[18:18] <smoser> and it installs (maas-import-pxe) things for *those* values of subarch.
[18:19] <smoser> and we have no place currently (see that download.json) that would describe those values.
[18:19] <smoser> and we dont want to leave that to user configuration.
[18:19] <rbasak> I imagined that there would be a separate stream that would tell you what arches and subarches are available.
[18:20] <RoyK> xibalba: rssh won't help you there
[18:20] <rbasak> That list would change as new hardware gets enabled
[18:20] <smoser> i'm not talking about the future
[18:20] <rbasak> It would also map detection codes
[18:20] <smoser> i'm talking about now
[18:20] <smoser> where should the string 'generic'
[18:21] <smoser> or 'highbank' come from / go.
[18:21] <rbasak> From that separate stream.
[18:21] <smoser> for saucy maas we wont hae that. i don tthink.
[18:21] <rbasak> Unless we can embed that data into the existing stream.
[18:21] <rbasak> It could start off static.
[18:21] <smoser> i dont think the data you want embeds well into that stream.
[18:22] <rbasak> Note that I'm worried I'm missing your point here.
[18:22] <smoser>  * saucy maas reads data from https://maas.ubuntu.com/images/ephemeral/releases/streams/v1/com.ubuntu.maas:download.json and imports ephemeral images from there.
[18:22] <rbasak> If we can't implement the stream right now, we'll need to hardcode it, like it was in quantal.
[18:23] <smoser> ok. i think that is what i was after
[18:26] <rbasak> smoser: isn't this exactly the same problem as determining what hwe kernels are avialable?
[18:26] <smoser> well, sort of. thats why we came here.
[18:27] <smoser> i didn't want to design a solution for mutiple subarch that did not include hwe
[18:27] <rbasak> OK
[18:28] <rbasak> Sorry this conversation seemed tedious. I didn't intend it to. It's a complicated and confusing issue :-/
[18:36] <smoser> rbasak, how much do i need to worry about subarch ?
[18:36] <smoser> can i just (for saucy maas) call it 'generic' ?
[18:37] <rbasak> smoser: that'll work for highbank, but I think there might be another subarch coming up that will not be generic.
[18:37] <rbasak> s/will/may/
[18:37] <smoser> how did maas use the subarch ?
[18:38] <smoser> sure there will be machines that wont work with the 'generic' kernel flavor.
[18:38] <smoser> but how did the user say 'boot that system with 'subarch' = highbank'
[18:38] <smoser> and will we break that if we only register 'generic' kernels in saucy maas.
[18:38] <rbasak> In quantal it needed to know to download the different kernel flavour from the installer pxe images
[18:38] <rbasak> So I arranged to pass that identifier all the way through
[18:39] <rbasak> (since at the time I expected more arm subarches that would have different identifiers)
[18:39] <smoser> oh wait. where did it get that? did it get it from ports.ubuntu.com ?
[18:39] <rbasak> Yes
[18:39] <rbasak> for pxe-files. I guess that's gone now
[18:39] <rbasak> or going
[18:40] <smoser> well the 'd-i' path for downloading is to be replaced with simplstreams data giving same information.
[18:40] <smoser> which i probaly need yoru help on.
[18:40] <rbasak> Right, but we still need to map the different subarches all the way through
[18:40] <rbasak> So I think maas still needs a field in its table of nodes with the subarch in it
[18:41] <rbasak> Which can remain highbank, even if the kernel flavour is now -generic. In fact it needs to do this, so that users can boot different releases on the same node.
[18:41] <rbasak> smoser: I have some time for a G+ if that would be easier?
[18:42] <rbasak> (or voip or whatever)
[18:47] <smoser> rbasak, let me try to write some stuff down.
[19:02] <smoser> rbasak, so... i just had a thought.
[19:02] <smoser> the mounting of /lib/modules over nfs is as simple as putting an entry in /etc/fstab actually
[19:03] <rbasak> smoser: but the fstab entry has to be dynamic, right?
[19:03] <rbasak> smoser: and arranging an NFS server on the MAAS server is another subsystem and another point of failure
[19:04] <smoser> well, it doens't have to be dynamic.
[19:04] <smoser> it has to be set on the system that shares the iscsi block device
[19:04] <smoser> to point *somewhere*
[19:05] <smoser> but that just points to one big share of lib/modules/ that has every possible version
[19:06] <rbasak> smoser: yeah but the NFS server IP will change
[19:08] <smoser> not necessarily.
[19:08] <smoser> yeah, it would have to change. so that system needs to have it set.
[19:08] <smoser> i dont know why it would be dynamic though
[19:16] <kyze> hey guys, i just built my raid1 but it seems parted is now not usable any more. print list only gives me an error about the primary gpt talbe being corrupt. any tips ?
[19:18] <mr_lou> Help. I'm in big trouble. Running Ubuntu JEOS as a server at work, and ran the  do-release-upgrade command today. But it had so many problems upgrading that I now have a broken install. It refuses to let me do anything.
[19:19] <mr_lou> apt-get -f install stops at python2.7-minimal
[19:19] <mr_lou> Typing python in the command shows that I'm running python 2.6.5
[19:19] <mr_lou> So I'm thinking maybe it stops because it expects python2.7 which it can't find because it was never installed.
[19:20] <mr_lou> But I can't remove anything, and I can't force install anything either.
[19:20] <mr_lou> Always stops me from doing anything, because of dependencies problems.
[19:20] <kyze> rollback to an older backup ?
[19:21] <mr_lou> No image backup. Only database backup and php files backup. I could format an reinstall the server, and then copy back those. But that's a last resort.
[19:21] <kyze> so apt-get -f install fails ?
[19:21] <mr_lou> Yes
[19:21] <mgw> Is it normal for a dpkg (the package uses dh_installinit) to overwrite the /etc/default file on upgrade? Looking at the man page for dh_installinit, I don't see that it wouldn't… but at the same time, I don't recall other packages doing that.
[19:22] <rbasak> mgw: it should ask you. It's a policy violation to overwrite your changes.
[19:22] <kyze> mr_lou: sorry im pretty new at this so i cant really help you :(
[19:22] <mr_lou> np
[19:23] <mr_lou> kyze, I'm newer. ;-P
[19:23] <mgw> rbasak: thanks… I must be doing something wrong in my rules
[19:23] <kyze> mr_lou wanna take a bet ? :D
[19:26] <mgw> rbasak: this is the relevent part of rules: https://gist.github.com/mgwilliams/2f0ef3d6e76d6760d9ec
[19:26] <mgw> and I have two files — nsq.upstart and nsq.default in debian/
[19:33] <Diegonat> hi guys? what would you say it is the best management tool for KVM ??
[19:45] <hallyn_> SpamapS: not yet (sorry, at plumbers)  though i did talk to one person who thought that it should be at least usable with just tcp (no netlink).
[19:45] <hallyn_> SpamapS: what exactly are you doing?  you're not mounting it on the host and then bind-mounting it the container I assume?
[19:50] <SpamapS> hallyn_: we just want to have an lxc container mount iscsi targets.
[19:54] <beebs> Hey All. I'm trying to set up a local juju server.  The documentation is a little light on the network setup prior to installing juju, mongodb and bootstrapping. Anyone have experience with this or know of a good walkthrough. I've been googleing and many step-by-steps seem old and outmoded.
[19:58] <SpamapS> beebs: you probably want #juju, but anyway, the bootstrap command actually installs mongo for you.
[20:01] <beebs> SpamapS: actually i'm thinking before I even get there, the host network setup. I have three NICs, and am wondering what the best practices are for setting those up, and any VLANs I should be setting up on the switch.
[20:11] <HSaka> Hello, how can i delete my raid? the disc are /dev/sdc /dev/sdd /dev/sde
[20:11] <HSaka> it's raid 5
[20:12] <henkjan> HSaka: mdadm --stop /dev/mdX
[20:13] <smoser> rbasak, still around?
[20:13] <henkjan> mdadm --zero-superblock /dev/sdX wil wipe your raid signatures from teh disk
[20:13] <HSaka> ty
[20:13] <mgw> rbasak: http://www.mail-archive.com/busybox@busybox.net/msg10671.html
[20:14] <HSaka> hmm how about if i want to add one of my disc back to raid 5, but it says one of them has no superblock?
[20:14] <mgw> so it looks like dpkg -i by default overwrites conffiles
[20:14] <HSaka> example /dev/sdc
[20:14] <mgw> which isn't a big deal for me, since we don't use dpkg -i in production — but it sure threw me off
[20:28] <Diegonat> guys whats that command similar to ls -l ?? something like "lls" . Do u know what im talking about?!
[20:40] <shauno> Diegonat: 'll' is a default alias on redhat systems, which is the closest I know of (alias ll='ls -l --color=auto')
[21:25] <swaT30> @zul or any other Ubuntu OpenStack folks, any plans on pushing https://review.openstack.org/#/c/32679/1/nova/network/security_group/quantum_driver.py into the Cloud Archive?
[21:31] <Guest72430> Question what -t stand for or do in this command  sudo mkfs -t ext4 /dev/sdb1       ???
[21:32] <sarnold> Guest72430: check the mkfs(8) manpage: mkfs [options] [-t type fs-options] device [size]
[21:32] <sarnold> Guest72430: in this case it tells mkfs the type of filesystem to create
[21:33] <zul> swaT30:  yes
[21:34] <swaT30> @zul any ETA? I'm trying to determine if we should apply the patch ourselves in the meantime, or wait for an update
[21:34] <swaT30> :)
[21:34] <Guest72430> thanks
[21:35] <zul> swaT30: not sure the grizzly SRU just went through and need people to test so once it goes through our qa process then it will hit the CA
[21:36] <swaT30> zul: ok great, I'll apply it manually and the update should just overwrite. Thanks :)
[22:11] <Diegonat> shauno, thanks
[22:11] <Diegonat> :)