=== dendrobates is now known as dendro-afk [04:46] morning all [04:51] evening [05:30] kim0, hello [05:30] hi flaccid how have you been? [05:33] busy as [05:35] kim0, ping me once your here, got some good news :) [05:35] flaccid, hmm [05:38] whats the goss [05:39] *goss? [05:39] aka gossip [05:39] whatever that news is [05:40] flaccid, :P [05:55] meh trolls. i'm going to dinner. [06:01] flaccid, lol [08:02] Morning folks [08:02] koolhead17: hey o/ [08:22] hi kim0 :) === daker_ is now known as daker [12:00] kim0, [12:00] koolhead17: hey [12:01] kim0, how was it :D [12:01] koolhead17: very good :) [12:03] k [12:43] * rwat starts to tear hair out. [12:44] any idea where to look for clues as to why my instances go straight from pending to terminated? [12:44] NC is able to run kvm [12:45] ok they don't quite go straight. They sit at pending for a few minutes [12:45] then terminate - no sign of any running [12:45] like something times out === dendro-afk is now known as dendrobates [13:33] rwat: check out nc.log in /var/log/eucalyptus === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [14:39] kim0: I just went to try to launch an instance so I could tail nc.log, and I'm now getting: FinishedVerify: Not enough resources (0 < 1: vm instances.Not enough resources (0 < 1: vm instances. [14:39] which is a bit different [14:40] try maybe we could then propose a sprint in Europe. Where we [14:40] meet for a few days to discuss plans for 11.10. Canonical would be open [14:40] to funding travel and accommodation for 2-3 x2go folks for this. [14:40] oops nvm that [14:41] rwat: try this .. euca-describe-availability-zones verbose [14:42] it lists the IP address of my cloud controller and some vm types all showing 0000 out of various amounts [14:43] like 0000/0002 0000/0002 etc [14:43] so if 0000 are free then it can't find anywhere to run the image? [14:43] yes [14:44] rwat: basically I think your NC is not registered [14:44] the cloud thinks it's out of resources [14:44] ok [14:44] try deregistering .. and reregistering [14:44] ok [14:44] the nc node [14:44] and I am not quite sure how to do that :) [14:45] woudl that be euca_conf --deregister-nodes ? [14:45] deregistering gives me: SUCCESS: removed node '192.168.101.70' from '//etc/eucalyptus/eucalyptus.local.conf' [14:46] registering gives : Trying rsync to sync keys with "192.168.101.70"...done. [14:46] euca-describe-availability-zones verbose still gives 0000/anything [14:47] ok if I stop the node controller that goes to 0000/0000 rather than 0000/0002 [14:48] I restart the NC and it goes back to 0000/0002 [14:48] so they are talking [14:49] rwat: do you restart eucalyptus with CLEAN=1 parameter [14:49] not so far [14:50] that's sometimes needed to flush the cache [14:50] please try that [14:50] how do I do that? [14:50] on all services [14:50] /etc/init.d/eucalyptus restart CLEAN=1 [14:50] ok [14:50] and for other services too [14:50] google that [14:50] kim0, euca uses upstart [14:50] no ? [14:50] restart eucalyptus CLEAN=1 [14:50] ew no idea :) [14:51] rwat: ^^ [14:51] doesn't upstart jobs support old style too [14:51] dont* [14:51] kim0: smoser I think you can use both [14:51] * kim0 nods [14:52] * kim0 can't get over tab completion :) [14:52] i dont know that the CLEAN=1 paramas will get through the init.d wrapper. they probably will [15:05] smoser: yes they will [15:05] our init.d scripts uses cleanstart and cleanstop to clean [15:05] not CLEAN=1 [15:05] this was done for upstart [15:07] kim0: ok well now I have 0002/0002 m1.small types free [15:08] wow I guess that's an improvement :) [15:08] and now it lets me get a pending instance [15:08] obino: Hi o/ [15:08] feel free to jump in and help :) [15:08] hi kim0 [15:09] rwat: tailf nc.log and let's see what happens [15:09] congratz again for the cloud days session: I'm going through the session I couldn't attend :) [15:16] obino: awesome, thanks man .. really appreciate it [15:17] you presenting that is [15:18] anytime :) [15:18] :) === dendrobates is now known as dendro-afk [15:29] kim0: OK the instance goes into pending, and then gets terminated. After that m1.small only shows 0001/0002 [15:29] so it seems that not only does it not start properly, it somehow fails to reset its counter [15:30] what does nc.log say then [15:30] nothing excetp [15:30] [Fri Mar 25 15:30:41 2011][005969][EUCADEBUG ] doDescribeResource() invoked [15:30] [Fri Mar 25 15:30:41 2011][005969][EUCADEBUG ] doDescribeInstances() invoked [15:31] repeated a few times [15:31] don't paste it here [15:31] use paste.ubuntu.com [15:31] ok [15:31] but there's no more than that over and over [15:31] eh [15:31] not too useful [15:31] rwat: try looking at other log files [15:32] yeah, I've been trying to wade throuhg them but there's such a huge load of them [15:32] rwat: is "sudo kvm-ok" happy on NC node ? [15:45] kim0: yes it tells me KVM acceleration can be used [15:46] weirdo .. then I'm fairly lost [15:46] yeah, well I know I am [15:47] what should normally be in nc.log? [15:48] ahh now I'm getting a lot of errors in axis2c.log about invalid keys [15:48] great :) [15:49] http://paste.ubuntu.com/585450/ [15:49] I have to say that is entirely greek to me [15:51] rwat: did you create a new (non admin) user and are using that ? [15:51] that's usually helpful :) [15:51] that error is greek to me too [15:52] ok I'll create a new user [15:53] what is axis2c? [15:53] rwat: with the new user, don't forget to euca-add-keypair for it [15:53] I just love java :) [15:54] javajava [15:57] ok I'm trying to start an instance with the new credentials [15:57] * rwat hates java [15:59] * kim0 was being sarcastic :) === dendro-afk is now known as dendrobates [16:07] kim0: and my instance just went from pending to terminated again [16:07] * rwat prepares to throw cloud out of window [16:08] * rwat remembers he has a pulled muscle and stops [16:08] rwat: hmm maybe ask in #eucalyptus [16:14] rwat: is the kvm process started on the NC at all? For me it was always useful to peek at the nc.log on the NC with grep -v doDescribe /var/log/eucalyptus/nc.log [16:14] rwat: if nothing is displayed when you do a run-instances, then the cc or clc is broken [16:28] TeTeT: kvm doesnt seem to start [16:28] check your bios [16:29] crazed: done that - its all fine [16:29] rwat: how far does it get? Is the image transferred to the NC? [16:29] can you boot a regular VM with kvm? [16:30] TeTeT: i don't believe it is, but then I'm not quite sure what I'm looking for [16:30] crazed: I haven't tried to run a regular image with kvm, I could give that ago [16:30] rwat: do the tail -f /var/log/eucalyptus.log | grep -v doDescribe [16:32] TeTeT: there aren't any lines created in nc.log that don't start with doDescribe [16:32] or don't have doDescribe in them [16:33] it doesn't look to me as though it gets as far as the NC [16:34] rwat: yeah, it's blocked by the CC, CLC, not sure what to do then, sorry [16:34] definitely try to launch a vm with kvm [16:34] always start at the hypervisor and work your way up :) [16:40] Hello Kimo [16:41] Yuvaraj: hey [16:41] Yuvaraj: o/ [16:42] could you paste the boto url, [16:42] ah yeah sure [16:42] http://boto.cloudhackers.com/ec2_tut.html [16:42] there is some fault with my system, its not at all saving the history [16:43] and missed the UCD yesteraday :( [16:43] Yuvaraj: read the logs :) [16:43] due to system fault [16:43] Yuvaraj: http://foss-boss.blogspot.com/2011/03/ubuntu-cloud-days-day-2-summary.html [16:43] yes sure :) [16:43] thank you :) [16:43] You're welcome [16:45] smoser: that udev/cloud-init bug is nasty [16:46] you've made some progress ? [16:46] reproducing ~ 4% of the time is no fun [16:46] smoser: only in understanding how weird udev is [16:46] and i don't know that i've ever seen it on a reboot [16:46] The bitwhere /dev/null can't be opened is weird too, but I think a red herring [16:48] smoser: all those Connection Refused messages suggest that there's something wrong in the sequence of fork/listen on the workers [16:48] but, they communicate over pipes.. so.. I'm having to look now into how a pipe even returns ECONNREFUSED [16:50] definitely one of those times where I wish I had ctags working properly. :-/ [16:52] * SpamapS realizes now that he can fix that in 10 seconds.. and does [18:15] crazed: ok, my NC seems to run ordinary kvm just fine [18:17] hm well that's a good first start, but unfortunately my knowldege of eucalyptus is not that great. i primarily use cloudstack at this time [18:23] crazed: is cloudstack ready for eral use? [18:23] real [18:29] i'd say it is [18:29] but it runs on rhel6 if you want to use kvm [18:30] it's got some ghettoness about it though.. [18:30] really i think nimbula is the most promising === daker is now known as daker_ === dendrobates is now known as dendro-afk [20:24] kim0: not trying to dodge your bug, i just really think it's a bug in debootstrap :) [20:24] all right, now why won't winxp install in kvm under lucid for me? [20:24] yeah I agree :) [20:26] hallyn: switched the bug to debootstrap [20:26] let me know if I should do something else [20:27] kim0: oh, i'd already done that myself :) [20:27] ah ok :) [20:27] hallyn: is that the lxc deboostrap amd64 bug? [20:27] kim0: well, i *was* wondering if there is some undocumented debootstrap config file [20:28] zul: it's the one where it doesn't install netbase [20:28] hallyn: i think the amd64 iso images had a similar problem afaik colin fixed it [20:28] kim0: maybe try a different mirror [20:29] zul: what exactly was the problem? The mirror didn't have the package available? [20:29] zul: colin fixed it today ? [20:29] coz I just tried it a few hours back [20:38] zul: doh, there's a bug in the actual templates :( [20:38] comes from looking at the same time at upstart jobs and writing bash [20:44] smoser: so the udev error comes because sendmsg couldn't send on the PF_NETLINK that udev uses to talk to the kernel about device events