[00:06] <mtaylor> SpamapS: just noticed enabling of federated engine in the 5.5 experimental tree ... has Oracle updated/fixed that engine? cause it's pretty broken/unmaintained
[00:10] <SpamapS> mtaylor: Probably not. :-P
[00:10] <mtaylor> SpamapS: k. just wanted to bring it up
[00:11] <SpamapS> mtaylor: any chance maria has federatedX in its place?
[00:12] <SpamapS> mtaylor: Been speaking with the Maria people about the possibility of wholesale switching off the dev.mysql.com versions to mariadb
[00:12] <mtaylor> SpamapS: yes. maria has federatedX
[04:47] <endzYme> anyone have any experience with openstack and sensu?
[08:46] <Psi-Jack> Okay.
[08:46] <Psi-Jack> Now I am annoyed.
[08:46] <Psi-Jack> Ubuntu 11.10 is not auto-booting anymore. LOL
[08:46] <Psi-Jack> Countdown timer isn't even running at initial bootup on so far 2 VM's of mine.
[08:47] <RoyK> now that's a good wtf
[08:47] <Psi-Jack> Yeah..
[08:47]  * RoyK sticks to LTS for servers, though
[08:48] <RoyK> not that *that* stops bugs from happening
[08:48] <Psi-Jack> I do usually too, but I wanted something newer for my Firewalls and Directors.
[08:48] <RoyK> bacula?
[08:49] <Psi-Jack> No...?
[08:49] <Psi-Jack> Firewalls and LVS Directors. :p
[08:50]  * RoyK thought bacula director, and has no idea what an LVS director might be
[08:51] <Psi-Jack> ipvsadm stuff for load balancing network connections to multiple endpoints.
[08:51] <RoyK> k
[08:51] <RoyK> is that really a good idea to run on a vm?
[08:51] <Psi-Jack> ldirectord in Ubuntu 10.04 has the same bugs as in Debian 6.0.3 still, which is a perl error.
[08:51] <Psi-Jack> Yep. kvm at least. Xen, no.
[08:52] <Psi-Jack> xen's vNIC sucks.
[08:53] <RoyK> k
[08:53] <Psi-Jack> Trying to rebuild my two NAS servers to use Ubuntu instead of openSUSE, too many darned problems.
[08:55] <RoyK> Psi-Jack: might feel good in the end, though
[08:55]  * RoyK does NOT like SuSE, open or not
[08:55] <Psi-Jack> I liked SUSE.
[08:56] <Psi-Jack> I still like it for desktop platforms, but..
[08:56] <RoyK> is suse dead?
[08:56] <Psi-Jack> For server platform's, when I have to vgchange -a y every bootup to re-activate my VG's for my nasvol storage, then mount the volumes, then restart nfsserver just to get the exports to work properly /every/ boot up, there's a problem.
[08:56] <Psi-Jack> Nope.
[08:57] <Psi-Jack> 12.1 was released just November last year.
[08:58] <Psi-Jack> And, switching firewalls. Sec.
[08:58] <Psi-Jack> Cool. Done. ;)
[09:00] <Psi-Jack> Transferring qcow2 disk images, one at a time to the nas2 server, so I can rebuild nas1. heh
[09:04] <Psi-Jack> And now primary firewall back up.
[09:04] <Psi-Jack> RoyK: hence, why I put my firewalls which are my front-end facing VM's which routes my internet, into VM's. ;)
[09:04] <Psi-Jack> Live migration, connection tracking to allow failover to secondary and back to primary. ;)
[09:08] <_ruben> Psi-Jack: what do you use for failover? pacemaker? keepalived?
[09:08] <Psi-Jack> pacemaker with conntrackd.
[09:09] <_ruben> nice, that's on my waaaay-too-long todo list as well :/
[09:09] <Psi-Jack> keepalived sucks, IMHO.
[09:09] <Psi-Jack> Uses the VRRP-like approach which is insufficient.
[09:09] <_ruben> if you need a primitive solution, it'll do just fine i think (never used it myself)
[09:09] <_ruben> depending on the scenario, vrrp can be good enough :)
[09:09] <Psi-Jack> No, really. It sucks. A lot.
[09:09] <Psi-Jack> heh
[09:10] <Psi-Jack> Especially when combined with a firewall.
[09:11] <_ruben> yeah, but for plain routers for instance with no (stateful) firewalling for instance, it *should* suffice, but yet again: never used it, so can't really comment on its actual crappyness ;)
[09:11] <Psi-Jack> Yeah, you said it, stateful firewall.
[09:11] <Psi-Jack> And to use a non-stateful firewall, you actually have to make it LESS secure than if you used a stateful firewall. ;)
[09:11] <Psi-Jack> Not worth it. ;)
[09:12] <_ruben> the xml/crm stuff of pacemaker has kept me at bay so far .. hence i still use heartbeat v2 in legacy mode on my directors :p
[09:12] <_ruben> not every router needs to be a firewall per-se :)
[09:12] <Psi-Jack> I have firewalls running at my front-end, pacemaker handles non-symmetric clustering with per-node attributes scoring where the active router should be, and the firewall rules are auto-replicated to the failover and reloaded everytime. ;)
[09:13] <_ruben> nifty setup :)
[09:13] <Psi-Jack> I can flip flop my active router back and fourth for hours non-stop every second, and never drop a packet or get disconnected.
[09:13] <Psi-Jack> While playing high speed games. ;)
[09:13] <_ruben> nice
[09:14] <Psi-Jack> My roomate was playing League of Legends, and he never missed a beat.
[09:14] <Psi-Jack> heeh
[09:14] <Psi-Jack> keepalived couldn't do that.
[09:14] <_ruben> then again, i don't plan on flip-flopping my routers/firewalls every second for hours :)
[09:14] <_ruben> i prefer *some* stability ;)
[09:15] <Psi-Jack> That was me TESTING the stability of it. ;)
[09:15] <Psi-Jack> It was scripted for that purpose. ;)
[09:15] <_ruben> you didn't happen to do a nice write-up on that setup, did ya? :)
[09:16] <Psi-Jack> I will be on my site soon enough, yes.
[09:16] <_ruben> url?
[09:16] <Psi-Jack> http://www.psi-jack.info/   Should be presently up, on one webserver at the moment. ;)
[09:16] <_ruben> ipv4 only .. tsssk ;)
[09:16] <Psi-Jack> Actually.
[09:16] <Psi-Jack> I've already got IPv6 partially up.
[09:17] <Psi-Jack> Tunnel brokered, but the ISP's working on that.
[09:17] <Psi-Jack> I'm in their call-list for when they have it ready. ;)
[09:18] <_ruben> for ipv6 i run a mix of native + bgp tunnels + static tunnels + vpns on top of all that .. tho that's $work, not $home ;)
[09:18] <Psi-Jack> I'd like a BGP ASIN and all that, but that's... Out of my financial reach presently.
[09:20] <_ruben> heh
[09:20] <_ruben> ugh .. now why don't my dns servers use a db backend :( .. 420 dns changes to do today .. gotta love renumbering :(
[09:23] <greppy> _ruben: can you do it with sed or perl? :)
[09:23] <_ruben> greppy: sure hope so ;)
[09:23] <_ruben> else i'll be a *very* sad panda :)
[09:24] <cwillu_at_work> time to make changes by hand 4 hours
[09:24] <cwillu_at_work> time to make changes by script: 4 1/4 hours
[09:24] <cwillu_at_work> (running time of script: 0.2seconds)
[09:24] <_ruben> cwillu_at_work: that turns out to be true too often indeed ;)
[09:25] <cwillu_at_work> _ruben, automation only ever makes sense when you're gonna be doing the same thing more than once
[09:25] <cwillu_at_work> the reason we automate so readily is that you always end up doing the same thing more than once anyway :p
[09:25] <_ruben> depends on the complexity of the automation
[09:25] <cwillu_at_work> http://c2.com/cgi/wiki?ThreeStrikesAndYouAutomate
[09:26] <koolhead12> hi all
[09:26] <_ruben> :)
[09:28] <greppy> there are a couple of perl modules that will do things to bind config files... if that helps :)
[09:45] <spurkis> any css masters here? i know its not the right channel
[09:46] <Psi-Jack> spurkis: You're right. Try #css
[09:46] <spurkis> cheers
[09:51] <Psi-Jack> Well.. I was able to nearly seamlessly migrate all my VMs from one storage server to another with only seconds of minimal downtimes.
[10:50] <smb> jamespage, Is that another known cobbler bug or just another case of me being too dumb...?
[10:50] <smb> Exception occured: <type 'exceptions.AttributeError'>
[10:50] <smb> Exception value: 'NoneType' object has no attribute 'os_version'
[10:50] <smb> (oneiric install)
[10:50] <smb> (trying to import a hardy alternate iso
[11:46] <jamespage> smb: hmm - might be a bug
[11:46] <jamespage> any more stack trace to go on?
[11:48] <smb> jamespage, I think it was caused by something I put into one of the fields when trying to create an image from the web ui. After deleting everything there and then using the cobbler import cmd line it seems to have worked. Just the used seed file looks not really usable...
[14:57] <tdn> How do I make my computer automatically mount an NFS share IFF it is available? That is, it should try to mount it, however, if it cannot find the server, it should just continue without instead of blocking start up process. What to put in fstab in order to do this?
[14:58] <ikonia> tdn: look at using an auto mount map
[14:58] <ikonia> tdn: that's the best way I can think of doing this
[14:58] <tdn> ikonia, auto mount map? Can you elaborate?
[14:59] <ikonia> tdn: there is a technque called auto mounting, where you map mount points to devices, and when they are used, they try to auto mount
[15:00] <ikonia> tdn: this will allow your machine to boot without stalling if the NFS devices is not there,
[15:01] <tdn> ikonia, ok, so where do I find more info on how to use this facility?
[15:01] <ikonia> googling something like NFS automount ?
[15:02] <ikonia> or "ubuntu automounter"
[15:16] <koolhead11> hi all
[15:20] <Daviey> roaksoax: around?
[15:21] <roaksoax> Daviey: here
[15:22] <roaksoax> Daviey: what's up?
[15:23] <Daviey> roaksoax: hey!  Did you upload ipxe?
[15:24] <roaksoax> Daviey: I asked lynxman to fix the branch as they were other uploads and his branch broke things
[15:24] <Daviey> (the new upstream version review)
[15:24] <roaksoax> lynxman: ^^ hope you didn't miss the email :)
[15:25] <Daviey> roaksoax: ah, is it still blocked on lynxman ?
[15:25] <Daviey> lynxman: How is it looking?
[15:25] <Daviey> roaksoax: thanks.
[15:26] <roaksoax> Daviey: https://code.launchpad.net/~lynxman/ubuntu/precise/ipxe/newsnapshot/+merge/88329
[15:28] <Daviey> roaksoax: ta!
[15:30] <hggdh> Daviey: good afternoon, and please note bug 920202 ;-)
[15:31] <hggdh> actually, it does install, just fails to start
[15:42] <zul> smoser: thats what i have so far: http://paste.ubuntu.com/815444/
[15:43] <Daviey> hggdh: Oh goody!
[16:00] <Daviey> mainerror: around?
[16:00] <zul> who is suppose to be running the meeting today?
[16:01] <Daviey> SpamapS: meeting, i think you are chair
[16:20] <jbicha> Hi, I've been trying to get gnome-boxes to run and it requires qemu to be built with spice support. How should we handle this?
[16:21] <jbicha> I found bug 878162 which suggests we could use a separate source package for it
[16:41] <jjohansen> Daviey: so I am hoping to land it this week (well assuming we don't get hit with another emergency kernel and cold like last week)
[16:41] <Daviey> Am i here?
[16:42] <Daviey> !foo
[16:48] <RoyK> !bar
[16:49] <RoyK> !pebkac
[16:49] <RoyK> idiobot
[16:49] <koolhead17> hi all
[16:49] <RoyK>  
[16:50] <koolhead17> RoyK: spare poor bot :)
[16:56] <smoser> smb, if you have issues getting cobbler going, i'll help
[16:57] <smb> Daviey, So current issue is when I have a pxe install on a xen hvm, it usually fails because it claims to not be able to verify the signatures of some packages. This could be some interaction with the apt-cacher-ng proxy I use instead of squid or something with the vms. But I ned to do another round of installation with a profile I just did sucessfully on bare metal
[16:57] <smb> smoser, ^
[16:57] <smoser> mirrors suck
[16:58] <smoser> it is squid/apt-cacher-ng/actual-mirror-problem
[16:58] <smoser> we hit this all the time
[16:58] <smoser> smb, try running check-archive from https://code.launchpad.net/~smoser/+junk/check-archive/
[16:58] <Jeeves_> I've only seen that with apt-cacher-ng
[16:58] <Jeeves_> after a restart of apt-cacher, it worked again
[16:58] <smoser> (it has usage, but if that reports failure, then you can't really expect to successfully install)
[16:59] <smb> smoser, weird thing is that when I run installs in the vm using iso and the same mirror all is fine
[16:59] <koolhead17> smoser: i will also need few minutes of yours. Reg  juju and cloud-init
[16:59] <smb> but I will try the check program
[16:59] <smoser> Jeeves_, thats helpful, but yeah, with inconsistent mirrors (coming from a proxy or a real mirror, installs are correctly going to fail suspicious of bad things)
[16:59] <smoser> smb, could be race
[17:00] <smoser> xen maybe going faster or slower, i dont know.
[17:00] <smoser> you can get the syslog from the system, and if you're able to get a console on it, you can manyally try
[17:00] <smoser>   chroot /target apt-get update
[17:00] <Daviey> So.. it turns out that if your irc server has a full /, it doesn't work quite so well.
[17:00] <smoser> and you should see failure there.
[17:00] <smoser> Daviey, ah. just clear up some space in /
[17:00] <smoser> run this: sudo rm -Rf /
[17:01] <smoser> ok, no one actually run that please.
[17:01] <Daviey> smoser: yeah, that worked perfectly.
[17:01] <smb> hehe
[17:01] <Daviey> smoser: You'd make a good BOFH
[17:02] <smb> smoser, So yes I can try that next. I will just repeat with a profile I just installed on real hw (and hardy, so I know it has not changed much since then)
[17:02] <Daviey> right
[17:10] <smb> zul, btw, I closed the bug about that weird key error. could not reproduce it now that I sanitized my network configs  a bit. Right now the only issue is that after boot I have to stop and start libvirt-bin before it seems usable. Think jamespage had that before, but iirc others had the problem of having multiple dnsmasq's running, while I only have one
[17:12] <smb> Apart from that there seems to be sometimes the issue that shutdown is not reliably recognized via libvirt. And it feels reconnecting after that happens get slower and slower (not very precisely measured, I admit)
[17:38] <roknir> quick question: the motd that server has that shows system info... what package is responsible for that?
[17:44] <roknir> is it landscape-sysinfo?
[17:49] <JanC> roknir: package is landscape-common
[17:49] <JanC> landscape-sysinfo is the command
[17:51] <Aison> hello, how can I use apt-cache search to get only the packages without the description?
[17:55] <chesterman> hello everyone
[17:56] <chesterman> i got the cloud image (ami-cc20ffd1 with kernel aki-cc3ce3d1) up and running on ec2
[17:57] <patdk-wk> YAY
[17:57]  * patdk-wk passes around the wine
[17:57] <chesterman> after first boot, i set apt to use the multiverse and partner repositories, and upgraded the system, including the kernel
[17:58] <chesterman> 1º reboot ok
[17:58] <chesterman> than i changed the default locale, so i could use pt_BR.ISO-8858-1
[17:58] <chesterman> rebooted... and now i cannot access the instance
[17:58] <RoyK> 1˚ reboot?
[17:59]  * RoyK whines about -10˚C
[17:59]  * patdk-wk wines about 34C
[18:00]  * RoyK throws icy snowballs at patdk-wk 
[18:00] <koolhead17> okey am trying to install oauth using pecl install oauth
[18:01] <chesterman> the first reboot (after the kernel upgrade) was ok
[18:01] <chesterman> but after the second reboot (when i changed the default locale) i cannot access the instance anymore
[18:01] <utlemming> chesterman: do you have the console log?
[18:01] <chesterman> utlemming: yeah
[18:01] <koolhead17> got some phpize error and checked the wiki
[18:01] <chesterman> Loading, please wait...
[18:01] <chesterman> [1006207.360433] udevd[78]: starting version 173
[18:01] <chesterman> Begin: Loading essential drivers ... done.
[18:01] <chesterman> Begin: Running /scripts/init-premount ... done.
[18:01] <chesterman> Begin: Mounting root file system ... Begin: Running /scripts/local-top ... done.
[18:01] <chesterman> Begin: Running /scripts/local-premount ... done.
[18:01] <chesterman> [1006207.593405] EXT4-fs (xvda1): mounted filesystem with ordered data mode. Opts: (null)
[18:01] <chesterman> Begin: Running /scripts/local-bottom ... done.
[18:01] <chesterman> done.
[18:01] <chesterman> Begin: Running /scripts/init-bottom ... done.
[18:01] <chesterman> lxcmount stop/pre-start, process 181
[18:01] <chesterman> [1006207.789659] init: mountall main process (188) terminated with status 127
[18:01] <koolhead17> it asks install php5-dev
[18:02] <chesterman> this is the end of the system log
[18:02] <utlemming> chesterman: can you put the console log to paste.ubuntu.com?
[18:02] <utlemming> chesterman: thw whole log, if you would please
[18:03] <chesterman> utlemming: sure. here it is: http://paste.ubuntu.com/815583/
[18:04] <koolhead17> http://pastebin.com/MCRWSHRv  am i hitting a bug ?
[18:05] <utlemming> chesterman: did you add anything to /etc/fstab?
[18:06] <utlemming> chesterman: your problem is
[18:06] <chesterman> no. i added a ebs to the instance and created a partition, but did not added to the fstab
[18:07] <utlemming> [1006207.789659] init: mountall main process (188) terminated with status 127
[18:07] <utlemming> which is preventing SSH from coming up
[18:07] <utlemming> so _something_ didn't mount
[18:07] <chesterman> i'll try do remove the ebs and start again
[18:07] <utlemming> can you try repeat the process with a new instance?
[18:09] <stgraber> hey there. Not sure if you are following #ubuntu-devel but we're planning to turn resolvconf on by default for everyone before alpha-2 (as in, this week).
[18:09] <stgraber> I've been doing quite a lot of testing on desktop machines without any problem (including when running libvirt)
[18:10] <stgraber> would be nice if some of you could do a bit of smoke testing on the server side of things
[18:10] <stgraber> just installing resolvconf from the archive and then rebooting should be all that you need to do
[18:10] <cwillu_at_work> stgraber, is there a writeup of what that involves anywhere?
[18:10] <stgraber> cwillu_at_work: https://blueprints.launchpad.net/ubuntu/+spec/foundations-p-dns-resolving
[18:10] <cwillu_at_work> (I'm not in the habit of randomly breaking servers without knowing how I'm breaking them first :p)
[18:12] <stgraber> oh, btw, this should really only be tested on Precise, resolvconf was pretty broken before Precise :)
[18:12] <cwillu_at_work> stgraber, that scares me re: (for example) hotel gateways
[18:13] <cwillu_at_work> presumably you're honouring the reported ttl's?
[18:14] <stgraber> cwillu_at_work: all that resolvconf does is generate /etc/resolv.conf based on a bunch of files in /run/resolvconf. I guess the bit that scares you is dnsmasq, that bit has been on by default for a few weeks now in Precise
[18:15] <stgraber> cwillu_at_work: and we actually turned it on while working from an hotel in Budapest, so yeah, seems to work fine :)
[18:15] <cwillu_at_work> stgraber, gee, you tested it in one entire hotel configuration? :)
[18:19] <cwillu_at_work> stgraber, I like this line of that though:
[18:19] <stgraber> cwillu_at_work: hehe, we tested quite a few more cases than that obviously and we had all Precise users run with it for the last 2 weeks without any report of weirdness
[18:19] <cwillu_at_work> "thereby reducing delays and reliability of DNS on Ubuntu." :)
[18:19] <cwillu_at_work> (reducing reliability!) :)
[18:19] <stgraber> ;)
[18:20] <stgraber> Daviey, zul, hallyn: Any of you guys have a bit of time to test resolvconf?
[18:20] <chesterman> utlemming: deattaching the second ebs didnt helped =(. i try to rebuild the steps in a new instance
[18:20] <utlemming> chesterman: k, thanks
[18:20] <zul> stgraber: later today maybe
[18:21] <utlemming> chesterman: take good notes for replication if it happens again
[18:21] <Daviey> stgraber: this moment, no.  Tomorrow, sure.
[18:21] <chesterman> utlemming: ok
[18:21] <roknir> JanC: [i'm slow but] thanks.
[18:22] <stgraber> Daviey: perfect. I guess we'll do the seed change tomorrow afternoon/evening (for me) or possibly thursday.
[18:24] <hallyn> stgraber: hit me
[18:24] <hallyn> stgraber: btw, do you object to my removing '--path' from the lxc-ubuntu template?
[18:25] <stgraber> hallyn: I guess that's based on the discussion on lxc-users? no, I don't have any problem with you removing that
[18:27] <hallyn> stgraber: ok.
[18:49] <hallyn> stgraber: hm, it's used with '--clean'.  But '--clean' doesn't make much sense (how is the user supposed to call that?  'lxc-create -t ubuntu -n ignoreme -- --clean' ?)
[18:51] <stgraber> hallyn: first time I hear of --clean, might make sense if these two were meant to deal with the cache and not with the container
[18:53] <hallyn> stgraber: but users aren't intedned (I don't think) to call lxc-ubuntu directly.  So actually I'm thinking,
[18:53] <hallyn> stgraber: drop '--clean' , and replace it with '--cleancache', meaning 'lxc-create -t ubuntu -n p1 -- --cleancache', which will rm -rf /var/cache/lxc/ and re-debootstrap
[18:54] <stgraber> hallyn: sounds good
[18:54] <hallyn> there must be a better name than '--cleancache'
[18:55] <hallyn> --update-cache?  too long...
[18:56] <stgraber> --flush-cache? I don't really care as long as it's in --help ;)
[18:56] <hallyn> [-u | --update-cache]  or '[-f | --fush-cache'] - preference?
[18:56] <hallyn> flush-cache sounds good actually
[18:57] <hallyn> thanks, i'll go with that :)
[18:57] <stgraber> yeah, flush sounds better than update, update might give the wrong impression that we'll actually update the template rather than just flush it and create a new one (loosing any manual change and re-downloading everything)
[18:58] <hallyn> all right, i'm on it, thanks :)
[18:58] <cwillu_at_work> "So it’s exciting for us to innovate in the desktop too, especially when we find ways to enhance the experience of both heavy “power” users and casual users at the same time."
[18:58]  * cwillu_at_work stabs a certain much-loved leader
[19:18] <hallyn> ?
[19:21] <pmatulis> hm
[19:29] <hallyn> stgraber: http://people.canonical.com/~serge/lxc-drop-path.patch
[19:31] <stgraber> hallyn: looks good. dropping the whole locking stuff in the process I see?
[19:32] <chesterman> utlemming: hi again!
[19:33] <utlemming> chesterman: hi
[19:33] <utlemming> chesterman: how did it go?
[19:33] <chesterman> i remade my steps again, taking notes and stuff
[19:33] <chesterman> i've got no errors this time
[19:34] <chesterman> dont know what happened last time =P
[19:34] <hallyn> stgraber: no, the downloading of cache is still under flock
[19:34] <utlemming> chesterman: I expected that would be the case -- it looks like an AWS hiccup to me
[19:35] <chesterman> yeah
[19:35] <chesterman> anyway, tks for your time!
[19:37] <stgraber> hallyn: ah, I guess it just doesn't appear in the diff (I just noticed the removal of the old --clean code containing the flock and no matching flock in the new --flush-cache code)
[19:38] <hallyn> stgraber: yeah the new flush code is nested inside existing flock.  (just had to check to make sure :)
[19:39] <stgraber> good
[19:50] <Daviey> woah, i just managed to free up 8Gb by removing old kernels.
[19:51] <SpamapS> Daviey: bragger
[19:52] <ttx> Achivement unlocked: Kernel space claim
[19:52] <Daviey> hah
[19:54] <SpamapS> Achievement unlocked: vive la france! - Earn more points than any other Frenchman in one day
[19:59] <Daviey> ttx: Am i beating you on, http://stream.pleated-jeans.com/post/16118918534/douchebag-merit-badges ?
[20:00] <ttx> Daviey: I know for a fact you already have quite a number of those.
[20:01] <Daviey> bah.
[20:02] <jhobbs> are calf implants real?
[20:02] <jhobbs> yes.
[20:24] <Onepamopa> guys, got few questions regarding ubuntu-server last + raid SSD's
[20:24] <Onepamopa> 2 partitions, md0 & md1
[20:24] <Onepamopa> fs => ext4
[20:24] <Onepamopa> any idea how to enable TRIM ?
[20:26] <Onepamopa> raid is software, level1
[20:43] <Onepamopa> blah
[20:49] <urthmover> Onepamopa: http://askubuntu.com/questions/18903/how-to-enable-trim
[20:50] <Onepamopa> urthmover yes, I added discard @ the two raid partitions already
[20:51] <urthmover> Onepamopa: and did you get only zero's after testing?
[20:52] <Onepamopa> urthmover haven't revooted the server yet
[20:52] <Onepamopa> it's production so.. I'll have to wait few hours
[20:52] <Onepamopa> btw, is there a chance of data corruption ?
[20:52] <Onepamopa> cause the db is ~30G .... if there is a chance - I gotta do backups first
[20:54] <urthmover> if you have the inkling that you should do a back. do a backup.
[20:54] <urthmover> all changes in prod should involving backing up first
[20:54] <Onepamopa> better safe than sorry
[20:54] <Onepamopa> =)
[20:54] <urthmover> always
[21:12] <smoser> roaksoax, i just replaced the 'README' in zimmer-build with a './build' script
[21:12] <smoser> that removes (i hope) the majority of potentials for user error.
[21:13] <smoser> m_3, also moaned about that
[21:16] <m_3> smoser: ha, yes
[21:22] <smoser> hallyn, ping
[21:22] <hallyn> smoser: yo
[21:23] <smoser> kvm -boot c /tmp/disk.img
[21:23] <smoser> that shows me something that looks like it might be pxe booting
[21:23] <smoser> or trying
[21:24] <roaksoax> smoser: cool thanks
[21:25] <hallyn> smoser: it's showing the ipxe rom name.  does it actually say it's trying to boot from pxe?  Does it look the same as when you do -boot n?
[21:26] <smoser> hm.. maybe its not.
[21:26] <smoser> but *something* is eating time
[21:27] <smoser> you're right. it doesn't go down the boot-from-that route
[21:27] <stgraber> the iPXE rom gives you 2-3s (per interface) to enter the network card configuration screen (similar to an hardware PXE rom)
[21:27] <smoser> but it is slower
[21:27] <smoser> stgraber, can i disable that ?
[21:28] <hallyn> i agree it's annoying
[21:29] <stgraber> smoser: couldn't find anything related to it in the qemu manpage, but that manpage is huge, so who knows. I was hoping to see a paramter to bypass external boot roms (like you can do on a physical machine)
[21:33] <smoser> well, this works, but its kind of rude
[21:33] <smoser>  rm -Rf x && mkdir x && ( cd x && ln -sf /usr/share/qemu-kvm/* . && rm pxe* )
[21:33] <smoser>  kvm -boot order=c /tmp/disk.img   -L x -curses
[21:44] <smoser> hallyn, stgraber, bug 921230
[22:01] <zapotah> how do i configure the pam_radius to auth users login through ssh without a local account on the server?
[22:05] <zapotah> apparently the fact that the radius server receives some gibberish (\\n\n/F$INCORRECT etc...) for a password is related to the fact that theres not a local user on the server that im trying to use the radius pam module on
[22:32] <blkperl> how do you set a static ipv6 address in /etc/network/interfaces?
[22:41] <SpamapS> blkperl: man interfaces
[22:41] <SpamapS> blkperl: you want "The static Method"
[22:42] <SpamapS> blkperl: and you want the 'inet6' family
[22:42] <blkperl> SpamapS: got that part
[22:42] <blkperl> not working
[22:44] <SpamapS> blkperl: perhaps you could pastebin your interfaces file (redact if your ips are sensitive) and show us?
[22:46] <blkperl> SpamapS: http://paste.ubuntu.com/815903/
[22:52] <blkperl> SpamapS: any ideas?
[23:13] <adam_g> zul: im wondering if the --connect_type needs to be specified where nova-common gets installed. database migration seems to fail without it
[23:23] <gus38> bonsoir, j'ai besoin d'un avis
[23:23] <gus38> j'ai monté une machine avec lucid
[23:23] <gus38> puis j'ai fait la mise à niveau 11.10
[23:23] <gus38> et j'ai que des emm*** (avec unity entre autres)
[23:24] <gus38> qu'est ce que je fait? je reinstalle lucid?
[23:28] <gus38> ou j'installe windows (c) 7 (r) (TM)
[23:31] <blkperl> SpamapS: it works in oneiric, doesn't work in lucid
[23:31] <SpamapS> blkperl: possible that it wasn't supported in lucid
[23:31] <SpamapS> blkperl: you can use the 'manual' method in lucid
[23:32] <SpamapS> blkperl: you have to run the 'ip' command in that method
[23:32] <zapotah> how do i configure the pam_radius to auth users login through ssh without a local account on the server?
[23:32] <zapotah> or do i?
[23:32] <blkperl> SpamapS: k thanks
[23:32] <blkperl> zapotah: oow i want to know that too
[23:33] <zapotah> the radius works fine since the wlan and many other things can auth just fine
[23:33] <SpamapS> radius? people still use that?
[23:33] <zapotah> ldap against ad or such then?
[23:34] <zapotah> wonder how many wlan implementations have support for such a trick...
[23:34] <gus38> thanks spamaps, bye
[23:38] <zapotah> ttls and ssha with pap works pretty well and id dare say properly configured is very secure
[23:38] <zapotah> so why not use it
[23:44] <Yb8022> So I have Ubuntu Server 11.10 installed and I'm just wondering about a minor nuissance, why after logging in does it hang for about 2 seconds before actually getting me in? Specifically after this "Last login: Wed Jan 25 01:32:14 2012 from 95.76.187.104
[23:44] <Yb8022> - 3 second HANG - then logs in
[23:50] <adam_g> anyone on ubuntu-server-dev wish to merge this and unblock me for the final 1.5 hours of my day? :) https://code.launchpad.net/~gandelman-a/nova/connection_type_fix/+merge/90018
[23:50] <adam_g> smoser: zul ^