Wednesday, 2011-02-09

*** Hayati has quit IRC00:00
*** masumotok has joined #openstack00:03
*** CloudChris has joined #openstack00:04
masumotoktermie: around? I have a question about nova/db/sqlalchemy/migrate_repo/versions/. when I propose merge request to cactus, my branch(live migration branch) add some columns to Service. can I add some lines to or create 003_cactus? I think you are responsible on it..00:08
kpepplemasumotok termie : i have the same question ... i've been using 003_cactus00:09
*** iammartian has quit IRC00:09
*** alekibango has quit IRC00:10
*** kashyapc has quit IRC00:11
masumotokkpepple: do you already make your branch public? I want to see just for my reference if possible..00:12
*** kashyapc has joined #openstack00:12
kpepplemasumotok: yes, here it is in lp:~ken-pepple/nova/dynamicinstancetypes00:14
kpepplemasumotok: this works but i am unsure if it is against best practices to seed the database in this way (lines 59 - 78)00:14
*** mlai has left #openstack00:15
*** joearnold has quit IRC00:17
masumotokkpepple: thanks!  actually we're going to find out unsure thing if core devs around :)00:17
*** reldan has quit IRC00:18
*** msassak has quit IRC00:18
*** kashyapc has quit IRC00:23
*** joearnold has joined #openstack00:24
*** kashyapc has joined #openstack00:24
*** jfluhmann has joined #openstack00:25
*** maplebed has quit IRC00:30
*** maplebed has joined #openstack00:32
*** kashyapc has quit IRC00:34
*** bwalker7125 has joined #openstack00:37
*** reldan has joined #openstack00:37
*** kashyapc has joined #openstack00:46
*** grapex has left #openstack00:48
Ryan_Laneis there any easy way to add display name to an image? seems the euca tools can't do it00:51
Ryan_LaneAh. I can edit the info.json file directly, and restart nova-objectstore. (that isn't terribly ideal ;) )00:55
nelsoncreiht: does the "B" release of swift have this code removed? # Hook to ensure connection resets don't blow up our servers.00:55
nelsoncreiht: Because, Ubuntu 10.04 has the version of Eventlet that has ECONNRESET00:56
*** rlucio has quit IRC01:00
nelsonIt looks like if you don't create cert_file, then swift doesn't use SSL, correct? I didn't see that in any documentation I've read so far; I've pulled that out of the source.01:01
*** openstackhudson has quit IRC01:02
*** dfg has quit IRC01:03
*** openstackhudson has joined #openstack01:05
*** reldan has quit IRC01:09
*** opengeard has quit IRC01:14
*** dirakx has joined #openstack01:26
*** opengeard has joined #openstack01:27
creihtnelson: ahh that hasn't been pulled out yet01:33
*** dragondm has quit IRC01:33
creihtand if cert_file isn't defined then it doesn't use ssl01:33
nelsonI guess it's more a function of whether you want to require Ubuntu 10.04 ... which I think you do.01:33
creihtyeah we require a recent enough version of eventlet that it should be fixed now01:34
creihtoh I guess it doesn't explictly say the behavior of ssl01:34
nelsoncreiht: yes, I see that the parameter is described .... right.01:34
creihtgood catch01:35
nelsonPretty sure I read that documentation once before, looking for a way to turn ssl off.01:35
creihtyeah, we don't use the ssl functionality, so it doesn't get as much attention in the docs01:35
*** westmaas1 has joined #openstack01:44
uvirtbotNew bug: #715535 in swift "Remove ECONNRESET patch from proxy server" [Low,Confirmed]
*** bwalker7125 has quit IRC01:48
*** vvuksan1 has quit IRC01:54
*** jtimberman has quit IRC02:03
*** jtimberman has joined #openstack02:04
*** HugoKuo has quit IRC02:06
*** RobertLaptop has joined #openstack02:06
*** HugoKuo has joined #openstack02:07
*** HugoKuo has joined #openstack02:09
*** prewarmoney has joined #openstack02:19
*** rlucio has joined #openstack02:24
*** vvuksan has joined #openstack02:31
*** opengeard has quit IRC02:35
*** prewarmoney has quit IRC02:41
*** fysa has joined #openstack02:42
*** dirakx has quit IRC02:44
*** joearnold has quit IRC02:44
*** opengeard has joined #openstack02:49
*** maplebed has quit IRC02:53
*** dirakx has joined #openstack03:00
*** baldben has quit IRC03:11
*** Guest85402 has joined #openstack04:00
*** guynaor has joined #openstack04:02
*** guynaor has left #openstack04:04
*** jtimberman has quit IRC04:15
*** ccustine has quit IRC04:16
*** littleidea has joined #openstack04:20
*** Guest85402 has quit IRC04:21
*** dirakx has quit IRC04:21
*** pothos has quit IRC04:21
*** lionel__ has quit IRC04:21
*** jbarratt has quit IRC04:21
*** annegentle has quit IRC04:21
*** uvirtbot has quit IRC04:21
*** brainproxy has quit IRC04:21
*** _cerberus_ has quit IRC04:21
*** klumpie has quit IRC04:21
*** smoser has quit IRC04:21
*** jaysun has quit IRC04:21
*** letterj has quit IRC04:21
*** sunech has quit IRC04:21
*** HugoKuo has quit IRC04:21
*** EdwinGrubbs has quit IRC04:21
*** drico has quit IRC04:21
*** notmyname has quit IRC04:21
*** Jbain has quit IRC04:21
*** chmouel has quit IRC04:21
*** jeevan_ullas has quit IRC04:21
*** dovetaildan has quit IRC04:21
*** jarrod has quit IRC04:21
*** lool has quit IRC04:21
*** antonym has quit IRC04:21
*** miclorb has quit IRC04:21
*** MarkAtwood has quit IRC04:21
*** pandemicsyn has quit IRC04:21
*** Daviey has quit IRC04:21
*** Gh0sty has quit IRC04:21
*** husimon has quit IRC04:21
*** rossij has quit IRC04:21
*** opengeard has quit IRC04:21
*** openstackhudson has quit IRC04:21
*** kashyapc has quit IRC04:21
*** tr3buchet has quit IRC04:21
*** imsplitbit has quit IRC04:21
*** londo has quit IRC04:21
*** sirp_ has quit IRC04:21
*** dabo has quit IRC04:21
*** blamar has quit IRC04:21
*** Adri2000 has quit IRC04:21
*** CloudChris has quit IRC04:21
*** iRTermite has quit IRC04:21
*** anticw has quit IRC04:21
*** aimon has quit IRC04:21
*** mtaylor has quit IRC04:21
*** clayg has quit IRC04:21
*** gholt has quit IRC04:21
*** RJD22 has quit IRC04:21
*** yosh has quit IRC04:21
*** stewart has quit IRC04:21
*** ctennis has quit IRC04:22
*** hggdh has quit IRC04:22
*** sandywalsh_ has quit IRC04:22
*** hazmat has quit IRC04:22
*** RobertLaptop has quit IRC04:22
*** filler has quit IRC04:22
*** arun_ has quit IRC04:22
*** ianweller has quit IRC04:22
*** tomo_bot has quit IRC04:22
*** timrc has quit IRC04:22
*** zul has quit IRC04:22
*** jfluhmann_away has quit IRC04:22
*** taihen has quit IRC04:22
*** jaypipes has quit IRC04:22
*** piezo has quit IRC04:22
*** soeren has quit IRC04:22
*** flashn has quit IRC04:22
*** karmabot has quit IRC04:22
*** ioso has quit IRC04:22
*** Gh0sty has joined #openstack04:24
*** Guest85402 has joined #openstack04:25
*** dirakx has joined #openstack04:25
*** opengeard has joined #openstack04:25
*** HugoKuo has joined #openstack04:25
*** RobertLaptop has joined #openstack04:25
*** openstackhudson has joined #openstack04:25
*** kashyapc has joined #openstack04:25
*** CloudChris has joined #openstack04:25
*** ctennis has joined #openstack04:25
*** miclorb has joined #openstack04:25
*** MarkAtwood has joined #openstack04:25
*** EdwinGrubbs has joined #openstack04:25
*** pandemicsyn has joined #openstack04:25
*** hggdh has joined #openstack04:25
*** drico has joined #openstack04:25
*** arun_ has joined #openstack04:25
*** stewart has joined #openstack04:25
*** sandywalsh_ has joined #openstack04:25
*** tr3buchet has joined #openstack04:25
*** zul has joined #openstack04:25
*** jfluhmann_away has joined #openstack04:25
*** iRTermite has joined #openstack04:25
*** hazmat has joined #openstack04:25
*** taihen has joined #openstack04:25
*** jaypipes has joined #openstack04:25
*** Daviey has joined #openstack04:25
*** pothos has joined #openstack04:25
*** imsplitbit has joined #openstack04:25
*** lionel__ has joined #openstack04:25
*** filler has joined #openstack04:25
*** anticw has joined #openstack04:25
*** londo has joined #openstack04:25
*** ianweller has joined #openstack04:25
*** sirp_ has joined #openstack04:25
*** dabo has joined #openstack04:25
*** blamar has joined #openstack04:25
*** aimon has joined #openstack04:25
*** notmyname has joined #openstack04:25
*** mtaylor has joined #openstack04:25
*** tomo_bot has joined #openstack04:25
*** jbarratt has joined #openstack04:25
*** RJD22 has joined #openstack04:25
*** husimon has joined #openstack04:25
*** piezo has joined #openstack04:25
*** soeren has joined #openstack04:25
*** jaysun has joined #openstack04:25
*** sets mode: +vvv pandemicsyn notmyname mtaylor04:25
*** Jbain has joined #openstack04:25
*** Adri2000 has joined #openstack04:25
*** clayg has joined #openstack04:25
*** chmouel has joined #openstack04:25
*** annegentle has joined #openstack04:25
*** jeevan_ullas has joined #openstack04:25
*** rossij has joined #openstack04:25
*** flashn has joined #openstack04:25
*** sunech has joined #openstack04:25
*** letterj has joined #openstack04:25
*** smoser has joined #openstack04:25
*** klumpie has joined #openstack04:25
*** _cerberus_ has joined #openstack04:25
*** brainproxy has joined #openstack04:25
*** uvirtbot has joined #openstack04:25
*** antonym has joined #openstack04:25
*** lool has joined #openstack04:25
*** jarrod has joined #openstack04:25
*** dovetaildan has joined #openstack04:25
*** gholt has joined #openstack04:25
*** yosh has joined #openstack04:25
*** karmabot has joined #openstack04:25
*** timrc has joined #openstack04:25
*** ioso has joined #openstack04:25
*** sets mode: +vvv letterj _cerberus_ antonym04:25
*** opengeard has quit IRC04:25
*** openstackhudson has quit IRC04:25
*** kashyapc has quit IRC04:25
*** tr3buchet has quit IRC04:25
*** imsplitbit has quit IRC04:25
*** londo has quit IRC04:25
*** sirp_ has quit IRC04:25
*** dabo has quit IRC04:25
*** blamar has quit IRC04:25
*** Adri2000 has quit IRC04:25
*** mtaylor has quit IRC04:25
*** mtaylor has joined #openstack04:25
*** RJD22 is now known as RJD22|away04:27
*** arcane has quit IRC04:29
*** opengeard has joined #openstack04:29
*** openstackhudson has joined #openstack04:29
*** kashyapc has joined #openstack04:29
*** tr3buchet has joined #openstack04:29
*** imsplitbit has joined #openstack04:29
*** londo has joined #openstack04:29
*** sirp_ has joined #openstack04:29
*** dabo has joined #openstack04:29
*** blamar has joined #openstack04:29
*** Adri2000 has joined #openstack04:29
*** arcane has joined #openstack04:29
*** mtaylor has quit IRC04:32
*** mtaylor has joined #openstack04:32
*** ChanServ sets mode: +v mtaylor04:32
*** joearnold has joined #openstack04:33
*** RJD22|away is now known as RJD2204:34
*** rlucio has quit IRC04:47
*** joearnold has quit IRC04:51
*** dirakx1 has joined #openstack04:54
*** dirakx has quit IRC04:54
*** Guest85402 is now known as sateesh05:00
*** dirakx1 has quit IRC05:01
*** RobertLaptop has left #openstack05:01
*** dirakx has joined #openstack05:02
*** hadrian has quit IRC05:04
*** vvuksan has quit IRC05:04
*** dirakx has quit IRC05:07
*** littleidea has quit IRC05:16
*** kashyapc has quit IRC05:17
*** opengeard has quit IRC05:31
*** adjohn has joined #openstack05:34
*** kashyapc has joined #openstack05:38
*** opengeard has joined #openstack05:44
*** adjohn has quit IRC05:47
*** omidhdl has joined #openstack05:55
*** wedjat has quit IRC06:05
*** opengeard has quit IRC06:12
*** wedjat has joined #openstack06:12
*** wedjat has joined #openstack06:12
*** adjohn has joined #openstack06:13
*** Hayati has joined #openstack06:14
*** westmaas1 has quit IRC06:21
*** kpepple has quit IRC06:36
*** kpepple has joined #openstack06:36
*** baldben has joined #openstack06:41
*** f4m8_ is now known as f4m806:41
*** kpepple has quit IRC06:42
*** kpepple has joined #openstack06:43
*** kpepple has quit IRC06:48
*** kpepple has joined #openstack06:49
*** kpepple has quit IRC06:50
*** kpepple has joined #openstack06:50
*** Hayati has quit IRC06:51
*** miclorb has quit IRC06:54
*** kpepple has quit IRC06:55
*** kpepple_ has joined #openstack06:55
*** kpepple_ has quit IRC06:56
*** kpepple has joined #openstack06:56
*** kpepple has joined #openstack06:57
*** Hayati has joined #openstack06:59
*** kpepple has quit IRC07:01
*** kpepple has joined #openstack07:02
*** CloudChris has quit IRC07:05
*** Hayati has quit IRC07:07
*** omidhdl has quit IRC07:07
*** kpepple_ has joined #openstack07:09
*** kpepple has quit IRC07:10
*** kpepple has joined #openstack07:14
*** kpepple_ has quit IRC07:14
*** kpepple_ has joined #openstack07:19
*** kpepple has quit IRC07:19
*** aimon has quit IRC07:22
*** aimon has joined #openstack07:27
*** kpepple_ has quit IRC07:27
*** kpepple has joined #openstack07:28
*** jsgotangco has joined #openstack07:29
*** kpepple has joined #openstack07:30
*** berendt has joined #openstack07:30
*** RJD22 is now known as RJD22|away07:33
*** kpepple has quit IRC07:34
*** kpepple has joined #openstack07:35
*** littleidea has joined #openstack07:35
*** miclorb has joined #openstack07:39
*** CloudChris has joined #openstack07:42
*** thimble has joined #openstack07:46
uvirtbotNew bug: #715618 in nova "fix logging message if AMQP server is not reachable" [Undecided,New]
*** kpepple has quit IRC07:49
*** kpepple has joined #openstack07:49
*** jsgotangco has quit IRC07:49
*** kpepple has quit IRC07:50
*** kpepple has joined #openstack07:51
uvirtbotNew bug: #715620 in nova "fix logging message if database server is not reachable" [Undecided,New]
*** kpepple has quit IRC07:55
*** kpepple has joined #openstack07:56
*** MarkAtwood has quit IRC07:59
*** citral has joined #openstack08:02
*** RJD22|away is now known as RJD2208:05
*** kpepple has quit IRC08:06
*** Hayati has joined #openstack08:07
*** kpepple has joined #openstack08:07
*** kpepple has quit IRC08:12
*** kpepple has joined #openstack08:12
*** calavera has joined #openstack08:14
*** fabiand_ has joined #openstack08:23
*** kpepple has quit IRC08:24
*** kpepple_ has joined #openstack08:24
*** kpepple has joined #openstack08:26
*** allsystemsarego has joined #openstack08:30
*** allsystemsarego has joined #openstack08:30
*** kpepple has quit IRC08:30
*** kpepple has joined #openstack08:31
*** perra has joined #openstack08:32
*** kpepple has quit IRC08:35
*** kpepple has joined #openstack08:36
RJD22is there someone who has experience with the deploy script?08:39
*** thimble has quit IRC08:39
*** miclorb has quit IRC08:41
*** cdbs has joined #openstack08:41
*** cdbs has left #openstack08:44
*** kpepple has quit IRC08:50
*** kpepple has joined #openstack08:51
*** kpepple has quit IRC08:55
*** kpepple has joined #openstack08:56
*** kpepple has quit IRC09:00
*** kpepple has joined #openstack09:01
*** reldan has joined #openstack09:07
*** kpepple has quit IRC09:09
*** kpepple has joined #openstack09:10
*** Nacx has joined #openstack09:10
*** DigitalFlux has joined #openstack09:11
*** Hayati has quit IRC09:15
*** kashyapc has quit IRC09:16
*** kpepple_ has joined #openstack09:17
*** kpepple has quit IRC09:17
*** thimble has joined #openstack09:19
*** ramkrsna has joined #openstack09:19
*** ramkrsna has joined #openstack09:19
*** kpepple_ has quit IRC09:21
*** kpepple has joined #openstack09:22
*** kashyapc has joined #openstack09:24
*** kpepple has quit IRC09:26
*** kpepple has joined #openstack09:27
*** kpepple has quit IRC09:28
*** kpepple has joined #openstack09:29
*** irahgel has joined #openstack09:29
*** jsgotangco has joined #openstack09:30
*** kpepple has quit IRC09:37
*** kpepple has joined #openstack09:38
*** miclorb_ has joined #openstack09:40
*** reldan has quit IRC09:41
*** kpepple has quit IRC09:42
*** kpepple has joined #openstack09:43
*** kashyapc has quit IRC09:45
*** befreax has joined #openstack09:46
*** befreax has quit IRC09:52
*** kpepple has quit IRC10:01
*** kpepple has joined #openstack10:01
*** cclien has quit IRC10:06
*** cclien has joined #openstack10:06
*** photron_ has joined #openstack10:09
uvirtbotNew bug: #715674 in nova "unable to access users table" [Undecided,New]
*** cclien has quit IRC10:12
*** cclien has joined #openstack10:17
*** kpepple has quit IRC10:19
*** kpepple has joined #openstack10:20
*** jsgotangco has quit IRC10:23
*** kpepple has quit IRC10:24
*** kpepple has joined #openstack10:25
*** clopez has joined #openstack10:27
*** Hayati has joined #openstack10:29
*** flashn is now known as flshn10:30
*** littleidea has quit IRC10:32
rccttx , soren: may you please put some insight on ? devin suggests a split in 2 bugs, one for the retrofitting the BadParameter exception and the current to solely deal with the paramater handling10:32
*** reldan has joined #openstack10:33
*** citral has quit IRC10:35
ttxI think his point is that if we introduce BadParamater we should make sure it's used wherever appropriate10:36
ttxso either your branch should fix that as well (use BadParameter everywhere appropriate) or another bug should be filed so that we remember to do it in the future10:37
ttxI think I prefer option 1 (do everything right at the first time)10:37
ttxrcc: ^10:37
*** kpepple has quit IRC10:37
rccok, so, what kind of use cases you mean? dealing bad parameters for other nova-manage commands i guess...10:38
ttxyes, I think so10:38
*** kpepple has joined #openstack10:38
rcchmm, i think i could ask in the bug for feedback regarding what people think we should check in nova-manage, use cases, etc10:39
ttxsounds good to me, maybe others will have different opinions though10:40
wedjathello everyone10:40
rcck, let's wait for soren opinion, i'll also ask vish10:40
wedjatfor swift, does someone know how to set the referer header http to be allowed by ACL suh as "" ?10:41
wedjati'm sending http requests with headers like "Referer :" but doesn't work10:42
*** miclorb_ has quit IRC10:45
wedjatoh ok, swift doesn't accept Referer Headers which don't start by 'http://*'10:47
*** kpepple has quit IRC10:53
*** kpepple has joined #openstack10:53
*** adjohn has quit IRC10:56
*** kpepple has quit IRC10:58
*** kpepple_ has joined #openstack10:58
*** kpepple_ has quit IRC11:03
*** kpepple has joined #openstack11:03
*** kpepple has quit IRC11:14
*** kpepple_ has joined #openstack11:15
*** reldan has quit IRC11:15
*** kpepple_ has quit IRC11:16
*** thimble has quit IRC11:16
*** kpepple has joined #openstack11:16
*** thimble has joined #openstack11:16
*** thimble has quit IRC11:23
*** stewart has quit IRC11:25
*** kpepple has quit IRC11:30
*** kpepple has joined #openstack11:31
*** drico has quit IRC11:41
*** zykes- has quit IRC11:42
cjreynI have a question about FlatDHCP. In eucalyptus NAT is done on the compute node. In openstack I'm seeing dns,asq running on the CC. Is this correct?11:43
*** zykes- has joined #openstack11:43
berendtcjreyn: nova-network has to be started on every nova node11:53
berendtcjreyn:   this graphic is better11:54
cjreynnova-network is running on both nodes, but dnsmasq only on CC11:54
*** CloudChris has quit IRC11:58
*** CloudChris has joined #openstack11:59
sorencjreyn: dnsmasq doesn't necessarily start right away.12:11
cjreynit never starts on the compute node12:14
*** metoikos has joined #openstack12:14
sorenIs nova-network running on it?12:14
sorenLook in nova-network's log file.12:15
cjreynno errors, and no mention of dnsmasq12:16
sorencjreyn: When you say never.... What do you mean?12:16
sorenI'm not suggesting that it may wait for a set time before it starts. It doesn't start until "it's needed".12:17
cjreynbefore or after starting instances, dnsmasq never runs on the compute node12:17
cjreynonly on the CC12:17
soren"When it's needed" may not be *completely* obvious, but it's event driven.12:17
berendtsandywalsh_: looks good.. thank you for the graphic12:17
cjreynok, but presumably its needed when an instance gets started, so it can supply an IP?12:17
sorencjreyn: Yes. At that time it will definitely be started.12:18
sorenperhaps sooner, but not later.12:18
sandywalsh_berendt, np. I have the OpenOffice originals if you need them. More here:
cjreynok, it never starts, even after an instance is launched. Hence my instance can never get an IP.12:19
*** Nacx has quit IRC12:19
cjreyndnsmasq starts on the CC, and listens on the br100 IP12:20
*** Nacx has joined #openstack12:20
*** irahgel has quit IRC12:20
*** reldan has joined #openstack12:21
sorencjreyn: I would love to help you debug it, but a) I'm pretty sure vishy can solve it in an instant once he turns up, so it would be a bit of a waste of time, and b) it would probably take so long that vishy would turn up in the mean time :)12:21
cjreynha ha, ok. I'll wait. But cheers anyways12:22
*** reldan has quit IRC12:22
sorenFor instance, I've just finished probably a good 1½ days of debugging. The result is an 11 character patch.12:22
*** fabiand_ has quit IRC12:23
*** westmaas1 has joined #openstack12:24
*** reldan has joined #openstack12:24
*** irahgel has joined #openstack12:25
*** reldan has quit IRC12:25
sorenttx: This is pretty cool:,distro=lucid/38/console12:26
sorenttx: I can pass a branch url to that job and it'll build a package out of it and install it on my test boxes.12:27
*** ctennis has quit IRC12:27
sorenttx: ...and then I can run the test battery against it.12:27
ttxpassword-protected, but I can imagine12:27
sorenOh, is it?12:27
ttxIt is12:30
* soren adjusts12:30
sorenor tries to. Hm..12:33
sorenttx: Try again.12:35
ttxsoren: looks good12:38
*** eikke has joined #openstack12:39
*** Hayati has quit IRC12:41
*** ctennis has joined #openstack12:41
*** reldan has joined #openstack12:45
*** gaveen has joined #openstack12:50
*** gaveen has joined #openstack12:50
*** reldan has quit IRC12:51
*** reldan has joined #openstack12:51
*** westmaas1 has quit IRC12:53
*** hadrian has joined #openstack12:54
*** reldan has quit IRC12:55
*** reldan has joined #openstack12:59
*** CloudChris has quit IRC13:10
*** CloudChris has joined #openstack13:21
*** stewart has joined #openstack13:23
berendtcan somebody please take a look at this graphic: i want to know if that's ok for a general overview of the components13:27
*** maplebed has joined #openstack13:28
*** DanEllis has joined #openstack13:30
ttxberendt: I think nova-manage speaks directly to the DB, not to the openstack API. To speak the openstack API, you need a client lib, like the modified cloudservers from SandyWalsh13:32
berendtttx: ok, i'll fix this13:32
*** Hayati has joined #openstack13:35
*** hadrian is now known as nothrvailble113:36
*** nothrvailble1 is now known as hadrian13:37
*** ramkrsna has quit IRC13:38
*** masumotok has quit IRC13:47
*** sebastianstadil has quit IRC13:48
*** thimble has joined #openstack13:49
*** dprince has joined #openstack13:50
*** vvuksan has joined #openstack13:58
*** vvuksan has quit IRC14:00
*** ugarit has joined #openstack14:02
*** clopez has quit IRC14:03
jaypipesberendt: nova-manage doesn't speak to the OpenStack API. Only to the EC2 API and to the underlying common database for certain things (like projects, quotas, users, etc...)14:06
jaypipesberendt: what *does* speak to the OpenStack API is the external python-novatools library located on github...and mirrored to Launchpad.14:07
jaypipesberendt: python-novatools is a slightly modified fork of the python-cloudservers code that speaks to the Rackspace Cloud Servers API 1.014:07
jaypipesberendt: also, the "Services" block should be renamed to something else, since Nova has a concept of a Service (nova.service.Service) internally. Services are worker processes that listen to a RabbitMQ for messages coming from other services. The internal Nova services include the Scheduler, the Network controller, the Volume controller, and the Compute worker.14:10
jaypipeshope that helps..14:10
daboberendt: there is another layer between the 'Services' box and the hosts - the scheduler. It's responsible for determining which host is used.14:10
dabojaypipes: stop typing faster than I can!14:10
jaypipesberendt: also, I'd rename "Cloud Controller" to Nova API node(s)14:11
jaypipesdabo: :)14:11
berendtthe scheduler is only waiting for incoming messages on the queue?14:14
DigitalFluxSo by using python-novatools, i can do the same action that i do with nova-manage ?14:16
DigitalFluxcreate users/projects/networks ?14:16
daboberendt: yes. Currently it picks a host at random; we're going to be adding logic to that process soon, though.14:16
berendtdabo: nova-scheduler needs access to the database?14:17
*** cdbs has joined #openstack14:17
daboberendt: It will have its own db that only has the information about the hosts (or soon, zones) for which it is responsible14:18
berendtbut the "stupid" nova nodes need access to the database? (sorry, i didn't read all code at the moment)14:19
daboberendt: what database do you mean?14:19
berendtthe mysql database used on the api host14:19
DigitalFluxCan someone please direct me to a doc/conversation to the differences between the EC2 model and Rackspace ?14:20
daboberendt: not to my understanding. What information would it need?14:20
DigitalFluxThe difference is only in the way of imaging the templates and the separation of compute/storage ?14:20
rcciirc there was a wiki page comparing openstack to ec214:21
rcclet me check14:21
berendtdabo: jaypipes:
daboberendt: I'm still unclear on the need for 'distributed storage'; what exactly is that used for?14:23
berendtdabo: the volumes in the virtual machines14:24
berendtdabo: nova-volume14:24
rccnot quite what you asking tho14:24
jaypipesberendt: cool. I would do 2 more things: in the block of Nova (Node)s, I would instead put the names of the nodes (network, compute, volume)14:24
jaypipesberendt: and also move MySQL into a more central place and put arrows from scheduler and the nova node group to MySQL..14:25
daboberendt: I haven't dealt with nova-volume, since we're not using it here (Rackspace), so I'll take your word on that14:25
berendtevery nova (node) is one set of network, compute, volume...14:25
berendttoo many details for the use of this graphic..14:26
berendtnext try... :)
DigitalFluxGuys, please answer me :)14:35
DigitalFluxif i use python-novatools against openstack bexar, i can get instances with persistant storage ?14:36
*** ppetraki has joined #openstack14:37
daboDigitalFlux: yes, that's the idea14:38
DigitalFluxdabo: OK Great, i will test it right away14:41
DigitalFluxAlthough i have to say that i didn't find any docs about that ..14:41
DigitalFluxdabo: Any docs around ?14:42
daboDigitalFlux: not sure where14:42
dabo..or if :)14:42
*** pvo has joined #openstack14:42
dabojaypipes: I'm getting glance/plugin errors creating an instance. Can you look at this xensource.log excerpt and let me know if it tells you anything?
jaypipesdabo: looking14:44
annegentleDigitalFlux: I don't know of docs on persistent storage either, it's a gap.14:44
jaypipesdabo: what are the errors you are getting?14:45
dabojaypipes: first, the instance isn't created :)14:45
daboon the compute side, it's just: Failure: ['XENAPI_PLUGIN_FAILURE', 'non-zero exit', '', '']14:46
dabowhich is why I'm looking at xensource.log14:46
*** f4m8 is now known as f4m8_14:47
*** gondoi has joined #openstack14:47
jaypipesdabo: well, the xensource.log doesn't give me anything more than the above... "Got exception XENAPI_PLUGIN_FAILURE: [ non-zero exit;..."14:47
dabojaypipes: ok, then, where can I dig deeper?14:47
daboit happens every time14:48
daboI'm using trunk for both nova and glance, btw14:48
jaypipesdabo: copy_kernel_vdi in the glance plugin is what is raising it (or returning non-zero). See if you can put some debugging statements in there...14:48
jaypipessirp_: got a few cycles to assist dabo?14:49
dabojaypipes: yeah, I saw the error at, but didn't know exactly what that was doing.14:49
jaypipesdabo: yeah, I don't know what all that stuff means either :(14:50
dabojaypipes: lotta good you are14:50
jaypipesdabo: yeah, I know, right? :)14:50
dabothey must keep you around for your looks. :)14:51
DigitalFluxannegentle: I just sent you a tweet about that, I think the only place that discussed the differences between EC2 and OpenStack/RackSpace was the mailing list, no docs mentioned persistant storage at all ..14:51
jaypipesdabo: take a look at /plugins/xenserver/etc/xapi.d/plugins/glance lines 46-7114:52
dabojaypipes: way ahead of ya14:52
jaypipesdabo: looks like all that is going on is some simple file movements. But there is no try/except block at all and no error handling. Not surprised there is no information in the logs ;)14:52
cjreynhey all. So I finally got an instance to boot, and can ping it. Its an ubuntu image, but ssh doesn't come up in the boot sequence till after it attempts to connect to which fails. Any ideas why this might be14:53
jaypipescjreyn: why is it trying to connect to that IP?14:54
jaypipescjreyn: during boot?14:54
cjreynfrom what I can gather, theres an init script called that attempts to connect to that IP for data passed to the instance on boot14:56
jaypipesberendt: good, but I'd still specify Network, Compute, and Volume in the Nova Nodes box... :)14:56
berendtjaypipes: i'll think about it.. :)14:56
jaypipesberendt: :)14:56
berendtjaypipes: dabo: thank you for the input.. i'll talk to my manager that i can distribute the graphic in the wiki under a free license14:57
cjreynjaypipes: does Openstack use the same method as Eucalyptus for passing data to an instance at boot itme?14:57
jaypipescjreyn: no.14:57
cjreynjaypipes: ok, so this may be an issue with the UEC image, as that's where I got it. How does Openstack pass paramters to instances at boot time?14:58
jaypipescjreyn: looks like you are trying to start up a Eucalyptus image on Nova. That won't work AFAIK.14:58
jaypipeszul: around? question about UEC above ^^14:59
cjreynjaypipes: links to a non tty image, i.e. one that accepts IP address injection on boot?14:59
zuljaypipes: hi reading14:59
jaypipescjreyn: I think it depends on the type of image being booted and the hypervisor used.15:00
berendtis it possible to use an other queuing system talking amqp, not only rabbitmq, activemq for example15:00
jaypipesvishy: could you elaborate on the above?15:00
zuljaypipes: i never tried using user-data on Openstack myself15:00
jaypipesberendt: Nova uses Carrot, so AFAIK, yes. :)
*** dfg has joined #openstack15:02
cjreynso is passing user data even possible with Openstack, or is it a feature to be added later?15:03
jaypipescjreyn: I'm not sure, to be honest. Was hoping vishy or someone more familiar with the virt layer would have an answer.15:04
annegentleDigitalFlux: ok, hi, thanks for mapping IRC>Twitter :) the wiki page shows EBS-like features are in both Eucalyptus and OpenStack. Is that what you mean?15:04
*** tleruitte has joined #openstack15:04
cjreynjaypipes: okey dokey, I'll wait around. Are there any weblinks to working ubuntu images?15:05
jaypipeszul: ? ^^15:05
ttxcjreyn: user-data works with Nova15:06
*** onxis has joined #openstack15:06
tleruittehi all! i have an issue using the script ( I can't ssh the VM run by the script. Anyone could help?15:06
cjreynttx: how?15:06
ttxcjreyn: same as EC2 or Eucalyptus, through a metadata service15:07
cjreynttx: the VM connects using http to this service?15:07
onxis...does anyone have an opinion on AoE (CORAID) when combined with Arista (Low Latency 10G) and Nexenta (ZFS)...15:07
ttxwhat you pass as user-data is exhibited by the metadata service for the instance to query15:07
ttxcjreyn: that's for the VM to implement. Ubuntu cloud images do query this for their cloud-init15:08
*** abecc has joined #openstack15:08
cjreynttx: how is this query done. On UEC its
ttxthe ubuntu UEC images work on Nova. They just query the metadata service at for user-data (among other things)15:09
ttxand Nova provides that (as EC2 and Eucalytpus do)15:09
cjreynttx: and Openstack provides this capacity, i.e. a webserver on ?15:09
cjreynttx: i tried pinging this address and it doesn't exist15:09
ttxcjreyn: it's a bit more complicated. That address is routed from the VM to the API server15:10
zulcjreyn: http://uec-images.ubuntu.com15:10
ttxcjreyn: should work if you query it FROM a vm15:10
ttx(the ubuntu cloud images don'ty boot if they can't access the metadata service)15:11
cjreynttx: yeah the one I got from has this problem. It times out, with Errno 113 No route to host15:12
annegentletleruitte: I think that the script downloads which may not have ssh installed... let me double-check that theory.15:12
cjreynttx: so if the VM can't access this service, where should I be looking to fix the problem?15:13
*** blueadept has joined #openstack15:13
DigitalFluxannegentle: kinda, a dialog on this issue on the documentation would be better though, with a tutorial on how to use the cloudservers command with Nova15:13
DigitalFluxannegentle: the discussion happening right now would be great to document too :-) (user-data)15:13
ttxcjreyn: what Network mode are you using ?15:13
cjreynttx: Flat15:14
ttxcjreyn: some network modes set up the redirection for you, some others don't15:14
annegentleDigitalFlux: yep, taking notes :)15:14
* ttx digs some doc15:14
ttxcjreyn: I'll have to defer to the networking specialists, but with Flat you might need to set up the redirection manually15:17
ttxI know it works automatically with VlanManager15:17
berendtjaypipes: carrot.. ok.. so i think the flag for setting the amqp server (rabbit_host) is a little bit confusing and should be renamed..?15:17
cjreynttx: ok. Who's the networks guru!?15:17
jaypipesberendt: sure, it is. feel free to file a bug on that. :) I think you're more than familiar with the bug-filing process by now :P15:18
ttxcjreyn: vishy.15:18
cjreynttx: he's on lunch ;-) ??15:19
ttxcjreyn: should be around in a couple hours15:19
cjreynttx: bonza. Cheers for your help15:20
ttxcjreyn: sorry I can't be of more help, using FlatManager hasn't been a success story for me either15:20
cjreynttx: I guess with a vlan tagging switch it all just works out of the box?15:21
ttxtechnically simpler, but more things you have to setup yourself15:21
berendtjaypipes: i think so.. ;)15:21
ttxcjreyn: that's the idea15:22
* ttx pauses15:22
tleruittehi all! i have a VM instances which is stuck in a 'networking' state (instead of scheduling or running). Does anyone know what can i do to get it running? Is it a bad network configuration?15:23
*** rnirmal has joined #openstack15:25
*** rnirmal has quit IRC15:25
annegentletleruitte: usually when that happens, it means nova doesn't know how to route the traffic to the bridge - which network setting is in your nova.conf?15:29
annegentletleruitte: more info on networking:
*** dprince has quit IRC15:30
tleruitteannegentle: i'm using novascript (, so everything should run on the same machine. I will take a look at your link, thanks for that!15:30
uvirtbotNew bug: #715831 in nova "rename gflag --rabbit_host" [Undecided,New]
*** dprince has joined #openstack15:32
*** rnirmal has joined #openstack15:34
*** msassak has joined #openstack15:35
*** dirakx has joined #openstack15:35
*** Hayati has quit IRC15:36
*** glenc has quit IRC15:39
*** jtimberman has joined #openstack15:41
*** dirakx has quit IRC15:44
*** glenc has joined #openstack15:44
DigitalFluxannegentle: those IRC logs are logged somewhere ?15:44
annegentleDigitalFlux: yup, see also you can use Google site: to search within those logs15:46
*** rnirmal has quit IRC15:49
*** rnirmal has joined #openstack15:50
*** reldan has quit IRC15:50
*** rnirmal has quit IRC15:50
*** som has joined #openstack15:51
*** som has left #openstack15:52
Ryan_Lanecjreyn: I can vouch for user-data using UEC images in nova15:53
Ryan_Lanecjreyn: I'm using it right now15:53
Ryan_LaneI'm using flatdhcp networking mode, though15:53
*** vvuksan has joined #openstack15:53
*** som_ has joined #openstack15:54
*** vvuksan has joined #openstack15:54
cjreynRyan_Lane, I had problems with FlatDHCP15:54
Ryan_Laneit's working well for me so far15:55
Ryan_Lanedo you have multiple NICs?15:55
som_is this OpenStack community??15:55
Ryan_Lanesom_: yep15:55
som_Greeat.. i landed in the right place..15:55
cjreynRyan_Lane, Yes, the compute node had two Nics. But dnsmasq never started on the compute node15:56
Ryan_Lanecjreyn: I have a controller node running nova-scheduler, nova-objectstore, and nova-api15:56
Ryan_Lanednsmasq runs on the network node15:56
Ryan_Laneoh, my controller is running nova-network too15:56
cjreynRyan_Lane, and nova-network on the controller?15:56
*** citral has joined #openstack15:56
Ryan_Lanelet me post my config15:56
cjreynRyan_Lane, that would be ace15:56
vvuksancjreyn: hey15:56
som_I mailed to Mr. Stephen Spector. I am new to this community...15:56
cjreynvvuksan, hey15:57
vvuksancjreyn: flatDHCP is tricky15:57
vvuksancjreyn: I learned myself :-)15:57
cjreynvvuksan, so I reverted to Flat mode. I have ping connectivity, but now problems with re-direction for UEC images15:57
Ryan_Laneit works better than flat, and is easier than vlan though :)15:57
*** dirakx has joined #openstack15:57
vvuksanwhat redirection15:58
vvuksanRyan_Lane: some of us don't have the VLAN option :-(15:58
Ryan_LaneI don't either15:58
Ryan_Laneso, I'm using flatdhcp15:58
vvuksanme too15:58
cjreynvvuksan, When the UEC images boot, they look for user data from a metadata service. For some reason the VM can't connect to this service15:58
vvuksancjreyn: that would indicate that 169.254 mapping is not setup correctly15:59
cjreynRyan_Lane, the advantages of DHCP are?15:59
vvuksanon the compute node15:59
vvuksanif you do iptables -t nat -L15:59
*** dragondm has joined #openstack15:59
vvuksandoes it show the 169.25415:59
cjreynvvuksan, yeah, thats it. How to solve?15:59
vvuksanand more importantly does it point to the API server15:59
Ryan_Laneit automatically sets up the metadata service ip for the instances, it handled the NAT for you15:59
Ryan_Laneerr handles15:59
som_I am new can anybody help mw out??16:00
Ryan_Laneit also handles assignment of floating IPs (via NAT)16:00
vvuksansom_: what's your question?16:00
Ryan_Lanesom_: ask a question, don't ask to ask16:00
*** reldan has joined #openstack16:00
cjreynRyan_Lane, ok, I'm sold, but as I said it never started dnsmasq on the compute node16:00
vvuksanRyan_Lane: about floating IPs16:00
Ryan_Lanednsmasq doesn't run on the compute node ;)16:00
Ryan_Laneit runs on the network node16:00
vvuksanRyan_Lane: can you have floating IPs assigned at euca-run-instance ?16:00
Ryan_Lanevvuksan: I don't believe so16:01
som_I want to contribute i am new I know java is there any ways I can contribute here??16:01
Ryan_Lanevvuksan: it's a two step process of running the node, then assigning the IP16:01
*** ugarit has quit IRC16:01
vvuksanRyan_Lane: right16:01
Ryan_Lanevvuksan: the nodes get a private IP address16:01
Ryan_Lanebut the private IP has access to the public network via NAT16:01
Ryan_Lanesom_: pretty much everything is in python16:01
Ryan_Lanesom_: I have a web interface project in php16:01
vvuksanRyan_Lane: link to web interface :-)16:02
som_So notting in Java??16:02
Ryan_Lanesom_: take a look at the available blueprints, and see if there's anything you want to do16:02
Ryan_LaneI don't think there's anything in java16:02
Ryan_Lanecjreyn: so, in my posted config....16:02
som_Thank u ..16:02
cjreynvvuksan, the iptables print doesn't list the redirection (its all for the libvirt interface)16:02
som_one more hing..16:02
vvuksanRyan_Lane: what no screenshots :-P ?16:03
*** maplebed has quit IRC16:03
*** dirakx has quit IRC16:03
Ryan_Lanecjreyn: is the controller, with network, api, mysql, objectstore, etc16:03
som_How can i build the apps using ur OpenStack??16:03
Ryan_Lanebasically everything but compute16:03
som_any resouses i can start with?/16:03
Ryan_Lanevvuksan: I'm going to make a screencast soon :)16:03
vvuksanRyan_Lane: awesome :-)16:03
Ryan_Lanevvuksan: I'm still working on it16:03
Ryan_Lanecjreyn: my_ip is set to the IP of each node...16:04
citralhi, I am trying to start my first instance, but it won´t come out of the `scheduling´ status, i.e. it won´t launch. Anything tips?16:04
som_Ryan: How can i develop app??16:04
Ryan_Lanesom_: program against the apis16:04
vvuksancitral: are you running the nova-api , nova-compute in the foreground ?16:04
Ryan_Lanesom_: nova provides the ec2 api, so you can program against ec2 if you want to16:04
som_Ryan: any resourses I can start with?16:05
*** reldan has quit IRC16:05
vvuksanRyan_Lane: the trouble then that I see with floating IPs is this. You tell nova to assign certain IPs to certain hosts16:05
Ryan_Lanethe wiki16:05
vvuksanRyan_Lane: right ? So when you do allocate-address you get some random IP16:05
Ryan_Lanevvuksan: actually, you assign them to the network host ;)16:05
Ryan_Laneyeah, you get a random IP16:05
Ryan_Lanesom_: google?16:05
citralvvuksan: yes16:06
som_I wanna develop apps on ur platform..16:06
cjreynRyan_Lane, yeah got that. What I'm really after now is the redirection for i.e. the metadat service16:06
vvuksancjreyn: nova-compute should do that16:06
vvuksancjreyn: setup redirect for 169.25416:06
som_Ryan: can I do that being a java developer (develop a app)??16:06
Ryan_Lanesom_: you can develop in whatever you want16:06
cjreynvvuksan, how?16:06
Ryan_Lanesom_: I'm using php, for instance16:06
Ryan_Lanesom_: you are developing against an API16:07
vvuksancjreyn: it should do it at instance startup16:07
cjreynvvuksan, remember I've reverted to Flat, i.e. no DHCP16:07
vvuksancjreyn: let me get this straight. a VM is kicked off on the instance but when you look at iptables -t nat -L16:07
Ryan_Lanecjreyn: you can ping the instance, when it is created, right?16:07
som_Ryan can i have ur mail ID??16:07
vvuksancjreyn: 169.254 is missing ?16:07
som_to be in touch..16:07
Ryan_Lanesom_: it's online. just search for it. but there isn't really much reason to contact me directly16:07
cjreynyeah, using Flat (no DHCP!!), I can ping, but its a UEC image, and fails when trying to contact the metadat service16:08
Ryan_Laneyeah, just seeing if the instance has network :)16:08
vvuksancjreyn: I understand. Key is that iptables needs to have that entry16:08
som_ll b available here in this channel??16:08
*** perra has quit IRC16:08
vvuksancjreyn: otherwise UEC images will fail16:08
Ryan_Lanesom_: eh? what are you talking about?16:08
*** jfluhmann has quit IRC16:09
Ryan_Laneeh? my iptables doesn't have that entry on the compute node16:09
som_Ryan.. wat i nmeant was ll u here.. if i face any prob that i a can reach u in the channel..16:09
cjreynvvuksan, ok, and with FlatDHCP this is done automatically? I tried pinging the IP of the metada service but it doesn't exist16:09
vvuksanRyan_Lane: that doesn't seem right16:10
vvuksanRyan_Lane: if you are gonna use metadata service you have to have that entry16:10
*** RobertLaptop has joined #openstack16:10
*** tleruitte has left #openstack16:10
*** MarkAtwood has joined #openstack16:10
Ryan_Lanevvuksan: that's on my network node16:10
*** som_ has quit IRC16:11
Ryan_Lanemy compute nodes only have rules for the instances. no metadata service16:11
Ryan_Lanethe network node has this: DNAT       tcp  --       tcp dpt:80 to:
vvuksangot that16:11
cjreynRyan_Lane, what is --routing_source_ip?16:11
Ryan_Lanecjreyn: the public IP on the network node16:12
Ryan_Laneit's the rule used for SNAT16:12
Ryan_LaneSNAT       all  --            to:
Ryan_Laneso that the nodes have access to the outside world16:12
Ryan_Laneso in dhcp mode, when the instances get an IP via dhcp, they also get a gateway set16:13
Ryan_Lanetraffic routes through the network node16:13
Ryan_LaneI gotta go16:14
Ryan_Laneback in like an hour :)16:14
*** dirakx has joined #openstack16:14
cjreynRyan_Lane, ok, im with you thanks16:14
*** Ryan_Lane is now known as Ryan_Lane|away16:14
vvuksancjreyn: back to your problem16:14
vvuksancan you confirm whether you see the 169.254 rule on the compute server ?16:14
*** Hayati has joined #openstack16:14
*** thimble has quit IRC16:15
*** h1nch has quit IRC16:15
cjreynvvuksan, aye. I guess its some ip tables magic to add that rule. No its not there right now. So 169.254 gets mapped to a differnt IP16:15
vvuksancan you post your nova.conf16:15
vvuksanlet's look at that16:15
*** DigitalFlux has quit IRC16:15
vvuksanah yeah16:16
vvuksani think i remember now16:16
vvuksanadd this16:16
cjreynvvuksan, so my controller has a public (eth0) IP of, the backend network is
cjreynvvuksan, on both controller and compute node?16:17
vvuksanrestart everything16:17
vvuksanyeah i'd put it everywhere16:17
cjreynvvuksan, this will update iptables on instance creation?16:19
vvuksani believe so16:19
vvuksankick off an instance16:19
vvuksanand watch iptables16:20
*** dirakx has quit IRC16:20
*** sateesh has quit IRC16:20
cjreynvvuksan, nope16:21
cjreynvvuksan, so does 169.254 get mapped to an IP that Openstack hosts the metadata service on?16:23
*** anticw has quit IRC16:23
vvuksancjreyn: yes it gets mapped to the API host16:23
*** anticw has joined #openstack16:23
cjreynvvuksan, I see16:23
vvuksanthat's the code that does it16:24
vvuksanswitch to flatdhcp16:25
vvuksansee if that does it16:25
vvuksanlooking through the code that may be the thing16:25
vvuksani gotta go be back later16:25
*** RJD22 is now known as RJD22|away16:26
*** vvuksan has quit IRC16:30
*** baldben has quit IRC16:30
*** reldan has joined #openstack16:30
*** littleidea has joined #openstack16:31
*** dirakx has joined #openstack16:33
*** Daviey has quit IRC16:36
*** Daviey has joined #openstack16:39
*** westmaas has quit IRC16:42
*** citral has quit IRC16:43
*** calavera has quit IRC16:43
*** dfg_ has joined #openstack16:49
*** lool- has joined #openstack16:51
*** dfg_ has quit IRC16:53
*** reldan has quit IRC16:57
*** dirakx has quit IRC16:57
*** grapex has joined #openstack16:58
*** reldan has joined #openstack16:58
*** kashyapc has joined #openstack17:01
*** CloudChris has quit IRC17:01
*** pvo has quit IRC17:02
*** lool- has quit IRC17:02
*** littleidea has quit IRC17:05
*** jero has quit IRC17:07
*** MarkAtwood has quit IRC17:08
*** DanEllis has quit IRC17:11
*** DanEllis has joined #openstack17:12
*** dirakx has joined #openstack17:12
*** dirakx has joined #openstack17:13
*** h1nch has joined #openstack17:13
*** h1nch has left #openstack17:13
*** evtoews has joined #openstack17:14
*** kashyap_ has joined #openstack17:15
*** maplebed has joined #openstack17:15
*** onxis has quit IRC17:16
*** justinsb has joined #openstack17:16
justinsbHello again all!  Can someone remind me of the rules for when proposed merges get merged?  Trying to figure out why this one hasn't auto-merged:
*** littleidea has joined #openstack17:18
edayjustinsb: reviewers should be marking them as 'resubmit', the proposer would actually perform a resubmit using another button. :)  But besides that, looks like just waiting for Devin to approve17:19
edayjustinsb: once there are 2 core approves, it will be marked for merge17:20
*** baldben has joined #openstack17:21
edaysandywalsh_: hey! just wrote you a lengthy email, lemme know if you want to chat (voice) at all17:21
*** littleidea has quit IRC17:21
sandywalsh_eday, thanks ... will give it a read shortly!17:21
justinsbeday: Thanks.  But oops ... so my 'resubmit' will block it?17:21
devcamcarjustinb: just marked lp:~justin-fathomdb/nova/san as approved17:22
edayjustinsb: nope17:22
edayjustinsb: just for the future17:22
justinsbeday: Thanks!17:22
justinsbdevcamcar: Thanks for the approve!17:22
sandywalsh_eday, re: dynamically notifying the schedulers: good suggestion. I'm on a call now, let's skype after. How long you around for?17:24
*** MarkAtwood has joined #openstack17:25
edaysandywalsh_: all day!17:27
sandywalsh_eday, great ... ping you shortly17:27
*** h1nch has joined #openstack17:28
devcamcaroops, looks like hudson is broked: /bin/sh: /var/lib/hudson/ not found17:28
justinsbdevcamcar: I presume that's Hudson, not my patch?17:29
edaymtaylor: ^ ?17:29
*** prewarmoney has joined #openstack17:29
mtayloreday: oops. can fix. one sec17:29
edaydevcamcar: hudson is no more, meet jenkins :)17:29
devcamcarjustinb: yes, not your patch17:30
edaydevcamcar: he was stealing too much silverware, so we had to fire him17:30
devcamcareday: yea, damn oracle17:30
prewarmoneyhey guys, I just upgraded our cluster to the latest maverick ppa release(655) and am getting errors contacting the metadata server from the instances again...17:31
cdbsIs there some problem with hudson? It marked my merge back to needs-review, and reported /bin/sh: /var/lib/hudson/ not found17:31
*** Ryan_Lane|away is now known as Ryan_Lane17:31
vishylaunchpad people: if a patch gets merged that fixes the same line as your patch, what is best status.  Do you mark the patch rejected?17:32
prewarmoneyI know I need to add a SNAT chain in my head-node's iptables, but can't remember the specifics17:32
devcamcarcdbs: hudson is broked at the moment somehow, mtaylor is looking at it17:32
*** grapex has left #openstack17:32
mtaylordevcamcar, eday: should be fixed  now tray again17:32
*** sebastianstadil has joined #openstack17:32
devcamcarmtaylor: thanks17:32
mtaylorvishy: I'd say "merged"17:32
devcamcarcdbs: can you link the merge that failed?17:32
mtaylordevcamcar: of course, let me know if I'm wrong17:32
cdbsdevcamcar: just a sec17:32
vishymtaylor: even though that patch didn't actually get merged?17:33
devcamcarmtaylor: just re-approved justinb's branch17:33
devcamcarcdbs: re-approved17:33
mtaylorvishy: well, the change is a good change though ... rejected seems weird17:33
mtaylorvishy: I'm honestly not sure - I'm not sure there's a _good_ thing there17:34
vishyok I'll just mark it merged with a note17:34
*** maplebed has quit IRC17:34
cdbsdevcamcar: thanks17:35
*** rlucio has joined #openstack17:41
*** dirakx has quit IRC17:42
vishyjustinsb: looks like you need to add your other email to mailmap17:43
justinsbvishy: Thanks - it's not another email address - I was just experimenting with using github & termie cool bzr bridge, but then I had to create a new project so I could fork from it, and it got into a big mess17:44
*** joearnold has joined #openstack17:44
justinsbvishy: Should I just resubmit a trivial patch with the right email address, or should I add it to mailmap?17:45
vishyadd it to mailmap and push17:45
justinsbvishy: OK - thanks!17:45
vishyotherwise you have to revert the commit with that author17:45
vishyand it is kind of a pain17:45
*** hadrian has quit IRC17:46
*** evtoews has quit IRC17:48
berendtAn IPv4 address space walks into a bar: "A strong CIDR please. I'm exhausted." ... :)17:48
berendt(found on
prewarmoneyDo I have to add any nat rules on the compute hypervisors to get the instances talking with the cc_host metadata port?17:50
*** hadrian has joined #openstack17:51
*** photron_ has quit IRC17:52
*** reldan has quit IRC17:53
*** evtoews has joined #openstack17:53
*** dirakx has joined #openstack17:56
*** RJD22|away is now known as RJD2218:02
justinsbvishy: Added my mailmap alias.  Can I get a reapprove please when you have a minute?
*** cynb has joined #openstack18:03
*** cynb has left #openstack18:05
evtoewshi all, in there is a line that reads "There is however one catch; since the pip-approach installs a wrong version of boto, you need to do..."  What is the right version of boto?18:10
*** prewarmoney has quit IRC18:14
*** alekibango has joined #openstack18:18
devcamcarevtoews: 1.9b118:18
annegentleprewarmoney: yes - I think so. check out the last part of
devcamcarevtoews: sorry typo'd, its actually 1.9b18:19
annegentleevtoews: 1.9b18:19
annegentlewah slowest typist :)18:19
annegentleevtoews: from
devcamcarjustinb: re-approved your san branch18:20
uvirtbotNew bug: #715936 in nova "ValueError when limit/offset = non-number in OpenStack API" [Undecided,New]
*** Daviey has quit IRC18:21
evtoewsdevcamcar annegentle: thx!18:21
*** vvuksan has joined #openstack18:21
*** rabbityard has joined #openstack18:22
*** iammartian has joined #openstack18:23
*** fabiand_ has joined #openstack18:25
*** pvo has joined #openstack18:27
*** blakeyeager has joined #openstack18:29
vishyjustinsb: you got caught by the pep8 police18:29
justinsbvishy: Doh!  I even ran pylint - must have missed one :-(18:29
vishylooks like there are a few double spacing errors18:30
justinsbWhat tool is it that is enforcing pep8?  I'd obviously like to run the same tool :-)18:31
eikkewhat'd be the preferred setup to run Swift using a shared FS?18:31
rabbityardjustinsb: just ./ and it does it at the end of the test run18:31
eikkeas in -> have multiple proxy servers, and storage on a shared FS, so using multiple object servers as well I guess18:32
justinsbrabbityard: OK, thanks!18:32
rabbityardjustinsb: i think there is a bzr-commit hook that will do it for you also18:32
justinsbrabbityard: Thanks!  The script seems to run the pep8 program, so I'm installing that and seeing if I can hook it into my editor18:35
vvuksanvishy: you around ?18:37
vishyvvuksan: I'm lurking18:37
vishyjustsb pep818:37
vishyjustinsb: ^^18:37
vvuksanvishy: i have put some more thought into my flatdhcp setup18:38
vishy0.6.118:38 does a pep8 test18:38
vvuksanam I correct to assume that it's impossible to associate specific IPs to specific machines18:38
vvuksanbasically in my setup I have host1 and host2 running flatdhcp18:39
vvuksandifferent DHCP ranges18:39
vvuksanif I say euca-allocate-address I will get a random address anywhere18:39
vishyfor floating ips?18:39
vishyyes that is true18:39
vishyit randomly selects a host18:39
vishythere is no real ability for the code to know where the ip will be used when it is allocated18:40
vvuksanthat's too bad :-(18:42
vishywhat use case are you trying to solve?18:42
vvuksanmy use case is that netops handles the network stuff18:43
vvuksanthis is all internal stuff18:43
vvuksani have couple hosts18:43
vvuksani can get ranges of internal 10.x IPs to use18:43
vvuksanbut I'm not allowed to run DHCP since that would interfere with DHCP that netops runs18:43
vvuksanwhat I was going with is18:44
*** CloudChris has joined #openstack18:44
vishy(it doesn't actually need to conflict, dnsmasq will ignore requests that aren't in its conf file)18:44
vvuksanset up br100 with private IPs ie. for host1 for host218:44
*** bcherian has joined #openstack18:44
vvuksanvishy: that is true18:44
vvuksani'm just worried :-)18:45
vishyvvuksan: so your problem is that you need 10.x range to be shared between both hosts18:45
vishyso you really need allocate to just allocate an ip from the pool even if a host isn't assigned18:45
vishyand when you associate the address it assigns the network host18:45
vvuksanbasically at run-instance18:46
vishyi don't think there is any problem with that happening18:46
*** MarkAtwood has left #openstack18:46
vishyyou would need to change allocate to pull from the entire pool of floating_ips18:46
vvuksanwhat I would need allocate to do is allow me to specify destination host18:47
vishyand to set the host in the floating_ip record when you do associate18:47
*** baldben has quit IRC18:47
vishyhmm i suppose it is a little complicated18:47
vvuksanit's tricky18:47
vishybecause associate_address would have to go through compute18:47
*** RobertLaptop has left #openstack18:47
vvuksanyou'd have to supply instance name to allocate18:48
vishyvvuksan: you could modify euca2ools to pass in a host paramater18:48
vvuksanbut that may seem as too much of hack18:48
vvuksanif it's only for me18:48
vishyvvuksan: the problem is that amazon doesn't support this18:48
vishyvvuksan: so we could add it to another api18:49
vvuksanI think it makes sense as a good feature18:49
vishyreally though i think the "right solution" is to make allocate address go through compute18:49
*** pvo has quit IRC18:49
vishyand wait to assign the network host in the table until associate18:50
vvuksanalthough I suspect/realize there may be security implications if you e.g. have set pool of floating IPs assigned to the same host18:50
vvuksanso it's tricky18:50
vvuksanthis may not matter to private clouds but to public it may not be as good18:51
*** nelson has quit IRC18:51
*** nelson has joined #openstack18:51
vvuksanvishy: one thing I realize doing it if I bridge eth0 with br100 and run DHCP. There may be a race condition when the instance requests the IP who answers the request dnsmasq or Netops DHCP18:52
vvuksanlikely dnsmasq but you never know18:52
*** irahgel has left #openstack18:53
*** CloudChris has quit IRC18:54
*** reldan has joined #openstack18:56
*** blakeyeager has quit IRC19:01
*** evtoews has quit IRC19:02
*** joearnold has quit IRC19:03
*** fabiand_ has quit IRC19:04
*** fabiand_ has joined #openstack19:04
jarrodwhen installing ubuntu on a compute node where the drives will also be used for volumes, should I use LVM?19:05
vishyvvuksan: yeah it may not work in most deployments19:06
vishyvvuksan: I don't know if a provider will have a pool of ips that can truly float through their entire system.19:06
vishyjarrod: probably19:07
jarrodi was thinking volume managing required lvm for logical volume sizing19:07
vvuksanvishy: i doubt it either unless they have one or couple huge boxes that do only that19:07
vishyi suppose we could separate nova-network (fixed) from nova-network (floating)19:08
vvuksanso what is fixed and how does it work ?19:09
vishyand just make sure the nova-network (floating) host has a route to all of the compute hosts19:09
vishyfixed is the stuff that controls dnsmasq and such19:09
vishyvvuksan: in your system, what if you just give one box floating ips19:10
vishyand make sure it has a route to the other box19:10
vvuksanit does have a route19:10
vvuksanthey are on same network segment19:10
vvuksanunless I created my own "network"19:10
vvuksanon top of existing network19:10
vvuksanwhich I suppose is possible19:11
vvuksanbut i still don't see what that would buy you19:11
vvuksansince VMs are still behind the "private" bridge19:11
vvuksanhow do e.g. eucalyptus or guys solve this ?19:12
vvuksando they ?19:12
*** Daviey has joined #openstack19:13
*** evtoews has joined #openstack19:14
*** joearnold has joined #openstack19:15
openstackhudsonProject swift build #192: SUCCESS in 27 sec:
openstackhudsonTarmac: Use sqlite's new WAL mechanism as a replacement for .pending files.19:16
*** reldan has quit IRC19:16
*** Daviey has quit IRC19:17
*** dirakx has quit IRC19:17
*** Daviey has joined #openstack19:18
*** fabiand_ has quit IRC19:18
*** Hayati has quit IRC19:19
*** burris has quit IRC19:23
*** kbringard has joined #openstack19:23
justinsbvishy: Thanks for the near-instant approve on my PEP8 fixes!19:23
* vishy blows the smoke off the muzzle of his six-shooter19:23
*** reldan has joined #openstack19:26
justinsbvishy: I don't know how you do it.... you're running faster than Launchpad!19:30
*** dirakx has joined #openstack19:33
*** dirakx has quit IRC19:38
*** mdomsch has joined #openstack19:39
openstackhudsonProject nova build #495: FAILURE in 1 min 31 sec:
openstackhudsonTarmac: Added support for 'SAN' style volumes.  A SAN's big difference is that the iSCSI target won't normally run on the same host as the volume service.19:42
vishyyou keep ninja-patching while i approve19:42
vishywh00t: patch is in19:45
justinsbwh00t indeed - finally merged!   Thanks vishy19:50
justinsbBut a build failure - is that normal?19:50
*** maplebed has joined #openstack19:52
*** dirakx has joined #openstack19:52
sirp_jaypipes: if you get a sec, could you take a look at
justinsbLooks like Jenkins failed with a rsync error, hopefully not the patch19:54
vishyjustinsb: yes looks like an unrelated error19:54
jaypipessirp_: approved, sorry, I thought I'd done that earlier, but it was devcamcar :)19:55
sirp_jaypipes: excellent, thanks19:55
*** berendt has quit IRC19:56
*** brd_from_italy has joined #openstack19:56
*** dirakx has quit IRC19:56
*** joearnold has quit IRC19:57
*** pvo has joined #openstack19:57
*** sebastianstadil has quit IRC19:57
*** rabbityard has quit IRC19:58
*** pvo has quit IRC20:02
*** maplebed has quit IRC20:03
*** mdomsch has quit IRC20:04
*** dirakx has joined #openstack20:11
RJD22is there someone who used the deploy script?20:11
RJD22I need some help with it20:11
*** Nacx has quit IRC20:13
*** eikke has quit IRC20:16
annegentleRJD22: which script are you using?
*** gaveen has quit IRC20:26
*** hggdh has quit IRC20:26
*** hggdh_ has joined #openstack20:26
*** kpepple has joined #openstack20:27
* soren loves graphs20:28
*** eikke has joined #openstack20:28
soren is my current favourite.20:28
kbringardvvuksan: did you ever sort out that truncate problem on CentOS 5?20:30
sorentermie, jaypipes: Ok, so you guys are doing some sort of automated integration testing work too, right?20:35
sorenjaypipes: I saw a Hudson^WJenkins job that suggested you had something, at least.20:35
jaypipessoren: haven't had much time to work on it, maybe only a few hours.  the work I have is up on lp:openstack-ci, but I haven't done the integration work with hudson yet.20:37
*** reldan has quit IRC20:37
*** dirakx has quit IRC20:37
sorenjaypipes: Ok.20:37
openstackhudsonProject nova build #496: STILL FAILING in 1 min 30 sec:
openstackhudsonTarmac: Pass timestamps to the db layer in fixed_ip_disassociate_all_by_timeout rather than converting to strings ahead of time, otherwise comparison between timestamps would often fail.20:37
jaypipessoren: the drizzle-automation test suite I wrote for Drizzle that has been running near constantly for the past 3 years was the basis for a lot of ideas, but been focused on Glance for the past 2 months.20:37
sorenjaypipes: What I have so far is quite basic.20:37
jaypipessoren: I think termie and others have some smoketests running against a cluster, but I've yet to see any details, code, and/or the test machines on hudson.20:38
sorenjaypipes: It creates a keypair, fires up an instance, opens the firewall for ssh, ssh's in, checks that user-data got passed through correctly, and logs back out, closes the firewall, and deletes the keypair.20:38
vvuksankbringard: yeah I compiled coreutils by hand :-/20:39
sorenjaypipes: It runs on real hardware, a box runnning each of Lucid, Maverick, and Natty.20:39
vvuksankbringard: then copied the binary to /usr/bin20:39
sorenjaypipes: It runs the tests against either what's in the trunk ppa, what's in the release ppa, or a random branch I point it at.20:39
*** gaveen has joined #openstack20:40
sorenjaypipes: That's about as far as I've gotten so far.20:40
sorenjaypipes: It's only with this patch landed that it's stable enough that I really want to start working out from that.20:40
vishyany core members around that want to do a quick review of:
jaypipessoren: My experience with hudson is that hudson is a great tool for handling the upper-layer of job queuing, mostly good for pushing jobs to servers for testing installation and configuration, but it falls over when you need to do anything more complex than that (for instance, fire X clients against lp:nova, setup a glance server from a tarball, run tests, collect performance numbers, and then store that in a database for20:41
jaypipes historical review like shown here:
termiejaypipes: i have the opposit experience20:41
* soren is on the hudson failures..20:41
jaypipestermie: course, I'm happy to have others prove me wrong.20:41
jaypipestermie: I just have 3 years experience of working with drizzle testing and regression automation and it's been my experience.20:42
*** eikke has quit IRC20:42
jaypipestermie: and when you email me saying "your code is a waste of time because it does some things that Hudson does", that doesn't exactly get me in the right mood ;)20:43
*** gaveen has quit IRC20:44
sorenvishy: approved20:44
termiejaypipes: i obviously didn't use those words20:44
termiejaypipes: so i would place emphasis on that interpretation being your own20:45
*** openstackhudson has quit IRC20:45
termiejaypipes: but nonetheless the email was asking why you are reinventing a lot of stuff20:45
termiejaypipes: you have some answers to that as is shown by your paragraph above20:46
*** openstackhudson has joined #openstack20:46
termiejaypipes: though at this point i still don't agree20:46
jaypipestermie: I spoke to all the anso people except for you and vishy about openstack-ci and even demoed the drizzle-automation stuff at the last summit. Saying to me "nobody on this end has even been aware that OpenStack-CI has even existed until the email the other day" is disingenuous.20:47
jaypipestermie: all I'm saying is that you could have stated your concerns in a more productive way.20:49
termiejaypipes: i can only speak from my experience, there was knowledge that you were working on testing code and that was it20:49
termiejaypipes: i am sorry that the email was not somehow couched in better language20:50
termiejaypipes: it was short and to the point20:50
termiejaypipes: and i was expecting it to be the beginning of a conversation20:50
*** dirakx has joined #openstack20:51
vishyjaypipes, soren, termie: regardless of what has been done, we need to unite our dev efforts on testing.20:51
*** hub_cap has joined #openstack20:51
jaypipestermie: fine, water under the bridge. for the future, a short way of starting the conversation would have been "hey, this is what we've worked on, and these are our plans and designs. what are yours?"20:51
termiejaypipes: that is almost exactly what i did20:51
* soren takes notes20:51
*** grapex has joined #openstack20:51
termiejaypipes: no offense was meant, it is just code20:52
RJD22annegentle: I'm using deply.py20:52
openstackhudsonYippie, build fixed!20:52
openstackhudsonProject nova build #497: FIXED in 1 min 35 sec:
openstackhudsonTarmac: Fixes the ordering of init_host commands so that iptables chains are created before they are used.20:52
openstackhudsonThis fixes an issue that arises if nova-network is started with flushed iptables after floating ips have been assigned.20:52*20:52
*** sebastianstadil has joined #openstack20:53
jaypipestermie: ok, let's move on. I'm over it. I've said my piece and I appreciate your response.20:53
jaypipestermie: so, where can I see the code that anso uses for automated testing? is it available?20:53
RJD22I'm having some troubles with puppetmaster and puppet since I never installed it and can't find any good tutorials on howto configure it correctly20:53
*** sebastianstadil_ has joined #openstack20:54
*** Ryan_Lane has quit IRC20:54
*** Ryan_Lane_ has joined #openstack20:54
annegentleRJD22: I don20:55
*** CloudChris has joined #openstack20:55
* annegentle starts over20:55
annegentleRJD22: I don't have direct experience with it, but this page is the existing doc for it
RJD22annegentle: I know20:55
RJD22but I'm having troubles with puppetmaster20:55
RJD22annegentle: do you have experience on how to configure puppetmaster?20:56
*** abecc has quit IRC20:56
RJD22or are you using a different guide to install nova?20:56
vvuksanRJD22: what kind of problems are you having ?20:56
termiejaypipes: sorry, in person conversation going on20:56
vvuksanRJD22: you can always ask on #puppet for help20:56
jaypipestermie: no worries, I understand.20:56
termiejaypipes: we don't have any code specific to this effort20:56
jaypipestermie: k20:57
*** sebastianstadil has quit IRC20:57
*** sebastianstadil_ is now known as sebastianstadil20:57
jaypipestermie: was the plan to set up host machines to fire, basically, smoketests/*.py against?20:57
annegentleRJD22: sorry, no puppeteering for me20:57
termiejaypipes: we have been trying ot connect with you before we started to see what there already was and get access to hudson20:57
annegentleRJD22: the deploy tool is for hundreds (thousands) of nodes, do you just want to try out Nova?20:58
jaypipestermie: yes, understand that. sorry, been busy with Glance stuff. not intentionally ignoring the efforts.20:58
termiejaypipes: yes the plan is to pretty much to run smoketests against a variety of deployment strategies20:58
jaypipestermie: another issue we're going to have is that I really don't feel like entering into Hudson manually the 200+ host machines that comprise huddle 48.  Was the plan to do that?20:58
annegentleRJD22: Shigetoshi (the woman who created the deploy tool) is a grad student in Japan. She'd be happy to help but it's the middle of the night there :)20:58
sorenjaypipes: Because you don't want to do the typing or because you think it's overkill?20:59
jaypipessoren: heh, bit of both :)20:59
RJD22annegentle: No I want to install multiple nodes and the deploy tool seems to help in that way20:59
sorenjaypipes: I'm sure we could work out the former. :)20:59
sorenjaypipes: We could outsource it to Amazon's Mechanical Turk.21:00
termiejaypipes: that isn't the goal, but there are a variety of IPs that will need to be typed in, certainly not 200 at this point21:00
sorenjaypipes: Or write a script. Not sure.21:00
*** Ryan_Lane_ is now known as Ryan_Lane21:00
termiejaypipes: PMd you, btw21:00
jaypipestermie: while I think that executing smoketests against a small subset of hosts is perfectly good for testing some simplistic usage scenarios in a multi-node install, I was thinking that we should have a way of firing a workload against a production huddle that represents a real, production set of requests. Openstack-CI was supposed to be that effort. I've got raw data now from RS production sysadmins for API node log fil21:00
jaypipeses that I'm stripping into a set of API requests that compose a "workload" that will be fired by the openstack-ci project against th ehuddle 48 machines.21:00
annegentleRJD22: okay, another script that may help that doesn't require puppet knowledge is at
RJD22annegentle: ok cool21:01
annegentleyou can install a cloud controller with a script and then multiple compute nodes with a 2nd script (compute node doesn't run the database, just runs the VMs)21:01
RJD22annegentle: is there any good doc on how to do that ?21:02
annegentleRJD22: I'm updating doc now in fact for the 2 scripts, but check out for the quick lowdown21:02
jaypipestermie: also, as far as "getting access to hudson", I'm not the admin there... I think soren or mtaylor is?21:03
RJD22annegentle: nice I'm always watching the wiki's recent changes :P21:03
annegentleRJD22: I just ran it a few minutes ago and trunk isn't working right now (I got a creds error)21:03
termiejaypipes: i think that sounds like a great test as well21:04
sorenmtaylor: You changed the e-mail in the ppa script? :(21:04
termiejaypipes: though, again, i would be expecting to do that with hudson21:04
annegentleRJD22: just so's ya know. Oh wait if you run bash instead of sudo ./ it should work. Updated the wiki page :)21:04
termiejaypipes: much of the code is the same regardless21:04
*** pothos_ has joined #openstack21:04
*** decramy has joined #openstack21:05
* RJD22 reruns ubuntu install21:05
sorenjaypipes: I can manage hudson, yes.21:06
sorenjaypipes: Err... Jenkins, I mean.21:06
*** pothos has quit IRC21:06
soren(that's going to take some getting used to)21:06
termiesoren, jaypipes: most of the hudson access was just to get at any info in there, it isn't necessary right this moment to do anything to it21:06
*** pothos_ is now known as pothos21:06
jaypipestermie: for large test clusters, how would you have Hudson perform a "wipe"? We need to be able to reset the bare metal machines back to "pure" state that represents the production envirnment before any guests get installed. I can do that via Hudson by issuing a series of requests from Hudson out to all the test cluster hosts, but it would be much faster to have that configuration information in a file and have the host r21:07
jaypipesunning openstack-ci fire off scripts that "wipe" the hosts in the test cluster.21:07
RJD22annegentle: I'm acctually setting up a cloud for scalr now since they need a platform to test :)21:07
jaypipestermie: I have a simple script from antonym that could do this, but it may be more efficient to have the host running the openstack-ci program issue a request to RS' internal Autohost client to do that...21:08
sorenjaypipes: I do that with Jenkins. I find it pretty simple, but perhaps you're trying to do something trickier than I.21:09
sorenjaypipes: I have a job that I configure to run on every box that does the cleanup/reinstall.21:09
*** colinnich_ has joined #openstack21:09
termiejaypipes, soren: sounds about like what i had in mind21:09
sorenjaypipes: It's pretty simple for me, since my slaves are configured to only run one job at a time.21:09
jaypipestermie: I don't disagree with you that some of the pieces may look like they duplicate Hudson functions.21:09
sorenjaypipes: I guess it's trickier if they are allowed to run multiple jobs at once.21:10
jaypipestermie: and I'm focusing my efforts right now in openstack-ci on the clients that execute the test workloads against the test cluster. The code that's currently in the trunk branch, a lot of it is base code that came over from drizzle-automation and some other projects...21:11
sorenmtaylor: Would you mind terribly if I changed it back? It's the only way I can think of that I can get notified about build/upload failures.21:11
annegentleRJD22: ah cool.21:11
*** dirakx has quit IRC21:12
*** eikke has joined #openstack21:12
termiejaypipes: that sounds like a decent focus, though i don't really like that it is a separate project21:12
edayjaypipes, termie, soren: Sounds like the biggest Q is should jenkins be a dependency for openstack-ci? Or does openstack-ci use some other job distribution mechanism?21:12
termieeday: i don't think openstack-ci should exist21:13
sorentermie: What does that mean?21:13
termiei think that test code should be part of nova21:13
termieand that hudson should be used to automate the tests21:13
jaypipestermie: you may have forgotten that OpenStack != Nova :)21:13
edaytermie: ok, then should nova ci test code depend on jenkins?21:13
termieeday: ideally not, i expect them to be runnable scripts21:14
jaypipestermie: we need a test runner that tests more than Nova. Because our production clusters have Glance + Nova + Swift.21:14
kbringardvvuksan: awesome, that sounds like fun :-)21:14
edaytermie: ok, then it sounds like jay's method of not using jenkins for coordinating this is preferred21:14
vvuksankbringard: :-)21:15
termieeday: why is that?21:15
edayjaypipes: package that all projects can import?21:15
vvuksankbringard: i fixed your faulty documentation :-)21:15
edaytermie: because you said it shouldn't be a dependency21:15
kbringardthanks, I noticed that :-)21:15
jaypipeseday: I suppose that's a perfectly fine option, too, sure.21:15
termieeday: that has nothing to do with it21:15
termieeday: the testing automation depends on jenkins21:16
termieeday: the bulk of hte test code itself does not need to21:16
jaypipestermie: that way I think of it, openstack-ci will be importing nova.smoketests.admin_test (or something like that) as *one of* the tests a client could run against the test cluster...21:17
jaypipestermie: but openstack-ci also needs to have its test clients also run tests against swift and glance.21:17
termiejaypipes: i have not forgotten that there are more projects, but splitting the code offers no benefit at the moment21:17
jaypipestermie: hope that makes more sense...21:17
*** colinnic- has joined #openstack21:17
termieswift and glance specific tests live in their codebase21:17
vvuksankbringard: i have been talking to Centos guys about making an openstack RPMS repo. they are interested21:17
jaypipestermie: why would I put a test runner that tests swift stuff in nova?21:17
*** bencherian has joined #openstack21:18
*** mdomsch has joined #openstack21:18
edaytermie: so, when I ask if they should depend on jenkins, I mean can I run this on my 3 machine cluster without having to install hudson to coordinate, or can I just give it a list of hosts and go (perhaps with setting up ssh keys or some other queue to coordinate)21:18
* jaypipes will be right back...dogs are bugging the crap out of him for dinner...21:18
edaytermie: obviously we can always kick this off with a jenkins job21:19
*** rcc has quit IRC21:19
termieeday: ideally you can run some specific portion of tests against your own cluster assuming you have set it up beforehnad21:19
*** colinnic| has joined #openstack21:20
*** colinnich_ has quit IRC21:20
termieeday: i doubt you'd be able to easily run all the tests in a continuous manner without extra infrastructure21:21
edaytermie: what if I just have 3 fresh boxes with latest ubuntu image? we'll want to have things automated to assume very little (so we don't need to manually perform any manual setup)21:21
*** bcherian has quit IRC21:21
* jaypipes back21:21
termieeday: that automated setup is what jenkins is in charge of21:21
termieeday: at this point i am assuming a chef script gets kicked off21:21
edaytermie: so it will be a dependency to bootstrap the tests21:22
termieeday: after whatever code re-images the machine21:22
jaypipestermie: agreed with you on that, like I said, I have no problem using Hudson for lots of stuff, including setting up and tearing down test environments.21:22
eday(for the record, i have no opinion, just trying to understand)21:22
*** joearnold has joined #openstack21:23
vishyeday, termie, jaypipes: so we have three sections 1. Image (pxe-boot) 2. Install (chef/puppet) 3. Test21:23
jaypipesvishy: 4. Record results for regression reporting.21:24
edayvishy: do you use any other multi-node coordination besides chef/puppet currently? (queue, jenkins, ssh, ...)21:24
jaypipesvishy, termie: I'm assuming you want to use Hudson for 1 and 2, right?21:24
*** colinnic- has quit IRC21:24
* vishy knows very little about hudson21:25
termiejaypipes: as well as for 3 and 421:25
*** colinnic| has quit IRC21:25
*** drico has joined #openstack21:25
*** dprince has quit IRC21:26
jaypipestermie: well Hudson may fire the test script, which may in turn record the test results, but Hudson doesn't actually spawn clients and issue the test requests, that's what I was getting at.21:26
vishyI would hope that the actual tests for nova are primarily in nova (at least the ones that are not using integrated features)21:26
jaypipesvishy: no problem with that at all.21:26
* soren calls it a day21:26
*** dirakx has joined #openstack21:26
termiejaypipes: hudson runs the script that spawns clients and issues requests, and then records the results21:27
termiejaypipes: i agree that code needs to exist to spawn those clients21:27
jaypipestermie: and where does that code live? I think it should live in openstack-ci, which is the purpose of that project.21:27
termiei agree that the code is applicable beyond just nova, however at the moment the splitting of projects is a pain point and i would rather the decision be made in some more general way21:29
termiei agree that there is a need for openstack-common, but i want a plan before we make it21:29
termiei am happy to move the code into openstack-common once there is such a plan21:30
edaytermie: we discussed candidates for openstack-common quite a bit at the summit, so there is a good plan. wish you hadn't missed it :)21:30
termiebut for now i think the most benefit is in keeping the codebase together21:30
*** dirakx has quit IRC21:30
JordanRinkesomeone said I was missing some chat about testing here?21:30
termieeday: a pity indeed but it still seems to have failed a bit21:31
JordanRinkeYou guys know I am working on putting together a fairly larger community test bed right?21:31
*** miclorb_ has joined #openstack21:31
termieJordanRinke: i have heard rumors of such things21:31
termieJordanRinke: we are discussing the automated tearing down and setting up of different deployment strategies21:31
vishyJordanRinke: how are the hosts going to be provisioned? pxe-booting?21:31
edaytermie: I think when there is obvious code duplication with glance and nova we should be sticking it in openstack-common packages, should have been doing this for a while. it's easy enough to just have a PPA dep package for it21:31
termieeday: it is not easy enough to do that21:31
jaypipeseday: there's more code in common between glance and swift now than glance and nova...21:32
termieeday: we have codebases that have to move in some lockstep at the moment21:32
jaypipesI disagree, but not sure where this fits into the testing discussion.21:32
edayjaypipes: if common testing/ci code belongs in openstack-common21:33
edayjaypipes: ahh, intersting re: swift/glance21:33
jaypipeseday: I don't care whether it goes into openstack-common or openstack-ci, really.21:33
edaytermie: all the more reason to work out those APIs and version issues now so it can be shared more easily21:33
edaytermie: the longer they diverge, the more difficult it will be, IMO21:34
termieeday: agreed, though there would be no divergence if it is in the same tree21:34
*** bencherian has quit IRC21:34
termieeday: and i don't think today is the day to hash it out21:34
edaytermie: I don't think putting swift, glance, and nova in one tree is a good idea :)21:34
termieeday: not suggesting swift21:35
termieeday: but glance yes21:35
jaypipestermie: glance is not going back into nova. let's move on.21:35
edaytermie: as jay just mentioned, there is a fair amount of shared code between swift/glance though, so there would be divergence there21:36
termieeday: the important part of my statement was that there is not divergence when the code is developed together21:37
JordanRinkevishy: we are working on some tech that should be announced in a few weeks, basically a really slick pxe bootstrap / chef integrated full bare metal to functional system platform21:37
JordanRinkeso we can rebuild a lab per release, once a week etc21:37
termieJordanRinke: we're sort of talking about rebuilding small parts of a lab every couple hours21:38
JordanRinkewe haven't hashed it out yet and it is totally open for discussion21:38
JordanRinkemy initial thought was just to have the lab in 2 zones that are rebuilt once a week each21:38
*** greenisus has joined #openstack21:38
JordanRinkeOnce I have hardware online the management will be up to the community, whatever works best is the schedule we will adhere to21:39
termieJordanRinke: that is good for the general case of having a deployment to manually work with21:39
JordanRinkefor dev test and for functional test21:39
termieJordanRinke: but we are also looking for a continuous integration platform21:39
JordanRinkeif we have enough units I would like to yeah...21:39
JordanRinkekeep a small section of it as a continous platform21:39
termieJordanRinke: exacto21:39
JordanRinkeI don't have actual number yet but I am hoping to be able to re-use some old gear and have about 400 systems. they will be old and only capable of running 2-3vm's per but would be good for testing21:40
JordanRinkeI am hoping to just get 10-20 online in the next few weeks and then regularly add systems as possible21:40
JordanRinkenothing solid yet, hopefully not jumping the gun talking about it before it is built :-D21:41
edaytermie: (sorry, was afk) agree, but the reality is things are diverging and them being developer together is very unlikely. sooner we can get them sharing through another package the better21:42
JordanRinkeso my question to you guys is ideally what would be the best config for a lab?21:43
JordanRinkesay I start with 20 boxes, whats the best way to use those?21:43
*** CloudChris has quit IRC21:45
*** reldan has joined #openstack21:45
*** dirakx has joined #openstack21:45
jaypipesJordanRinke: add them into as available build machines that termie and others can fire off jobs against.  At least, that would be my first thoght...21:45
*** dirakx has quit IRC21:46
jaypipesJordanRinke: that way, we can at least add some of the deployment, installation, and configuration tests that are sorely missing from our automation.21:46
creihtJordanRinke: where are these machines being setup (out of curiosity)?21:47
jaypipescreiht: your basement. didn't you know?21:47
creihtjaypipes: there are no basements in Texas :)21:48
JordanRinkecreiht: initially the couple I can get hands on with in the short term will be SAT4 but beyond that we are looking at putting them in a number of non rackspace facilities to make community use more interesting and introduce geographical dispersion for testing/federation.21:48
creihtJordanRinke: You know that SAT4 doesn't have outside access right?21:48
JordanRinkethat is what they say now...21:49
jarrodshould I use LVM on the compute nodes if I wish to use their drives for the VM block storage that will be controlled by volume manager?21:49
vishyjarrod: i think you asked that earlier and i responded probably21:49
JordanRinkebut seriously we know that it isn't allowed currently but we are going to try to work with NetSec/RTS to come up with a reasonable solution to that problem21:49
vishyjarrod: unless you want to allocate some number of drives to local storage for compute21:50
jarrodprobably does not instill confidence21:50
vishyjarrod: how about this then21:50
vishythat is how we do it at NASA21:50
vishymore confidence?21:50
jarrodmuch :)21:50
jarrodthanks vishy :)21:50
creihtJordanRinke: ok, if you figure out how to get them to do that, will you let me know? :)21:50
termiemy battery is about to die :/21:51
termieattempting to locate an outlet21:51
JordanRinkei thought pretentious coffee shops had outlets at every table?21:54
JordanRinkeso, in your guys minds... what do we need in terms of hardware for a functional community lab?21:56
*** mray has joined #openstack21:58
*** mray has joined #openstack21:58
*** maplebed has joined #openstack21:59
*** brd_from_italy has quit IRC22:01
*** joearnold has quit IRC22:01
*** ctennis has quit IRC22:01
*** rpedde has joined #openstack22:03
*** MarkAtwood has joined #openstack22:04
jaypipesJordanRinke: whatever are the most common things that people are trying to deploy Nova on (and having trouble with) would be the priority, IMHO.22:04
jaypipesJordanRinke: you may want to post to the mailing list asking what the community lab should first have available.22:05
*** DanEllis has quit IRC22:05
*** vvuksan has quit IRC22:06
*** allsystemsarego has quit IRC22:07
kpeppleJordanRinke: at least a multinode environment of every hypervisor that we support22:08
*** hub_cap has quit IRC22:08
*** dsockwell has joined #openstack22:09
termiejaypipes: my computer is dying and i need to leave really soon, but i was chatting witht he anso guys and i think we have a proposal you can get behind22:10
*** dubsquared has joined #openstack22:10
termiejaypipes: so i will write an email later today22:10
termiejaypipes: and maybe we can make everybody happy22:11
termiejaypipes: :D22:11
dubsquaredbetter check with anotherjesse before you send22:11
jarrod[    2.953802] Brought up 16 CPUs -- yay ubuntu22:11
dsockwellgood afternoon all.  i'm from eucalyptus land, but openstack looks really promising.  can anyone tell me if the project plans to incorporate SAN storage in the future?22:12
alekibangojaypipes: are you the right person to talk to about automated functional nova tests?22:12
evtoewshi all, i've installed bexar from these instructions and now i'm trying to fire up an instance with these instructions but i've realized that neither sets of instructions actually start any nova-* services. ps -elf | grep nova returns nothings.  am i missing something or is something missing from the instructions?  w22:12
evtoewshat services need to run?22:12
dsockwellalso do instances in openstack survive reboot of the host machine or entire cluster?22:12
alekibangoevtoews: which OS?22:12
evtoewsalekibango: ubuntu 10.1022:13
alekibangowhich packages did you install ? where did you took them?22:13
*** joearnold has joined #openstack22:14
alekibangosorry for asking those :) but it will help22:14
*** ctennis has joined #openstack22:14
alekibangoor from source?22:14
alekibangoah, reading the manual you gave me22:14
evtoewsalekibango: np.  the tarball i used was
uvirtbotLaunchpad bug 714678 in nova "Missing items in Bexar release tarball" [Critical,In progress]22:15
alekibangoso tarball.  but there imho will be some bexar deb packages for your ubuntu...22:16
alekibangothat might be a lot easier for you22:16
alekibangoyou need to run some nova services, like compute, api etc22:17
evtoewsalekibango: right.  i'm looking for the definitive list.22:17
alekibangosoren: are there some ubuntu 10.10   bear packages ?22:17
alekibangoevtoews: running nova is ... to be used when your services are running22:18
alekibangoevtoews:  look here
alekibangoand look on source of the nova script22:19 i mean22:19
alekibangothere you will see how to run it manually22:19
alekibangobut its a lot  easier to just use packages22:19
alekibangothere are some, surely...  just i do not use them... i build my own :)22:19
*** greenisus has quit IRC22:19
alekibangoevtoews: look on this:
alekibangothose services you want are  here   -- read lines .... installing $SERVICE script to /usr/local/bin22:21
alekibangoyou need to configure and run them :)22:21
alekibangoannegentle: there is something missing in bexar intallation steps :) how to run the services22:22
evtoewsalekibango: so every single one of those services in should run?  that's a different list from
*** flacoste has joined #openstack22:23
*** flacoste has left #openstack22:24
alekibangowell,  it depends... nova is very configurable22:24
*** BK_man has joined #openstack22:24
alekibangoyou can use just some services... and run them on different computers...22:25
alekibangoor you can use just one22:25
evtoewsalekibango: i see the source code of  guess i'll go with that for now.  to be honest i've been avoid the deployment scripts in order to get a better feel for what's going on in nova and what it really takes to configure it.22:25
annegentlealekibango: ah yes22:25
alekibangoevtoews: i would say look for bexar deb packages for ubuntu22:25
alekibangothere must be some :)22:25
alekibangoits really much easier to get it running22:26
dsockwellah -- SAN question partially answered, i found the blueprints page22:26
evtoewsalekibango: i haven't seen any referred to in the documentation but if there are some that would be pretty great.22:26
alekibangodsockwell: instead of SAN i wouldsay try   sheepdog22:26
*** burris has joined #openstack22:27
alekibangoevtoews: i am sure some, maybe for now unofficial, packages exist22:27
alekibangothere are for sure packages of trunk22:27
evtoewsalekibango: where could i find packages for bexar?  trunk isn't stable enough for us.22:28
annegentleevtoews: if you're on Maverick, check out
dsockwellalekibango: okay, thanks. i don't necessarily need the SAN to have a command interface with openstack if the next release does live migration.  do you know if a xen instance on openstack will persist if the host machine is rebooted?  eucalyptus scrubs instances on boot.22:29
alekibangodsockwell: xen has no sheepdog support ATM22:30
alekibangobut i believe it will come in next few months22:30
*** ppetraki has quit IRC22:30
annegentleevtoews: oh sorry, the latest rev that worked was 650. And ummmm how do you get a particular rev?22:30
alekibangobut to be fair, maybe there some is, i just dont know, i prefer KVM22:30
evtoewsannegentle: right...what can i replace ppa:nova-core/trunk with for bexar instead?22:31
alekibangoevtoews: maybe you can build your packages using bexar source22:31
alekibangoand build scripts...22:31
alekibangoi will give you link22:32
alekibangoah thanks22:33
dsockwellso if I start a VM on openstack, and then reboot the machine that the VM lives on, will the VM be destroyed like it would in eucalyptus?22:34
alekibangodsockwell: it should be running :)22:35
*** kbringard has quit IRC22:35
justinsbdsockwell: I just pushed some SAN code.  Solaris was merged this morning; HP/LeftHand SAN patch just went up22:35
dsockwelljustinsb: are you in santa barbara?  i can have a six-pack on your doorstep in 3 hours.22:36
alekibangoif you do now own some SAN, rather use sheepdog... its cheap, fast, nice, and hopefully also reliable :)22:36
justinsbdsockwell: Actually, it's my last name :-)  I'm in San Francisco.22:36
justinsbdsockwell: Thanks though!  What hardware are you looking to support?22:37
dsockwellan openindiana box22:37
dsockwellso your solaris work is interesting for me22:37
justinsbdsockwell: Well, the Solaris code should work.  It relies on COMSTAR, which I think OI has?22:37
justinsbWell let me know if anything doesn't work right!22:38
dsockwellokay.  once i'm in a position to test it out i will be in touch.22:38
alekibangodsockwell: build your own packages... its the best way22:39
rlucioalekibango: awesome timing, i was actually just looking into where the newest packaging branch was, thanks22:41
alekibangoannegentle: ^^ :)22:42
evtoewsannegentle alekibango:  having never done packaging before how do i ensure that the packages i build are the bexar release?  is there a reason this wasn't already done for ubuntu/bexar?22:42
alekibangoevtoews:  try those:
alekibangothey are prebuilt for you already22:42
*** fcarsten has joined #openstack22:43
evtoewsalekibango: thx.  i'll give those a shot.22:44
alekibangostill,  i suggest sheepdog 2.1 (not yet in ubuntu) if you will use it22:44
alekibangoevtoews: just make sure they are bexar release :)22:46
evtoewsalexibango: nova 2011.1~bzr645-0ubuntu0ppa1~lucid1 OpenStack Hudson (2011-02-03).  that's the release date for bexar so it should be.  :)22:47
*** troytoman has joined #openstack22:47
alekibangoi will try to write debian specific instructions... when i will double test everything :)22:48
dsockwellah, sheepdog looks like it uses the IP protocol for transport.  that shiny gigabit switch is outside my budget atm, but it looks promising22:49
*** mdomsch has quit IRC22:50
*** matiu has joined #openstack22:51
*** btorch has joined #openstack22:54
fcarstenswift/cloudfiles: What is the purpose of the zero sized "directory marker" files? It seems to me that all necessary path information is already in the item names?22:55
JordanRinkeanyone know if there is a new nova all in one image post Bexar release?22:56
*** rpedde has quit IRC22:58
*** blueadept has quit IRC23:00
*** freeflying has quit IRC23:00
alekibangoall in one image? you mean something to run in virtual machine?23:01
*** ranger57 has joined #openstack23:04
*** freeflying has joined #openstack23:04
*** pvo has joined #openstack23:05
*** MarkAtwood has quit IRC23:05
*** Hayati has joined #openstack23:06
openstackhudsonProject dashboard-tarmac build #2,277: FAILURE in 35 sec:
openstackhudsonProject nova-tarmac build #59,898: FAILURE in 36 sec:
*** mray has quit IRC23:07
*** dirakx has joined #openstack23:08
*** prewarmoney has joined #openstack23:08
*** pvo has quit IRC23:08
JordanRinkealekibango: yeah a VM using UML or something23:08
prewarmoneyhey guys, just caught a possible bug in the launching of a uec image... I get this in the console output: "Caught exception reading instance data: http:\/\/\/2009-04-04\/meta-data\/placement\/availability-zone"23:09
*** bradshaw1 has joined #openstack23:09
alekibangoJordanRinke: i dont know...23:10
alekibangoinstall your own :)23:10
openstackhudsonProject dashboard-tarmac build #2,278: STILL FAILING in 36 sec:
openstackhudsonProject nova-tarmac build #59,899: STILL FAILING in 36 sec:
prewarmoneywe're running the 2011.2~bzr655-0ubuntu0ppa1~maverick223:14
creihtfcarsten: directory markers are for backwards compatibility with the older system, you can get the same behavior with prefix/delimeter queries23:16
fcarstencreiht: cool, thanks.23:16
openstackhudsonProject dashboard-tarmac build #2,279: STILL FAILING in 35 sec:
Ryan_Lanehmm, looks like sheepdog has been added to launchpad, any possibility this will get added to a ppa? :)23:16
ranger57hi, I bundled a maverick image by using uec-publish-tarball, but whenever I start an an instance, I get a mounting error from euca-get-console-output23:17
openstackhudsonProject nova-tarmac build #59,900: STILL FAILING in 35 sec:
ranger57has anyone else encountered this?23:17
*** Hayati has quit IRC23:17
*** pvo has joined #openstack23:17
*** gondoi has quit IRC23:20
dsockwellso just to clarify (and ask again and be annoying) instances on openstack do survive reboots of their host?23:21
dsockwellinstead of being destroyed in eucalyptus23:21
Ryan_Lanedsockwell: I believe so. can always test and find out :)23:21
openstackhudsonProject dashboard-tarmac build #2,280: STILL FAILING in 36 sec:
openstackhudsonProject nova-tarmac build #59,901: STILL FAILING in 35 sec:
dsockwellah, but i'd rather pester people online at this stage23:22
fcarstendsockwell: Wouldn't that depend as much on what hypervisor you use as the cloud-stack?23:22
dsockwellno, i don't know of any hypervisor that destroys instances on reboot.  it was a eucalyptus 'feature' that bugged me, and i wanted confirmation that openstack doesn't have it23:23
dsockwelli suppose xen can be told to23:23
BK_manhi all23:25
BK_manGridDynamics would like to announce availability of RHEL 6.0 build of Nova:
dsockwellsince nobody is stepping in with 'gosh thats annoying' i'll presume it's safe23:26
openstackhudsonProject dashboard-tarmac build #2,281: STILL FAILING in 35 sec:
* jk0 is not impressed with the launchpad maint right now23:26
*** pvo has quit IRC23:26
jaypipestermie: cool, thx termie. I stepped away for a while too... I'll look forward to your email. cheers.23:26
openstackhudsonProject nova-tarmac build #59,902: STILL FAILING in 35 sec:
JordanRinkeBK_Man: you on twitter etc so I can retweet that or something?23:27
jaypipesalekibango: still around? yes, I'm one of the people to talk to. others are termie, vishy, and devcamcar23:27
* JordanRinke is winning a twitter argument with jk023:27
BK_manJordanRinke: yep
JordanRinkebah, i was just messing with you jk0 takes the fun out of it when you don't bicker with me lol23:28
jarrodcan i follow all of the multi-node-installation procedures, but using bexar for the nova ppa install to accomplish multi server setup?23:29
jarrodI cannot find /etc/nova created in order to specify my cloud controller, databases, etc23:29
*** MarkAtwood has joined #openstack23:30
openstackhudsonProject dashboard-tarmac build #2,282: STILL FAILING in 35 sec:
*** bradshaw1 has quit IRC23:32
openstackhudsonProject nova-tarmac build #59,903: STILL FAILING in 36 sec:
dubsquaredbk_man:  nice, reading now!23:34
BK_man2:35AM in Moscow. I'm probably heading myself to a bed :)23:35
*** bradshaw has joined #openstack23:35
openstackhudsonProject dashboard-tarmac build #2,283: STILL FAILING in 35 sec:
alekibangojaypipes: i am using  to install nova... it takes just few minutes for whole cloud.... and i would like to have some way to test the functionality23:36
openstackhudsonProject nova-tarmac build #59,904: STILL FAILING in 35 sec:
*** sebastianstadil has quit IRC23:37
alekibangoi could setup similar thing for ubuntu easily23:37
alekibango(one install machine -> installs debian or ubuntu cloud)23:37
*** troytoman has quit IRC23:37
alekibangothat could be nice to extend with automated functional tests23:38
alekibangoand check every nova build on few operating systems23:38
alekibangolater even with other distros23:38
*** BK_man has quit IRC23:38
jaypipesalekibango: that URL returns a 404.23:38
*** westmaas has joined #openstack23:38
alekibangoheh :)
alekibangoits a system for automated installation of machines, primarilly debian23:39
alekibangoi can help by setting this up on some servers23:40
alekibangoso you can extend it with tests and we can run them on every merge request23:40
alekibangoor build :)23:40
alekibangoon all systems, for start with ubuntu and debian23:40
jaypipesalekibango: and this would run from Hudson? in other words, we'd kick off a FAI job from Hudson that would wipe and install some base image with Xen on, say, Ubuntu, on the test cluster(s)?23:41
openstackhudsonProject dashboard-tarmac build #2,284: STILL FAILING in 35 sec:
alekibangojaypipes: we can start it by any way23:41
alekibangowe can write script which will act on incoming email23:42
openstackhudsonProject nova-tarmac build #59,905: STILL FAILING in 35 sec:
alekibangoor whatever23:42
*** MarkAtwood has quit IRC23:42
jaypipesalekibango: k. well, since we already use hudson, it would make sense to use it for this, too. :)23:42
alekibangojaypipes: if you will get me some machines (real ones, impi is advantage), i can help by configuring them23:42
dubsquaredDoes anyone know where there is some good nova-volume notes?  I've started playing around, but wanted to know if there was any doc before I just start hacking at this ...23:43
alekibangos/impi/ipmi/ :)23:43
alekibangodubsquared: hardly any.... but try sheepdog !23:44
alekibangodubsquared: if you find some, tell me23:44
jaypipesalekibango: perhaps JordanRinke can provide a machine or ten. Not sure. I have no ability to procure anything, unfortunately. :(23:45
alekibango1 for server, 3+ for nova machines23:45
dubsquaredI want 10 more servers, JordanRinke23:45
dubsquaredgive give give!23:45
alekibangothe 1 can be cheap server23:45
*** bradshaw has left #openstack23:46
openstackhudsonProject dashboard-tarmac build #2,285: STILL FAILING in 36 sec:
alekibangojaypipes: i would like to contribute fai config to nova...23:46
alekibangoso others can use it23:46
alekibangoand expand it23:46
openstackhudsonProject nova-tarmac build #59,906: STILL FAILING in 35 sec:
alekibangoand find more usage cases... those functional tests can serve also as usage examples23:47
dubsquaredYou know the 'most interesting man in the world' commercials …. when JordanRinke gives out servers, which he doesn't do often, he prefers high end shiz23:47
alekibangopeople are constantly fighting with it :)23:47
alekibangohaving few example uses with all config would be GREAT23:47
alekibango(i do only my one)23:48
alekibangojaypipes: one server is installed (including partitioning and formatting drives) in 3-5 minutes23:48
*** prewarmoney has quit IRC23:49
*** dubsquared has quit IRC23:50
jaypipesalekibango: awesome. I've put in an email to our production folks about getting machines. I will let you know what I find.23:50
alekibangok :)23:50
openstackhudsonProject dashboard-tarmac build #2,286: STILL FAILING in 35 sec:
jarroddang failed to register maverick-server-uec-amd64-vmlinuz-virtual.manifest.xml23:51
openstackhudsonProject nova-tarmac build #59,907: STILL FAILING in 35 sec:
jarrodsounds like the xml is borked23:52
*** evtoews has quit IRC23:54
*** dfg has quit IRC23:55
openstackhudsonProject dashboard-tarmac build #2,287: STILL FAILING in 35 sec:
openstackhudsonProject nova-tarmac build #59,908: STILL FAILING in 36 sec:

Generated by 2.14.0 by Marius Gedminas - find it at!