[00:42] <Hajuu> Hey guys - What would be the most likely place to find a log of why my server keeps dropping its network connection?
[00:42] <Hajuu> :)
[00:43] <Rafael_> Executing: rsync.exe  -v -rlt -z --chmod=a=rw,Da+x --delete --exclude="/cygdrive/D/My Documents/Charts/Search" "/cygdrive/D/My Documents/" "192.168.1.11::backuptest/My Documents/"
[00:43] <Rafael_> sending incremental file list
[00:43] <Rafael_> rsync: send_files failed to open "/cygdrive/D/My Documents/Charts/Search/Data/Applications/Windows/MSS.log": Device or resource busy (16)
[00:43] <Rafael_> anybody can tell me why is tha msitake
[00:43] <Hajuu> msitake lol
[00:45] <qman__> Rafael_, the file is currently open in a program and can't be copied
[00:45] <Rafael_> qman__ that is why is being excluded..i want to exclude that fodler
[00:46] <Rafael_> --exclude="/cygdrive/D/My Documents/Charts/Search"
[00:46] <Rafael_> qman__ any idea?
[00:50] <qman__>  --exclude="/cygdrive/D/My Documents/Charts/Search**"
[00:50] <qman__> for more info, see the manual
[00:51] <Hajuu> Hey guys - What would be the most likely place to find a log of why my server keeps dropping its network connection?
[00:51] <qman__> Hajuu, dmesg
[00:53] <Hajuu> even after rebooting it?
[00:53] <Hajuu> cause I really dont want to haul out a monitor and connect it etc
[00:56] <qman__> I think it logs dmesg by default
[00:56] <qman__>  /var/log/dmesg
[00:57] <Hajuu> sweet lemme check it out
[00:58] <Hajuu> hmm
[00:58] <Hajuu> nothing particularly helpful that I can see
[00:59] <Hajuu> the last line is;
[00:59] <Hajuu> [42949408.800000] eth0: no IPv6 routers present
[00:59] <Hajuu> but im using ipv4, so thats not a problem
[00:59] <Hajuu> It always works if I restart the server again
[00:59] <Hajuu> is there maybe like, some way to make a connection auto-reconnect?
[01:00] <Hajuu> or would I have to make some daemon script for that?
[01:01] <Hajuu> theres a possibility its as simple as my router making its lease expire and it isnt getting a new one
[01:01] <Hajuu> like, automatically
[01:01] <Hajuu> but I obviously cant ssh in to restart networking
[01:10] <qman__> well, if there's no log there, as far as the kernel is concerned, everything is working
[01:10] <Hajuu> well, I definately cant connect :P
[01:10] <qman__> check the dhcp logs, and if there's nothing there, I'd suggest a driver or hardware issue
[01:10] <Hajuu> meh I dont see how itd be a driver or hardware issue (except the router maybe)
[01:11] <Hajuu> cause like, it works when it first starts up
[01:11] <Hajuu> and I bet if I did /etc/init.d/networking restart itd probably come back up
[01:11] <Hajuu> ill check anyway.
[01:11] <Hajuu> I don't seem to have a dhcp log
[01:13] <qman__> looks like that goes in /var/log/daemon.log
[01:13] <qman__> I'm also seeing some in /var/log/syslog
[01:14] <Hajuu> only logs in daemon.log are about mysql doing stuff
[01:14] <Hajuu> and I couldnt see anything in syslog
[01:14] <Hajuu> ill paste what it gives me
[01:14] <Hajuu> (not here, obviously :P)
[01:16] <Hajuu> http://pastebin.com/xgZvM6wx
[01:16] <Hajuu> thats the last few lines of my syslog
[01:16] <Hajuu> so it obviously shows it connecting
[01:16] <Hajuu> then nothing after that
[01:17] <qman__> when I do a grep -i dhcp, I get stuff like this
[01:17] <qman__> Apr 22 18:03:41 mediaopty dhclient: DHCPREQUEST of 192.168.1.7 on eth2 to 66.227.157.119 port 67
[01:17] <qman__> Apr 22 18:03:41 mediaopty dhclient: DHCPACK of 192.168.1.7 from 192.168.1.1
[01:17] <qman__> in daemon.log, and a few in syslog
[01:18] <Hajuu> how do you do that command, sorry?
[01:18] <qman__> grep -i dhcp /var/log/daemon.log
[01:18] <qman__> or
[01:18] <qman__> grep -i dhclient /var/log/daemon.log
[01:19] <Hajuu> not getting anything for either of those in daemon.log or in syslog
[01:19] <qman__> then dhcp is probably not running
[01:19] <Hajuu> which is strange, since I am using dhcp to issue a static address
[01:19] <qman__> what's your /etc/network/interfaces look like?
[01:19] <Hajuu> as far as I remember
[01:19] <qman__> you mean a dhcp reservation
[01:20] <Hajuu> yeah
[01:20] <Hajuu> auto eth0
[01:20] <Hajuu> iface eth0 inet dhcp
[01:20] <qman__> yeah, you should be getting some dhclient messages somewhere then
[01:20] <qman__> try grep -iR dhclient /var/log
[01:20] <qman__> see if you get anything back
[01:21] <Hajuu> ahh yeah there we go
[01:21] <Hajuu> for some reason in /var/log/installer/syslog
[01:21] <Hajuu> oh
[01:21] <Hajuu> and in /var/log/daemon.log.0 too
[01:21] <Hajuu> lots in there
[01:21] <qman__> ok, so it was working at one time
[01:22] <qman__> it just hasn't gotten any since the last log rotation
[01:22] <qman__> head /var/log/daemon.log to see when it rotated
[01:22] <Hajuu> man
[01:22] <Hajuu> it wont let me paste these log lines lol
[01:22] <Hajuu> /var/log/daemon.log.0:May  5 05:24:48 liveSRV dhclient: No DHCPOFFERS received.
[01:22] <Hajuu> /var/log/daemon.log.0:May  5 05:24:48 liveSRV dhclient: No working leases in persistent database - sleeping.
[01:23] <qman__> that's the most recent?
[01:23] <qman__> also, may 5?
[01:23] <Hajuu> well like I say
[01:23] <Hajuu> theres no logs at all for that in daemon.log
[01:23] <Hajuu> only in daemon.log.0
[01:23] <Hajuu> its date is probably wrong
[01:23] <Hajuu> lol
[01:23] <qman__> ok
[01:24] <qman__> well, date and time is pretty important, especially for SSL
[01:24] <Hajuu> ssl is all working
[01:24] <Hajuu> everything works
[01:24] <Hajuu> except every few days
[01:24] <Hajuu> it becomes disconnected
[01:24] <Hajuu> and I have to either haul out a keyboard and monitor
[01:25] <Hajuu> or restart the thing (which is what I do)
[01:25] <Hajuu> its just a dev server
[01:25] <Hajuu> but its inconvienient to not just be able to fire up my dev stuff and get to work
[01:40] <handheldCar> anybody know which mail server ubuntu server installs during the software selection?
[01:40] <qman__> handheldCar, postfix+dovecot
[01:40] <nealmcb_> handheldCar: typically I think postfix is the favorite
[01:40] <persia> It's up to the user, but postfix is recommended
[01:41] <qman__> tasksel uses postfix and dovecot
[01:41] <qman__> exim is also a supported configuration
[01:41] <qman__> but most people use postfix
[01:41] <handheldCar> That's what I want.
[01:41] <nealmcb_> it is nicely integrated with other mail services in tasksel
[01:42] <qman__> I like postfix, it's really easy to set up with debconf, and lots of people use it
[01:42] <qman__> so it's easy to find information on special configurations
[01:44] <handheldCar> Does anyone use some alternative package instead of vim-nox? Last I checked, it wasn't available.
[01:52] <smoser> hggdh, here makes more sense
[01:52] <hggdh> heh
[01:52] <smoser> so... ther eisn't by chance another dhcp server running in that lab somewhere is ther e?
[01:52] <smoser> did you see my comments https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/566792
[01:52] <smoser> that was the root cause of many (unfortunately not all) of my failures.
[01:52] <hggdh> I do not know... up to now I do not know the whole net setup there :-(
[01:53] <hggdh> yes, I saw it, and I wondered also
[01:53] <hggdh> one way is to sniff for DHCP server offers
[01:53] <smoser> right
[01:53] <smoser> easy enoug to do on the CC
[01:53] <smoser> it would see all of them
[01:53] <hggdh> yeah
[01:54] <smoser> well, maybe it wouldn't though
[01:54] <hggdh> want me to start one? I am running a test there
[01:54] <smoser> are responses broadcast
[01:54] <smoser> ?
[01:54] <smoser> it might not see the NCs response
[01:54] <smoser> let me look in my logs
[01:54] <hggdh> as long as they go through the same subnet we are listening to, yes, we can get the responses
[01:56] <hggdh> smoser: there is also a bit of news -- I moved back to topo1; I have run, so far, 120 instances, and *no* errors
[01:56]  * hggdh wonders
[01:56] <smoser> hm..
[01:56] <smoser> topo1 is all-in-one + NC
[01:56] <smoser> ?
[01:57] <hggdh> I am in the middle of a 400-instance right now
[01:57] <hggdh> and yes, topo1 is CLC+CC+SC, and 5 NCs
[01:57] <hggdh> er, walrus also
[01:58] <smoser> what topo did we see the success and blank instance-id from ?
[01:59] <hggdh> topo2
[01:59] <smoser> what is that ?
[01:59] <hggdh> one machine each for CLC, walrus, CC and SC, 2 for NC
[02:00] <SpaceBass> lost power, hard drop and now my software raid5 won't come back - seem to have lost superblocks, any way to recover?
[02:00] <hggdh> brb -- time to eat
[02:35] <mathiaz> smoser: hggdh: there is a dhcp server on the UEC network but it should *not* hand out lease to unknown hosts
[02:36] <smoser> mathiaz, thanks.
[03:02] <hggdh> and, meanwhile, 100% success on topo1
[03:04] <hggdh> except memory usage is growing
[03:47] <flyback> is it just me
[03:47] <flyback> or is squashfs a fucking failure
[03:47] <flyback> every machine I seen it on
[03:47] <flyback> eventually starts spitting out massive errors
[03:47] <flyback> cdrom or flash
[03:52] <persia> !ohmy | flyback
[03:52] <persia> But generally, I think it's only best to use for short-term read-only stuff.
[03:53]  * flyback laughs
[03:54] <flyback> you know what you get for a lifetime of being nice to others
[03:54] <flyback> fucked over
[03:54] <flyback> you get me
[04:00] <handheldCar> u get 2 stand up 4 yourself every now and then.
[04:02] <persia> standing up for oneself is fine.  Seeking support or to improve collaboratively-maintained software is best done by being nice to other folks with whom one collaborates.
[04:02] <persia> The catching more flies with honey than vinegar thing.
[04:03]  * handheldCar was just speaking in regards to flyback being a push over, taking kindness to extreme.
[04:05] <persia> Yeah.  flyback /parted though :)
[07:18] <Lantizia> Hey UEC, why?
[07:19] <Lantizia> I know it's a very generic question, but I've read as much information I can find and I don't see why I should bother.... Surely theres OpenVZ and KVM
[07:20] <twb> Lantizia: they divide up the resources of a single host.
[07:20] <Lantizia> so does OpenVZ
[07:20] <twb> I was talking about VZ and KVM.
[07:20] <twb> UEC doesn't do that
[07:21] <Lantizia> ok confused
[07:21] <twb> Lantizia: the purpose of a cloud is to distribute a virtual host across a number of physical hosts, theoretically making it easier to scale resource allocation up/down as demand requires, and possibly improving reliability.
[07:22] <Lantizia> again... done... you can migrate servers without downtime with either of those 2
[07:22] <twb> I'm assuming that clouds run a virtual host on multiple hosts *CONCURRENTLY*.
[07:23] <Lantizia> hmm thats an advantage I guess
[07:23] <twb> http://en.wikipedia.org/wiki/Cloud_computing
[07:25] <Lantizia> twb, by that definition then... even VMware's ESX isn't cloud
[07:25] <Lantizia> because it can't run a machine concurrently on two heads at once and keep them synced
[07:27] <ttx> twb: you have a strange definition of cloud computing
[07:28] <ttx> cloud computing is just "utility computing"
[07:28] <persia> That's also a bit of a strange definition, I think.
[07:28] <Lantizia> ttx, i.e. run an "appliance" "up there"
[07:29] <Lantizia> lol
[07:29] <persia> There's an assuption of some abstraction in "cloud": e.g. storage-as-a-service+cpu-as-a-service combining.
[07:29] <ttx> On Demand Service, Ubiquitous Network Access, Location Independent Resource Pooling, Rapid Elasticity, Measured Service
[07:29] <Lantizia> can I just chalk "cloud" computing as a silly sales buzz word and think nothing more of it except it's just virtualisation
[07:29] <ttx> Lantizia: no
[07:30] <persia> Often also combined with multiple layers of service, so one has an app that restfully coordinates with other apps, without any knowledge of how many hosts serve the apps, etc.
[07:30] <ttx> It's a technology transition
[07:30] <Lantizia> buzz buzz buzz
[07:30] <ttx> not really, all technologies go through that type of transition, why not computing ?
[07:30] <persia> Lantizia: Note really.  It's about persistance.  So if I have a server (real, virtual, whatever), I typically expect it to run for a while, and do cronjobs, etc.
[07:31] <Lantizia> so if it's not just a buzz word... then it's twb's definition
[07:31] <ttx> But I'd agree it's difficult to see clear through the hype
[07:31] <Lantizia> by the very fact it's "cloud" computing it means it's not limited to just 1 location
[07:31] <persia> If I have a cloud app, it's quite likely to be composed of a demand-based set of hosts running each layer of an app, where the servers are created and destroyed in minutes based on use.
[07:31] <Lantizia> but by that I don't just mean high availability either... thats something else lol
[07:32] <ttx> Lantizia: you can build HA on top of cloud computing... but using cloud won't make all your things HA by magic
[07:32] <ttx> it's just a different way to offer and consume computing resources
[07:33] <Lantizia> fluff, lots of fluff
[07:33] <ttx> technology maturity makes it more and more of a commodity
[07:33] <twb> Lantizia: ESX is certainly not cloud computing.
[07:33] <ttx> like electricity
[07:33] <ttx> twb: right, ESX is just managed virtualization
[07:33] <Lantizia> i don't get why you've argued against twb's position then ... when he was right
[07:33] <ttx> think KVM + a shiny and powerful adminGUI
[07:34] <Lantizia> already got that... Proxmox
[07:34] <ttx> twb said "clouds run a virtual host on multiple hosts *CONCURRENTLY*."
[07:34] <Lantizia> I'd swap "on" for "over"
[07:34] <ttx> no, it still runs a VM in a single place
[07:34] <Lantizia> but otherwise he is right
[07:35] <twb> With cloud computing you give your service provider a diskless VM image which runs, say, a bunch of shitty Java or PHP in apache, talking to their database service.
[07:35] <persia> Or really good code: doesn't matter
[07:35] <twb> And the service provider decides on what host(s) the VM will be started/stopped, and how much memory and such it'll have
[07:36] <twb> persia: IME enterprise-level software is automatically bad code :-P
[07:36] <Lantizia> bored now, bye
[07:36]  * persia goes back to trying to figure out how to enable emulation on a PPC 7447A
[07:37] <ttx> twb: then your service provider is adding value on top of cloud... nothing will make your "shitty Java" automatically scalable
[07:37] <ttx> anyway, I see there is still a lot of confusion, and more advocacy is needed
[07:38] <ttx> I only recently understood it's not hype, and it's not new
[07:38] <persia> Targeted advocacy as well: there's lots of different folks saying different things are "cloud computing".
[07:39] <persia> A managed service offering is a critical component, but at least from what I see, it's insufficient to achieve the overall goals often used as examples of why the old model is outdated.
[07:39] <twb> I confess I haven't tried to get it working.
[07:39] <ttx> persia: I can only recommend Simon Wardley's talk (Situation normal, everything must change), it's clarifying where all the others are confusing
[07:40] <twb> I don't approve of any app that requires a graphical web browser as its base platform.
[07:40] <persia> twb: Who said you need that?
[07:40] <ttx> twb: I run a private cloud every day. It's been a few months since I used a web browser to do any interaction with it
[07:41] <ttx> so I wouldn't call it "base platform"
[07:41] <twb> Well, it'd be a bit silly to run something like Emacs on a cloud VM
[07:42] <ttx> ah, you mean most cloud apps would end up being served by HTTP ?
[07:43] <persia> twb: Wrong model.  Consider M-x-plagarism which makes a call to some service that distributes a text match against the entire contents of the internet on N servers, and returns scheme that auto-marks significant blocks of text that might be plagarism.
[07:44] <twb> ttx: in my head I lump SOAP and REST as "tacked onto the browser model as an afterthought"
[07:44] <twb> persia: I guess...
[07:44] <persia> cloud doesn't need to be REST: REST just makes it easier to write the services.
[07:46] <twb> I suppose you could also use it for stuff like NNTP servers or game servers.
[07:47] <ttx> you can, and you will, use it for anything. It's just a question of time :)
[07:47] <persia> ttx: You don't happen to know whether if I call `qemu ...` it will auto-invoke kvm where appropriate, do you?
[07:48] <ttx> persia: no, I admit never have run "qemu", only kvm
[07:48] <persia> heh, OK.
[08:30] <ttx> persia: Daviey might know the answer
[08:31] <ttx> otherwise #ubuntu-virt
[08:33] <persia> Daviey already gave me the virtsh answer, but it didn't work perfectly for me, but part of that seems to be issues with qemu running how I want, which in turn appears to be related to a build-failure.
[08:34] <persia> And from what I read upstream, I might run into other issues due to potentially not having a working kernel target and very much expect to have an issue due to not having any firmware to feed qemu)
[08:34] <persia> So it's more of a long-term thing :)
[09:07] <Daviey> persia: Is this for a funky arch?
[09:08] <ttx> funky funky
[09:08] <Daviey> persia: fwiw, i've tried to use qemu to emulate arm.. and USING it is slow.. compiling on it is impossible :)
[09:08] <persia> Daviey: Of course :)
[09:08] <persia> Works for me.
[09:09] <persia> `mk-sbuild --arch=armel lucid; sbuild -d lucid-armel foo.dsc`
[09:09] <persia> Speed seems similar to native.
[09:12] <Daviey> persia: odd, i last tried it in the karmic cycle..  I wasn't using sbuild, and it made me pull my hair out.  I was tempted to use an n810 as a build box :)
[09:12] <persia> There were massive improvements in lucid: thank lool.
[09:12] <persia> It got good enough that I added support in mk-sbuild and pbuilder-dist
[09:13]  * ttx hugs sbuild
[09:14] <Daviey> persia: that is great news.
[09:14] <persia> The thing that was bugging me today is that when I run virt-manager, it offers to connect to my powerpc server as a Virtualisation Host, but qemu-kvm FTBFS on powerpc.  I haven't been able to sort out whether a PPC 7447A (my chip) can use it.
[09:14] <persia> Next game will be to find a way to have qemu/kvm be able to auto-launch against schroots.
[09:16] <Daviey> persia: cowbuilder might be appropriate?
[09:16] <persia> why?
[09:16] <Daviey> i'm assuming you are using schroots for the enviroment?
[09:17] <persia> Yep.
[09:17] <persia> sbuild is designed to build against schroots with configurable overlay mechanisms.
[09:18] <persia> Used to mostly be LVM with snapshots, but smoser and I added support for other sorts of overlays in lucid, so it can do directory chroots with aufs (which is *fast*) or tarball-based schroots kinda like pbuilder.
[09:20] <Daviey> persia: that is interesting.. My local build server uses cowbuilder and some support for qemubuilder, i might switch it.
[09:21] <persia> Soyuz uses sbuild :)
[09:21] <persia> (mind you, it's a different fork of the original sbuild than the fork that later became the packaged sbuild)
[09:22] <Daviey> :(
[09:23] <persia> Nah, it has different requirements.  I think I know what they are now, but I didn't discover them early enough for lucid, which means we can't merge until LTS+1
[09:23] <Daviey> Well my project for next week is rebuilding the build server :)
[09:24] <persia> How do you dispatch/collect builds?
[09:31] <Daviey> persia: It's not elegant; dput over scp -> fsniper for detecting uploaded source files (other ionotify solutions could do this) -> I can't remember if the publishing is done by apt-ftparchive or mini-dinstall.
[09:32] <Daviey> It's not elegant, but i hate compiling locally and often the packages i test are for other boxes (not my dev box), so having a bunch of local debs is a PITA when i really want a repo.
[09:33] <persia> Makes sense.  I'm fiddling a bit with some tools to easily separate work and build as a developer: once I get a bit closer, I might ask you to help test.
[09:33] <Daviey> persia: happy to!
[09:34] <persia> Because I'm *not* planning on doing serious compilation on a 700MHz netbook, even if I find it nice to have enough battery to last all day.
[09:36] <eagles0513875> hey guys im just wondering is xen precompiled with the kernel or does the kernel require recompilation
[09:36] <Daviey> persia: Oh, and powernap is an obvious benefit :)
[09:39] <persia> eagles0513875: It needs a special kernel, but that kernel isn't compiled by default (at least for amd64)
[09:39] <eagles0513875> ok but its included in the main stream kernel just not compiled right
[09:40]  * persia doesn't know
[09:40] <eagles0513875> ok im hearing 2 different things
[09:41] <eagles0513875> persia: ignore my question
[09:42] <persia> There used to be a xen flavour of the kernel offered by default.  It went away.  I suspect it's still required.  kvm is the recommendation.
[09:42] <persia> Daviey: which powermap?
[09:45] <Daviey> persia: powernap, kirklands auto hibernate magic.
[09:45] <persia> Oh well.  I'm confused.  I won't fix qemu-kvm/powerpc for lucid ( http://launchpadlibrarian.net/45034888/buildlog_ubuntu-lucid-powerpc.qemu-kvm_0.12.3+noroms-0ubuntu8_FAILEDTOBUILD.txt.gz )
[09:46] <persia> Daviey: Oh.  Doesn't affect me: my netbook lasts all day without hibernate (Sharp PC-Z1).
[09:46] <Daviey> persia: sorry, i mean for a dedicated build server.
[09:46] <persia> (plus it has no hibernate suppoort in the kernel)
[09:48] <persia> Hmm.  My servers run at ~60W, but I might look into that, since one of them could be asleep most of the time.
[09:49] <Daviey> persia: Is that buildd healthy for ubuntu-lucid-powerpc.qemu-kvm_0.12.3+noroms-0ubuntu8, somewhat suprised to see a bunch of packages that "cannot be authenticated"
[09:49] <persia> (well, except for the 15W one, but that has enough other issues that fiddling with power management is likely to cause pain)
[09:49] <Daviey> heh
[09:50] <persia> I think it's only sorta healthy.  There was some issue that caused the powerpc buildds to die hard in February, and lamont hacked something up: I think they are looking forward to clean lucid installs.
[09:51] <Daviey> ahh
[09:51] <persia> But I trust ftpmaster.internal enough to expect it's unlikely to fall victim to a MiM attack.
[09:51] <Daviey> oh aye, i just wondered if it was an indicator that something else wasn't clean in the enviroment
[09:51] <persia> soren: So, do you actually know what causes the issue, if it's not the prototype?
[09:52] <soren> persia: Yes. Hang on.
[09:52] <persia> Daviey: I can replicate the issue in my chroots, so while it might not be clean, it's not better than what I have.
[09:52] <persia> Err, not worse.
[09:52] <Daviey> sure
[09:53] <soren> persia: So, the problem is that qemu-kvm and our kernels are out of sync for powerpc/kvm.
[09:54] <persia> Hrm?
[09:54] <soren> http://git.kernel.org/?p=linux/kernel/git/avi/kvm.git;a=commitdiff;h=ec3c11aa5f9d0a7f48f46d6790c33ccc654fd6ec
[09:54] <soren> We don't have that patch, for instance.
[09:54] <persia> Then why does it only fail for powerpc?
[09:54] <soren> The kvm code for powerpc attempts to access a member of a struct which doesn't exist in the version of the struct we have in our kernels.
[09:54] <persia> Ah!  The patch is only for powerpc
[09:55] <soren> Yes.
[09:55] <persia> Thanks!
[09:55] <Daviey> persia: looks like there might be a fix in upstream kernel
[09:55]  * persia hunts for a bug, prepared to file a new one
[09:55] <Daviey> looking at the git logs
[09:55] <soren> persia: The fix (for Lucid) is to disable kvm for powerpc.
[09:55] <soren> persia: (in qemu-kvm's build)
[09:56] <persia> Why?
[09:56] <persia> I'd rather leave it FTBFS, and do a rebuild in a couple weeks once the kernel SRUs.
[09:57] <soren> persia: What's the kernel going to be updated to?
[09:58] <persia> Oh, probably -updates.  RIght.
[09:58] <persia> Err, -proposed
[09:59] <soren> No no, I mean which version?
[09:59] <persia> Just cherrypicks, I think.
[10:00] <persia> I've been told that a config change will happen to enable LVM booting on powerpc, for example.
[10:00] <soren> persia: Ok. Because that pvr member did not land in that struct in 2.6.32.
[10:01] <persia> soren: Should that patch apply against 2.6.32?
[10:01] <soren> persia: ...so someone needs to figure out what to cherry pick to make this work.
[10:01] <persia> Ah, so it's more than just that bit.
[10:01] <soren> persia: I don't know if that patch alone will fix it.
[10:01] <soren> persia: In fact, I'm reasonably sure it won't.
[10:02] <soren> persia: http://www.mail-archive.com/kvm-ppc@vger.kernel.org/msg00856.html
[10:02] <soren> persia: It's a pretty hefty patch set.
[10:03] <soren> persia: Take it from someone who's had to do this way too many times: Don't bother.
[10:03]  * persia stops the bug filing process, and looks forward to being able to actually use the "Virtualisation Host" on the powerpc server for maverick
[10:04] <persia> soren: Yeah.  I'd have to actually care more than "Huh: be nice to do virtual test installs on that server once in a while".
[10:05] <soren> persia: So, again: The fix for Lucid is probably just just disable kvm for powerpc.
[10:05] <persia> I think you're right.  I can't upload that.  Feel like pushing a fix?
[10:05] <soren> You can't? Why?
[10:05]  * persia isn't core-dev
[10:05] <soren> That's just silly.
[10:06] <soren> It's probably the same problem for ia64, by the way.
[10:06]  * soren looks at build logs..
[10:07] <soren> Uh... No.
[10:07] <soren> :)
[10:07]  * persia has yet to find a small, quiet, low-power ia64 or sparc server, so has limited interest in those arches.
[10:07] <soren> persia: I'm crazy busy these days.. If you whip up a patch, I'd be more than happy to sponsor it.
[10:08] <soren> Who here is going to UDS, by the way?
[10:08]  * soren raises his hand
[10:09] <soren> Daviey: Are you coming this time?
[10:09] <Daviey> soren: Oh aye o/
[10:09] <soren> \o/
[10:10] <Daviey> \o/
[10:11] <persia> soren: `sed -i 's/any/i386 amd64 sparc armel/' debian/control; dch -i "Don't build on architectues where there is no kernel support"`
[10:11] <soren> persia: Oh, no no.
[10:11] <soren> persia: Just pass --disable-kvm to configure if attempting to build on powerpc.
[10:12] <persia> That gives us qemu on powerpc?
[10:12] <soren> persia: That's the idea.
[10:12] <persia> Oh nifty.  I'll actually spend time on a patch that does that.
[10:12] <soren> Well, hypothesis is more accurate, I guess.
[10:12] <bluethundr> Why do I not see a  05-domain_id in /etc/amavis/conf.d on my Ubuntu mail server and why do I care/not care? http://pastebin.com/mjnL8uef
[10:13] <soren> persia: Confirmed: http://www.mail-archive.com/qemu-devel@nongnu.org/msg23996.html
[10:13] <soren> persia: (If anyone knows, it's agraf)
[10:13] <persia> heh.
[10:14] <persia> What's the ia64 fix?
[10:19] <soren> persia: I'm not sure... Let me check.
[10:24] <soren> persia: To not try at all, I think.
[10:27] <soren> persia: Yup. ia64 host support didn't exist at all up until about a month ago.
[10:40] <persia> soren: Thanks.  I'll adjust debian/control for that.
[10:41] <soren> persia: Or we could just leave it as ftbfs. It's meant to work eventually.
[10:41] <persia> Long as I'm cleaning up: it can be re-enabled for maverick, and it looks nicer.
[10:41] <soren> persia: Certainly no reason to go and fix PaS at least. And if we don't do that, it's going to turn up as a ftbfs anyway.
[10:43] <persia> soren: I was given an example earlier that implied Soyuz was smart enough to not need P-a-s for lots of stuff anymore, and am a bit interested in trying it out :)
[10:43] <soren> persia: I was under the impression that it was intentional.
[10:46] <persia> soren: What, ignoring .dsc input?  I think it was originally.
[10:50] <soren> persia: I see. We finally realised it was a wart?
[10:50] <persia> I think so.
[10:51] <soren> Great. I've always thought it was. :)
[11:08] <pths> I'm setting up DRBD and the notify scripts there use the mail command. Any suggestions for a pacakge that provide that command and would require the least possible configuration?
[11:50] <joebike> >	i'm trying to set up source control, will i run into any problems if I use apache on my local and then nginx for my production environment? both on ubuntu
[11:52] <twb> You don't need an httpd to run a VCS
[12:35] <ewook> I so do not like appliance-boxes.
[13:28] <NCommander> coffeedude: ping?
[13:59] <persia> soren: https://bugs.launchpad.net/ubuntu/+source/qemu-kvm/+bug/568904
[14:00] <persia> Tested and built successfully on powerpc.  Feel free to clean up the changelog if you have better explanations.
[14:00]  * persia dislikes 2 minutes of coding taking 4 hours to check
[14:31] <ttx> kirkland: about bug 532733, do you agree we should move it to potential SRU target rather than potential pre-release candidate ?
[14:31]  * kirkland looks
[14:31] <ttx> zul: same question about bug 541439
[14:31] <kirkland> ttx: yes, absolutely
[14:32] <ttx> zul: did you ping the release team for bug 533029 ?
[14:32] <ttx> me asks questions again
[14:33] <ttx> zul: about bug 541439, do you agree we should move it to potential SRU target rather than potential pre-release candidate ?
[14:33] <zul> ttx: yep
[14:33] <zul> ttx: what about it?
[14:33] <ttx> zul: did you ping the release team for bug 533029 ?
[14:33] <zul> ttx: yes
[14:33] <zul> ttx: yep no response from slangasek
[14:33] <ttx> ok
[14:33] <persia> kirkland: You're doing qemu-kvm stuff now, right?  What do you think about bug #568904?
[14:33] <ttx> i'll keep it on the "potential" list then
[14:33] <zul> if he did i missed it though
[14:34] <ttx> smoser/kirkland: i'm keeping bug 565018 and bug 566792 on the release radar, even if I doubt we'll find anything by release day
[14:34] <zul> ttx: im trying to tack down a php cgi crasher as well
[14:35] <ttx> smoser: could any of those be explained by a DHCP conflict ?
[14:36] <ttx> mathiaz: I keep bug 563829 as a potential pre-release fix, since you're on it... I suspect bug 423252 is quite unlikely and shouldbe moved to the "potential SRU" category ?
[14:37] <mathiaz> ttx: bug 563829 - yes I have something ready to rool
[14:37] <smoser> ttx, for "not reachable via ssh" i can't come up with an explanation that would be caused by dual dhcp. additionally, other people see it and don't have the dhcp symtoms. kirkland saw it several times last night. i had him enable debug and kick off  arun
[14:37] <mathiaz> ttx: roll
[14:37] <ttx> kirkland, mathiaz, hggdh, please update status for your remaining work items on http://people.canonical.com/~pitti/workitems/canonical-server-ubuntu-10.04.html
[14:37] <mathiaz> ttx: I was planning to get slangasek review
[14:37] <smoser> for the other, that originally came from the data center, where we have no reason to suspect dual dhcp
[14:37] <ttx> mathiaz: sounds good.
[14:37] <ttx> smoser: ok
[14:37] <smoser> and also, there we saw "200 OK" response from the server, with empty data
[14:37] <ttx> which cannot be right.
[14:38] <SpaceBass> hey folks
[14:38] <smoser> which i guess could be routing... if there was another server answering
[14:38] <smoser> some server that thought it was that funky address
[14:38] <SpaceBass> any software raid gurus? My system crashed in the middle of running fsck on my raid - now I cannot mount it, although it seems to be running cleanly
[14:42] <ttx> mathiaz: what about bug 423252 ?
[14:43] <mathiaz> ttx: sru
[14:44] <ttx> ack
[14:45] <e-DIO-t>  yO!
[14:49] <SpaceBass> false alarm... found it... system is still running fsck on the drive...must have resumed after the reboot?
[14:54] <ttx> mathiaz, kirkland, smoser, zul: please have a look at https://wiki.ubuntu.com/ServerTeam/ReleaseStatus and confirm it reflects our current status -- meeting in 1 hour
[14:54] <kirkland> ttx: doing now
[14:55] <zul> ttx: looks good
[14:55] <smoser> ttx, yeah, looks good here.
[14:55] <smoser> i hope that hggdh can run a topo2 test today
[14:55] <smoser> kirkland, you have results form your last night debug run ?
[14:55] <ttx> the idea is "what we might try to sneak into release", given that everything that can be safely SRUed will be, to leave some room in the build q
[14:56] <kirkland> ttx: the two UEC bugs under "might try to fix by release" ... those are almost certainly SRUs
[14:56] <ttx> the two cloud image issues ?
[14:57] <zul> ttx: can you poke slangasek about autofs5 in the meeting today
[14:57] <ScottK> ttx: There's a clamav upload that might yet get in.  It's got an unfortunately large and complex diff due to problems in our current package's debconf handling.  What's in queue now is what Debian went with.
[14:57] <zul> ttx: my large pointy stick doesnt seem to be pointy enough
[14:57] <ttx> ScottK: you have a bug number ?
[14:58] <ScottK> ttx: No bugs written, just some stuff ends up configured in ways that would suprise users if they noticed.
[14:58] <ScottK> AFAIK no one noticed yet.
[14:58] <ttx> ScottK: ok, added
[14:59] <ttx> ScottK: I also have a "universe" section for last minte opportunity fixes on that page
[14:59] <ttx> ScottK: following up on your ML post
[14:59] <ScottK> Great
[14:59] <ScottK> Once the language packs are done and it's clear there aren't major Main uploads coming I'll push more of that
[15:04] <oru_work> can someone please remind me where DNS servers are defined ?
[15:04] <klaas> resolv.conf
[15:04] <klaas> in /etc
[15:05] <hggdh> smoser: I will run a topo2, yes. As soon as I find out wht cempedak rebooted in the middle of the night
[15:11] <oru_work> i see
[15:31] <nxvl> zul: i'm uploading fix for Bug #566803 to debian
[15:31] <nxvl> mathiaz: ^^
[15:31] <zul> nxvl: mind if i just cherry pick it?
[15:31] <nxvl> if you want to
[15:32] <zul> i really dont want to do a FFE for the version in debian
[15:32] <nxvl> ok, will paste the patch in a bit
[15:32] <nxvl> let me finish uploading
[15:33] <zul> thanks
[15:34] <nxvl> paste.ubuntu.com/421067/
[15:34] <nxvl> http://paste.ubuntu.com/421067/
[15:39] <crazygir> ssh known_hosts question.. how do you identify which host is which?
[15:39] <crazygir> I'm cursing these damn uuid conventions
[15:39] <crazygir> I have to remove one host from my known_hosts file, as the system has changed.. but knowing which host is which isn't so straight forward with uids
[15:42] <raphink> crazygir: you can use real names if you prefer
[15:42] <raphink> use HashKnownHosts no
[15:43] <raphink> if your ssh config
[15:43] <raphink> s/if/in/
[15:44] <ScottK> crazygir: When you fail to connect to the host, the error gives you the line number of the offending entry.
[15:44] <raphink> also
[15:50] <crazygir> thanks!
[15:57] <zul> nxvl: fix uploaded
[15:59] <nxvl> :D
[16:02] <kamusin> I was installing RC of Lucid and I get https://bugs.edge.launchpad.net/ubuntu/+source/debian-installer/+bug/569035, have you seen something similar?
 kirkland, mathiaz, hggdh, please update status for your remaining work items on http://people.canonical.com/~pitti/workitems/canonical-server-ubuntu-10.04.html -- or I'll assume it's the current state
[16:15] <kirkland> ttx: i just need to have a short conversation with hggdh to sign off on the rc tests
[16:15] <ttx> kirkland: ack
[16:16] <kirkland> ttx: actually, i can sign off on the RC-candidate tests
[16:16] <hggdh> ttx: I cannot perform the upgrade tests, cannot run KVM yet on my box
[16:16] <kirkland> hggdh: do you have mumble working yet?
[16:16] <hggdh> kirkland: yes, plugging in
[16:17] <kirkland> hggdh: jump in my channel for a couple of minutes, and let's hammer out the sign-off bits
[16:17] <hggdh> kirkland: ack
[16:18] <hggdh> kirkland: of course, if I do not sounds like a chipmonk
[16:30] <\sh> hmmm...I wonder why couchdb needs libasound somehow and several X libs
[16:30] <\sh> + libavahi on a server
[16:31] <\sh> oha...erlang and it's X crap
[16:35] <\sh> oh damn...it's xulrunner dep and I have a server with a desktop installation ;)
[16:39] <npope> greetings all.
[16:40] <npope> I am trying to move /var (which is currently on /) to its own disk, lvm.  Whenever i drop to single user mode and copy /var/* over to /new_var/ update fstab and reboot. it fails saying can not mount /var/lock and /var/run (however both directories both exsist on /new_var).  Any suggestions?
[16:52] <smoser> https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/564355
[16:57] <Eric^-> Hey gues
[16:57] <Eric^-> guys*
[16:57] <Eric^-> Does dell have a diffrent cord to external screens?
[17:05] <hggdh> smoser: topo2 logs from 04/23 are at tamarind:/home/cerdea/rig-topo2-logs.tar
[17:24] <jdstrand> mathiaz: hi!
[17:25] <jdstrand> mathiaz: why does the mysql error log have these permissions:
[17:25] <jdstrand> /var/log/mysql/error.log -rw-rw---- mysql adm
[17:26] <jdstrand> ttx: also, do you know why the tomcat webapps directory has these permissions:
[17:26] <jdstrand> /var/lib/tomcat6/webapps drwxrwxr-x tomcat6 adm
[17:27] <jdstrand> 'adm' seems on odd choice for the group in both...
[17:28] <jdstrand> zul: this looks odd:
[17:28] <jdstrand> /var/lib/php5 drwx-wx-wt root root
[17:28] <mathiaz> jdstrand: hm - it may be related to the umask
[17:28] <mathiaz> jdstrand: are you running a specific script?
[17:29] <jdstrand> mathiaz: I'm confused
[17:29] <jdstrand> mathiaz: what do you mean? I used find and discovered these
[17:29] <zul> jdstrand: not sure why?
[17:29] <mathiaz> jdstrand: sorry - you've just started to list a lot of permission problems
[17:30] <mathiaz> jdstrand: I thought you were a specific security scripts that look for these issues
[17:30] <mathiaz> jdstrand: that's all I meant
[17:30] <jdstrand> zul: world writable for /var/lib/php5?
[17:30] <zul> jdstrand: hmmm...
[17:30] <jdstrand> mathiaz: I am using the QRT get_file_info.sh script
[17:30] <mathiaz> jdstrand: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/lucid/mysql-dfsg-5.1/lucid/annotate/head%3A/debian/mysql-server-5.1.mysql.upstart
[17:30] <mathiaz> jdstrand: ^^ there is a umask settings there
[17:31] <mathiaz> jdstrand: I'm not sure what's the impact of it
[17:31] <jdstrand> mathiaz: I'm more concerned about the group 'adm' having write permission
[17:31] <mathiaz> jdstrand: so for the mysql error log the unusual part is that adm has write permision?
[17:31] <jdstrand> mathiaz: that is highly unusual for a log
[17:31] <mathiaz> jdstrand: right
[17:32] <zul> jdstrand: well i know the /var/lib/php5 is used for garbage collection at least
[17:32] <jcastro> mathiaz: so puppet is sending michael dehann, who worked as cobbler upstream for a while.
[17:32] <jdstrand> mathiaz: do you consider this a bug (I do)? if so, I can file it
[17:32] <mathiaz> jdstrand: if the umask is set to 007 could that produc such a permission?
[17:32] <jcastro> mathiaz: so if you care about that you can get 2 birds with 1 stone
[17:32] <mathiaz> jcastro: great - thanks
[17:32] <mathiaz> jdstrand: please file a bug
[17:33] <mathiaz> jdstrand: and outline why it's an issue
[17:33] <Scunizi> How do I list the contents of a directory on a samba share using cli?  ie.. "ls smb://<ip>/<share>" ??
[17:34] <jdstrand> mathiaz: possibly-- it could maybe be fixed with 027, but I don't know what impact that would have on other files created
[17:34] <jdstrand> mathiaz: I'll mention that in the bug
[17:34] <mathiaz> jdstrand: great - thanks
[17:34] <mathiaz> jdstrand: I don't think this is releace critical though
[17:34] <jdstrand> mathiaz: no, it isn't
[17:34] <Pici> Scunizi: You'll either need to mount it or use smbclient.
[17:34] <jdstrand> mathiaz: possibly, SRU, but that is your decision
[17:35] <Pici> Scunizi: smbclient probably has a batch-like mode where you can just issue an ls.
[17:36] <Scunizi> Pici: ok.. I'll man smbclient for info .. I've been messing with an Fstab line to auto mount the share but have been unsuccessful at this point.. manually mounting it is no problem
[17:42] <jdstrand> mathiaz: filed as bug #569085. I'll let you decide how to triage it
[17:44] <jdstrand> mdeslaur: do you know why php5 has this:
[17:44] <jdstrand> /var/lib/php5 drwx-wx-wt root root
[17:44] <jdstrand> mdeslaur: that looks very wrong to me...
[17:44] <jdstrand> mdeslaur: zul speculated garbage collection
[17:45] <mdeslaur> jdstrand: well, it's where php stores stuff like session information
[17:46] <mdeslaur> jdstrand: what exactly do you think is wrong with the permissions?
[17:47] <jdstrand> mdeslaur: I don't know what it is used for-- if it is only session info, then fine. if it is part of include_path, far less fine ;)
[17:49] <jdstrand> mdeslaur: I guess I was thinking sessions should be /var/cache/php5 or something... /var/lib seemed like it suggests something else (but may not)
[17:49] <jdstrand> hardy is the same btw...
[17:53] <jdstrand> mdeslaur: well, I've added it to my list of further investigations
[17:55] <mdeslaur> jdstrand: it's the default session.save_path
[17:56] <jdstrand> mdeslaur: ok cool. thanks :)
[19:20] <ivoks> kirkland: have a minute?
[19:21] <kirkland> ivoks: shoot
[19:21] <ivoks> kirkland: i'm investigating why qemu image won't boot with drive option boot=on (default and only option in libvirt)
[19:21] <ivoks> kirkland: and i noticed different instructions from seabios with boot=on, boot=off
[19:21] <ivoks> with boot=off, seabios resets ata drive
[19:22] <ivoks> while with boot=on it doesn't
[19:22] <kirkland> ivoks: interesting; i've not seen this
[19:22] <ivoks> kirkland: have you tried booting without kvm? :)
[19:23] <ruben23> hi any knows how to used phpsysinfo
[19:23] <ivoks> kirkland: http://pastebin.com/KbUhXJqW
[19:24] <ivoks> i've modified seabios's source to print debug info (level 8)
[19:24] <kirkland> ivoks: you mean just through qemu?
[19:24] <ivoks> kirkland: yes
[19:24] <kirkland> ivoks: i have not tried that
[19:24] <kirkland> ivoks: perhaps ask aliguori in #ubuntu-virt ?
[19:24] <ivoks> simple command: /usr/bin/qemu -M pc-0.12 -no-kvm -m 512 -smp 1 -name blabla3 -no-acpi -boot c -drive file=/home/ivoks/VM/blabla3/tmpojECnn.qcow2,if=ide,index=0,boot=off -serial file:/tmp/bla2.log -parallel none -usb -vga cirrus
[19:25] <ivoks> this works, but with boot=on it doesn't
[19:25] <Dr_Alien> Hi guys
[19:26] <Dr_Alien> i want to deploy a web server with an ftp function. ive been trying to install ebox but appamour fails and blocks the installation, any alternatives or ideas on what i should do? many thanks
[19:31] <nealmcb_> Dr_Alien: what is the apparmor error message?
[19:32] <Dr_Alien> just FAILED according to my putty system.
[19:32] <mathiaz> hggdh: I've just created an LP project for uec-testing-scripts
[19:32] <mathiaz> hggdh: I've also created a team uec-testing-scripts-dev to host the bzr branch
[19:32] <mathiaz> hggdh: I've added you as a member of the team
[19:33] <mathiaz> hggdh: which means you should have write access to the branch
[19:33] <mathiaz> hggdh: lp:uec-testing-scripts/
[19:35] <hggdh> mathiaz: thank you. And, BTW, the scripts are marvelous!
[19:35] <mathiaz> hggdh: thanks
[19:35] <mathiaz> hggdh: is the debugging info enough?
[19:35] <hggdh> yes, just received an email about the project
[19:35] <zul> mathiaz: ping for the testcase for 292971 be applied to the older versions as well?
[19:36] <mathiaz> zul: bug 292971 ?
[19:36] <hggdh> mathiaz: I think so. The only thing to look at is make sure the logs have the correct {debug,info,warning, etc} settings
[19:38] <mathiaz> zul: you may not be able to use the openldap-dit project for an older version of ubuntu
[19:38] <mathiaz> zul: especially for hardy
[19:39] <zul> mathiaz: i was afraid of that ;)
[19:39] <mathiaz> zul: you may be able to go back in the bzr branch history
[19:39] <mathiaz> zul: there should be a version in the history of the openldap-dit project that was targeted at hardy
[19:39] <mathiaz> zul: so you could use that version for testing
[19:40] <zul> mathiaz: ok thanks ill go pick away at it
[19:40] <mathiaz> zul: great - thanks
[19:41] <Dr_Alien> Neal ill run diagnostics in a sec and see if i can do it again
[19:42] <Dr_Alien> im not sure how to install ebox really..
[19:42] <Dr_Alien> i cant FTP files up
[19:42] <Dr_Alien> so i have to use ssh to get the files.
[20:02] <MattCampbell> I configured a virtual machine using libvirt, using the NAT networking option, and now I went to set up some port forwarding, e.g. so connections to port 25 on one of the host's IP addresses are forwarded to port 25 in the VM.  But I want the VM to see the real source IP.
[20:02] <MattCampbell> My guess is that I should use iptables to do this.
[20:08] <Dr_Alien> does anyone have issues configuring ebox?
[20:17] <guntbert> Dr_Alien: no offense - but don't you think that in the time you've put into ebox you would have been able to configure your server completely to your needs?
[20:17] <Dr_Alien> Guntbert, im having issues with adding new uses to an FTP server.
[20:17] <Dr_Alien> if you could tell me how to do this and seccure my ftp box that would be prue gold.
[20:19] <guntbert> Dr_Alien: that is nothing I have done - but for a start: what ftp server are you using?
[20:19] <Dr_Alien> vtftp i think
[20:20] <guntbert> Dr_Alien: sorry "I think" is not good enough - the configuration files and methods are different between servers
[20:21] <Dr_Alien> Ok let me see.
[20:21] <Dr_Alien> its http://doc.etherpad.org/F7Txtk7IeB
[20:21] <Dr_Alien> ops
[20:21] <Dr_Alien> sorry
[20:21] <Dr_Alien>  vsftpd
[20:23] <guntbert> Dr_Alien: please use my nick when you are talking to me to highlight me - I'm participating in several channels - let me look into the docu - I'll be back
[20:24] <Dr_Alien> Sure. thanks
[20:27] <hggdh> smoser: do you need topo3 on the rig for tests?
[20:28] <ruben23> hi nayone can help how to used phpsysinfo..?
[20:29] <smoser> did you already run them ?
[20:29] <smoser> hggdh, ?
[20:32] <hggdh> I ran the stress on topo3, 93% success. I will reconfigure to topo4 is you do not need the rig now
[20:33] <hggdh> smoser: ^
[20:33] <smoser> sure.
[20:33] <smoser> you can have it
[20:33] <smoser> can i see the topo3 results some where ?
[20:33] <smoser> and 93% of what
[20:33] <hggdh> k, topo4 on the way now ;-)
[20:34] <jeeves_Moss> how can I make a startup script that runs ONCE @ startup (ie. server startup) under a specific user?  (ie., I need to get 2 game servers to start @ startup, but have them run as a non priv user)
[20:34] <hggdh> smoser: 93% success on creating, using, and destryoying instances
[20:35] <smoser> T * .93 = S
[20:35] <smoser> what is the value of T
[20:35] <smoser> hggdh,
[20:35] <guntbert> Dr_Alien: please have a look at http://howto.gumph.org/content/setup-virtual-users-and-directories-in-vsftpd/  -- its written for debian but I see no reason why it wouldn't apply to ubuntu - ask if questions remain :-))
[20:36] <hggdh> smoser: 400
[20:36] <FFF666> I've started an image in UEC, it's my first run.  Then I use the "watch -n5 euca-describe-instances" command to see the current state of the image. After a while the image doesn't start
[20:36] <hggdh> smoser: 400 total instances, 384 OK, 16 KO
[20:37] <smoser> yeah. can you post console logs of 16 failures ?
[20:37] <FFF666> smoser: are you talking to me?
[20:38] <hggdh> smoser: I will upload everything (uec_test.py results, euca logs) to tamarind
[20:38] <smoser> FFF666, no, i wasnt.
[20:38] <smoser> but what is "a while"
[20:39] <FFF666> 3 minutes
[20:39] <FFF666> but the state now says terminated
[20:39] <smoser> my first guess is that image is to big to fitin your --instance-type
[20:39] <FFF666> i put small
[20:40] <smoser> whats your image ?
[20:40] <FFF666> ahhhhh
[20:40] <FFF666> ubuntu 9.10
[20:40] <smoser> yeah, that wont fit in 2G
[20:40] <smoser> the lucid ones do
[20:40] <smoser> so you can either
[20:40] <smoser> a.) run lucid
[20:40] <smoser>  (image)
[20:40] <smoser> b.) modify instance type of small to 3G
[20:40] <smoser> c.) use --instance-type c1.medium
[20:41] <FFF666> ok, I'll try
[20:41] <hggdh> smoser: tamarind:/home/cerda/rc-topo3.tar
[20:42] <hggdh> smoser: sorry, /home/cerdea
[20:42] <funkycat90210> I'm having a problem whereby 8.04LTS packages are too old for newer apps I want to install. So I want to move from 8.04LTS to 9.10 or even 10.4 non-LTS, can I upgrade remotely or should I do a fresh install?
[20:43] <npope> funkycat90210: upgrade remotley with dist-upgrade
[20:44] <ScottK> funkycat90210: Upgrading is generally reasonably safe, however to upgrad to 9.10 you need to do it in steps: 8.04 -> 8.10 -> 9.04 -> 9.10.
[20:44] <funkycat90210> npope, thx
[20:45] <npope> funkycat90210: you need to follow ScottK's steps though, you cant just go from 8.04 to 9.10
[20:45] <guntbert> funkycat90210: LTS -> LTS goes in one step - every other combination: only step by step
[20:45] <npope> can you go from 8.04 to 10.04 yet?  (10.04 is not released yet?)
[20:45] <ScottK> Also use update manager to do it.
[20:46] <ScottK> npope: It's not recommended unless you know what you are doing.
[20:46] <funkycat90210> guntbert, which is fine assuming the steps work, some reports on the web suggest that doing so results in an unbootable system requiring manual intervention so I may have to drive to the datacenter to do this
[20:46] <ScottK> sudo do-release-upgrade is the command
[20:46] <npope> ScottK: dist-upgrade?
[20:46] <ScottK> npope: Going direct from 8.04 -> 10.04 before it's release is not recommended for general use.
[20:47] <FFF666> smoser: I installed the cloud UEC in the vmware  virtual machine, is that ok?
[20:47] <ScottK> do-release-upgreade is better for going from release to release than just dist-upgrade.
[20:47] <funkycat90210> ScottK, i dont have access to the gui but would do-release-upgrade do it?
[20:47] <npope> ScottK: oh I know thats why I posed the question
[20:47] <ScottK> Yes
[20:47] <ScottK> funkycat90210: Yes.  That's all cli
[20:47] <funkycat90210> ScottK, nice
[20:47] <guntbert> funkycat90210: well - in my experience upgrades only have problems with things like sound or video - which you don't need in this case :-)  but of course my experience is limited
[20:49] <funkycat90210> guntbert, thx.. I'll upgrade a machine and see how that works, in theory I can upgrade all 6 or so machines via do-release-upgrade
[20:50] <smoser> FFF666, i'm not sure if you can install all components on a single system or not.  i know its on-trivial. lifeless was working on trying that.
[20:50] <funkycat90210> so do-release-upgrade 8.10 then 9.04 then 9.10
[20:50] <guntbert> funkycat90210: yes but either you wait for the release of lucid or you do it step by step :-)
[20:50] <smoser> additionally you need vt extensions on the node controllers (ie, they must run 'kvm-ok')
[20:50] <smoser> i'm not sure if that is possible in a vmware guest or not
[20:51] <smoser> it *could*, but its not going to be fast.
[20:51] <funkycat90210> guntbert, so when lucid comes out i'll be able to do do-release-upgrade 10.04 directly?
[20:52] <FFF666> smoser: I followed this guide https://help.ubuntu.com/community/UEC/CDInstall.
[20:52] <FFF666> I have two ubuntu servers installed
[20:52] <guntbert> funkycat90210: according to documentation yes - thats one of the points of LTS releases
[20:53] <ScottK> You CAN do it now, it's just not recommended or supported.
[20:53] <funkycat90210> guntbert, well in that case I'll go from 8.04LTS to 10.10LTS and I want to move away from LTS
[20:53] <smoser> FFF666, you have 2 virutal servers installed as vmware guests ?
[20:53] <funkycat90210> err 10.4 LTSi ment to say
[20:53] <guntbert> Dr_Alien: there will be one or two adaptions to the tutorial
[20:54] <funkycat90210> Can I go from 10.4LTS to 10.10non-LTS?
[20:54] <smoser> FFF666, what does 'kvm-ok' show on the node controller
[20:54] <FFF666> smoser: yes I've two
[20:55] <guntbert> funkycat90210: of course you will be able to do that - but from  then  on you will need to upgrade step by step
[20:55] <FFF666> smoser: kvm-ok  is a coomand?
[20:55] <smoser> y
[20:56] <FFF666> it says: your cpu does not support KVM extentions
[20:56] <FFF666> KVM acceleratoin can NOT be used
[20:57] <funkycat90210> guntbert, got it.
[20:57] <FFF666> Do I have to used two real compuers?
[20:58] <smoser> FFF666, then without some hacking you wont be able to do this.
[20:58] <smoser> FFF666, the requirement is for vt extensions on the cpu.
[20:58] <smoser> most likely that means "real computers"
[20:59] <FFF666> but Ihave a "new" pc, it's a core 2 duo 8600,
[20:59] <FFF666> but Ihave a "new" pc, it's a core 2 duo P8600,
[20:59] <FFF666> I'll check the bios
[21:00] <smoser> because your host supports VT does not mean your guests do
[21:00] <smoser> kvm has support for nested vt with amd processors
[21:00] <smoser> it would appear that vmware does not support nested vt with intel
[21:00] <smoser> it *could*
[21:00] <funkycat90210> ohh so I should wait for 10.4 LTS, upgrade to that from 8.04 LTS, then when 10.10 comes out, edit /etc/update-manager/release-upgrades, change prompt from lts to normal then do do-release-upgrade
[21:01] <funkycat90210> sounds right or do i need more coffee?
[21:02] <funkycat90210> http://www.howtoforge.com/how-to-upgrade-ubuntu-8.04-to-ubuntu-8.10-desktop-and-server has info on upgrading out of LTS
[21:06] <sirninja> I've put ubuntu server 9.10 on an old laptop that I don't use any more. How do I get it to automatically connect to a wireless network on startup?
[21:07] <sirninja> and I know I don't need restricted drivers because in the desktop version, the wireless card was detected automatically
[21:09] <funkycat90210> sirninja, years ago I set up a custom wpa_supplicant command to do this, it seems like distros assume only people logging into a gui will use wireless.
[21:09] <sirninja> funkycat90210: if it's really difficult, it's not that big of a deal to run a cable, will that be easier?
[21:11] <funkycat90210> sirninja, much easier
[21:11] <funkycat90210> sirninja, just set up dhcp/static ip and it will work on bootup
[21:12] <sirninja> funkycat90210: I found wicd supports wireless networks. Would that work?
[21:12] <FFF666> smoser: are you there?
[21:12] <smoser> yeah
[21:13] <FFF666> smoser: the image state is terminated, it doesn't work
[21:13] <smoser> right.
[21:13] <smoser> its not going to.
[21:13] <FFF666> smoser: why?
[21:14] <smoser> because you have to have VT support in the node controller
[21:14] <smoser> ie, 'kvm-ok' needs to say : KVM acceleration can be used
[21:15] <FFF666> so, that is a requerimient?
[21:15] <AtomicSpark> Does anyone know what power profile Ubuntu Server defaults to? Ubuntu Desktop defaults to on demand where it scales back the cpu freq when it's not needed.
[21:15] <smoser> *or* you'd need to hack it so that it would use kvm without acceleration, whic hwoudl give you terribly slow guest performance (guest performance without vt acceleration is bad on real metal in kvm)
[21:15] <smoser> FFF666, yes, it is a requirement.
[21:15] <funkycat90210> sirninja, non familiar with wicd
[21:15] <AtomicSpark> smoser: which is just qemu.
[21:16] <smoser> AtomicSpark, right.
[21:16] <AtomicSpark> indeed.
[21:16] <AtomicSpark> I just joined, let me guess, someone wants to libvirt without cpu extentions?
[21:16] <smoser> this could in theory work.  with amd processors, you could run 2 guests on the same virtual network using kvm as the hypervisor.
[21:17] <smoser> then, those guests would see vt support also.
[21:17] <smoser> it'd be slow. but it would (in theory) work.
[21:18] <smoser> AtomicSpark, well, yes, but at a higher level. FFF666 wants to run UEC (which uses libvirt) inside VMware guests.
[21:18] <smoser> it isn't an uncommon request.. it makes sense for trying something out.
[21:18] <smoser> its just not supported.
[21:18] <AtomicSpark> So basically a VM within a VM?
[21:19] <FFF666> yes
[21:19] <smoser> yeah
[21:19] <AtomicSpark> Interesting.
[21:19] <smoser> nested vt is supported with kvm and amd
[21:19] <AtomicSpark> I did not know that.
[21:19] <smoser> http://www.linux-kvm.com/content/kvm-nested-virtualization-works
[21:19] <AtomicSpark> Makes sense I suppose.
[21:20] <FFF666> smoser: so, I have to install UEC in a real PC?
[21:21] <smoser> the simple answer is yes
[21:21] <smoser> on 2 real pcs
[21:21] <smoser> that have VT extensions
[21:22] <FFF666> I've in my laptop VT extension, it's a core2 duo P8600
[21:22] <smoser> yeah. anything recent other than netbooks its the norm
[21:22] <smoser> and if its amd64 you'd be hard pressed to not find it
[21:22] <AtomicSpark> It's tricky with Intels. They didn't put VT on everything.
[21:23] <FFF666> but why the kvm-ok command says that I don't have acceleration?
[21:23] <smoser> or, rather, they put VT on everything and disable it control pricing more effectively :)
[21:23] <AtomicSpark> FFF666: Did you enable it in BIOS? Did you know you have to power off *and* remove the battery?
[21:23] <smoser> FFF666, because the virtual machine that you're running that command in does not have vt
[21:24] <FFF666> ok, I'll check the bios
[21:24] <AtomicSpark> smoser: Something like that, yes. Business 101: Make a single thing with a set of features then clone it and disable features. Instant tiered pricing!
[21:25] <FFF666> Will you both still there?
[21:25] <FFF666> 	Will you both still here?
[21:25] <smoser> i'll be here for a bit.
[21:25] <AtomicSpark> Maybe. I don't lurk in here often, but I'll wait for you.
[21:25] <smoser> FFF666, most likely your laptop does have it
[21:26] <smoser> you ran kvm-ok on your laptop ?
[21:26] <smoser> "bare metal" ?
[21:26] <FFF666> ok, if you havent noticed I'm learning english
[21:26] <FFF666> jaaj
[21:26] <FFF666> or haha, as you laugh
[21:26] <FFF666> :)
[21:26] <AtomicSpark> smoser: i didn't know about kvm-ok. what is that packaged with? the wiki just tells you to grep the cpu info file.
[21:27] <smoser> in lucid it is packaged with cpu-checker
[21:28] <AtomicSpark> Hmm.
[21:29] <smoser> previously it was in kvm.  which didn't make it easy to use. i *think* (kirkland would know better) it was moved to a different package, that was tiny and easily installed (possibly by default), so that you could run it without installing kvm
[21:29] <AtomicSpark> Makes sense.
[21:29] <AtomicSpark> I use virt-manager to remotely manage/install virtual machines on my headless proliant server. Works *very* well.
[21:42] <hggdh> kirkland: topo3 complete, 400 total runs, 16 failures, available at tamarind:/home/cerdea/rc-topo3.tar
[21:42] <hggdh> kirkland: now setting up topo4
[21:45] <hggdh> BTW -- the package is cpu-checker, and -- IIRC -- it is installed by default
[22:16] <sirninja> how do I set up a static ip?
[22:17] <AtomicSpark> sirninja: /etc/network/interfaces
[22:17] <AtomicSpark> sirninja: man interfaces
[22:17] <AtomicSpark> sirninja: or read the ubuntu server guide
[22:18] <AtomicSpark> It's pretty simple.
[22:19] <AtomicSpark> With that said, I usually use static DHCP for all of my servers so I can have a single place to reassign them. Changing many computers would be a pita without it.
[22:20] <sirninja> AtomicSpark: That's a good idea. I didn't think of doing that
[22:20] <AtomicSpark> It's a very good idea. Unless your DHCP server goes down. :)
[22:21] <AtomicSpark> But most people seem to use hostnames anyways for uris. So if your DHCP/DNS server goes down, you're screwed either way.
[22:22] <sirninja> AtomicSpark: well, I'm just setting up a torrent box, so it may just be easier to set it on the actual torrent box
[22:22] <AtomicSpark> Ah. I still use static dhcp. Very nice. Same place to set up port forwarding too.
[22:23] <sirninja> AtomicSpark: I'm looking to see if my router has an option for that, but I don't see any
[22:23] <AtomicSpark> Shame.
[22:23] <AtomicSpark> Mine is called static leases?
[22:23] <AtomicSpark> I use the Tomato firmware on my WRT54GL
[22:24] <sirninja> I found it
[22:24] <AtomicSpark> Yay!
[22:24] <sirninja> I'm using dd-wrt
[22:24] <AtomicSpark> Boo. :P
[22:24] <sirninja> it's the only thing my router could run besides the default firmware sadly
[22:24] <AtomicSpark> I know. :(
[22:25] <AtomicSpark> I forgive you for your gpl violation.
[22:25] <AtomicSpark> Dinner time.
[22:33] <ruben23>  hi guys how do i ssh command form my linux box
[22:50] <ecrane> Hi can someone advise me how to start ubuntu's NFS server in verbose mode? Need to debug why other unix boxes can connect to my NFS shares, but windows can't....
[22:57] <konqrunner> @ecrane: Did you check the 'Troubleshooting' section at https://help.ubuntu.com/community/NFSv4Howto ?
[23:00] <ecrane> yeah but I saw nothing in there for logging for nfsd, just for logging on gssd/kerberos
[23:10] <konqrunner> @ecrane: Looks like you have to start logging differently for the different daemons.
[23:10] <konqrunner> I found this site: http://docs.hp.com/en/5992-0715/ch08s05.html where they explain logging for rpc.mountd, rpc.statd and rpc.lockd
[23:11] <konqrunner> You'll probably have to adjust paths etc. to your distro
[23:12] <konqrunner> I guess with connection problems, rpc.mountd would be the natural choice, right?
[23:29] <ecrane> konqrunner: good stuff, ty.
[23:30] <konqrunner> you're welcome. HTH
[23:33] <Jeeves_Moss> what am I doing wrong with my apache file?  http://pastebin.com/zzEriT4m  I can't get my v-hosts to work
[23:36] <konqrunner> Jeeves_Moss: What do you mean by 'can't get my vhosts to work'? What kind of error do you get?
[23:37] <Jeeves_Moss> konqrunner, it won't direct to the proper public_html directory
[23:40] <Jeeves_Moss> konqrunner, so, I don't think that it's directing to the proper directory baised on domin name
[23:42] <konqrunner> Jeeves_Moss, Did you enable the config file for the vhost you're trying to access?
[23:45] <konqrunner> Jeeves_Moss: Check /etc/apache2/sites-enabled. There needs to be a symlink to the vhost config file in /etc/apache2/sites-available/
[23:45] <konqrunner> Then you'll have to restart httpd, naturally
[23:46] <konqrunner> If that doesn't help: Do you have any server management software installed. Esp. Plesk is very unnerving if it comes to overriding manual configs...
[23:49] <Jeeves_Moss> konqrunner, nope, virgin install
[23:50] <konqrunner> How about that symlink?
[23:50] <Jeeves_Moss> ???
[23:51] <konqrunner> With apache2, you can configure as many vhosts as you like, without taking up resources
[23:51] <konqrunner> So, if you want to actually run a vhost, you need to tell apache that you want to enable it
[23:52] <Jeeves_Moss> well, that's a copy of the vhost config file for one of the domains
[23:52] <konqrunner> you do so by placing a symlink in /etc/apache2/sites-enabled, which points to the configuration file in /etc/apache2/sites-available.
[23:52] <Jeeves_Moss> yes, it's set up like that
[23:53] <konqrunner> But you do get a page served when you call up the vhost's URL?
[23:54] <Jeeves_Moss> I get the default "it works!" page
[23:54] <bogeyd6> !apache2
[23:54] <bogeyd6> hmm
[23:54] <bogeyd6> !apache
[23:55] <bogeyd6> https://help.ubuntu.com/8.04/serverguide/C/httpd.html
[23:56] <Jeeves_Moss> konqrunner, any ideas?
[23:56] <konqrunner> Hang on, I had a similar error a while back. Trying to figure out, what it was...