=== pug_1 is now known as DaveR [01:00] Hmm, I'm trying to mount an iscsi volume via UUID, but neither blkid nor listing /dev/disk/by-uuid seem to list it. [01:18] I've got a server, and I've connected via xdmcp. both are on my lan. I've plugged a usb hd in to the server, an icon appears on my xfce desktop. When I click it, it says failed to mount, not authorized. I've installed gvfs, autofs, all ntfs packages. I've added my username to plugdev, disk, gvfs. I've rebooted. [01:19] I'm using xfce4 [01:22] adam_g: ill take a look tomorrow [01:25] I also have almost all privs under user privileges [01:25] still can't mount!!\ [01:26] option to automount is set in dconf-editor [01:51] Hi Guys [01:53] I have Ubuntu 12.04.1 to be installed on directly on Compellent SAN. While installing it is not able to see SAN. While CentOS is able to see that [01:57] decci: you probably need some driver or package [01:57] blkperl: What I did is trying installing Ubuntu 12.04.1 first on local disk and then trying installing open-iscsi and see if it detects the SAN [01:58] blkperl: Let me send paste the output [01:59] blkperl: http://paste.ubuntu.com/1614757/ [02:00] blkperl: Can you look into this? I am unable to see SAN under /proc/partitions [02:01] you have a local disk installed? [02:01] decci: is it not sda? [02:02] patdk-lap: sda is my local disk [02:02] blkperl: sda is my local disk [02:02] decci: anything in the logs? [02:03] my iscsi knowledge is weak === xnavor__ is now known as xnavor [02:04] what about: iscsiadm -m session -P 3 [02:04] do I recall that the last time someone had problem with a compellent SAN that they might not have authorized the new machine on the SAN controller or something similar? [02:05] sarnold, well, he has two issues kindof [02:05] first, his i350 is setup for hardware iscsi, and centos sees it as hardware hba [02:05] but ubuntu doesn't seem to see it [02:05] second, is him playing with soft-iscsi [02:06] patdk-lap: correct [02:06] patdk-lap: sudo iscsiadm -m session -o show output is there [02:06] that is not what I asked for though [02:07] patdk-lap: let me try this command [02:23] blkperl: My VPN is creating issue. I will come back after 4-5 hours [02:23] blkperl: Will you be here to see those output and logs [02:26] * patdk-lap will be long asleep [02:33] patdk-lap: I am back [02:34] patdk-lap: here is your output: http://paste.ubuntu.com/1614802/ [02:36] patdk-lap: Here is the complete http://paste.ubuntu.com/1614805/ [02:40] patdk-lap: Are you there? [02:40] blkperl: u there [02:41] decci, people go in and out of the channel at times [02:41] I think if I insert driver for openm-iscsi into Ubuntu 12.04.1 image it should work [02:41] escott: ok [02:41] decci, if you want to give a quick recap others can take a look [02:41] I have Ubuntu 12.04.1 to be installed on directly on Compellent SAN. While installing it is not able to see SAN. While CentOS is able to see that [02:41] decci: patdk-lap said one of your problems might have been trying to use the software iscsi support when you've got a hardware card that should do it for you ... assuming you could get drivers for it loaded? [02:41] http://paste.ubuntu.com/1614757/ [02:42] sarnold: I did tried that right...installing Ubuntu locally and then attaching it to SAN rather than directly installing onto the SAN [02:43] sarnold: All I am facing issue is with /proc/partitions while it is detecting the compellent SAN [02:44] sarnold: http://paste.ubuntu.com/1614805/ [02:44] escott: Any views? [02:44] decci, i dont know anything about iscsi [02:45] escott: oh [02:45] escott: Can you help me how to insert driver into ubuntu image? [02:46] decci, if you just need a module in the initramfs you put it in in /etc/modules.d and then update-initramfs [02:49] I think it's more like /etc/initramfs-tools/modules [02:50] genii-around, cool thanks. i never knew about that file [02:57] escott: Can you forward me what steps I wouldneed === shirgall is now known as jrp-afk === smw_ is now known as smw [04:57] newbie here - sorry... how do i figure out the ip address at which to access teh webhost running on a vm? === shang__ is now known as shang === smb` is now known as smb [09:34] jaimef: it would be nice to see all mounts between running the command. Also one should be calling $ udevadm settle , between lvm commands in such a script due to racing with udev. [09:36] hi guys... any help [09:37] i have a mysql datasbed server which on critical Filesystem Size Used Avail Use% Mounted on /dev/mapper/opendatabase 128G 112G 9.6G 93% / <------------------------storage is almost full 93 pecent , any suggestion how do i resolve this issue, im afradi i might break anything with existing databased, dont know what todo [09:48] damnit nodejs, always too old for anything [09:51] goin for chris-lea ppa [09:52] ruben231: is it on lvm? [09:52] can you give it more space or need to compress? [09:54] hey, not sure this is the best place to discuss cloud images, but wanted to mention that http://cloud-images.ubuntu.com/releases/lucid/release is gone or 403s since a couple of days, and I wondered whether this was accidental or intended [10:03] ztane: yes lvm [10:08] lool, I would suspect accidental, but I am not in the position to do anything. [10:12] Ok; I'll file a bug against ubuntu-on-ec2, hopefully that's the right place [10:13] utlemming: ^ [10:13] lool, I wrote an email to Ben/Scott which I think may be in the position to change stings [10:14] things [10:15] Daviey, Since both of them are not really in this tz, I suspected email works better. :) [10:16] smb: You underestimate the power of scrollback highlights :) [10:16] But sure, thanks for sending it [10:16] Daviey, You underestimate the power of "meh, this is long ago" :-P [10:17] Maybe not so bad for them, but I usually only find someone having Samba issues... [10:17] smb: ah thanks; I filed https://bugs.launchpad.net/ubuntu-on-ec2/+bug/1117181 [10:17] Launchpad bug 1117181 in ubuntu-on-ec2 "http://cloud-images.ubuntu.com/releases/lucid/release gone since a couple of days" [Undecided,New] [10:19] * smb subscribed Ben [10:20] the sad part is that I notice some changes that might be bugs because it breaks for the lucid files I was mirroring, but the changes appear to date back to 10.10 timeframe because I hadn't updated my mirror, oh well [10:23] ruben231: if you have space and the filesystem is resizeable, you can use lvextend -L +10G /dev/mapper/opendatabase [10:24] ruben231: then if it is ext4 say, resize2fs /dev/mapper/opendatabase [10:29] ztane: so i just add HDD..? coz during install i just used the default setting --> entire disk with LVM [10:29] would this do..? [11:03] need a expert advice on web server (ubuntu) should i host the virtual hosts in /var/www or home/username ? [11:04] /home/username is pretty much how everyone does it [11:04] ruben231: the question is whether you used all the space already [11:05] ok another issue i have with that is when i assign a user to home directory e.g. /home/ben the user can access oher home directories [11:05] I just did it and ssh into the system [11:05] define "access" [11:05] Yizi: there are patches to kernels where the users cannot read others files even though they would have "permissions" [11:05] being able to read/write files [11:06] shouldn't be able to write files for sure [11:06] right my first job is to move the site from /var/www to /home/ [11:06] you could look into ACLs [11:06] why? [11:06] i need to host 3 sites on one server [11:07] setup ubuntu and did the apache virtual host [11:07] https://help.ubuntu.com/community/FilePermissionsACLs [11:07] installed phpmyadmin with php all worked great, can see the site too but the only issue i have no is giving access directory access to the web design company [11:09] i followed this guide for setting up the virtual hosts https://www.digitalocean.com/community/articles/how-to-set-up-apache-virtual-hosts-on-ubuntu-12-04-lts [11:11] Yizi: you also need php in safe mode / cgi if you really do not want the users to read the other users homedirs using the www-data account [11:11] a really good idea is don't host people you cannot trust [11:11] ok i will put the php in safemode now [11:12] which will then mean that nothing really works :D [11:12] huh? [11:12] almost everything good requires safe mode to be off === matsubara-afk is now known as matsubara [11:13] which means if the sites need to run anything substantial, you cannot use safemode [11:13] im pretty confused with this, i think i'll just give up on ubuntu really did try [11:13] and instead need to configure php in some cgi mode or fastcgi or I do not actually know... so that the php actually runs with the user's account [11:13] if you're going to be doing shared hosting, you should trust all the people on it [11:14] otherwise, use vms [11:14] Ben64: not necessary [11:14] its impossible to be 100% secure [11:14] but 98 % secure [11:14] meaning: if the users are not really malevolent... [11:14] so if they roll 20, critical hit and your server dies [11:14] but really you wouldnt want [11:15] do things like have 1 bug on 1 site mean that everyone can without any problems breach to all the other sites too [11:17] you are right the user is not a sudoer so it cant create anything or modify in other home direcories [11:19] it would just make it a lot easier if i could allow a user to access /var/www/domain.com so when they ssh they only have access to the folder [11:20] uh...i don't think its what you want.... http://www.binarytides.com/setup-a-jail-shell-with-jailkit-on-ubuntu/ [11:22] how about setting up a FTP and assigning a user to access /var/www [11:23] whats your obsession with /var/www [11:23] because i spend 2 days setting up the site in /var/www [11:23] theres nothing you can do there you can't do in /home [11:31] ok guys does this sounds better now [11:32] created a home user [11:32] moved the whole site to their [11:32] renamed the virtual to point to the new location [11:32] so the site is now under /home/ben/domain.com [11:48] another home user can see the files on another home users directory e.g. cat index.php but isn't able to edit it, how can i stop them view the files [11:48] is there a simple permission file? [11:55] Daviey, jamespage, hi, what's best way to debug a hook? i tried with debug-hooks and running the hook there as it's shown in the manual, but i receive that error: subprocess.CalledProcessError: Command '['unit-get', 'private-address']' returned non-zero exit status 2 === Pupeno_w_ is now known as Pupeno_w [12:34] hrmf. seems if I want to snapshot a vm, it's suspended for the time the snapshot takes, which is several minutes [12:39] yeah [13:47] jamespage, in virtualmaas http://bazaar.launchpad.net/~virtual-maasers/charms/precise/virtual-maas/trunk/view/head:/scripts/setup-maas how did you make the nodes search .virtualmaaas.com [13:48] ie, when they boot i dont see how they'd have that in their search... or, maybe they dont, but they only end up with fqdn i guess. [13:54] zul, Daviey. Found a new problem with ceilometer. It has a requires for requests<1.0, but our version is 1.1.0-1, what should we do? do we pin that to the old version? [13:55] hgwehhwefhoiergw'jeoigboiegbkjpgjopgohigopjgekop [13:55] yolanda: nope submit a patch upstream requrests<=1.1 [13:55] smoser: not quite sure I understand the question? [13:55] do you mean the physical server or the kvm instances? [13:55] ok [13:56] zul: this caused some issues, https://review.openstack.org/#/c/21221/ [13:56] instances [13:56] jamespage, i'm sorted, i think. [13:56] maas dhcp doesn't send 'search' or 'domain-name' entries in ists responses. [13:56] so the nodes can only resolve by fqdn. [13:57] zul: it's fixed with, https://review.openstack.org/#/c/21258/ [13:57] but thats ok, because theey only get fqdn (ie, when juju connects them to another node, they get the fqdn of the other node) [13:58] zul / yolanda: Therefore.... we are potentially screwed. [13:58] Daviey: gimme a sec [14:00] fuckers [14:01] oh, so we should stay under 1.0? [14:03] yolanda: thinking [14:14] HELP, Trying to ssh to one of my ubuntu-server installed machine 1 from ubuntu installed machine 2, I am not able to ssh , the message says, "ssh_exchange_identification: Connection closed by remote host". while I can ssh from other computers. [14:18] Daviey/yolanda: we can do something like this for python-keystoneclient [14:24] zul: crazy idea.. patch ceilometer to work with 1.1 ? [14:25] SpamapS: its is a crazy idea === ggg is now known as w0rmie === slank_away is now known as slank [14:36] Daviey/yolanda: there is a mini transition going on there is a transition going on i think, novaclient (https://review.openstack.org/#/c/20119/) and python-cinderclient (https://review.openstack.org/#/c/21222/ although the fix is not correct), and python-keystoneclient got reverted but im working on a proper fix. [14:36] sarthor: running denyhosts or something? [14:36] maybe we can put that on hold then? [14:36] RoyK: no [14:37] yolanda: it will make openstack uninstallable on ubuntu imho [14:37] sarthor: I've seen that if the host is in /etc/hosts.deny - otherwise, check the ssh logs === megha is now known as baba [14:38] yes, problem is on ceilometer but also on the other packages i see... [14:39] from my side i can test ceilometer with newest requests version, [14:41] yolanda: yeah im going through the other packages, novaclient was fixed yesterday in the upload i did to our openstack-ci [14:41] keystoneclient will be fixed soon, and keystone will be fixed soon as well [14:42] RoyK: I checked /etc/hosts.deny, there is entry of two IPs, sshd 192.168.1.82 and sshd 192.168.1.2, that 192.168.1.2 is the computer, from which i am tryin to access. I removed that two lines. but after a minute the lines again there. Why? and how to solve. [14:42] i must double check ceilometer deps again, this wasn't added and i found the problem when ceilometer wasn't starting [14:43] sarthor: check if denyhosts is running [14:44] sarthor: dpkg -l | egrep 'denyhosts|fail2ban' [14:45] RoyK: denyhosts 2.6-10 Utility to help sys admins thwart ssh crackers [14:45] bingo [14:46] cd /var/lib/denyhosts [14:46] grep for your ip there [14:46] that is [14:46] stop denyhosts first [14:46] then remove your ip from the files there [14:46] add it to allowed-hosts [14:46] remove it from /etc/hosts.deny [14:46] start denyhosts again [14:48] in which file I have to check my IP. [14:48] all of them [14:54] sarthor: adding your ip to allowed-hosts will whitelist it permanently [14:55] RoyK: there is hosts-valid file, no allowed hosts. [14:55] make it [14:56] and I deleted from all the files there. and then added to hosts-valid file , restarted denyhosts again.. this time I connected but the password did not work. and now added again to that /etc/hosts.deny [14:56] I said allowed-hosts [14:57] also, stop denyhosts, then clean up the files, then add to allowed-hosts, then start denyhosts [14:57] RoyK: there is no file name allowed-hosts in that directory [14:58] 15:55 < RoyK> make it [15:02] RoyK: Worked great. Thanks friend. [15:02] Thanks to All of you. [15:13] Daviey, poke [15:13] ogra_: poke yourself :) [15:13] haha [15:14] I need to dash for 10. === matsubara is now known as matsubara-lunch === ggg is now known as w0rmie === ggg is now known as w0rmie === ahs3` is now known as ahs3 [15:53] hi, what's syntax in debian/control to specify a min and max version for a package? [16:04] yolanda: https://code.launchpad.net/~zulcss/python-keystoneclient/fix-requests-2/+merge/146887 [16:05] ok [16:06] magical patch [16:07] done === matsubara-lunch is now known as matsubara [16:21] jamespage: ping keystone testsuite runs fine now (needed to fix the python-requests building of python-keystoneclient) [16:26] zul: great! [16:28] HI, please help me. .pl URLs are downloading instead of displaying [16:34] sarthor: Do you have libapache2-mod-perl2 installed and enabled? [16:37] When i use ssh with my DNS Tunnel (dns2tcp) i always get "Connection to localhost closed by remote host." by the ssh server on the DNS server host... [16:39] sarthor: eg: sudo apt-get update && sudo apt-get install libapache2-mod-perl2 ... then make sure that the file perl.load is in /etc/apache2/mods-enabled and do sudo service apache2 restart [16:44] zul https://code.launchpad.net/~yolanda.robla/ceilometer/grizzly/+merge/146899 [16:50] yolanda: have you tested this? [16:50] zul, i've done a build [16:51] and it isntalls? [16:53] you have to manually install pecan, wsme... [16:53] but after all is fixed, it installs [16:55] yolanda: i think the tools/test-requires for ceileometer is not complete because you need pep8 for the tools/hacking.py [16:55] let me check [16:56] it's not listed in tests requires? [16:56] and not on pip requires even [16:57] should i submit a patch? [17:00] yolanda: please [17:00] ok [17:09] Hi guys, does anyone know whether a 512MB Ubuntu server is big enough memory to run a very small web server and a znc? [17:09] As well as a few other things I might want to add in the future [17:10] Note: that meets the minimum requirements: https://help.ubuntu.com/12.10/serverguide/preparing-to-install.html#system-requirements [17:11] Note: I don't see why you'd have a problem for a minor web server [17:11] Ok, thanks :) [17:20] zul: http://people.canonical.com/~agandelman/folsom/keystone-2012.2.1-0ubuntu1.1~cloud0/ [17:20] jamespage: ^ [17:20] adam_g: +1 [17:20] adam_g: i swear i was going to get to that today ;) [17:21] where is the proper way to edit iptables rules on ubuntu-server ? [17:27] sweettea, not to [17:27] use like ufw or something [17:29] I dont want to install ufw [17:30] I was just reading about it [17:30] if I make iptables edits, is there a iptables save function [17:30] that will reload on boot, etc? [17:32] https://help.ubuntu.com/community/IptablesHowTo [17:32] iptables-save [17:32] :) [17:33] zul https://review.openstack.org/21338 [17:37] ahs3: woohoo, thanks [17:38] zul, saw the comments for dhellmann? is that right if that is added in tox.ini? should we also consider that file? [17:38] yolanda: yeah it needs to be test-requires as pep8==1.3.3 [17:39] although it's already listed in other files as dhellmann says? [17:40] hallyn: np [17:41] yolanda: yes [17:41] ok [17:41] i'll modify the proposal and answer [17:42] why is that pinned to that version? [17:45] yolanda: its standard across openstack projects [17:46] zul, can you check the openstack-metering channel? i don't have arguments to explain that [17:47] why is that needed also in test-requires, and not only in tox [17:52] Hello guys, I have a ZNC running for IRC, but when I connect it sets my hostname to HEX.me (.me is the my domains tld), but on the network i'm connecting my VNC to, I have a vhost, my znc idents me but if I ident myself then my vhost activates, but if i just let my ZNC ident me it doesn't activate, why is this? [17:59] zul, so the MP should be ok then? [18:00] yeah go ahead [18:01] Hello guys, some got an idea about this: In our OpenStack installation, booting a Ubuntu Cloud Image fails at creating ssh keys. So we're unable to connect to the VM. [18:02] We check the logs of the VM and saw that the sshd is not able to use any key exchange [18:06] Hi, .pl is downloading not opening apache. HELP me please [18:06] ubuntu server using [18:09] sarthor_: is mod_perl enabled?> [18:09] sarthor_: you probably want to search for mod_perl [18:10] seriously doubt it is going be mod_perl compatable [18:10] probably just want normal cgi support [18:10] * Pici shrugs [18:10] Pici: how to enable that ? I am not an expert [18:10] patdk-wk: mod_cgi then? === ggg is now known as w0rmie [18:11] probably just missing the mapping that .pl should use cgi [18:11] sarthor: I explained this to you earlier [18:12] http://www.thesitewizard.com/archive/addcgitoapache.shtml [18:13] genii-around: I asked that time i was somwhere else. I left from that place. the machine is still on there. but know I am miles away from that place. So I was unable to see your answer genii-around [18:13] but now* [18:14] sarthor: eg: sudo apt-get update && sudo apt-get install libapache2-mod-perl2 ... then make sure that the file perl.load is in /etc/apache2/mods-enabled and do sudo service apache2 restart [18:14] I would be very careful about mod_perl, as it's doesn't run perl scripts the same way you do at the command line [18:15] adam_g: looks like nova-manage config convert went away so should we just remove it? [18:15] patdk-wk: Aah. I don't use perl on apache so wasn't aware of that. cgi runs them the way they were intended however? [18:16] ya [18:16] mod_perl runs it fine, but it doesn't *clean state* between runs, the program is always active, so you can easily have state pre-existing, or leak memory easily when using it [18:17] also, perl running in the apache process address space. a bit gross. :) [18:17] ya, everything should be fastcgi :) [18:18] genii-around: I did as your said. perl.load file is also there. i did restart apache2. checked again .still dowlnoading [18:19] I'm using DNS Tunnel (dns2tcp + ssh dynamic port forwarding) but my ssh tunnel just crashes after a few seconds (Connection to localhost closed by remote host.) [18:20] I just don't understand what is wrong? It works only for few seconds, than the server just shuts down the connection... :( [18:20] i tried also different dns-tunnel server clients, but the result is the same, it's just something with the ssh server... [18:20] This is my sshd_config file on the "fake" DNS server http://pastebin.com/XQvTwqH6 [18:20] dunno, maybe some of you guys have some experience with this [18:21] ya, there is a lot of other setup you need to do to activate mod_perl [18:21] and it needs to be done, per perl script [18:23] patdk-wk: will that be too tough to do? In case i am not expert. I just used to read instructions and do as instructed. every I do not understand them. [18:23] not a good idea [18:23] yea... but I am a learner [18:24] then learn from the link I supplied [18:24] zul: https://code.launchpad.net/~yolanda.robla/ceilometer/grizzly/+merge/146899 [18:24] can you approve it? [18:24] unless the script you are using says it was MADE for mod_perl, and gives you instructions for it, I wouldn't even attempt it [18:24] I think now I remember why I don't like perl under apache [18:25] but I think in that start I am doing this stuff which is tough for me. may be it will make me run from learning.. :) [18:26] also too much green chilli and lemon on beans.. Muuuh....Several battles to win [18:31] hallyn: do we actually have anything to do for https://bugs.launchpad.net/ubuntu/precise/+source/lxc/+bug/1060404 in precise? [18:31] Launchpad bug 1060404 in lxc "update-grub runs and fails in containers" [High,Fix released] [18:32] stgraber: i don't think so. bug is in grub [18:32] hallyn: right, let me get rid of the lxc task then === ggg is now known as w0rmie [19:02] genii-around: Hi again, as you told me I did like for the .pl files downloading issue. as you told me in the previous message. I did like that and restarted the apache2 but no change. Then I restarted the machine, and know there is not downloading issue. but the files are just opening as text files in the browser. Any solution please? [19:02] hi, would someone be able to help me with specifying a preseed file for automated install? [19:03] sarthor: if you set the content-type to text/plain, it should work out nicely [19:04] RoyK: I am using mozilla and chrome. on cromium is like txt file. but on mozilla firefox still that .pl download file. [19:04] sarthor: try, from the commandline, HEAD http://url [19:04] see what content-type is set to [19:07] RoyK: that says Content-Type: text/plain === matsubara is now known as matsubara-afk [19:07] sarthor: url? [19:07] 500 Can't connect to url:80 (Bad hostname) [19:09] I am in that machine where sql-ledger is installed. and I put this command on that machine command prompt. HEAD http://localhost [19:09] hallyn: Does /etc/dnsmasq.d-available/ work for you or do you have a better name suggestion? (I don't particularly like the long name but that's the best I could come up with) [19:09] Content-Type: text/html [19:09] RoyK: Content-Type: text/html [19:13] stgraber: yeah that's fine with me [19:13] obvious enough if you look at all [19:13] and reminiscent of /etc/apache2/sites-available [19:14] hallyn: yeah, I took the name from what apache and freeradius do with their config [19:17] hallyn: ok, I have a fix for lxc in raring. Will do some more testing, if it works as expected, I'll upload lxc in raring, quantal and precise with that fix and any other pending SRU. [19:21] stgraber: cool, i'll copy that into libvirt then, thanks :) === ggg is now known as w0rmie [19:30] hallyn, did we lose the man page to kvm in raring? [19:30] or is it just me [19:31] * RoyK checks [19:34] hi...I just installed gcc4.7 from a ppa, but now this is not the default gcc (which is 4.6)...how can I change that? [19:34] smoser: well, kvm is just a compatibility symlink at this point so you'd technically want to check the manpage of qemu-system-x86_64, but we can probably provide a symlink to make that easier [19:35] smoser: looks so === ggg is now known as w0rmie === ggg is now known as w0rmie === ggg is now known as w0rmie === ggg is now known as w0rmie === ggg is now known as w0rmie === ggg is now known as w0rmie [20:05] smoser: looks like it [20:06] stgraber: actually kvm will become a wrapper, not just a symlink [20:06] have i not pushed that whole merge? FEH [20:06] root@castiana:~/data/vm/lxc# ls -l /usr/bin/kvm [20:06] lrwxrwxrwx 1 root root 18 Jan 29 12:55 /usr/bin/kvm -> qemu-system-x86_64 [20:06] yeah i never finished that merge [20:06] I think I updated my laptop this morning [20:07] stgraber: yeah the merge isn't finished [20:07] it'll have to wait until monday at this point [20:08] hallyn, i opened a bug. [20:09] bug 1117636 [20:09] smoser: Error: Could not parse data returned by Launchpad: The read operation timed out [20:09] smoser: thanks, fixed in my git tree (git://github.com/hallyn/qemu #de.jan28.ubuntu2) === ggg is now known as w0rmie [20:10] meanwhile debian exp has moved to 1.4.0 upstream, but I'm not sure I want ot do that this late [20:10] so i may stick with 1.3 [20:11] kvm will be a wrapper ? [20:11] what will it do [20:12] my guess is, detect kvm support and call qemu with --enable-kvm when found [20:12] smoser: it will do "qemu-system-x86_64 -enable-kvm "$*"" :) [20:12] actually [20:12] exec qemu-system-x86_64 -machine accel=kvm:tcg "$@" [20:12] always qemu-system-x86_64 ? [20:12] yup [20:14] hallyn: hey.. Do you know where we seem to be seeing different cpu types in kvm instances? [20:15] hallyn: http://pb.daviey.com/982y/ , http://pb.daviey.com/c9g4/ , http://pastebin.ubuntu.com/1617691/ [20:15] Daviey: yeah the switch from qemu-kvm to qemu makes some changes, but they shouldn't be much === ggg is now known as w0rmie [20:16] hallyn: the qemu64 standardisation was good for live migrations [20:16] hallyn: the variance seems to be causing woe. [20:16] (this is precise) [20:17] Daviey: wait what, you're seeing unexpected result in precise, or the qemu64 default in precise was good? [20:17] Daviey: you had a patch specifically to default to qemu64 right? I'm sure we can write a new patch for qemu to do the same, if that would suffice. (not sure if it would) === ggg is now known as w0rmie [20:19] hallyn: no, this oddity is on precise [20:19] simple question, how can you boot a new ubuntu server image, and have it execute some scripts you want after boot up.. ? aka like a script outside of the image that it loads and runs, is this possible? [20:19] Daviey: then no, i have no idea. [20:19] And i don't think my patch forced it to qemu64.. simply exposed vmx in the already default qemu64 [20:19] somethign like , lets say i have a script that says build me a haproxy server, then another script that says build me db server, so we can automate our stack to launch, a new image, update all software on the server, install what we need and configure files [20:20] Daviey: but you did just remind me that your patch to expose nested vmx to qemu64 instances was never upstreamed [20:20] right [20:20] Daviey: and we don't (yet) have that patch in raring's qemu. [20:20] hallyn: looking at, https://github.com/aliguori/qemu/blob/master/target-i386/cpuid.c .. it seems it's not using the default qemu64 for some reason [20:21] I wonder if this is something in openstack doing it.. [20:21] Daviey: precise is on qemu-kvm, not qemu, === ggg is now known as w0rmie [20:21] hallyn: a similar patch was proposed upstream and lazily NAK'd [20:22] Daviey: yeah... [20:22] Daviey: i'm just commenting that i need to remember to re-add that to raring [20:22] Daviey: but so you're saying that in lucid through oneiric, it woudl default to qemu64, and in precise and later it does not? [20:22] super :) [20:23] hallyn: No.. I'm saying that using an openstack deployment, spanning multiple nodes.. and seeing different cpu types. :/ [20:23] My hunch is openstack is overriding the cpu type now. [20:23] which raises a big *sigh* if it is indeed the case [20:24] Daviey: ok, so nothing for me to do in precise's kvm then - just wanted to make sure [20:25] it would seem (to a naive goon like me) like exposing the unerlying cpu type by default would make sense :) [20:25] how can I change the default gcc version? [20:25] i realize that doesn't make sense wrt migration, === ggg is now known as w0rmie [20:26] hallyn: *sigh* - https://bugs.launchpad.net/nova/+bug/1003373 [20:26] Launchpad bug 1003373 in nova "support to Nova to allow exposing specific CPU models to the guest, including the host CPU native model." [Wishlist,Fix released] [20:28] Daviey: ok but support for that != making it the default [20:28] anyone have any ideas? [20:29] hmm [20:30] Gateway69: look at cloud-init userdata. But juju will set up a stack to your liking automatically for you, to save yourself a lot of trouble. [20:30] Or you can reinvent it yourself if you wish :) [20:31] haha, juju, yea.. i started looking into that.. ideally internally be able to launch our stack and test it locally, then push any code changes to the cloud.. does juju work with VM? .. [20:32] hallyn: https://review.openstack.org/#/c/8527/6/nova/virt/libvirt/driver.py does suggest it's disabled by default.. [20:32] my head hurts [20:32] hello loverly people :) im stuck trying to set up my firs fileserver at home.... (yes newbie incomming). the question. i got access to my server via ssh, i m remotely on it. but i cannot access internet from the server. as i get no response when i do ping. i got a hiint that it might be /etc/network/intercades but i got my gateway correctly set. so it might be resolv.conf. but on that i get this warning ive never seen before. [20:33] Dynamic resolv.conf(5) file for glibc resolver(3) generated by resolvconf(8) [20:33] # DO NOT EDIT THIS FILE BY HAND -- YOUR CHANGES WILL BE OVERWRITTEN [20:33] Shogoot: see /etc/resolvconf/* files for explanations there :) [20:34] Shogoot: when testing network connectivity, check both ips and hostnames; 4.2.2.1 and 8.8.8.8 both respond to pings, so try those.. [20:34] dang... 8.8.8.8 works [20:35] but i cannot do apt-get update... why is that? [20:35] Shogoot: try ping 91.189.91.15 [20:36] (that's an IP for us.archive.ubuntu.com) [20:36] working [20:36]