=== Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [01:24] jamespage, http://people.canonical.com/~agandelman/ca/grizzly/python-eventlet-0.12.1-0ubuntu1.1~cloud0/ [01:24] zul, ^ [01:25] adam_g: +1 [01:25] jamespage, also, i guess i broke the script last time i fixed it :) https://code.launchpad.net/~gandelman-a/cloud-archive-utils/better_version_query/+merge/181184 [01:25] zul, thanks [01:44] ah very nice to find this [01:44] trying my hand w/LVM for the first time, gernerally have avoided it [01:44] i hot-added a 2nd disk to my system, but i'm not seeing /dev/sdb [01:45] xibalba: check dmesg | tail, the kernel might pick a different name.. [01:45] ah i grepped for sdb [01:45] checking [01:46] doesnt appear to have shown up yet [01:46] virtual machine btw, not that it makes much difference [01:46] it should show up as a /dev/sd* though [01:46] right? [01:47] xibalba: maybe /dev/virtio*? I'm not positive there.. [01:47] hmm i think that'd be for KVM, though I will check. This is a vmware guest [01:48] rebooting to see if it pulls it in, though it shouldn't require a reboot [01:48] yeah now it shows, drats [01:48] so i want to expand my / volume onto /dev/sdb [01:49] k i need to part it first [01:54] PostgreSQL 9.1 on 12.04 insists on installing with default encoding LATIN1. What magic locale incantation do I need? [01:54] I've done update-locale en_US.UTF-8 [01:54] or LANG=en_US.UTF-8 rather [01:58] so when adding a 2nd disk to my LVM [01:58] do i need to make another logical volume for it? [01:59] how do i instruct lvcreate to use the whole disk [02:01] josePhoenix: this thread may be helpful: https://lists.ubuntu.com/archives/ubuntu-devel/2013-June/037216.html [02:01] thanks [02:03] sarnold: the thing is, I'm not seeing template1 with SQL_ASCII... or UTF-8 [02:03] I'm not sure where it's picking up LATIN1 [02:04] /etc/default/locale and /var/lib/locales/supported.d/local both reflect my choice of en_US.UTF-8 so I'm flummoxed :\ [02:05] josePhoenix: if you can apt-get purge postgresql and then re-install, that _might_ be a help. At least, that's how I'm reading this... [02:05] josePhoenix: it looks like it uses whatever locale was in place for the process that starts the install [02:06] josePhoenix: or, apparently, set the locale per-database in postgres once it's running? [02:06] I have done the purge and reinstall thing a couple of times after trying different ways to set the locale for dpkg [02:06] or apt-get rather [02:07] looks like I can create a UTF-8 db off template0? [02:07] josePhoenix: ooh, how about /etc/postgresql/version/cluster/environment ? [02:07] doesn't exist on here [02:07] any LVM experts here? [02:07] ah, sorry [02:08] no, I think template0/template1 are created at the same time that file is written :\ [02:08] josePhoenix: hey! comment #12: https://bugs.launchpad.net/ubuntu/+source/postgresql-common/+bug/969462 [02:08] Launchpad bug 969462 in postgresql-common "fails to start after install if invalid locale is set" [Low,Fix released] [02:09] josePhoenix: turns out nearly everything pitti said in that mail thread is incorrect now. sorry. [02:09] haha [02:09] no worrie [02:09] s [02:10] josePhoenix: so, I guess check /etc/environment and make sure that doesn't override /etc/default/locale, or the other way around.. [02:10] how do i grow the LVM after doing a lvextend ubuntu /dev/ubuntu/root [02:10] oy vey [02:11] josePhoenix: oh. that's unstable. debian unstable. wonder if that's made it to ubuntu yet? sheesh, maybe it's time for me to be done for the night. :) [02:11] well thanks for helping me look into this [02:11] it looks like some kind of voodoo no matter how you slice it :\ [02:12] i used resize2fs /dev/ubuntu/root 8G [02:12] but its telling me on-line resizing required [02:12] Maybe I need to actually reboot for locale changes to take effect? [02:12] That would not work if I'm trying to do this in a vagrant provisioning script :\ [02:12] josePhoenix: that'd be unfortunate but possible :( [02:13] well, worth checking at least [02:13] josePhoenix: you could probably set the locale in a preseed file [02:13] xibalba: did you resizefs before or after growing the physical space available? [02:13] i extended it to a 2nd disk [02:13] I'm not even sure how I'd do that :) [02:13] i just figured out one piece [02:13] did a lvresize -L +8GB /dev/ubuntu/root [02:14] was resized successfully, now need to grow the FS [02:14] If it'd require making a new box, I might as well override the locale on this db and call it a day [02:14] xibalba: aha; you can't resize filesystems that are mounted read-write, only read-only, and .. I'm not even sure which ones you can re-size online while mounted read-only... [02:14] so it said "extending logical volume root to 23.24GB" [02:14] oh damn i thought LVM can resize online [02:14] well i guess it can but not the filesystem [02:15] josePhoenix: sounds fair to me.. [02:15] xibalba: that's my understanding [02:15] fuckin a, windows can do it! =P [02:17] assuming the kernel supports on-line resizing. (As of this [02:17] writing, the Linux 2.6 kernel supports on-line resize for [02:17] filesystems mounted using ext3 and ext4.). [02:17] xibalba: looks like i was wrong. :D [02:17] hmm, but how is the grand question [02:18] resize2fs -p /dev/ubuntu/root 8G isn't working for me [02:18] just says on-line resizing required, can't find that mentnioned in the man page [02:18] mentioned* [02:18] running 3.2.0-29 [02:19] * xibalba hmmms loudly [02:19] xibalba: how about just "resize2fs /dev/ubuntu/root" ? [02:20] oh [02:20] well how aobut that [02:20] thanks buddy [02:20] * xibalba sends sarnold a virtual high five [02:20] woo! [02:21] xibalba: darn. now I wish I'd suggested trying without the 8G and then again without the -p. [02:21] which one was it? now we'll never know.. [02:30] without the -p for whatever reason [02:30] this is going to be sweet, not i can through this into a bash script w/vmware customization [02:30] and call the PREINIT portion of the customization, and resize the disks for my dingleberry devs on the fly [02:30] w/out their bitching and moaning about having to do anything sysadmin related [02:31] time for food, bbl [02:48] sarnold: well, it wasn't pretty, but I just overrode every locale setting I could when creating the db and that seems fine [02:48] thanks for your help :) === bitnumus is now known as Guest74220 [04:23] any gotchas or issues to be aware of when using LVM? [04:25] !lvm [04:25] Tips and tricks for RAID and LVM can be found on https://help.ubuntu.com/community/Installation/SoftwareRAID and http://www.tldp.org/HOWTO/LVM-HOWTO - For software RAID, see https://help.ubuntu.com/community/FakeRaidHowto [05:39] Does anyone here know how I could proxy connections from a server with no firewall to a server with a firewall (unable to portforward)? === markthomas1 is now known as markthomas === Guest74220 is now known as bitnumus === b0rg is now known as B0Rg666 === smb` is now known as smb [08:10] smb, hey - I have an openvswitch module built from upstream trunk with some fixes for 3.11 broadley working [08:10] but I'm getting a kernel oops when de-registering devices.... [08:11] Would it be worth opening a bug for it. Usually its simpler to have all output and info together [08:11] smb, I'll append to the existing bug report [08:12] smb, my plan is to fix this in upstream trunk then backport the required commits for our package [08:13] jamespage, Ok, and remind me of its number. Right, Oh and maybe a place to have your copy of upstream trunk to look at. [08:13] smb, sure [08:13] smb, https://bugs.launchpad.net/ubuntu/+source/openvswitch/+bug/1213879 [08:14] Launchpad bug 1213879 in openvswitch "kernel fault ovs 1.10.1 + linux 3.11" [Critical,New] [08:45] smb, I have an inkling as to where the problem is [08:47] jamespage, Well crashing on removal usually is some structures or resources used after free [08:47] smb, yeah [08:47] smb, its a WARN_ON [08:47] rather than BUG_ON [08:48] Oh, ok not as drastic as I thought then [08:48] smb, ovs uses its own workqueue implementation [08:48] smb, I'm wondering whether is something racey going on between the ovs workqueue and the main system_wq [08:49] jamespage, hm, i'd need to look closer at the code and the error/warning [08:50] Of course I find that any level of my machinery may get updated before proceeding === HisaoNakai_ is now known as HisaoNakai [09:17] hi! i have an ubuntu-cloud-image configured and running fine with qemu via cloud-localds. but booting lasts quite long when i don't provide the second disk with the cloud-init data. do i have to provide this disk all the time? [09:24] webwurst: cloud images usually look for cloud-init (userdata) data [09:24] webwurst: if they can't find it, they keep trying really hard for 2 minutes [09:24] then they give up [09:26] ivoks: is it possible to change this behavior after i have configured my image? [09:26] well, you can edit the image and remove cloud-init upstart jobs [09:27] but, what's the point of using cloud image then? [09:27] how would you loing into that machine then? [09:28] log in [09:28] * ivoks needs coffee [09:41] ivoks: it seems i can log in via ssh without cloud-init-userdata. after i started once with the data provided.. [09:42] whats the most secure IRC application server? [09:44] webwurst: oh, so you initially start with cloud-init, and then want to disable it? [09:44] ivoks: i just want to have stand-alone virtual ubuntu image with easy and fast setup. cloud-init seemed to be fine for that. will try to disable cloud-init upstart and see if it works for me [09:45] ivoks: yes :) [09:45] webwurst: in that case, after first boot, just disable cloud-init upstart jobs [09:56] smb, OK - so I tried my patch against the 1.12 branch and I don't see that kernel warning when running the test with 3.11 [09:57] smb, hmm- might consider a version jump to support 3.11 - its late but 1.12 is much closer to supporting this that 1.10 [09:58] jamespage, *sigh* and I still wait on the update of my VM... But agreed if 1.12 is actually already released upstream, it might be worthwhile to jump there [09:58] smb, yeah - gonna ask about status of 1.12 [09:59] it might just be a dev branch [10:21] jamespage, Yes as there seems not to be a tag of same name, but neither there would be for a 1.11 [10:21] smb, no tags - branch-1.12 [10:22] smb, I queired upstream anyway [10:22] jamespage, Yes, that was what I tried to say. :) Sounds good [10:22] smb, I'll prepare a package update to 1.12 on the assumption that even if we go with a snapshot, its a better place than we are right now with 1.10 on 3.11 kernel [10:22] thats OK [10:23] Yeah, at least this does not seem to cause warning/issues in your testing. [11:05] smb, lp:~james-page/ubuntu/saucy/openvswitch/1.12-snapshot if you want to try it [11:07] jamespage, I can look at compile and load/unload. Not making much use of it normally for what I do [11:13] smb, the package contains dep-8 tests that exercise the module - see debian/test/openflow [11:13] jamespage: erm adt failures? [11:13] jamespage, Ah ok, will do [11:14] zul, I just saw a heat one - importing yaml? wtf [11:14] *boggle* [11:17] zul, [11:17] import simplejson as json [11:17] ValueError: bad marshal data (unknown type code) [11:17] that sounds like a fundamental problem with simplejson to me [11:17] all on amd64 [11:17] ummmm.. [11:36] ivoks - whats up with the horizon packaging? [11:36] context - bug 1210253 [11:36] Launchpad bug 1210253 in cloud-archive "With Havana 2 installed, Launching horizon UI results in the error " NameError: name 'Dashboard' is not defined"" [Undecided,Confirmed] https://launchpad.net/bugs/1210253 === cyphermox_ is now known as cyphermox === dosaboy_ is now known as dosaboy_afk === cosmin is now known as hit === hit is now known as cosmin === psivaa is now known as psivaa-afk === cosmin is now known as hit === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [12:24] rbasak, hallyn_, zul, roaksoax, adam_g: if any of you have time I'd appreciate it if you could test bug 1213021 [12:24] Launchpad bug 1213021 in openvswitch-lts-raring "[SRU] openvswitch 1.4.0 not compatible with lts-raring HWE kernel" [High,Fix committed] https://launchpad.net/bugs/1213021 [12:25] its in precise-proposed awaiting verification [12:25] I'm using it on one of my machines OK [12:25] uvirtbot: i would if i had the hardware :( [12:25] zul: Error: "i" is not a valid command. [12:25] jamespage: i would if i had the hardware :( [12:25] zul, cloud instance? [12:26] jamespage: good point [12:29] jamespage, Ok, so fwiw the 1.12 snapshot compiles, succeeds in compiling the dkms package, and runs those tests with the 1.12 userspace installed [12:30] smb, w00t! [12:30] smb, OK _ I'm just waiting from a statement back from upstream on 1.12 status and I'll upload... [12:30] ack [12:30] smb, the package still has two patches cherry picked from trunk [12:30] well one from trunk and the other is my 3.11 support patch (not yet accepted!) [12:33] jamespage, I see. Ah yeah that one has the weird move of a hunk of code removed. The rest kinda looked like expected on a glance [12:36] jamespage, Hm, that comment about dom0 networking... Wonder whether that is always or only if allowing the xen scripts to configure bridges (which is one of those things deprecated but probably still possible) [12:48] jamespage: some files missing; zul is already on top of it === psivaa-afk is now known as psivaa [14:02] i've been compiling my kernels for 12.04 since the original 3.2.0 one. However, as soon as I go above 3.8, i.e. 3.9 in this case, I only get black screen after grub (unless i manually specify vgamode). what changed after 3.8 kernel that i have to do this? [14:02] same story for 3.10 [14:12] jamespage/roaksoax: https://code.launchpad.net/~zulcss/horizon/1210253/+merge/181311 [14:17] zul, +1'ed but still " 163. By zulcss@ubutu.com 7 minutes ago " [14:18] disposable, you know that Ubuntu supports the 3.5 and 3.8 kernels from quantal and raring on 12.04 as well right? [14:18] jamespage: crappers ill fix that after [14:18] zul, fix that before you merge it [14:18] otherwise the commit will have wonky info [14:19] jamespage: k [14:21] jamespage: yes, but the patches i need are only for 3.7 and 3.9. [14:22] disposable, are they bugs or features? [14:32] adam_g, hey - the ncc redux is installing the -agent packages for quantum plugin [14:32] it just need to install the -plugin package like it does for NVP [14:33] adam_g, also fwiw I think we should deprecate and ignore the quantum network configuration stuff [14:33] its awkward [14:33] maybe that should be policy [14:42] jamespage: it doesn't matter (it's linux-vserver), i just need to understand why booting up 3.9 kernel with 'gfxmode $linux_gfx_mode' gives me black screen and works without it [14:43] disposable, try in #ubuntu-kernel [14:43] jamespage: thanks, i didn't know such channel existed [14:47] anyone know the differences between libreadline6-dev & libreadline-gplv2-dev? [14:47] where did the gplv2 package come from? === freeflying is now known as freeflying_away [15:00] rbasak: howdy! [15:00] rbasak: could you please verify this sru bug #1172966 [15:00] Launchpad bug 1172966 in maas "SAY command in config.local.template breaks local boot on highbank" [High,Fix committed] https://launchpad.net/bugs/1172966 [15:01] roaksoax: I don't have access to a highbank maas installation any more [15:02] zul, bah - samba snuck in a new bd from universe on the last merge I did [15:02] rbasak: ack! thanks [15:03] jamespage: oh? [15:03] libtevent-dev [15:03] roaksoax: I was going to ask matsubara but I can't find him. He used to manage the MAAS QA lab - not sure if he still does. [15:03] rbasak: he does, but I think he is on vacation or something [15:03] Ah OK [15:04] jamespage: gah.. [15:09] jamespage: want me to do a MIR [15:09] zul, samba includes an embedded version [15:09] just looking at why that is no longer enough [15:09] jamespage: cool ok === Ursinha is now known as Ursinha-afk [15:16] jcastro: "Ubuntu Server Guide - LTS only ( Community )" might struggle for attendance with the clashes? [15:17] zul, yeah - might want to mir that [15:18] jcastro: If you could swap it with, say - Flavors Quality Assurance ( Community ).. I imagine you'll get better coverage [15:18] jamespage: cool ill do it this afternoon [15:18] Daviey: sure, what day is it? === Ursinha-afk is now known as Ursinha [15:18] zul, thanks v much [15:19] zul, should be trivial as its a breakout of embedded code to a sep library [15:19] Daviey: ok I see it [15:19] I'll ask to swap it [15:19] so its technically already in main [15:21] Daviey: Done! [15:21] zul, jamespage, adam_g, We have a "Openstack steps for saucy" and a "openstack charm for saucy".. I think it made sense to have both for the last vUDS, but do you think there is enough content for both at this one? [15:22] Daviey: im sure there is [15:22] Well, currently we are oversubscribed by one session.. If there is one session that can be cropped, that would be helpful.. Not not necessary if we use the 3rd room. [15:23] Daviey, I think we could squash into a single session [15:24] Yeah, that is what i was thinking.. [15:32] i though vuds was wed thur friday [16:12] hallyn_: ping [16:22] * hallyn_ lays low [16:23] zul: . [16:24] hallyn_: can you run lxc-create with the template? [16:24] what do you mean? [16:25] well say i have a config already and a fstab in my /var/lib/lxc/ and if i run lxc-create -n will it work? [16:25] zul: you're wanting to run *just* the template then? [16:26] hallyn_: right since i already have an image [16:26] mounted via loopback [16:27] zul: the template is the thing that creates the rootfs (and fills in some config details like devices.allow). If you alredy have both a config and a rootfs... why do you need to lxc-create at all? [16:27] thats what i was thinking [16:27] zul: is your config a full config? [16:28] hallyn_: yeah its going to be [16:28] then i guess just use what you've got :) [16:31] i intend to :) [16:31] then what was your question? [16:32] fg [16:52] hallyn_: you already answered it [16:52] zul: ok :) === psivaa is now known as psivaa_ === psivaa_ is now known as psivaa- === psivaa- is now known as psivaa [17:34] I'm trying to install ubunto server 12.10 onto a VM, got it up and running but having network issues, can only ping from vm->host and host->vm and am forced to use the vmware workstation console. My thought is if I can see the VM I could login to it via ssh, alas sshd is not installed on the server. I now have the iso mounted at /mnt/cdrom but how does one install a package like sshd from the iso image? [17:35] hallyn_, "I'm not sure. It might be worth it for users. But it's starting to sound like enough code that it might not be justified" [17:35] i'm willing to accept that. [17:36] rbasak's plan is to wrap lxc invocation, so we can put alias like things there. [17:36] tgunr, sshd is not on the server cdrom [17:36] 'apt-get install sshd' [17:37] can't do that, no network connection , see above [17:37] only from vm->host, host->vm [17:37] get the sshd package to the host [17:37] and scp it from the guest [17:37] or fix your networking :) [17:38] I'm trying to, but working thru this workstation console is painful [17:38] can;t scp if no network [17:39] rather, can;t scp if no sshd on VM [17:39] you said you can get to host [17:39] yes [17:39] log into vm via console [17:39] scp user@host:/tmp/*.deb . [17:39] sudo dpkg -i *.deb [17:40] * tgunr DOH! [17:40] * tgunr is an idiot, thinking inside the box [17:40] fwiw, your network is probably broken at the host level [17:40] you've problably configured host only networking or something. [17:41] yeah, but where, ip forwarding on, but i see no new config on interfaces like i supposed would be there [17:41] its a vmware config option [17:41] i haven't touched that in years [17:41] but there is a config. [17:41] hmm, ok, will look, thanks for tip [17:42] one would think that would kind of be up front as everything needs a network these days [17:43] http://www.vmware.com/support/ws55/doc/ws_net_configurations_hostonly.html [17:43] thats what i think you have ienabled. [18:00] smoser: the longer answer is, i'm going to work on converting all commands to api right now and then re-vsiit [18:46] jamespage: around? [18:46] jamespage: I'm seeing what looks like https://bugs.launchpad.net/ubuntu/+source/ubuntu-meta/+bug/1125408 still on 12.04.3 [18:46] Launchpad bug 1125408 in ubuntu-meta "12.04.2 minimal virtual install is oversized" [Undecided,New] [18:46] jamespage: except now the size is not fine, it's 741M after install, and the modules are eating 158M [18:50] jamespage: ah, wait, I'm remembering now... because there's no -virtual, that's why the module install footprint was still so large [18:56] jamespage: it does kind of raise the question of whether or not it makes sense to even have a minimal virtual install now though... === cp is now known as Guest90522 [20:55] dnssec-tools for 13.10 seems to be broken when signing with NSEC3; donuts reports bad NSEC data [20:55] Any DNSSEC experts around to help point me in the direction to debug/fix it? [21:27] finally got network up on 12.10 server, now trying to install sshd so i can get out of console on workstation but alas, http://polymicro.clarify-it.com/d/q29maj how do you install sshd?e [21:29] tgunr: does apt-get update run without error? [21:29] trying [21:30] YES! Thanks! [21:32] woo :) [21:32] i really need a nap [21:44] 13.10 server is looking quite stable for me so far; aside from this glitch with the DNSSEC tools [21:54] * Patrickdk never cared for dnssec tools [22:25] smoser, hallyn_: seems there is a bug in daily lxc (maybe the ubuntu-cloud template?) that causes the container to end up with hostname 'ubuntu' when using with juju local provider? [22:59] hello all [23:09] sidnei: meaning 'lxc.utsname' in the container config says 'ubuntu' ? [23:18] hallyn_: nope, that's correct: lxc.utsname = sidnei-local-machine-8, but /etc/hostname ended up as ubuntu inside the container [23:19] sidnei: can you put up a pastebin showing the basic commands juju does that i could use to reproduce? [23:19] (setting up a saucy container with the ppa in which to test) [23:21] hallyn_: i suspect it is not doing anything to set up the hostname, let me see if i find the cloud-init config [23:22] dinner, bbl [23:29] hallyn_: it's doing lxc-create -n ... -t ubuntu-cloud -f config-file -- --hostid -r precise --userdata (with https://pastebin.canonical.com/96192/) [23:31] i have a virtual machine, started with 8GB disk (testing, so dont ask me why so small) [23:32] i expanded it to 16GB, trying resize2fs now. doesn't seem to be expanding [23:32] did resize2fs /dev/sda1 8G [23:32] xibalba: hey, I thought you got that working last night? [23:33] i did with LVM [23:33] trying it w/out LVM now [23:33] =D [23:33] ah! :D [23:33] try again without the 8G? or try with 16G? [23:33] trying both, uno momento [23:34] or try spelling out $((16 * 1024)) [23:34] you can login to my box if you want and screen w/me [23:34] oh nm i nated it [23:34] oh, $((16 * 1024 * 1024)) -- sigh I'm confusing my megabytes and gigabytes these days. just this morning I confused gigabytes and terabytes. sheeeesh. [23:35] HW or SW megabyte?! [23:35] =P [23:35] or a network megabyte [23:35] "network megabyte"? :) [23:35] how many bytes till your bit in the ass [23:35] yeah in networking we use 1,000 like hardware/HDD guys [23:36] hang on ill move thit o my EXT network [23:39] sarnold, hop in [23:39] k i see you [23:40] sarnold, fdisk shows 17.2 [23:43] size shows that its option [23:45] oh f me [23:45] i didn't fdisk that shit [23:47] i dont think this is the route i want to go sarnold as it appears i can't do this while the system is live [23:49] Resizing single physical partition to occupy entire drive space is done using fdisk by first deleting your existing patition and then recreating it with EXACTLY SAME STARTING SECTOR NUMBER. [23:49] not gonna work for me [23:49] alrighty [23:49] throwing that idea out [23:49] thanks sarnold [23:50] xibalba: aww. is using lvm a suitable answer? [23:51] yea it is fine, just wanted to try the way i'm more use to which is non-LVM [23:51] lvm is fine though