Monday, 2011-01-24

*** littleidea has quit IRC00:02
*** littleidea_ is now known as littleidea00:02
*** MarkAtwood has joined #openstack00:08
*** larstobi has quit IRC00:11
*** mdomsch has joined #openstack00:20
*** jimbaker has joined #openstack00:37
*** westmaas has quit IRC00:57
*** localhost has quit IRC01:02
*** stewart has quit IRC01:13
*** adiantum_ has quit IRC01:15
*** westmaas has joined #openstack01:18
*** localhost has joined #openstack01:19
*** adiantum_ has joined #openstack01:20
*** ArdRigh has joined #openstack01:28
*** stewart has joined #openstack01:37
*** adiantum_ has quit IRC01:41
*** paulmillar has joined #openstack01:42
*** adiantum_ has joined #openstack01:43
*** stewart has quit IRC01:58
*** jimbaker has quit IRC02:00
*** mdomsch has quit IRC02:09
*** ArdRigh has quit IRC02:13
*** pvo_away is now known as pvo02:15
*** MarkAtwood has quit IRC02:17
*** adiantum_ has quit IRC02:18
*** adiantum_ has joined #openstack02:24
*** miclorb_ has quit IRC02:25
*** kpepple has quit IRC02:30
*** kpepple has joined #openstack02:41
*** hagarth1 has quit IRC02:51
*** Ryan_Lane has joined #openstack03:12
*** MarkAtwood has joined #openstack03:29
*** miclorb has joined #openstack03:38
*** opengeard has quit IRC03:43
*** stewart has joined #openstack03:44
*** kashyapc has joined #openstack03:51
*** adiantum_ has quit IRC03:55
*** adiantum_ has joined #openstack03:55
*** littleidea has quit IRC03:56
*** littleidea has joined #openstack04:01
*** mdomsch has joined #openstack04:16
*** arreyder has quit IRC04:16
*** arreyder has joined #openstack04:16
*** stewart has quit IRC04:17
*** stewart has joined #openstack04:21
*** adiantum_ has quit IRC04:23
*** stewart has quit IRC04:26
*** MarkAtwood has quit IRC04:26
*** MarkAtwood has joined #openstack04:29
*** adiantum_ has joined #openstack04:35
*** littleidea has quit IRC04:37
*** littleidea has joined #openstack04:41
*** omidhdl has joined #openstack04:45
*** stewart has joined #openstack04:47
*** opengeard has joined #openstack04:51
*** omidhdl has quit IRC05:01
*** omidhdl has joined #openstack05:01
*** magglass1 has joined #openstack05:03
*** stewart has quit IRC05:18
*** MarkAtwood has quit IRC05:24
*** zenmatt has quit IRC05:25
*** adiantum_ has quit IRC05:32
*** adiantum_ has joined #openstack05:38
*** omidhdl has quit IRC05:42
*** omidhdl has joined #openstack05:44
*** MarkAtwood has joined #openstack05:55
*** pvo is now known as pvo_away05:59
*** stewart has joined #openstack06:04
*** adiantum_ has quit IRC06:08
*** pvo_away is now known as pvo06:12
*** pvo is now known as pvo_away06:13
*** Cybodog has quit IRC06:13
*** adiantum_ has joined #openstack06:14
*** MarkAtwood has quit IRC06:18
*** MarkAtwood has joined #openstack06:18
*** kpepple has left #openstack06:24
*** stewart has quit IRC06:36
*** adiantum_ has quit IRC06:39
*** GasbaKid has joined #openstack06:42
*** adiantum_ has joined #openstack06:45
*** guigui has joined #openstack06:46
*** adiantum_ has quit IRC06:53
*** adiantum_ has joined #openstack06:59
*** miclorb has quit IRC07:12
*** littleidea has quit IRC07:14
*** littleidea has joined #openstack07:16
*** thimble has joined #openstack07:25
*** thimble has quit IRC07:32
*** thimble has joined #openstack07:33
*** MarkAtwood has quit IRC07:34
*** littleidea_ has joined #openstack07:47
*** littleidea has quit IRC07:51
*** littleidea_ is now known as littleidea07:51
*** adiantum_ has quit IRC07:51
*** adiantum_ has joined #openstack07:58
*** MarkAtwood has joined #openstack08:00
*** miclorb has joined #openstack08:00
*** littleidea has quit IRC08:01
*** brd_from_italy has joined #openstack08:02
*** GasbaKid has quit IRC08:02
*** GasbaKid has joined #openstack08:02
*** georgy has joined #openstack08:03
georgyI am facing a problem while generating project zipfile08:05
georgyOSError: [Errno 2] No such file or directory: '/usr/local/lib/python2.6/dist-packages/nova-2011.1-py2.6.egg/nova/..//CA'08:06
georgyI found that it is a bug being fixed... but is there a temporary work around for this so that I cna start working?08:07
*** MarkAtwood has quit IRC08:08
*** rcc has joined #openstack08:12
*** calavera has joined #openstack08:17
*** adiantum_ has quit IRC08:21
*** freeflying has joined #openstack08:21
*** stewart has joined #openstack08:23
*** adiantum_ has joined #openstack08:27
*** georgy has quit IRC08:30
*** georgy_ has joined #openstack08:30
georgy_hi all08:31
*** nati has joined #openstack08:32
*** stewart has quit IRC08:37
*** allsystemsarego has joined #openstack08:41
*** allsystemsarego has joined #openstack08:41
*** ibarrera has joined #openstack08:41
*** adiantum_ has quit IRC08:51
*** pquerna has quit IRC08:53
*** pquerna has joined #openstack08:53
*** fitzdsl has quit IRC08:55
*** fitzdsl has joined #openstack08:56
*** adiantum_ has joined #openstack08:58
*** adiantum_ has quit IRC09:06
*** MarkAtwood has joined #openstack09:09
*** ramkrsna has joined #openstack09:11
*** adiantum_ has joined #openstack09:12
*** arthurc has joined #openstack09:13
*** fabiand_ has joined #openstack09:28
fabiand_when creating rings, is the last part (sdb1) of the "url" (aka z1- i provide, the name of the rsyncd target? or does it have to be a real device?09:34
*** adiantum_ has quit IRC09:35
*** adiantum_ has joined #openstack09:47
*** fabiand_ has quit IRC09:48
*** ramkrsna has quit IRC09:49
*** fabiand_ has joined #openstack09:50
*** adiantum_ has quit IRC10:02
*** ramkrsna has joined #openstack10:03
*** rcc has quit IRC10:04
georgy_I am getting OSError: [Errno 2] No such file or directory: '/usr/local/lib/python2.6/dist-packages/nova-2011.1-py2.6.egg/nova/..//CA'. I understand that this is because th cas are not being copied automatically. Is there some temporary workaround that I can use to get nova runnung?10:20
georgy_I am currently unable to run nova-manage zipfile .... because I get this error10:22
*** miclorb has quit IRC10:22
*** x_click_x has joined #openstack10:23
sorenI'm not sure if the CA directory gets installed at all right now (by
sorengeorgy_: As a temporary measure, you could probably just copy the CA direcotry from the tarball into /usr/local/lib/python2.6/dist-packages/nova-2011.1-py2.6.egg/10:25
georgy_that worked thanks10:28
x_click_xAnyone who has installed the nova in ubuntu?10:28
x_click_xI followed the tutorial but I can't run the "nova-manage network create 1 64"10:30
sorenHow does it fail?10:30
x_click_xI get: network does not match any options:10:30
sorenOh, you're using the packages in Maverick?10:30
sorenThey're really, really old.10:31
sorenI mean.. they were new and fresh when Maverick came out, but that's a while ago.10:31
sorenWe have a PPA with the freshest stuff in it.10:31
x_click_xSo I should use this ppa10:34
x_click_xthanks soren10:34
ttxgeorgy_, soren: that would be bug 68313710:37
uvirtbotLaunchpad bug 683137 in nova "CA code doesn't get installed by " install"" [Medium,In progress]
ttxsoren: are you still working on that ? Or should we unassign you and give someone else a chance to pick it up ?10:38
*** georgy_ has quit IRC10:40
*** GasbaKid has quit IRC10:44
sorenttx: The way  I was meaning to fix it is deferred to Cactus for sure. It'll rework all of the CA code entirely.10:45
sorenttx: Just installing the CA directory should be simple, though.10:45
sorenttx: So yeah, just unassigning it would probably be good.10:45
ttxok, will do10:46
ttxand will also document the workaround.10:46
*** x_click_x has left #openstack10:56
* soren lunches11:12
*** MarkAtwood has quit IRC11:13
*** reldan has joined #openstack11:15
*** rackerhacker has quit IRC11:22
*** larissa has quit IRC11:23
*** larissa has joined #openstack11:23
*** rackerhacker has joined #openstack11:23
*** reldan has quit IRC11:25
colinnichfabiand_: the "sdb1" part is not (necessarily) a real device, but the name of the storage mount. By default, the name of each folder under /srv/node/11:29
*** reldan has joined #openstack11:31
*** BK_man has joined #openstack11:34
*** adiantum_ has joined #openstack11:36
*** miclorb_ has joined #openstack11:37
fabiand_colinnich: :) good day. thanks. and the /srv/node mount points are shared via rsync, so i suppose it _is_ the rsyncd "share"name?11:43
*** miclorb_ has quit IRC11:43
*** kashyapc has quit IRC11:44
*** ramkrsna has quit IRC11:45
colinnichfabiand_: yes. The location is also known to all the swift storage services if you were considering moving the mounts11:49
*** ramkrsna has joined #openstack11:52
*** ctennis has quit IRC11:58
*** kashyapc has joined #openstack12:01
*** ctennis has joined #openstack12:14
*** cynb has quit IRC12:15
*** paulmillar has quit IRC12:20
*** rcc has joined #openstack12:27
rcchi soren12:28
rccmay you please have a look at  before I propose for merge?12:28
uvirtbotLaunchpad bug 681417 in nova "All errors about quotas should include a unit" [Low,In progress]12:28
*** reldan has quit IRC12:41
*** stewart has joined #openstack12:43
*** adiantum_ has quit IRC12:45
fabiand_colinnich: the whole path (/srv/nodes/sdb1) is known to them?12:47
colinnichfabiand_: /srv/nodes is known in the config files, and sdb1 comes from the find config12:47
colinnichfabiand_: "find"? of course I meant "ring"12:47
fabiand_colinnich: :) ah okay ... I was thinking that parted or so was sued to discover local devices (like sdb1) ... it those ring device names are just pathnames ...12:50
colinnichfabiand_: yes12:50
fabiand_colinnich: nice. thanks for this comprehensive answer :)12:51
*** Daviey has quit IRC12:56
*** Daviey has joined #openstack13:03
*** GasbaKid has joined #openstack13:07
*** Daviey has quit IRC13:11
*** nelson has quit IRC13:11
*** Daviey has joined #openstack13:12
*** Daviey has quit IRC13:15
*** Daviey has joined #openstack13:18
sorenrcc: It's very possible that there's only one place where it's wrong. I just saw it in one place, didn't have time to check everywhere else, and phrased the bug more generally.13:18
*** Daviey has quit IRC13:19
sorenttx: Have you tried the Ubuntu Desktop UEC images?13:20
ttxsoren: no13:20
sorenttx: Ok.13:21
*** cynb has joined #openstack13:28
*** Daviey has joined #openstack13:29
*** Daviey has joined #openstack13:30
*** brd_from_italy has quit IRC13:31
*** Daviey has quit IRC13:34
*** reldan has joined #openstack13:34
*** zul has quit IRC13:35
*** jfluhmann_ has quit IRC13:35
*** zul has joined #openstack13:36
*** pvo_away is now known as pvo13:38
*** Daviey has joined #openstack13:39
*** xEU has joined #openstack13:46
*** brd_from_italy has joined #openstack13:47
*** aliguori has joined #openstack13:50
*** ramkrsna has quit IRC13:51
*** adiantum_ has joined #openstack13:53
*** brd_from_italy has quit IRC14:00
rccok soren, cos I grepped all strings of the source, and the only message I found that there's a unit lacking is volume/api.py14:00
rcci'll propose to merge then14:00
*** pvo has quit IRC14:01
*** pvo_away has joined #openstack14:01
*** pvo_away is now known as pvo14:01
*** pvo has joined #openstack14:01
*** ChanServ sets mode: +v pvo14:01
*** deshantm_ has quit IRC14:03
*** deshantm has joined #openstack14:04
jaypipesmorning all.14:11
sorenjaypipes: o/14:12
jaypipessoren: morning :)14:12
ttxjaypipes: 'lo14:13
jaypipesttx: bonjour!14:13
*** brd_from_italy has joined #openstack14:16
*** dprince has joined #openstack14:17
*** brd_from_italy has quit IRC14:20
*** brd_from_italy has joined #openstack14:21
*** brd_from_italy1 has joined #openstack14:33
*** brd_from_italy has quit IRC14:35
*** littleidea has joined #openstack14:35
*** brd_from_italy has joined #openstack14:36
jaypipesBK_man: around?14:37
*** brd_from_italy1 has quit IRC14:38
BK_manjaypipes: yep14:41
uvirtbotNew bug: #706940 in glance "Missing deps in ubuntu packaging" [Undecided,New]
jaypipesBK_man: heya. I released a new tarball for Glance yesterday, just wanted to let you know since I know you were looking into RPM packaging for glance...14:42
*** brd_from_italy has quit IRC14:43
jaypipessoren: could I ask for your assistance on
uvirtbotLaunchpad bug 706940 in glance "Missing deps in ubuntu packaging" [Undecided,New]14:43
jaypipessoren: I'm not sure how deb packaging works, sorry :(14:43
* jaypipes just spent a whlie figuring out the release process for tarballs/source packages...14:44
sorenjaypipes: Sure.14:44
BK_manjaypipes: thanks a lot! Already build a package using new tarball14:44
*** ppetraki has joined #openstack14:45
sorenttx: is glance-upload included in any of the packages?14:46
*** f4m8 is now known as f4m8_14:46
sorenttx: (re bug 706940)14:46
uvirtbotLaunchpad bug 706940 in glance "Missing deps in ubuntu packaging" [Undecided,New]
ttxsoren: it's in "glance"14:46
sorenttx: Really?14:46
jaypipesBK_man: sweet, thx! :)14:46
sorenttx: Oh, my grep missed the wildcard.14:46
jaypipessoren: grep fail. ;)14:46
ttx$ dpkg -L glance | grep upload14:46
BK_manttx: could it be a valuable to have glance-upload in separate package?14:47
ttxBK_man: certainly.14:47
ttxwell, a separate binary package from same source package, to be precise.14:47
jaypipesBK_man: small request... the description of Glance has changed (we removed Teller and Parallax to make things simpler to understand...). Can you use the description here?
BK_manttx: correct14:48
jaypipesBK_man, ttx, soren: glance-upload is a temporary tool that will be replaced early in Cactus, so I wouldn't put too much effort into it..14:48
*** littleidea has quit IRC14:48
ttxjaypipes: agreed.14:48
sorenjaypipes: I'm not :)14:48
jaypipesk :)14:48
BK_manjaypipes: sure, will update spec. Could you please also review deps and tell me which is missing and which is unnecessary?14:48
jaypipesjust making sure you knew...14:49
jaypipesBK_man: yep14:49
sorenjaypipes: Ok, I've fixed the packaging and will build new packages.14:49
BK_manjaypipes: ok, will leave glance-upload in openstack-glance14:49
*** rpedde has joined #openstack14:49
sorenjaypipes: It's in a bit of temporary state right now, until it lands in Ubuntu proper. I'd rather wait until that's settled before I walk you through it, if that's cool with you.14:49
sorenjaypipes: It's pending review.14:49
jaypipessoren: sure, that's cool.14:50
sorenrpedde: Hey man :)14:50
jaypipessoren: thx for your help!14:50
* rpedde waves14:50
sorenjaypipes: Sure thing.14:50
*** rnirmal has joined #openstack14:50
*** westmaas1 has joined #openstack14:53
*** westmaas has quit IRC14:56
*** blakeyeager has joined #openstack14:56
*** gondoi has joined #openstack14:58
ttxsoren/jaypipes: I'm trying to test Glance, but wonder if what I try is even supported: EC2 Api -> Glance -> Ubuntu cloud images14:59
ttxI tried glance-upload with kernel and image but it complains about "kernel and ramdisk required for machine image"15:00
ttxWonder if i should just try raw.15:00
*** omidhdl has quit IRC15:01
jaypipesttx: yes, use raw.15:01
jaypipesttx: otherwise you need to use --kernel and --ramdisk params.15:02
soren--ramdisk is mandatory for non-raw images?15:02
ttxjaypipes: ok. Never tried raw images with the EC2 API toolset15:02
*** zenmatt has joined #openstack15:05
*** zul has quit IRC15:06
*** zul has joined #openstack15:07
*** brd_from_italy has joined #openstack15:07
*** nelson has joined #openstack15:08
jaypipessoren: glance-upload --help ;)15:09
jaypipessoren: it's really just a testing tool, not a full-blown client...15:10
ttxjaypipes: when I try to run an instance from Glance, it fails curling the image from port :333315:13
ttxjaypipes: (nothing runs on port 3333 on the glance host, fwiw)15:14
ttxI tried to change glance_port to 9292 or 9191 without much success.15:14
*** brd_from_italy has quit IRC15:15
*** dendro-afk is now known as dendrobates15:15
ttxlooks like I'm falling in the very same hole I've described recently. How embarassing.15:16
jaypipesttx: did you resolve the issue?15:19
jaypipesttx, soren: made me laugh:
ttxjaypipes: well, no. That's why we need to keep objectstore around, no ? euca-run-instances tries to download image from an S3-like thing.15:20
sorenjaypipes: How we've gotten this far without a AssignStringInToOut is amazing.15:20
*** allsystemsarego has quit IRC15:20
jaypipessoren: :)15:20
*** Jordandev has joined #openstack15:20
ttx(hence the 3333 port)15:20
jaypipesttx: yep. :)15:21
*** troytoman has joined #openstack15:21
JordandevHello. from this wikipage ( ) there is a host agent and guest agent. What kind of agent are used here? technology?15:22
sorenjaypipes: Is Glance supposed to eventually support the EC2 style image handling (uploading images in small chunks along with a manifest and then pointing the image registry at the manifest)?15:22
ttxLooks like I could use a step-by-step tutorial on how to run an instance from the openstack API, if anyone has that in store15:22
jaypipessoren: eh, not sure. I suppose if that functionality is deemed critical. right now though, the focus isn't on doing things the EC2 way...15:23
jaypipesttx: one sec...grabbing link15:23
jaypipesttx: look for the Prepare Images section...should give you an idea.15:24
sorenjaypipes: Well, if I get bored and add it at somepoint, will it be accepted?15:24
ttxok, I'll just wear my XenAPI filter glasses15:24
ttxjaypipes: ^ thanks.15:24
jaypipesttx: it's not really very complete documentation, more notes than anything else... sorry :(15:25
jaypipessoren: code from you? any day. ;)15:26
sorenjaypipes: I just know that close EC2 adherence is important to Ubuntu and Ubuntu is important to me :)15:26
j05httx: feel free to update it!  its a wiki!15:27
ttxsoren: do we have plans to have the openstack-api-compatible cloudservers packaged/available in Ubuntu ?15:27
jaypipessoren: it's not that we don't like the EC2 API, it's just that we haven't prioritized it; been just trying to get XenServer snapshots working with Glance for Bexar.15:27
sorenjaypipes: ...and I'd really like it if the majority of the existing documentation on UEC would remain true if Nova replaces Eucalyptus.15:27
jaypipessoren: where is that docuemntation please?15:28
sorenjaypipes: Somewhere on the Ubuntu wiki. There's a *lot*.15:28
*** bradshaw has joined #openstack15:28
sorenttx: You should know this ^ :)15:29
ttxsoren: indeed.15:29
sorenAh, on help. Not wiki.15:29
ttxit's in a bit of a sorry state though.15:29
*** brd_from_italy has joined #openstack15:30
*** pharkmillups has joined #openstack15:30
sorenAnyways, I want as few surprises for existing EC2 users as possible. When they're ready to try the awesomeness of the OpenStack API, they can switch in due time, but the transition should be as smooth as possible.15:30
ttx <-- that's my reference15:30
sorenTransition from EC2 or UEC to Nova, that is.15:31
sorenttx: That should work now, I believe.15:31
ttxright, except the euca_conf --get-credentials bit.15:31
sorenttx: Right, good point.15:31
ttxsoren: do we have plans to have the openstack-api-compatible cloudservers packaged/available in Ubuntu ?15:32
zulyeah changing documentation on the ubuntu wiki would really really suck15:32
jaypipesmaybe it's just me, but I find the entire ec2/euca client tools interfaces overly complicated, not user friendly, and utterly implementation-focused (as opposed to interface-focused)...15:33
sorenttx: It's not on my TODO list.15:33
ttxzul: well, that doc is already outdated, so..15:33
sorenttx: Yet.15:33
sorenttx: ...but it should be, I suppose.15:33
* soren takes a (long) break... need to go buy groceries and stuff15:33
Jordandevno replay for my small question :)15:33
ttxJordandev: looking...15:34
jaypipesJordandev: oops, sorry!15:34
Jordandevfrom this wikipage ( ) there is a host agent and guest agent. What kind of agent are used here? technology?15:34
Jordandevno problem ;)15:34
ttxJordandev: hrm, that sounds a bit outdated. The "host agent" is probably the nova-compute node15:35
ttxThe "guest agent" is being worked on15:35
*** blakeyeager has quit IRC15:35
jaypipesJordandev: yes, as ttx said, the host agent is the nova-compute *worker* (see nova/compute/*)15:35
Jordandevi see15:36
Jordandevand the relation between nova-compute and the cluster controller are based on AMQP15:36
jaypipesJordandev: and the guest agent is not done at all... it will (possibly) be a small daemon that runs in the guest VM that provides, say, usage records as a stream to either the host agent or directly to an AMQP pipe...15:36
jaypipesJordandev: yep, that is correct.15:36
*** annegentle has left #openstack15:37
jaypipesJordandev: nova-compute listens for messages from the scheduler node15:37
Jordandevso all the relation between nova-compute and cluster controller are based only on AMQP15:37
jaypipesJordandev: the scheduler node is given API requests from the API node.15:37
jaypipesJordandev: yes.15:37
*** annegentle has joined #openstack15:37
jaypipesannegentle: morning doc-buddy. :)15:38
Jordandevperfect, i have a better overview of nova now15:38
annegentlejaypipes: good morning :)15:38
Jordandevthanks jaypipes, ttx15:38
jaypipesJordandev: any time.15:38
jaypipesannegentle: you in San Antonio this week?15:38
jaypipespvo: just about to pack for my trip to SA this week. What's the weather supposed to be like this week?15:39
annegentleI'll probably drive down Wed. or Thurs to SA15:39
jaypipesannegentle: cool.15:40
annegentleare we having a jamboree? :)15:40
j05hannegentle: where are you located normally?15:40
annegentlej05h: I'm in Austin, keepin' it weird.15:42
j05hawesome.  me too!15:42
jaypipesannegentle: nah, just coming down for the week to work with a number of folks in SA and some meetings.15:42
annegentlej05h: where do you sit? I'm under the IKEA leaves in the middle15:42
j05hi stand in my house!15:42
j05h<— not a rackspacer15:43
annegentlej05h: ha! cool - I work from home some days :)15:43
pvojaypipes: 60-65F15:43
pvolows mid-30s - 40sF15:44
*** Daviey has quit IRC15:44
*** littleidea has joined #openstack15:46
sorenRock! Glance (source package) was just accepted into Natty.15:46
sorenOk, now I'm really leaving.15:46
*** dprince has quit IRC15:46
*** dendrobates is now known as dendro-afk15:48
*** dendro-afk is now known as dendrobates15:49
jaypipespvo: cool, cheers.15:51
*** dprince has joined #openstack15:53
*** guigui has quit IRC15:56
*** cynb has quit IRC15:59
*** littleidea_ has joined #openstack16:01
*** pvo is now known as pvo_away16:03
*** littleidea has quit IRC16:04
*** littleidea_ is now known as littleidea16:04
*** jimbaker has joined #openstack16:06
*** jbaker_ has joined #openstack16:06
*** jimbaker has quit IRC16:06
*** Daviey has joined #openstack16:11
*** omidhdl has joined #openstack16:15
*** jaypipes has quit IRC16:18
*** pharkmillups has quit IRC16:18
*** brd_from_italy1 has joined #openstack16:19
*** bradshaw has quit IRC16:21
*** brd_from_italy has quit IRC16:21
*** opengeard has quit IRC16:25
*** dragondm has joined #openstack16:26
*** GasbaKid has quit IRC16:26
*** GasbaKid has joined #openstack16:26
*** larstobi1 has joined #openstack16:28
*** bradshaw has joined #openstack16:28
Ryan_LaneI just released a web interface for openstack:
*** Daviey has quit IRC16:29
*** brd_from_italy has joined #openstack16:32
*** Daviey has joined #openstack16:33
*** kashyapc has quit IRC16:33
j05hRyan_Lane: nice!16:34
*** brd_from_italy1 has quit IRC16:34
Ryan_LaneI'm still chasing down a couple of bugs :)16:35
j05haren't we all.16:35
Ryan_Lanethe only problematic one is if the wiki account is different from the shell account16:35
Ryan_Laneand I think that's a bug in nova16:35
Ryan_Lanegoing to see if I can smash that one today :)16:36
*** cynb has joined #openstack16:37
*** pvo_away is now known as pvo16:39
*** pharkmillups has joined #openstack16:40
*** sandywalsh has joined #openstack16:42
*** calavera has quit IRC16:42
*** matclayton has joined #openstack16:47
*** brd_from_italy has quit IRC16:48
thimbleRyan_Lane: this looks very promising. I´ll have a closer look tomorrow, I hope. Must set up mediawiki first :)16:51
*** kashyapc has joined #openstack16:51
Ryan_Laneit requires quite a bit of setup right now16:52
Ryan_Lanelet me make sure I post all pre-reqs on the documentation page16:52
Ryan_Lanetechnically, you can use it without puppet or powerdns, but the schemas have to be loaded16:52
thimbleokay. I´ll set it up without puppet for the first time. are the scheme files included in your package?16:53
Ryan_Lanenope. I can add them to svn though16:54
*** johnpur has joined #openstack16:54
*** ChanServ sets mode: +v johnpur16:54
thimblethat would be nice.16:54
Ryan_LaneI should add my LDAP support scripts too16:54
thimbleis it neccessary to run the extension with mediawiki on the host which serves nova? Or can it be standalone? I had no closer look yet16:57
Ryan_Lanecan be standalone16:57
Ryan_Lanenot that I've tested it, but it uses the EC2 API and LDAP for everything16:57
Ryan_LaneI've been testing on the same box as nova16:57
*** ibarrera has quit IRC16:59
thimbleokay. maybe I can test both tomorrow.16:59
*** blakeyeager has joined #openstack17:01
*** brd_from_italy has joined #openstack17:02
Ryan_Lanethimble: cool. thanks :)17:06
*** dprince has quit IRC17:06
*** brd_from_italy has quit IRC17:08
*** dprince has joined #openstack17:08
*** dprince has quit IRC17:12
*** matclayton has left #openstack17:14
*** dprince has joined #openstack17:15
*** dprince has joined #openstack17:15
*** thimble has quit IRC17:19
*** dirakx has quit IRC17:20
*** pvo is now known as pvo_away17:22
*** maplebed has joined #openstack17:24
*** omidhdl has quit IRC17:28
*** joearnold has joined #openstack17:28
*** dubsquared has joined #openstack17:29
*** omidhdl has joined #openstack17:30
*** pvo_away is now known as pvo17:31
Adri2000I may have done something wrong, but now I can't start nova-network anymore:17:32
*** dprince has quit IRC17:35
*** dprince has joined #openstack17:35
*** MarkAtwood has joined #openstack17:36
*** x_node has joined #openstack17:38
x_nodeHi, I need some help in my nova installation17:40
*** ccustine has joined #openstack17:40
x_nodeI installed nova in ubuntu lucid, and the I'm following the tutorial in documentation but when I run nova-manage project zipfile project user, I get "No vpn data for project test03"17:43
dubsquaredthat's OK…you can continue17:43
dubsquaredit's an error that doesn't keep it from working...17:44
x_nodebut my instance state is always in "lauching"17:44
dubsquaredwhat network manager are you using?17:45
*** dendrobates is now known as dendro-afk17:45
dubsquaredand can you paste the contents of /var/log/nova/nova-compute.log to paste.openstack.org17:45
*** lorinh1 has joined #openstack17:47
*** westmaas has joined #openstack17:48
zulttx: shouldnt the modprobing of nbd be done in the nova-volume upstart script?17:48
*** Tushar has joined #openstack17:49
*** fabiand_ has quit IRC17:50
*** westmaas1 has quit IRC17:50
x_nodeI see this error... "2011-01-24 14:16:35,493 ERROR nova.compute.manager [VWI-68PUGFXZLSG1FJ1B dani test02] instance 1: Failed to spawn"17:51
*** jdurgin has joined #openstack17:51
*** Jordandev has quit IRC17:52
*** bradshaw1 has joined #openstack17:52
x_nodepaste #55117:53
*** bradshaw has quit IRC17:54
lorinh1I'm trying to use quickstart to get a single-node instance up and running, but when I do "bin/nova-manage user admin admin", I get this error:
lorinh1The database is empty. How are the tables supposed to get created initially?17:54
annegentlelorinh1: sorry, we haven't update the quickstart yet I guess - you need to run nova-manage db sync17:55
annegentlelorinh1: before you create any users17:55
annegentlelorinh1: can you paste in the URL of the page you're referring to?17:55
annegentlelorinh1: I did a quick pass through the docs on Friday but still need to do another17:56
annegentlelorinh1: great, thanks17:59
*** ctennis has quit IRC17:59
*** dprince_ has joined #openstack18:02
x_nodedubsquared: do you know the reason for the problem?   thanks.18:02
*** dprince has quit IRC18:03
lorinh1annegentle: It's also missing from here:
*** daleolds has joined #openstack18:07
annegentlelorinh1: ok, thanks! Also, you can pick up the updated script from github - it's just not up-to-date in nova's souce yet18:08
lorinh1annegentle: thanks18:08
lorinh1Where does the code that Hudson uses to generate the PPA packages live?18:10
*** alekibango has joined #openstack18:11
*** johnpur has quit IRC18:11
*** dendro-afk is now known as dendrobates18:13
*** x_node has left #openstack18:15
*** Tushar has quit IRC18:19
dubsquaredx_node: sorry, got pulled into a call…ill bbiab18:20
*** dprince_ has quit IRC18:23
*** dprince has joined #openstack18:24
uvirtbotNew bug: #707049 in nova "the net_managers expect different args to create_networks, so nova-manage's call to net_manager.create_networks was changed to use named args to prevent argument mismatching." [Undecided,New]
openstackhudsonProject nova build #445: SUCCESS in 1 min 30 sec:
openstackhudsonTarmac: Fix describe_regions by changing renamed flags.  Also added a test to catch future errors.18:27
*** retr0h has joined #openstack18:28
*** retr0h has joined #openstack18:28
*** k0p_ has joined #openstack18:32
k0p_hi guys18:32
k0p_is swift a blobstore?18:32
edayk0p_: yeah18:33
*** jdurgin has quit IRC18:34
k0p_webservices api?18:34
k0p_like s3 ou googlestore?18:34
edayk0p_: something like s3, yes18:34
*** icarus901 has quit IRC18:35
uvirtbotNew bug: #707055 in nova "__dn_to_uid in ldapdriver should return the value of the uid attribute" [Undecided,New]
*** MarkAtwood has quit IRC18:37
openstackhudsonProject nova build #446: SUCCESS in 1 min 28 sec:
openstackhudsonTarmac: the net_managers expect different args to create_networks, so nova-manage's call to net_manager.create_networks was changed to use named args to prevent argument mismatching.18:37
*** dirakx has joined #openstack18:37
*** jdurgin has joined #openstack18:38
*** te has joined #openstack18:38
*** larstobi1 has quit IRC18:41
*** larstobi has joined #openstack18:41
*** BK_man has quit IRC18:44
*** x_node has joined #openstack18:44
*** pvo is now known as pvo_away18:47
*** joearnold has quit IRC18:49
*** JordanRinke has joined #openstack18:51
uvirtbotNew bug: #707072 in nova "Launchpad Translation not showing untranslated messages automatically if new log and error messages are added" [Undecided,New]
*** x_node has left #openstack18:52
*** reldan has quit IRC18:56
*** adiantum_ has quit IRC19:01
*** nelson has quit IRC19:02
*** nelson has joined #openstack19:02
*** dendrobates is now known as dendro-afk19:03
*** pharkmillups has quit IRC19:04
*** lorinh1 has left #openstack19:05
*** dragondm has quit IRC19:05
*** kbringard has joined #openstack19:06
Adri2000ok, I've "fixed" my problem with nova-network (fixed == reinstall :p)19:06
Adri2000though I still can't get the VM network to work between the compute nodes :(19:07
*** rnirmal_ has joined #openstack19:07
Adri2000I let nova do the network configuration itself, but looks like it needs some help19:07
*** rnirmal has quit IRC19:09
*** rnirmal_ is now known as rnirmal19:09
*** sandywalsh has quit IRC19:10
*** kpepple has joined #openstack19:12
*** sirp_ has quit IRC19:15
*** sirp- has joined #openstack19:15
kbringardhey guys, I'm trying to get OS running on CentOS 5.5... I followed the instructions and got everything built and installed, but it doesn't look like the DB got created19:16
kbringardI tried to create it, with nova-manage db sync19:16
kbringardbut it gave me a sqlalchemy error19:17
kbringardit gives the same error if I try to do nova-manage db version19:17
*** dendro-afk is now known as dendrobates19:17
kbringardImportError: No module named migrate.versioning19:17
kbringardi installed sqlalchemy 0.6.6 with no errors19:18
kpepplekbringard: what does your /etc/nova.conf file look like ?19:18
kpepplekbringard: can you post output of 'nova-manage db sync' on ?19:19
kbringardhmm, maybe I missed a step then19:20
*** mdomsch has quit IRC19:20
kbringard(I've got it all up and running on ubuntu... packages make life easier :-) )19:20
kbringardhas someone by chance written init scripts for rhel/centos, or should I do that myself?19:23
kpepplekbringard: i was waiting for the elusive CentOS 6 to ship … CentOS 5.5 is just a hatful of pain due to dependencies19:24
kbringardI'm sure :-/19:24
kbringardas I'm reiterating :-D19:25
*** MarkAtwood has joined #openstack19:25
kbringardwhen I try to run the tests it fails again with the migrate.versioning error19:26
uvirtbotNew bug: #707094 in nova "Add memcache support to the ldap driver" [Undecided,New]
*** zenmatt has quit IRC19:27
kpepplekbringard: assuming you are using the bzr source (trunk revno 604), try  `pip install sqlalchemy-migrate` or rebuild your virtualenv with `./ -f`19:28
*** ascovel has joined #openstack19:29
ascovelhey guys, I have a question about setting up networking with the fixeddhcp manager...19:29
kbringardkpepple: ah, this looks promising...19:30
kpepplekbringard: we've been having some issues with over the past few days with dependencies19:31
kbringardyaaay... pip installing the sqlalchemy-migrate seems to have resolved that issue... I'll forage forward and let you know how it goes :-)19:31
kpepplekbringard: no worries19:32
retr0hkpepple: would you have any problem with us getting the centos wiki a little more up to date?19:32
*** fabiand_ has joined #openstack19:33
ascovelI guess the question I want to ask is, why was the default bridge interface reassigned with an IP from the pool?19:33
kpeppleretr0h: yes, would like to help here … really was hoping that CentOS 6 was going to GA before Bexar RC … but with the CentOS 5.6 release that may be wishful speaking. Am also working on a Fedora dev/hacking setup …19:33
retr0hkpepple: its too bad about centos 5.4 and 5.5 and 5.6 :)  dependencies seem to suck...19:35
kpeppleretr0h: yes, yes they do … the python 2.6 can be overcome, but once you go down that path you just keep finding more dependencies that can't be met with CentOS packages19:37
*** reldan has joined #openstack19:38
*** omidhdl has quit IRC19:41
colinnichnotmyname: I've merged your code in19:42
*** zenmatt has joined #openstack19:42
*** mdomsch has joined #openstack19:46
*** fabiand_ has quit IRC19:51
*** GasbaKid has quit IRC19:52
*** GasbaKid has joined #openstack19:53
*** Daviey has quit IRC19:59
*** clauden_ has joined #openstack20:03
*** Daviey has joined #openstack20:04
*** ctennis has joined #openstack20:06
creihtmtaylor: btw, eventlet 0.9.14 was released today20:09
mtaylorcreiht: w00t!20:10
*** retr0h has quit IRC20:11
bradshaw1New admin here trying to setup nova. I am having some issues with the network configuration. I am using FLatDHCPManager as it sounded like what would be easiest. The issue I am seeing is that when I euca-run-instances the br100 gets reconfigured from its normal ip to that of the network of the new VM. It feels like there should be a new interface that is created and attached to the bridge? Am I expecting the wrong outcome?20:12
kpepplebradshaw1: usually your br100 should stay at (with package install at revno 604) and then your instances get,, etc20:14
kpepplebradshaw1: what build of nova are you using ? what does your nova.conf file look like ?20:15
*** jdurgin has quit IRC20:15
bradshaw1I am using what I believe is 604 from the PPA20:15
bradshaw1so if I am going to need 2 interfaces to get external connectivity when running in FlatDHCPManager mode?20:16
*** zenmatt has quit IRC20:17
kpepplebradshaw1: no, I only have one interface on my Ubuntu box and it's working fine for me20:17
kpepplebradshaw1: make sure you don't have the br100 interface defined in /etc/network/interfaces. make sure you /etc/nova.conf file is pretty spartan (mine is at
*** adiantum_ has joined #openstack20:20
*** BK_man has joined #openstack20:21
*** adiantum_ has quit IRC20:26
*** jfluhmann has quit IRC20:27
*** zenmatt has joined #openstack20:28
*** adiantum_ has joined #openstack20:32
*** adiantum_ is now known as adiantum20:32
*** j05h has quit IRC20:36
*** j05h has joined #openstack20:39
*** adiantum has quit IRC20:40
*** dragondm has joined #openstack20:44
*** fabiand_ has joined #openstack20:45
*** sandywalsh has joined #openstack20:48
*** kpepple has left #openstack20:50
*** adiantum has joined #openstack20:52
*** retr0h has joined #openstack20:55
retr0hI am following the mac instructions here :
*** stewart has quit IRC20:56
retr0hwhen running python tools/ I get an error about "ImportError: No module named sphinx.setup_command"20:56
sorenttx: Around, by any chance?20:57
creihtretr0h: You need to install python-sphinx20:58
retr0hcreiht: i did an easy_install sphinx .. is that the same/20:59
creihtretr0h: in theory that would seem so20:59
creihtwell, dunno, I haven't done dev on a mac in a long time21:00
j05hretr0h: do it as the man says.21:00
kbringardI've got another fun dep issue21:01
kbringardImportError: No module named json21:01
kbringardhowever, I have installed simplejson21:01
creihtkbringard: what version of python are you using?21:01
creihtjson is the built in version of json for python 2.621:01
creihtwell that is interesting21:01
*** jfluhmann has joined #openstack21:01
kbringardlet me take a step back21:02
kbringardthis is a CentOS 5 machine21:02
*** jdurgin has joined #openstack21:02
*** gdusbabek has quit IRC21:02
kbringardI have python 2.6 installed, and that is what is linked to in the scripts21:02
*** redbo has quit IRC21:02
kbringardbut this is when I run the tests21:02
*** sirp- has quit IRC21:02
* creiht also isn't a nova dev, so there may be some nova specific stuff that I am unaware of21:02
kbringardso perhaps it's using the wrong version in the tests?21:02
kbringardcreiht: naw, I'm with ya, I've been wrestling with cent deps all day :-)21:02
*** chmouel has quit IRC21:02
*** edleafe has quit IRC21:02
*** dubs has quit IRC21:02
*** JordanRinke has quit IRC21:02
*** cory__ has quit IRC21:02
*** h1nch has quit IRC21:03
creihtkbringard: my first guess is that perhaps somehow the tests aren't using the right version of python?21:03
kbringardyea, you're probably right... since the network manager starts fine21:04
kbringardand the network manager stuff is the only thing failing in the tests21:04
kbringardand it's *all* with that json import error21:04
creihtnova devs: btw, it might be useful to use simplejson in preference to json as it had some speed improvements after the version that was merged into python 2.621:05
*** troytoman has quit IRC21:06
*** pvo_away is now known as pvo21:07
*** stewart has joined #openstack21:08
*** dprince has quit IRC21:09
*** ctennis has quit IRC21:09
ttxsoren: I'm now21:12
*** dendrobates is now known as dendro-afk21:12
*** MarkAtwood has quit IRC21:13
*** pvo is now known as pvo_away21:13
*** dragondm has quit IRC21:14
*** dendro-afk is now known as dendrobates21:14
*** cynb has quit IRC21:15
*** brd_from_italy has joined #openstack21:18
sorenttx: Ok, so I was looking at bug 705914.. What exactly made you file the bug? What are the symptoms?21:18
uvirtbotLaunchpad bug 705914 in nova "nova-network should stop dnsmasq when it stops (or not)" [Low,Confirmed]
*** cynb has joined #openstack21:19
ttxsoren: when you ru nova-network, it spawns a dnsmasq to handle dhcp for the net21:19
*** dragondm has joined #openstack21:20
ttxsoren: when you stop  (or purge nova-network) that dnsmasq is still running21:20
ttxsoren: so you end up with something superfluous running21:21
sorenHow did you notice?21:21
sorenI think that's key.21:21
ttxsoren: hm. I think I purge, clean up some leftovers, reinstall21:22
*** pvo_away is now known as pvo21:22
ttxsoren: and then I hit some ""port already in use" thing21:22
sorenttx: And then you happened to notice it in ps' output?21:22
*** adiantum has quit IRC21:23
sorenWould all your concerns be alleviated if we could reliable adopt existing dnsmasq processes that were left over?21:23
*** ctennis has joined #openstack21:23
*** gdusbabek has joined #openstack21:23
*** redbo has joined #openstack21:23
*** h1nch has joined #openstack21:23
ttxsoren: well it would still be a bit unclean that after purging the packages you still have extra processes21:23
*** dabo has joined #openstack21:23
*** JordanRinke has joined #openstack21:23
ttxsoren: but I think vishy's arguments are stronger than mine21:23
sorenttx: I think there is another issue at play here.21:24
sorenttx: I'm not sure there's a bug report for it..21:24
*** pvo is now known as pvo_away21:24
sorenttx: The dnsmasq process should only really be there if there's an instance that needs it.21:24
*** chmouel has joined #openstack21:24
*** sirp- has joined #openstack21:25
ttxsoren: hah. Anyway, that's a minor thing. No need to spend cycles imrpving it just now21:25
*** stewart has quit IRC21:26
*** syah has joined #openstack21:26
ttxsoren: there are other things, like noticing VM status on nova-compute startup, that are a bit more pressing :P21:26
sorenttx: Well, the real kicker is that restarting nova-network stumbles because it finds this dnsmasq left over.21:26
ttxsoren: just restarting it, is ok, I think. Purging /var/lib and resintalling is not21:27
ttxsoren: that's a corner case to me21:27
sorenttx: Oh... Yeah, you might be right.21:27
*** adiantum has joined #openstack21:27
*** miclorb has joined #openstack21:28
ttxsoren: even if you just purge the package, it should still work, since our current buggy purge leaves /var/lib/nova around21:29
sorenAw... harash.21:29
ttxso it takes a maniac cleaner like me to trigger the bug ;)21:29
sorenharsh, even.21:29
sorenMy has OCD, too :)21:29
kbringardwe should start porting nova to python 3, so once centos/rhel 6 comes out with python 2.6 we can still complain about redhat :-p21:30
sorenttx: You know you can commit to that branch, too, right? :)21:31
ttxsoren: yes, and I did, just not that specific bug :P21:31
sorenttx: Then you forefeit your rights to complain :)21:31
*** jfluhmann has quit IRC21:32
ttxI didn't complain. I just pointed out the combination of two bugs that made things work for most people :P21:32
*** jfluhmann has joined #openstack21:33
*** pvo_away is now known as pvo21:34
*** bcwaldon has quit IRC21:34
*** bcwaldon has joined #openstack21:34
*** jtran has joined #openstack21:36
jtrananyone here?21:36
kbringardI am :-p21:37
*** reldan has quit IRC21:40
Ryan_Laneare there any common reasons why nova-compute would show as down in "nova-manage service list"?21:42
Ryan_Lanemine are showing as down, and I have no idea why21:42
sorenThere's a "nova-manage service list"? Cool!21:42
Ryan_Laneit kind of worries me you didn't know that :)21:42
sorenMe too! :)21:43
sorenAh, it's from the burnin-support branch.21:43
sorenYeah, didn't pay much attention to that one, I suppose.21:43
sorenRyan_Lane: Are the timestamps updating correctly for the compute workers?21:44
Ryan_Lanefrom the service list?21:45
sorenThey're in UTC, right?21:45
sorenLike 21:45-ish right now?21:45
Ryan_Laneabout a minute off on the compute nodes21:45
sorenThat may be your problem.21:46
kbringardanyone know where I can get python 2.6 libvirt-python rpms for cent 5.5?21:46
Ryan_Lanesoren: thanks21:47
sorenRyan_Lane: np21:48
Ryan_Laneand yes, nova-manage service list is neat :)21:48
Ryan_Lanesimilarly the ability to enable and disable the services21:49
sorenIt's my new favourite toy.21:49
*** fabiand_ has quit IRC21:49
*** mdomsch has quit IRC21:54
*** adiantum has quit IRC21:57
kbringardwhere are the machine types defined21:57
*** adiantum has joined #openstack22:03
*** westmaas has left #openstack22:06
*** k0p has joined #openstack22:07
*** ccustine has quit IRC22:10
*** dirakx has quit IRC22:10
*** k0p_ has quit IRC22:10
*** MarkAtwood has joined #openstack22:11
*** adiantum has quit IRC22:15
*** kpepple has joined #openstack22:17
*** pvo is now known as pvo_away22:18
*** Guest84379 has quit IRC22:18
vishykbringard: chuckle22:18
kbringardhehe, is that a good chuckle or a bad chuckle :-)22:19
vishysoren, Ryan_Lane: good thing we didn't default to disabled:)22:19
kpeppleis jay.pipes around ?22:20
vishyis anyone looking at the mysql disconnect bug?22:20
vishykbringard: that was in response to your red hat comment22:22
vishykbringard I was reading scrollback22:22
Ryan_LaneI'm having a problem with access to using flatdhcp networking...22:22
kbringardah, hehe22:22
vishykbringard: nova/compute/instance_types.py22:22
kbringardahhh, awesome, thanks vishy22:22
*** sandywalsh has quit IRC22:22
Ryan_LaneIt looks like the iptables rule is configured correctly, but I can't telnet to port 80 on that address22:22
vishyRyan_Lane: show me the rule?22:23
Ryan_LaneDNAT       tcp  --       tcp dpt:80 to:
Ryan_LaneI can connect to on 877322:24
*** GasbaKid has quit IRC22:25
*** GasbaKid has joined #openstack22:25
vishyRyan_Lane: cat /proc/sys/net/ipv4/ip_forward22:26
Ryan_Laneecho'd 1 into it. still can't connect :(22:26
uvirtbotNew bug: #707171 in nova "HyperV instance is created with .001% CPU limit" [Undecided,New]
vishyI think I've run into every possible misconfigured setting at some point dealing with nova...22:27
vishyRyan_Lane: are you trying via localhost?22:27
vishythat doesn't work iirc22:27
Ryan_Lanelet me try from another host then22:28
vishyyou need to be on a different host22:28
vishyyes other host + ip_forward should be it22:28
Ryan_Lanenope :(22:28
*** adiantum has joined #openstack22:29
JordanRinkeso, I am trying to fix a hyperv bug, I hope I am doing this bzr stuff right :/22:29
*** rnirmal has quit IRC22:29
Ryan_Lanenet.ipv4.conf.br100.forwarding = 122:30
Ryan_Laneso, it should let me try through the bridge on another host22:30
*** rnirmal has joined #openstack22:30
*** kbringard has quit IRC22:31
JordanRinkeso, I have this branch with the fix in it, do I just propose it for merge or is there anything else to do?22:31
vishyRyan_Lane :(22:32
sorenOh, I know what that is.22:33
soreni think.22:33
* soren reads scrollback22:33
*** westmaas has joined #openstack22:33
sorenYeah, that's not it.22:33
vishyRyan_Lane can you pastie me -L -n -v?22:33
sorenvishy, Ryan_Lane: I was this earlier today, too.22:33
sorenIt's not iptables.22:33
vishysoren oh?22:33
*** cynb has quit IRC22:33
sorenIt's because nothing is responding to arp who-has requests.22:34
Ryan_Laneis dnsmasq supposed to be doing that?22:34
sorenWell, it was for me, at least.22:34
sorenActually, it's doesn't really matter.22:34
Ryan_Lanethis makes sense, because I did have ip forwarding enabled before I rebooted22:35
sorenSomething just needs to do it so that the guest actually will send the packet.22:35
Ryan_LaneI hadn't added it to sysctl22:35
vishysoren: does this only affect flatdhcp?  I've never seen this issue with vlan mode...22:35
sorenvishy: It depends on the guest, really.22:35
sorenvishy: I saw it today with an ubuntu desktop image. Presumably because network-manager sets a link-local address for eth0.22:36
Ryan_Lanepastebin for iptables is here:
sorenvishy: Once there's a link-local address set, doesn't route packets destined for other link-local addresses.22:37
*** cynb has joined #openstack22:37
Ryan_Lanejust incase I have actually screwed something up :)22:37
vishysoren: oh- that was the issue i was seeing before with desktop version failing to hit metadata22:37
sorenvishy: Rather, it tries to send them directly. And that involves arp.22:37
sorenvishy: ...and nothing responds to arp.22:37 the packet never gets sent.22:37
vishyRyan_Lane, pastie the -t nat table as well?22:37
sorenI got our pastebin added to pastebinit this morning, by the way. If anyone cares.22:38
sorenIt'll be in the upcoming release (due out Friday).22:38
vishysoren: should we just have network node ip addr add
*** bradshaw1 has quit IRC22:39
sorenvishy: That would solve it.22:39
sorenvishy: I just wonder if we can do better.22:39
* vishy is weak on arp spoofing :)22:40
sorenvishy: If each compute node had it set instead, the arp packets never needed to leave that box.22:40
vishysoren: I actually think the compute node is a better place to the forward, since it would work in flat mode as well22:40
vishys/to/to do22:41
sorenvishy: Yup, let's do that.22:42
sorenSomething like:22:42
uvirtbotNew bug: #707180 in glance "glance 0.1.3 at PI doesn't contain revno 49 changes" [Undecided,New]
sorenip addr add scope link dev <interface name>22:42
* soren wanders off to bed22:43
pquernais opnestack setup to do GSoC?22:43
vishyRyan_Lane: your forwarding rule isn't getting hit, so i think your test is not working successfully.  You need to be on a box that is using the network node as a gateway. to test if the forward rule is set22:43
creihtpquerna: not that I know of, but that would be very cool22:44
Ryan_Lanevishy: I'm assuming the VMs get gateways via dhcp?22:46
Ryan_Laneshould I have the network_host setting set to in this situation?22:46
*** jtran has left #openstack22:47
vishyRyan_Lane yes, no network_host is the hostname that nova-network is running on22:47
Ryan_Laneok. I assumed as much22:47
*** ascovel has quit IRC22:49
vishyRyan_Lane if you go to one of your other boxes and do a route add -net gw <ip of nova-network>22:50
*** nati-1 has joined #openstack22:50
vishythen you should be able to curl and see it show up in the iptables rule and it should forward you properly22:51
*** ppetraki has quit IRC22:51
vishymight need net
vishynot sure how picky route is22:51
Ryan_LaneSIOCADDRT: Invalid argument22:51
annegentlepquerna: not that I know of, but thanks for the deadlines on GSoC. It's a great program.22:53
Ryan_Laneadding as host instead of net works22:54
Ryan_Lanethen I can telnet to the address/port properly22:55
*** cynb has quit IRC22:55
vishyok then it is working22:56
Ryan_Lanestrangely enough, I can't ping the VM's IP address either22:56
*** pvo_away is now known as pvo22:56
vishyprobably just needed the ips added22:56
Ryan_LaneI see the DHCP traffic22:56
vishy* ip_forward22:56
vishycheck nova-network log and verify the ip is reporting leased22:57
*** nati-1 has quit IRC22:57
vishyand the guest console log shows the ip being discovered22:57
Ryan_Lanedoesn't look like it22:58
Ryan_Lanethe syslog shows dnsmasq offering it22:58
Ryan_Laneagreed. I wonder if the image I'm using is having issues22:58
vishywhich image?22:58
Ryan_LaneI'm using a uec 10.04 amd64 image22:58
*** adiantum has quit IRC22:59
vishydid you start a new one after adding ip forward?22:59
Ryan_LaneI only need the forward on the network node, right?22:59
Ryan_Laneand yeah, after the forward22:59
vishyso dhcp fail...23:00
vishywhere is br100 bridging?23:00
Ryan_Lanehmm. there's two instances of dnsmasq running23:00
vishythere should be two23:00
Ryan_Laneah ok23:00
Ryan_Lanebr100 is bridged to eth223:00
Ryan_Laneall of which are connected to the same switch23:01
Ryan_Lanethe compute nodes and network-node are all using eth223:01
openstackhudsonProject swift build #172: SUCCESS in 26 sec:
openstackhudsonTarmac: Fix rate limiting (middleware and util function) to allow for catch up.  Configurable logging in auditor. Docs.23:01
Ryan_Lanebr100 is bridged to eth2 on all, and all are connected to the same switch23:01
Ryan_Lanenova is doing the bridging23:01
vishydoes br100 have an ip on network-node?23:02
Ryan_Lanenetwork-node? yes23:03
Ryan_Lanesorry :)23:03
vishycan you ping that ip from compute nodes?23:03
Ryan_Lanethe compute nodes don't have an ip address on the bridge23:03
Ryan_Lanebut if I add one, yes23:03
Ryan_LaneI'm seeing requests and responses via tcpdump on the compute node for dhcp23:04
vishybut somehow the response isn't getting back23:04
Ryan_Laneand on the network node23:04
vishyare you using NWFilter FW driver or IPTables?23:04
*** adiantum has joined #openstack23:05
Ryan_Lanehow do I check that?23:05
vishywhat is the address of nova-network on br100 and what is the range of ips you are assigning to guests?23:05
Ryan_Laneinet brd scope global br10023:06
Ryan_Laneassigning and higher23:06
vishyah ok you're doing private23:07
Ryan_LaneI'm seeing both the request and the response for the IP on the compute node23:07
Ryan_Laneso the dhcp server itself is fine23:08
Ryan_LaneI'm guessing the instance is screwed up somehow23:08
vishyso the response packet isn't getting back into the instance23:08
Ryan_LaneI'll connect directly to the vnet instead of the bridge23:08
vishyi would check compute host iptables rules23:09
vishyand tcpdump on the compute host to see if it is getting the packet23:09
Ryan_Lanewas using tcpdump on compute host23:09
Ryan_Laneon br10023:09
Ryan_Lanesaw dhcp traffic both ways23:10
Ryan_Laneno dropped traffic23:10
dubsquaredsuper late to the party, but is this with flatDHCP?23:11
vishycould be getting blocked by iptables rules on the compute host before it gets into the guest though23:11
retr0hstill stuck on  tools/ with "ImportError: No module named sphinx.setup_command" could there be a chicken/egg problem where glance is installed before sphinx?23:12
vishydubsquared: yes23:12
Ryan_Laneah. true23:12
vishyRyan_Lane: there are specific rules for dhcp, you should make sure that those are getting hit with an -L -n -v23:12
Ryan_Lanedo I need that in the security groups too?23:13
Ryan_Lanelooks like the dhcp reply didn't make it to vnet023:15
*** brd_from_italy has quit IRC23:15
*** cynb has joined #openstack23:16
Ryan_Laneyeah. getting filtered, for sure23:16
Ryan_Laneiptables on compute node:
Ryan_LaneI need to kill that stupid virbr0 device :)23:18
Ryan_Lane    6  1968            udp  --  *      *              udp spt:67 dpt:6823:19
Ryan_Lanethat should be allowing the traffic23:19
Ryan_Lanethough nova-local and nova-inst-20 aren't included in the INPUT chain for some reason23:20
*** nelson has quit IRC23:22
*** nelson has joined #openstack23:23
Ryan_LaneI bet libvirt is destroying the input rules23:23
*** adiantum has quit IRC23:26
Ryan_Lanehmm. why isn't nova adding any input rules?23:27
openstackhudsonProject swift build #173: SUCCESS in 29 sec:
openstackhudsonTarmac: In-depth documentation on paste.deploy configuration files.23:31
openstackhudsonget_logger now uses python's log routing to separate multiple log_names and log_facilities in the same process.23:31
*** adiantum has joined #openstack23:32
*** gondoi has quit IRC23:33
vishyRyan_Lane     6  1968            udp  --  *      *              udp spt:67 dpt:6823:34
vishylooks like it is working there23:34
vishyRyan_Lane: don't see a matching rule coming back though...23:35
Ryan_Lanethere's no input rules at all23:35
Ryan_Lanethey aren't being added23:35
vishynova-inst-20 got removed from the input chain23:36
Ryan_Laneinfact, when an instance is created, the input chain gets wipes23:36
Ryan_Lanethat has got to be a bug :)23:36
vishywait do you need input?23:36
*** reldan has joined #openstack23:36
vishyisn't forward enough?23:36
Ryan_Lanewould it be?23:37
vishyRyan_Lane: i thought so...23:37
*** cynb has quit IRC23:38
Ryan_Laneoh, it should be23:40
Ryan_Lanethere's no drop rule in the input chain anyway23:40
vishyso maybe you are correct about your image being borked23:40
vishytry an ami-tty23:40
Ryan_Laneah. true. I'm not injecting anymore, I can try that23:41
Ryan_Laneis that 32 or 64 bit?23:42
*** adiantum has quit IRC23:42
*** GasbaKid has quit IRC23:45
*** GasbaKid has joined #openstack23:45
openstackhudsonProject swift build #174: SUCCESS in 28 sec:
openstackhudson* Tarmac: Trivial output tweaks.23:46
openstackhudson* Tarmac: The default reseller_prefix in swift is "AUTH". The domain_remap feature fails in this case because browsers tend to convert urls to lowercase and so by the time the remapped url reaches proxy-server, it is not correct and the request fails.23:46
openstackhudsonChanged the code so that if the account starts with the wrongly cased version of reseller prefix, it is corrected. reseller_prefix read from proxy-server.conf.23:46
*** adiantum has joined #openstack23:47
*** miclorb has quit IRC23:49
anticwcreiht: wrt to PEP8 i can do that, but it's going to affect almost every line in the script23:50
anticwcreiht: so anything else pending will just me a merge pain23:50
creihtanticw: Not a pressing issue23:50
creihtIt is one of those scripts that wasn't very maintained or cared about until you started using it :)23:52
*** Ryan_Lane is now known as Ryan_Lane|away23:54
anticwcreiht: i'm only using it as far as it finds consistency problems23:54
creihtIt is good that it is getting some love, as it can prove to be useful23:55
*** rpedde has quit IRC23:55
*** rnirmal has quit IRC23:57

Generated by 2.14.0 by Marius Gedminas - find it at!