Thursday, 2011-04-07

*** reldan has joined #openstack00:02
*** MotoMilind has quit IRC00:03
*** adjohn has joined #openstack00:08
*** joearnold has quit IRC00:11
*** kakella has left #openstack00:16
*** reldan has quit IRC00:17
*** Zangetsu has quit IRC00:19
*** zul has joined #openstack00:24
*** reldan has joined #openstack00:27
*** herki has joined #openstack00:37
*** herki_ has quit IRC00:38
*** jeffjapan has quit IRC00:40
*** mahadev has quit IRC00:43
*** dendrobates is now known as dendro-afk00:51
*** jtran has quit IRC00:54
*** jtran has joined #openstack00:57
*** jtran has left #openstack00:57
*** j05h1 has quit IRC00:59
*** j05h has joined #openstack00:59
*** mahadev has joined #openstack01:04
*** reldan has quit IRC01:06
*** mahadev has quit IRC01:11
*** mray has joined #openstack01:11
*** Ryan_Lane has quit IRC01:15
*** dragondm has quit IRC01:16
*** mahadev has joined #openstack01:17
*** adjohn has quit IRC01:17
*** adjohn has joined #openstack01:19
*** AimanA is now known as HouseAway01:33
*** winston-d has joined #openstack01:36
*** Julio__ is now known as JulioBell01:43
*** jeffjapan has joined #openstack01:49
*** KnightHacker has quit IRC01:50
*** Pentheus has quit IRC01:52
*** jfluhmann_ has joined #openstack01:59
*** maplebed has quit IRC01:59
*** blamar has quit IRC02:01
*** RobertLaptop has quit IRC02:01
*** dfg has quit IRC02:01
*** jstinson has quit IRC02:01
*** _cerberus_ has quit IRC02:01
*** filler has quit IRC02:01
*** aloga has quit IRC02:01
*** jarrod has quit IRC02:01
*** camm_ has quit IRC02:01
*** dsockwell has quit IRC02:01
*** troytoman-away has quit IRC02:01
*** iRTermite1 has quit IRC02:01
*** lool has quit IRC02:01
*** smoser has quit IRC02:01
*** soosfarm_ has quit IRC02:01
*** kang___ has quit IRC02:01
*** pandemicsyn has quit IRC02:01
*** Beens_ has quit IRC02:01
*** mtaylor has quit IRC02:01
*** xtoddx has quit IRC02:01
*** jheiss has quit IRC02:01
*** deepy has quit IRC02:01
*** gdusbabek has quit IRC02:01
*** cclien_ has quit IRC02:02
*** Cen has quit IRC02:02
*** czajkowski has quit IRC02:02
*** devcamcar has quit IRC02:02
*** flashn_ has quit IRC02:02
*** vishy has quit IRC02:02
*** openstackjenkins has quit IRC02:02
*** _0x44 has quit IRC02:02
*** clayg has quit IRC02:02
*** taihen has quit IRC02:02
*** jtimberman has quit IRC02:02
*** antonym has quit IRC02:02
*** Cen has joined #openstack02:03
*** cclien_ has joined #openstack02:03
*** czajkowski has joined #openstack02:03
*** devcamcar has joined #openstack02:03
*** flashn_ has joined #openstack02:03
*** vishy has joined #openstack02:03
*** openstackjenkins has joined #openstack02:03
*** _0x44 has joined #openstack02:03
*** clayg has joined #openstack02:03
*** taihen has joined #openstack02:03
*** jtimberman has joined #openstack02:03
*** antonym has joined #openstack02:03
*** sets mode: +vv _0x44 antonym02:03
*** xtoddx has joined #openstack02:03
*** jstinson has joined #openstack02:03
*** blamar has joined #openstack02:03
*** RobertLaptop has joined #openstack02:03
*** dfg has joined #openstack02:03
*** _cerberus_ has joined #openstack02:03
*** filler has joined #openstack02:03
*** aloga has joined #openstack02:03
*** jarrod has joined #openstack02:03
*** camm_ has joined #openstack02:03
*** dsockwell has joined #openstack02:03
*** troytoman-away has joined #openstack02:03
*** iRTermite1 has joined #openstack02:03
*** lool has joined #openstack02:03
*** smoser has joined #openstack02:03
*** soosfarm_ has joined #openstack02:03
*** kang___ has joined #openstack02:03
*** pandemicsyn has joined #openstack02:03
*** Beens_ has joined #openstack02:03
*** mtaylor has joined #openstack02:03
*** deepy has joined #openstack02:03
*** jheiss has joined #openstack02:03
*** gdusbabek has joined #openstack02:03
*** sets mode: +v pandemicsyn02:03
*** deepy has quit IRC02:04
*** deepy has joined #openstack02:05
*** jfluhmann_ is now known as jfluhmann_bedtim02:15
*** rds__ has quit IRC02:24
*** throughnothing has joined #openstack02:39
*** clauden has joined #openstack02:49
*** dragondm has joined #openstack02:50
*** throughnothing has quit IRC02:54
winston-dBK_man : hi02:59
*** jfluhmann_bedtim is now known as jfluhmann_03:00
*** clauden has quit IRC03:00
winston-dhas anyone tried using glance as image service03:01
*** vernhart has quit IRC03:03
*** kashyap has joined #openstack03:05
*** thejcurious has joined #openstack03:06
*** mray has quit IRC03:08
*** dendro-afk is now known as dendrobates03:11
*** jim__ has joined #openstack03:56
*** nwww has quit IRC03:58
jim__I'm new to swift/cloudfiles and using the Python api.  Anyone here up for answering a question about etag handling?04:01
*** koolhead11|afk is now known as koolhead1704:01
HugoKuo_morning huys , what's your nova trunk version now ?04:01
*** jim__ has quit IRC04:03
*** hadrian_ has joined #openstack04:15
*** hadrian has quit IRC04:17
*** hadrian_ is now known as hadrian04:17
*** hadrian has quit IRC04:25
*** f4m8_ is now known as f4m804:44
*** dendrobates is now known as dendro-afk04:46
*** kashyap has quit IRC04:47
*** dendro-afk is now known as dendrobates04:47
koolhead17hi all04:52
*** benbenhappy has joined #openstack04:57
*** dittob has joined #openstack05:02
dittobdoes anyone has more detailed feature matrix than one on the wiki page?05:06
*** dirakx has joined #openstack05:11
*** ramkrsna has joined #openstack05:15
*** kashyap has joined #openstack05:19
*** dendrobates is now known as dendro-afk05:32
*** dendro-afk is now known as dendrobates05:33
*** jfluhmann__ has joined #openstack05:34
*** cole has quit IRC05:36
*** ramkrsna has quit IRC05:38
*** jfluhmann_ has quit IRC05:38
*** cole has joined #openstack05:46
termienotmyname: aye, complicated05:46
*** cole is now known as Guest8972605:47
termienotmyname: i think i can do a decent job, but i don't know the failure conditions as well as i do on nova05:47
termieso i tis more scary to demo05:47
*** dirakx has quit IRC05:47
*** zaitcev has quit IRC05:48
*** nwww has joined #openstack05:53
*** mahadev has quit IRC06:04
*** guigui has joined #openstack06:08
*** nijaba has quit IRC06:16
*** nijaba has joined #openstack06:16
*** dragondm has quit IRC06:21
*** dittob has quit IRC06:24
*** guynaor has joined #openstack06:36
*** guynaor has left #openstack06:38
ttxjk0: now, yes06:39
ttxjaypipes: time is getting short, but that sounds doable...06:44
*** lionel has quit IRC06:52
*** lionel has joined #openstack06:53
*** rcc has joined #openstack06:59
*** flopflip_ has quit IRC07:07
*** flopflip has joined #openstack07:09
*** jamesurquhart has joined #openstack07:11
*** vernhart has joined #openstack07:11
*** jamesurquhart has quit IRC07:13
*** srb has joined #openstack07:21
*** srb is now known as Guest6596807:21
Guest65968Hi !! Can u recommend a link for creating Custom UBuntu images07:22
*** mahadev has joined #openstack07:28
*** Nacx has joined #openstack07:29
*** mahadev has quit IRC07:33
*** MarcMorata has joined #openstack07:39
*** benbenhappy has quit IRC07:41
*** reldan has joined #openstack07:44
*** daveiw has joined #openstack07:44
winston-dGuest65968: please refer to
*** miclorb has quit IRC07:47
*** reldan has quit IRC07:48
sorenBK_man: I did test Ajaxterm. My conclusion is that it's not entirely there yet. I'm hoping I'll get on top of it today.07:50
*** miclorb has joined #openstack07:57
*** jfluhmann__ has quit IRC07:59
*** miclorb has quit IRC08:01
HugoKuo_should I install python-suds on all nova components hosts?08:10
uvirtbot`New bug: #753280 in nova "We should use policy routing for VM's." [Undecided,New]
sorenHugoKuo_: With current trunk, suds shoul no longer be necessary.08:25
HugoKuo_soren : hanks08:25
sorenHugoKuo_: sure08:26
HugoKuo_did you ever see the openstack nova topology on Stackops website?08:26
HugoKuo_   it looks complex...08:27
*** adjohn has quit IRC08:27
*** irahgel has joined #openstack08:28
*** jeffjapan has quit IRC08:30
*** reldan has joined #openstack08:33
*** Seoman has joined #openstack08:35
*** MarcMorata has quit IRC08:35
*** Seoman has quit IRC08:36
*** MarcMorata has joined #openstack08:36
*** reldan has quit IRC08:38
*** reldan has joined #openstack08:41
uvirtbot`New bug: #753292 in nova "Ping or ssh failed when libvirt_type is lxc" [Undecided,New]
*** bamedro has joined #openstack08:43
koolhead17HugoKuo_, thanks for the link.08:52
HugoKuo_u r welcome...08:53
HugoKuo_but don't you feel that's too complex ?08:53
koolhead17it has few information better than our existing docs08:53
*** nagyz has joined #openstack08:53
HugoKuo_why nova need three switch @@?08:53
nagyzis there an internal API to query registered services?08:54
*** lionel has quit IRC08:54
ttxHugoKuo_: it doesn't.08:54
nagyzor should I query this from the database directly?08:54
HugoKuo_I guess that doc is base on bexar release....08:54
*** lionel has joined #openstack08:54
HugoKuo_ttx : thanks08:54
ttxHugoKuo_: it's just one way to deploy it08:54
nagyz(I've already looked at the developer documentation, nova.service doesn't say anything about it)08:54
*** reldan has quit IRC08:55
HugoKuo_I'm trying to make it simpler .....:)08:55
koolhead17nagyz, nova-manage service list  is this what your looking for08:56
koolhead17HugoKuo_,  that stackops portal u mean?08:56
nagyzkoolhead17, ah, great, ty!08:57
HugoKuo_koolheadl7 : I want multi nodes like what you see on stackops , but I want everything in one network , include instances....08:57
nagyzjust getting to know openstack. :)08:57
koolhead17nagyz, welcome. so am i :)08:57
koolhead17HugoKuo_, give me your blog page to srart with :D08:58
HugoKuo_so maybe I don't need make it such complex . in my test there's only one switch and 4 machines08:58
HugoKuo_I choose plan B now........08:58
HugoKuo_but give up using deployment tool , just do everything manually...08:59
koolhead17HugoKuo_, i like those colorful fishes :P09:00
HugoKuo_man , you can feed them XD09:00
HugoKuo_they'll follow your arrow~~~09:00
koolhead17they were singing too i suppose HugoKuo_ :P09:02
uvirtbot`New bug: #752735 in nova "Volumes aren't zeroed out" [High,Confirmed]
*** nwww has quit IRC09:13
koolhead17can someone suggest me the reason behind these different shell under nova-manage09:18
soren"different shell"?09:25
*** Nacx is now known as nacx09:25
*** nacx has quit IRC09:25
*** nacx has joined #openstack09:26
*** daveiw has quit IRC09:31
HugoKuo_       I really need a suggestion about this issue..... I face it til now . did I miss any package ?09:33
HugoKuo_it's nova-network.log09:33
sorenHugoKuo_: You need to point lock_path to somewhere where Nova can create lock files.09:34
HugoKuo_is there any flag can do that in nova.conf?09:35
HugoKuo_bcz I face same problem on compute node while I run up a instance ..............09:35
HugoKuo_thanks man09:35
*** benbenhappy has joined #openstack09:36
koolhead17soren, yes09:42
koolhead17nova-manage shell ipython/bpyhton/python/09:43
sorenkoolhead17: They correspond to ipython, bpython and python.09:44
koolhead17so its like a choice provided09:45
sorenYeah. Just try them?09:45
koolhead17also nova-manage instance_type and flavor does same operation, if am not wrong09:48
sorenThat's correct.09:49
koolhead17so is instance09:49
koolhead17nova-manage instance09:49
sorenWhat about it?09:50
koolhead17what is it for , wanted to know09:50
sorenIt does stuff to instances. Live migration, for instance.09:51
soren(no pun intended)09:51
koolhead17soren, and nova-manage vm :P09:51
* koolhead17 will ask if nova supporting live migration :D09:52
sorenAh, sorry, apparantly instance only has one subsubcommand: list09:53
sorenvm does stuff to instances, like live migrate them.09:53
koolhead17soren, nova-manage instance09:54
koolhead17instance matched multiple options:09:54
koolhead17        instance_type09:54
koolhead17        instance09:54
uvirtbot`Launchpad bug 753339 in nova ""instance_type" subcommand in nova-manage masks "instance" subcmd" [Undecided,New]09:54
*** benbenhappy has left #openstack09:54
*** uvirtbot` is now known as uvirtbot09:54
* koolhead17 request annegentle to do sumthing with the section of which has wrongly described the creation of a nova project09:55
koolhead17soren, cool so instance and vm will be almost same :)09:57
*** reldan has joined #openstack09:58
koolhead17soren, shall i file the documentation error along with solution in nova-manage project ?09:59
sorenkoolhead17: Sure.09:59
uvirtbotNew bug: #753339 in nova ""instance_type" subcommand in nova-manage masks "instance" subcmd" [Undecided,New]
*** adjohn_ has joined #openstack10:03
*** Guest65968 has left #openstack10:05
*** adjohn_ has quit IRC10:06
*** adjohn has joined #openstack10:09
uvirtbotNew bug: #753352 in nova "Error in Documentation about nova project creation " [Undecided,New]
*** befreax has joined #openstack10:11
*** befreax has left #openstack10:11
HugoKuo_nova-network host must be the gateway of instance ?10:27
HugoKuo_there's only one NIC on my nova-network host (eth0, once nova-network service up , the br100 set to   , how come ?10:30
*** daveiw has joined #openstack10:31
*** rds__ has joined #openstack10:33
koolhead17HugoKuo_, i would suggest check the DB and see whats happening inside it10:34
koolhead17HugoKuo_, are you on lucid?10:35
HugoKuo_I know what's happen in DB  ...... bcz default gateway is first of fixed_range netowkr10:35
n1md4good morning.  interesting problem with my cloud, I have to scp /var/lib/nova/instances/_base/ to each node before the instance will boot, any ideas why?10:35
HugoKuo_I'm on maverick...10:35
HugoKuo_did you try to let nova-network in a single box ?10:35
HugoKuo_I always face this issue while I isolate nova-network in a single host10:36
koolhead17n1md4, you can do that only when you have an instance running. in same directory you will have a "instanceid" directory created with logfile of the specific node10:37
koolhead17HugoKuo_, what n/w mode your using10:37
*** adjohn has quit IRC10:38
koolhead17HugoKuo_, have not checked on that :(10:38
HugoKuo_so that I did not set any bridge manually in /etc/network/interfaces...10:38
HugoKuo_sorry @@10:40
koolhead17HugoKuo_, :P10:40
koolhead17HugoKuo_, suggest me some documentation which explains installation of glance10:45
n1md4koolhead17: Not sure what you mean, what can I only do with a running instance?10:50
koolhead17n1md4, you can ssh your node once it starts running isn`t it?10:51
*** ahmedelgamil has joined #openstack10:52
n1md4koolhead17:  I'm not sure.  I'm assuming the _base image is created after it's published, it's not copying a running instance, but the base image - is my guess.10:54
koolhead17n1md4,  tail -f /var/log/nova/nova-compute.log10:55
koolhead17  this might tell you complete story10:55
n1md4I've been over this quite a bit, and the only 'fix' seems to be copying the image to nodes
*** DigitalFlux has joined #openstack10:57
*** ahmedelgamil has quit IRC10:59
*** lionel has quit IRC10:59
*** ahmedelgamil has joined #openstack10:59
*** lionel has joined #openstack11:00
*** ahmedelgamil has quit IRC11:00
*** reldan has quit IRC11:03
*** nwww has joined #openstack11:09
*** ibarrera has joined #openstack11:12
HugoKuo_koolheadl7: I do not touch glance yet :|11:13
*** omidhdl has joined #openstack11:16
koolhead17soren,  nova-manage log request foo  it outputs notthing  but when i see the logfile it tries to check from /var/log/nova.log  but i don`t have such file everything about nova is inside /var/log/nova/*11:28
sorenYeah, that looks dreadfully broken in so many ways.11:30
*** benbenhappy has joined #openstack11:30
uvirtbotLaunchpad bug 753416 in nova "nova-manage log broken in several ways" [Undecided,New]11:32
*** bkkrw has joined #openstack11:34
*** ctennis has quit IRC11:34
*** morfeas has quit IRC11:40
*** fabiand__ has joined #openstack11:40
uvirtbotNew bug: #753416 in nova "nova-manage log broken in several ways" [Undecided,New]
HugoKuo_guys , in your knowing . should fixed_range netowork different from all hosts ?11:43
*** rsaidan has joined #openstack11:46
*** ctennis has joined #openstack11:54
*** ctennis has joined #openstack11:54
*** morfeas has joined #openstack11:54
*** rsaidan has quit IRC11:58
*** rsaidan has joined #openstack11:58
*** mray has joined #openstack12:02
*** guigui4 has joined #openstack12:05
*** guigui has quit IRC12:06
*** mray has quit IRC12:06
*** mgoldmann has joined #openstack12:11
*** h0cin has joined #openstack12:15
*** h0cin has joined #openstack12:15
*** hggdh has quit IRC12:16
koolhead17soren, we allready have tail -f doing the same :)12:20
sorendoing.... what?12:20
koolhead17for checking logs12:21
sorenSounds like a recipe for insanity. Those logs are pretty verbose.12:21
koolhead17those logs were responsible 4 me to understand many of nova-manage commands :D12:22
*** hggdh has joined #openstack12:26
*** hggdh has quit IRC12:29
*** hggdh has joined #openstack12:30
koolhead17soren, also nova-manage project and nova-manage account does same thing if am not mistaken12:32
sorenkoolhead17: Not sure why you are telling me this?12:38
*** fabiand__ has left #openstack12:38
*** benbenhappy has left #openstack12:38
koolhead17soren, because am trying to use all the available command and after testing the both command came to this conclusion12:39
nagyzis there a way to force a hostname?12:41
nagyzso it shows up right in mysql12:41
sorenkoolhead17: I still don't completely understand why you are telling me about it?12:42
sorenkoolhead17: If you think it's a mistake, you should file a bug.12:43
koolhead17soren, ok. i wanted to confirm it  :)12:43
sorenWe can do that in a bug report, too.12:43
sorenTrust me. I do not scale.12:43
sorenAll bugs should not go through me first.12:43
koolhead17soren, okey.12:44
*** avishay has joined #openstack12:47
*** lionel has quit IRC12:48
avishayHello. Can somebody please help me configure my Swift installation? 'stat' works, but uploads hang, and I can't figure out why from the logs.12:48
*** adjohn has joined #openstack12:50
*** lionel has joined #openstack12:52
*** guigui4 is now known as guigui12:55
*** mgoldmann has quit IRC12:57
*** HugoKuo__ has joined #openstack12:58
*** HugoKuo_ has quit IRC13:01
uvirtbotNew bug: #753477 in nova "nova-manage account and nova-manage project does same " [Undecided,New]
*** ppetraki has joined #openstack13:02
avishayCan somebody please help me fix my Swift installation?13:06
*** reldan has joined #openstack13:11
*** adiantum has joined #openstack13:12
sandywalsh_eventlet must magically monkeypatch itself right into socket. Is that correct?13:13
*** lvaughn_ has quit IRC13:13
*** mray has joined #openstack13:13
*** lvaughn has joined #openstack13:14
*** hadrian has joined #openstack13:20
*** omidhdl has left #openstack13:22
*** Zangetsu has joined #openstack13:23
*** avishay has quit IRC13:23
*** blamar has quit IRC13:25
*** dirkx has joined #openstack13:26
*** alex-meade has joined #openstack13:26
sorensandywalsh_: I don't think it does so if you don't actually tell it to.13:29
sorensandywalsh_: "Telling it to" might be as simple as calling eventlet.monkey_patch(), though.13:29
sandywalsh_soren, thx ... got it figured out. You're correct.13:30
sandywalsh_soren, and yes, we do tell it to :)13:30
*** mray has quit IRC13:31
*** guigui has quit IRC13:31
*** guigui has joined #openstack13:31
*** bcwaldon has joined #openstack13:34
*** kashyap has quit IRC13:40
*** adjohn has quit IRC13:41
*** RickB17_ has joined #openstack13:41
*** f4m8 is now known as f4m8_13:48
*** RickB17_ has quit IRC13:48
*** rchavik has quit IRC13:52
*** allsystemsarego has joined #openstack13:54
*** mdomsch has joined #openstack13:55
*** ironcame12 is now known as ironcamel213:57
*** adjohn has joined #openstack14:01
*** onlany has joined #openstack14:03
onlanydo u know about any limit, how many virtual machines can be managed with openstack?14:04
onlanyfor example vcenter can manage up to 1000 vms14:04
sandywalsh_annegentle, didn't get to docs yesterday, doing it today. :/14:05
sandywalsh_onlany, there is no hard limit. You can add more compute nodes to scale horizontally14:06
sandywalsh_onlany, and we're working on clusters via Zones14:06
*** adjohn has quit IRC14:06
sandywalsh_onlany, but the design is shooting for about 1mm compute nodes of about 60 (small) instances per compute14:06
*** [1]RickB17 has joined #openstack14:07
sandywalsh_onlany, ... and we have a ways to go :)14:07
*** gondoi has joined #openstack14:07
onlanythanks for the answer14:08
nagyznova-api has to run on only one node, right?14:08
nagyzso if I want to communicate with the compute nodes, I should do it by using rabbitmq?14:08
sandywalsh_nagyz, each service has an file that you should use to talk to the service with14:10
sandywalsh_nagyz, you can run as many nova-api nodes as you need14:11
nagyzbut at the end of the day, a nova-api node communicates with the actual nova-compute node(s) by mq, right?14:11
sandywalsh_nagyz, compute/, scheduler/, network/, etc14:11
sandywalsh_nagyz, correct14:11
nagyzok, ty14:11
nagyzI'm actually extending the rest API14:12
nagyzinside nova-api14:12
sandywalsh_which one, OpenStack API or EC2?14:12
nagyzadded a new one :)14:12
nagyzwe'll see if we want to push it back as a patch14:12
nagyzthat decision is not up to me. :)14:12
sandywalsh_is there a spec for it?14:12
sandywalsh_(not that it's critical)14:13
*** j05h has quit IRC14:13
nagyzno, not yet14:13
nagyzwe're evaluating openstack as a platform for TClouds14:13
sandywalsh_ah, not familiar with TClouds ... <looking>14:13
*** kakella has joined #openstack14:13
nagyzit's a research project :)14:14
sandywalsh_looks fun!14:14
nagyzmost of my problems so far are coming from python... I've always been a Java guy. :)14:15
sandywalsh_right ... those can be hard habits to break :)14:15
sandywalsh_enjoy your newfound freedom ;)14:15
nagyzI haven't looked at actually using mq inside the code, but does rabbitmq support some kind of synchronous messaging?14:15
sandywalsh_ping me if you have any questions about the OS API as a comparison14:15
sandywalsh_nagyz, we have an layer that does call-response14:16
sandywalsh_but it's not synchronous14:16
nagyzah, okay. what I'd like to do is basically send a command to each registered compute-node, and wait for all the responses14:16
sandywalsh_(no guarantee of order)14:16
koolhead17nagyz, it wont be that difficult. :P14:16
nagyzis there similar functionality somewhere?14:17
nagyzkoolhead17, as..? :)14:17
koolhead17nagyz, see jython14:17
sandywalsh_nagyz, there is a fan-out exchange per service you can use to broadcast to all services of a class14:17
nagyzkoolhead17, picking up a new language is always fun :)14:17
nagyzsandywalsh_, thanks, I'll have a look14:18
onlanynagyz, szia :)14:18
*** blamar has joined #openstack14:22
*** bcwaldon has quit IRC14:23
*** bcwaldon has joined #openstack14:24
*** j05h has joined #openstack14:26
*** bcwaldon has quit IRC14:26
*** bcwaldon has joined #openstack14:28
creihtjaypipes: morning14:36
*** adjohn has joined #openstack14:37
creihtcan you point me to that bug again that we are working around in babel?14:37
*** mahadev has joined #openstack14:46
*** Zangetsu has quit IRC14:53
creihtjaypipes: and why can't we use these:
*** hub_cap has joined #openstack15:00
*** mahadev has quit IRC15:04
*** Shentonfreude has joined #openstack15:05
*** dragondm has joined #openstack15:05
justinsbsoren: The XML schema have tests nows in the check-xsd branch.  Any chance you could re-review including the schema now please?
justinsbsoren: Sorry, meant now as in "now that has happened" rather than "at this very moment" :-)15:06
justinsbsoren: That'll teach me to IRC pre-coffee15:07
*** mahadev has joined #openstack15:11
*** troytoman-away is now known as troytoman15:11
*** kashyap has joined #openstack15:14
*** onlany has quit IRC15:14
uvirtbotNew bug: #753660 in nova "Cannot work with Nova using RightAWS" [Undecided,New]
*** fysa has joined #openstack15:19
nagyzif I'd want to add a new node to host VMs on, is it enough to run nova-compute on it, or do I need nova-network too?15:19
nagyzI guess nova-network too.15:19
nagyzin services list I see both hosts.15:24
*** alekibango has joined #openstack15:25
*** bkkrw has quit IRC15:26
*** smeier00 has joined #openstack15:28
*** mgoldmann has joined #openstack15:29
*** guigui has quit IRC15:34
jaypipescreiht: we do use those. only problem is they don't work properly with regular setuptools.setup. they work with
jaypipesmtaylor: do you remember the babel bug number we ran into for this?15:36
creihtjaypipes: I'm just trying to better understand the issues involved15:36
creihtso in what ways do they not work?15:37
jk0ttx: what are your thoughts on ? I recall deliberately renaming those a few months back15:37
uvirtbotLaunchpad bug 740605 in nova "OpenStack API v1.0 sends computeFault, not cloudServersFault" [Low,In progress]15:37
jaypipescreiht: absolutely, no worries. unfortunately, I have to fall back on mtaylor's memory for the specific bug.15:37
jaypipescreiht: it was mtaylor who worked it out...15:37
jaypipescreiht: the original code just used babel, but there was something regarding the message compilation process that did not correctly fire *automatically* during build...15:38
creihtjaypipes: do we have a doc somewhere on the expected workflow for this stuff?15:38
ttxjk0: do you have traces of that rename from a few months back ? Doesn't ring a bell15:38
jk0ttx: hm, not offhand. let me try and search for the bug15:38
jaypipescreiht: bit of a black magic thing right now, I know...15:38
jaypipescreiht: when mtaylor's head pops up in IRC today, let's the three of us discuss, ok?15:39
jaypipescreiht: it's coffee time for him right now I think ;)15:39
uvirtbotLaunchpad bug 680285 in nova "rackspace-specific "cloudserversFault" name should be generic" [Undecided,Fix released]15:40
justinsbjk0: I opened that bug, FYI.  If we want to be compatible with CloudServers API clients, you have to follow the XML schema15:40
jk0justinsb: right, I understand. I just want to figure out why we renamed them in the first place15:40
justinsbjk0: I read it as "branding issues"15:41
creihtjaypipes: and is there a reason why we put the .pot files in /po instead of the babel default /locale?15:41
creiht(just curious)15:42
[1]RickB17btorch: you there?15:42
ttxjk0: right, the original bug is that it was branded... but the APi 1.0 is branded15:42
jk0cool, just making sure15:44
jaypipescreiht: not sure about that one, sorry.15:44
*** alex-meade has quit IRC15:44
creihtjaypipes: and what exactly is build_i18n supposed to produce?15:46
creiht(running it doesn't seem to do anything obvious)15:47
*** mahadev has quit IRC15:48
*** adjohn has quit IRC15:51
jaypipescreiht: nothing is obvious because there are no translation files (.po files) in /po yet ;)15:54
ironcamelhey guys. why has nothing been merged today. lets pick it up!15:54
[1]RickB17can anyone point me towards some documentation for troubleshooting SWAUTH?  I was originall using devauth, but as per recommendations i'm trying to setup swauth. and it doesn't appear to be working.  No entries in the syslog, and when i run command like swath-prep it just hangs with no output.15:54
[1]RickB17btw i'm fairly new to swift.15:55
creiht[1]RickB17: swauth-prep is just hanging, and you have no logs, then it sounds like swauth-prep isn't able to connect15:56
jaypipescreiht: so, here's the steps that happen... once I setup LP to import a certain branch (lp:swift) for translation, LP looks for a .pot file (the po/swift.pot). It finds that template file (swift.pot) and takes all the English message definitions out of that and produce a screen for translators where they can translate those strings (see the same for Nova:
creihtis this on an saio, or multi-server install, and can you post how you are running swauth-prep?15:57
creihtjaypipes: right15:57
creihtany way I can get a .po file in there just for testing?15:58
[1]RickB17creiht: what are ways to verify configuration/connectionvity?15:58
jaypipescreiht: once translators have started translating those strings into a language, LP creates a $language.po file and inserts it into the po/ directory. This is then included in the source package for Swift. When build[_i18n] runs, those .po files are compiled into .mo message catalogs and *installed* into the /usr/share/locales15:58
jaypipescreiht: does that make a bit more sense as to why build_i18n doesn't *seem* to do much right now?15:58
creihthehe... just trying to wrap my head around what is supposed to be going on :)15:59
*** Guest89726 is now known as cole16:00
jaypipescreiht: you can create a .po for testing.. yes. you can use the babel commands you reference above. What *does not* happen correctly is the integration of those steps *automatically* when you run build. That's what the is for. mtaylor, can you verify that for me?16:00
jaypipescreiht: and believe me, I understand your frustration. I ripped my hair out trying to get it all working the first time round with Nova...16:01
creihtso what is so magicall about having to have it done through build?  is there a reason why we can't just run the commands manually16:01
jaypipescreiht: there's little to no documentation regarding the holistic process of doing i18n for packaging...16:01
*** hub_cap has quit IRC16:02
*** rcc has quit IRC16:02
jaypipescreiht: good question. I'll have mtaylor answer that. (that's the exact question I asked mtaylor when I ran into the issue the first time). I asked him what I needed to do in the build scripts to have those babel/setup commands run during the build. I believe mtaylor's answer in email was "damn it, why can't we ever have fully baked Python libraries. grrr." He then went off to tinker and came back with the DistUtilsExtra16:03
jaypipessolution ;)16:03
*** adjohn has joined #openstack16:04
nagyzsee ya tomorrow16:04
*** nagyz has quit IRC16:04
jaypipescreiht: yeah, I know :)16:04
*** adjohn has quit IRC16:04
jaypipescreiht: anyway, let's wait to mtaylor shows up before discussing any further. I don't want to dig my hole deeper if I've been describing the problem space incorrectly.16:05
creihtyeah no worries16:05
creihtI should make some lunch anyways16:05
jaypipescreiht: done with your training?16:05
jaypipescreiht: I'm sure you enjoyed it thoroughly.16:05
creihtI'm trying hard to supress those memories16:06
*** j05h has quit IRC16:06
*** hub_cap has joined #openstack16:06
*** dirakx1 has joined #openstack16:07
*** dirkx has quit IRC16:09
*** mahadev has joined #openstack16:09
uvirtbotNew bug: #753724 in swift "_make_request fails to catch timeouts" [Undecided,New]
medberryWhat happens at the design summit? Not much info on the registration site.16:12
*** fabiand__ has joined #openstack16:14
*** j05h has joined #openstack16:15
*** mgoldmann has quit IRC16:15
medberrynm, I found
*** CloudySunshine has joined #openstack16:18
CloudySunshineIs it possible to have three or more locations with more than one storage node per site and have each site replicate each other site only once per zone and not per server?16:20
CloudySunshineI want each server to have a replica of each other server, but I don't want the extra costs in bandwidth when it comes to a multihomed installation16:21
*** Zangetsu_ has joined #openstack16:22
CloudySunshineright now if I was to build three site two servers per site install with glusterfs no matter what I do any write will be commited to all 5 servers regardless of their location16:22
*** mahadev_ has joined #openstack16:22
*** mahadev has quit IRC16:22
*** reldan has quit IRC16:26
*** daveiw has quit IRC16:28
*** omidhdl has joined #openstack16:28
*** ibatanov has joined #openstack16:29
*** zul has quit IRC16:30
*** kashyap has quit IRC16:34
[1]RickB17i did some searching, but cannot find anythign about troubleshooting "Auth subsystem prep failed: 400 Bad Request" from swift-prep command.  no output in syslog.  Any ideads?16:34
[1]RickB17running swift 1.2.016:37
*** reldan has joined #openstack16:37
*** MotoMilind has joined #openstack16:39
*** fabiand__ has left #openstack16:39
*** deepy has quit IRC16:40
*** deepy has joined #openstack16:40
creiht[1]RickB17: are you running the saio, or a multi-server install of swift?16:43
*** kashyap has joined #openstack16:50
*** rnirmal has joined #openstack16:50
[1]RickB17creiht: mulitple, 1 proxy, 1 storage, 1 auth (that is not longer required from my understanding)16:50
mtaylorjaypipes, creiht: morning fellas16:53
mtaylorcreiht: so - IIRC, it wasn't just a problem of babel making the po files - I believe there was also a deficiency in installing them properly16:53
*** rsaidan has quit IRC16:57
*** joearnold has joined #openstack17:03
*** zaitcev has joined #openstack17:03
creiht[1]RickB17: Then the first step would be to check the logs on your proxy server17:04
creiht(assuming you are running swauth on your proxy server)17:05
creihtmtaylor: is there a reason why all the stuff has to get done in the build stage?17:05
*** hub_cap has quit IRC17:05
*** hub_cap has joined #openstack17:06
mtaylorcreiht: well, the main thing is that you want things to install into proper locations when you do python install - and specifically when you do that with an alternate root, you want it to work right17:06
mtaylorcreiht: I mean, if you can get the babel version of things working so that there is a defined repeatable process that I can run in packaging - then great! I just couldn't get that to happen17:06
creihtI'm just trying to wrap my head around everything17:07
mtaylorcreiht: as jay said earlier - my reaction before was much along the lines of "really? there isn't a good story for this?"17:07
mtaylorand then I cried for a little while17:07
creihtmtaylor: so is there a reason why the pot files have to be in the po dir, rather than the default locale dir?17:08
mtayloruh, I don't know what you mean?17:08
creihtwell jaypipes changed it on ours, so I figured there must be a reason17:08
mtayloryou mean in the source tree? I think distutils.extra just expects to find them there17:09
creihtok well let me poke around a bit17:12
*** adiantum has quit IRC17:13
*** mark has left #openstack17:15
*** _vinay has joined #openstack17:15
*** zul has joined #openstack17:15
_vinayI am running the unit-tests in nova/tests dir17:16
_vinaypython run_tests.py17:16
_vinayquestion is where the log level config stored17:16
_vinayI need to reduce the log-level17:16
_vinayIt dumps a lot of DEBUG info17:16
justinsb_vinay: Use ./ instead17:17
creihtmtaylor: so if I understand correctly, launchpad will dump the .po in the dir, but not the .mo17:18
mtaylorcreiht: yes17:18
creihtand you want install to build the .mo automatically17:18
[1]RickB17creiht:  is the only entry in the /var/log/syslog on the proxy server17:19
mtaylorcreiht: yes. (and also to perform any message-merging with a new .pot file if that's needed in between merging new .po files from launchpad)17:19
creiht[1]RickB17: can you paste the command you are running to run the prep?17:19
creihtmtaylor: ok so on install, you want the babel update_catalog and compile_catalog commands to be run17:20
*** nacx has quit IRC17:20
*** aliguori has quit IRC17:22
creiht[1]RickB17: Are you running the prep on the proxy server, or on a different server?17:24
vishysoren: ping17:25
[1]RickB17creiht: proxy17:26
*** adiantum has joined #openstack17:28
creiht[1]RickB17: what version of the code are you running?17:28
[1]RickB17creiht: 1.2.017:29
[1]RickB17from the repos17:29
creiht[1]RickB17: can you paste your proxy config?17:30
*** zul has quit IRC17:30
*** mdomsch has quit IRC17:31
_vinayjustinb : same thing17:31
_vinayI get a log file run_tests.log which has lot of DEBUG logs in it17:31
creiht[1]RickB17: and things were working before with devauth?17:32
*** dirkx has joined #openstack17:32
*** zul has joined #openstack17:33
creiht[1]RickB17: in your proxy config, your default_swift_cluster should be:17:34
creihtdefault_swift_cluster = local#
creihtwithout the extra stuff on it17:34
openstackjenkinsProject nova build #779: SUCCESS in 2 min 29 sec:
openstackjenkinsTarmac: Remove <addressSet> and <item> from AllocateAddress response, and fix bug #751176.17:34
uvirtbotLaunchpad bug 751176 in nova "AllocateAddress returns incompatible tags" [Medium,In progress]
creihtnot sure if that would mess up the prep, but may cause other issues down the road17:34
creihtgotta run to the dentist, bbl17:35
*** Ryan_Lane has joined #openstack17:36
*** irahgel has left #openstack17:36
jaypipescreiht: if would be awesome if we could have Nova and Swift be the same when it comes to handling i18n...17:40
jaypipescreiht: it's a pain for those of us who typically have to do the translation/i18n stuff if there's different processes/'s for each project. Makes documentation easier if it's all the same.17:41
*** mahadev has joined #openstack17:41
[1]RickB17creiht: ok i will try that change.  THanks17:42
jaypipescreiht: so since Nova uses /po, I would prefer that Swift also use /po to store the po files.17:42
*** mahadev_ has quit IRC17:42
*** rds__ has quit IRC17:43
*** reldan has quit IRC17:43
*** slyphon has joined #openstack17:44
creihtjaypipes: I agree with having things being similar, just not real happy with how it looks right now, and trying to figure out if there is a better way17:45
jaypipescreiht: can you explain why you don't like how it looks?17:46
[1]RickB17creiht: i think it's heading in the right direction, but
*** hub_cap_ has joined #openstack17:47
*** adiantum has quit IRC17:48
jaypipes[1]RickB17: think we'll have to talk to creiht post-dentist ;)17:50
*** hub_cap has quit IRC17:50
*** hub_cap_ is now known as hub_cap17:50
[1]RickB17jaypipes: thats true lol17:52
*** hub_cap has quit IRC17:54
*** hub_cap has joined #openstack17:54
*** rds__ has joined #openstack17:55
*** hggdh_ has joined #openstack17:55
*** hggdh has quit IRC17:55
*** imsplitbit has joined #openstack17:56
*** CloudySunshine has quit IRC18:01
*** dlafferty has quit IRC18:03
*** mahadev has quit IRC18:09
*** mahadev has joined #openstack18:10
*** kapil_ has joined #openstack18:13
*** hazmat has quit IRC18:13
*** zigo-_- has quit IRC18:14
*** kapil_ is now known as hazmat18:14
*** lionel has quit IRC18:14
*** hazmat has quit IRC18:15
*** hazmat has joined #openstack18:15
*** zigo-_- has joined #openstack18:15
*** lionel has joined #openstack18:15
*** joearnold has quit IRC18:16
vishyquiet in here today18:18
* eday drops a pin18:18
vishyit's like a sprint commercial18:19
*** Ryan_Lane has quit IRC18:19
*** dirkx has quit IRC18:21
zulmore like a verizon comerical18:22
zulcan you hear me now...goood18:22
kpeppleannegentle: you around ?18:26
*** aliguori has joined #openstack18:30
*** Ryan_Lane has joined #openstack18:34
*** smeier00 has left #openstack18:35
*** fabiand__ has joined #openstack18:36
zigo-_-How do I delete an uploaded ami?18:39
openstackjenkinsProject nova build #780: SUCCESS in 2 min 29 sec:
openstackjenkinsTarmac: fixed the way ip6 address were retrieved/returned in _get_network_info in nova/virt/xenapi/vmops18:39
kpepplezigo-_-: euca-delete-bundle maybe ?18:42
*** reldan has joined #openstack18:44
zigo-_-kpepple: What do I give as params?18:44
kpepplezigo-_-: -b <BUCKET>  and maybe imagename ? the man page is pretty hazy ...18:46
*** bcwaldon has quit IRC18:47
zigo-_-I tried that, and it did NOTHING... :(18:47
*** reldan has quit IRC18:47
*** fabiand__ has quit IRC18:47
kpepplezigo-_-: hmmm ... dunno on that one18:48
zigo-_-Does it takes time to do?18:48
*** Ryan_Lane has quit IRC18:48
*** bcwaldon has joined #openstack18:49
kpepplezigo-_-: it shouldn't ... try using the --debug option ... it might give you some insight18:50
*** rds__ has quit IRC18:51
*** MarcMorata has quit IRC18:54
kpeppledoes anyone have a code snippet about how i can access the specific sql-alchemy error out of the exception.DBError object ? I'm trying to make nova-manage a bit less cryptic where we are trying to insert duplicate keys into the DB.18:55
*** Ryan_Lane has joined #openstack18:55
*** fabiand__ has joined #openstack19:04
*** hub_cap_ has joined #openstack19:04
*** rackerhacker has quit IRC19:05
*** hub_cap has quit IRC19:06
*** HouseAway is now known as AimanA19:06
*** spectorclan has joined #openstack19:06
*** pandemicsyn has quit IRC19:07
*** notmyname has quit IRC19:08
openstackjenkinsProject nova build #781: SUCCESS in 2 min 28 sec:
openstackjenkinsTarmac: Bypass a potential security vulnerability by not setting shell=True in, using johannes.erdfelt's patch.19:09
*** coreym has joined #openstack19:10
*** pandemicsyn has joined #openstack19:10
*** ChanServ sets mode: +v pandemicsyn19:10
spectorclanHas anyone tried any open source Webinar tools like WebHuddle or OpenMeeting?19:11
coreymanyone familiar with setting up swift all in one?19:11
*** reldan has joined #openstack19:11
*** notmyname has joined #openstack19:12
*** ChanServ sets mode: +v notmyname19:12
*** fabiand__ has quit IRC19:13
btorch_coreym: sure what's up ?19:14
*** btorch_ is now known as btorch19:14
annegentlekpepple: ayup?19:15
kpeppleannegentle: quick doco question -- i added 'nova-manage vm' to the nova-manage man page ... anywhere else that these changes need to go ?19:15
*** Ryan_Lane is now known as Ryan_Lane|food19:16
*** hggdh_ is now known as hggdh19:17
*** rackerhacker has joined #openstack19:17
annegentlekpepple: great, thanks for doing that. There isn't a comprehensive command-line guide in the docs, but nova-manage commands are sprinkled in other how-tos. What does nova-manage vm do exactly? (Ha, I could read the man page!)19:18
kpeppleannegentle: live-migration and list (running instances)19:18
openstackjenkinsProject nova build #782: SUCCESS in 2 min 26 sec:
openstackjenkinsTarmac: adds a timeout on session.login_with_password()19:19
*** dprince has joined #openstack19:19
dprincejustinsb: Just removed the Fkey relationship on the id column in InstanceTypes.19:20
annegentlekpepple: ok, there's a live-migration information in the openstack-manuals project in doc/source/computeadmin.xml19:20
dprincejustinsb: That code ran functionally for me with MySQL either way.19:20
kpeppleannegentle: ok, i'll take a look there and propose a change if needed. thx.19:20
annegentlekpepple: but Kei already included the nova-manage vm command in there19:20
kpeppleannegentle: oh, ok.19:20
annegentlekpepple: so yeah, it might already be in the live migration doc19:20
*** omidhdl has left #openstack19:21
annegentlekpepple: wow, mark this day down in history, it's already documented? Crazy.19:21
annegentlekpepple: just merged that in this week, but still :)19:21
kpeppleannegentle: i'd mark it down as a win anyway :)19:22
annegentlekpepple: :)19:22
*** reldan has quit IRC19:22
*** hub_cap has joined #openstack19:22
*** mahadev has quit IRC19:22
*** joearnold has joined #openstack19:23
justinsbdprince: Thanks.  Do you know how to do a downgrade?  Should I just wipe my DB and start again?19:23
*** mahadev has joined #openstack19:23
*** fabiand__ has joined #openstack19:24
dprincejustinsb: something like:19:24
dprincejustinsb: python ./nova/db/sqlalchemy/migrate_repo/ downgrade 13 --repository=./nova/db/sqlalchemy/migrate_repo/ --url=sqlite:///tests.sqlite19:24
_vinayHi kpepple19:24
vishyzigo-_-: are you looking for removing the image? euca-deregister-image19:24
_vinayI am looking for logging config in nova source code19:25
_vinayany pointers19:25
dprincejustinsb: You'd have to use the correct Mysql string in place of sqlite.19:25
justinsbdprince: Cool - thanks!19:25
_vinayspecifically when I run tests (using I get lot of DEBUG from different modules19:25
_vinayI want to cut that down19:25
vishykpepple: for db error err.inner_exception will give you the original exception19:25
dprinceActually. I'm a bit dissapointed that all of the migrations don't actually support downgrades. Should I file a bug on that? I mean we should always support that right?19:25
_vinayso looking for cnfig where that is stored19:26
_vinaylike logging.config ??19:26
vishyjustinsb: I don't think we're going to swing opinion on the api validation in cactus19:26
justinsbdprince: No harm in filing a bug, but I think that there have been a few IRC chats and we're not likely to require downgrades19:27
vishy_vinay: there are a few logging related flags19:27
vishy_vinay: what are you trying to do?19:27
justinsbvishy: It's your call.  With great responsibility comes great power :-)19:27
dprincejustinsb: Sure. That would appear to be the case. In general they aren't too difficult to do though.19:27
vishy_vinay: ah i see now19:27
vishy_vinay: you could set debug level to info19:28
*** MotoMilind has quit IRC19:28
*** MotoMilind has joined #openstack19:28
justinsbdprince: I think we should try to have a best practice document at least for migrations.  But jay pipes has done the most on this and I believe the deeper you go the more you hate SA migrations...19:29
vishy_vinay nova/tests/ -> change FLAGS.verbose = True to False19:29
vishyjustinsb: I've already resigned to the fact that cactus v1 is not going to be deployable at scale using os api, but I think we can get it in early diablo and have a usable set of packages.19:30
_vinayI changed it to False.. but still same level of verbosity19:31
*** hub_cap has quit IRC19:31
_vinay2011-04-07 05:35:33,117 DEBUG nova.virt.xenapi.fake [-] Calling getter VM.get_record from (pid=5918) __getattr__ /home/vinagupta/nova-src/nova/nova/virt/xenapi/ 2011-04-07 05:35:33,117 DEBUG nova.virt.xenapi.fake [-] Calling getter VM.get_record from (pid=5918) __getattr__ /home/vinagupta/nova-src/nova/nova/virt/xenapi/ 2011-04-07 05:35:33,117 AUDIT nova.auth.manager [-] Deleting project fake 2011-04-07 05:319:31
_vinayHere are last few lines19:31
*** hub_cap has joined #openstack19:31
*** hadrian has quit IRC19:32
openstackjenkinsProject swift build #235: SUCCESS in 28 sec:
openstackjenkinsTarmac: Make it so the proxy's _make_request method's error handler catches timeouts.19:32
justinsbvishy: I think that's sad.  I think we could focus 100% on OS v1.0 with minimal extensions and get there in a week.19:32
justinsbvishy: But I agree with you - I'm looking to build something on OS and the only sensible approach is to use the EC2 APIs19:33
vishyjustinsb: Perhaps we should discuss a delay to see if we can get OS api up to snuff.19:33
vishyttx: ping19:33
justinsbvishy: Sounds like a good idea, though I think we could even do it within the existing framework...19:33
_vinayvishy: any idea about my question .. anywhere else I need to look19:34
justinsbvishy: Any change that doesn't touch anything outside of the Openstack API should be easy to get through the exception system19:34
*** mgoldmann has joined #openstack19:34
justinsb_vinay: I don't understand... when I run tests, I only see the green OK, unless a test fails, in which case I want all the debug output I can get...19:35
openstackjenkinsProject nova build #783: SUCCESS in 2 min 28 sec:
openstackjenkinsTarmac: Renamed computeFault back to cloudServersFault in an effort to maintain consistency with the 1.0 API spec. We can look into distinguishing the two in the next release. Held off for now to avoid potential regression.19:35
vishy_vinay: hmm apparently verbose is set manually in some of the tests19:35
_vinayjustinsb: but surely you get a run_tests.log file .. correcT?19:35
justinsbvishy, _vinay: I think all of them - in fake_flags19:35
_vinaythat one has lot of logs which I am trying to reduce19:36
vishyjustinsb: yes he just turned that one off19:36
justinsb_vinay: I've never looked at the logfile... what are you trying to accomplish?19:36
vishy_vinay: you know that you can run just one test at a time?19:36
justinsbvishy: Right sorry, just read back through the discussion19:36
_vinayI am trying to understand the code and the flow of things19:36
vishyfor example: test_cloud:CloudTestCase.test_describe_instances19:37
_vinayvishy : yes I am actually running just one test at a time19:37
vishy_vinay: which test are you running?19:37
_vinayeven just one test produces hundreds of debug log19:37
_vinayI am running VMTest19:37
justinsb_vinay: I think reading the Python code is a better way to understand what's going on19:38
_vinayjustinsb yes.. but also I want to understand how logging is done :)19:38
_vinayhence need to solve this puzzle19:39
vishy_vinay: sigh, it looks like 8 of the files set verbose to True at import, so until that is changed setting it in the fake_flags won't work19:39
vishy_vinay: you could manually set it back to false in setUp for the test you are interested in.19:39
justinsb_vinay: Well, I think that maybe a good way to understand logging and understand the code at the same time may be to fix the logging problem :-)19:39
_vinayvishy: oh... which 8 files are those.. I guess I can find out... grep etc19:39
vishyyes try this19:39
jaypipesjustinsb, dprince: feel free to look at Glance for examples of migration stuff. I think I know where most of the sharp edges are now at least.19:40
vishygrep -r FLAGS.verbose nova/tests19:40
_vinayvishy, justinsb got it... will try this and let you know.. thanks a lot19:41
jaypipesvishy: besides the volumes API, what are the specific things you are referring to that are currently preventing the OS API from being deployable at scale?19:42
*** reldan has joined #openstack19:42
*** adiantum has joined #openstack19:43
*** zul has quit IRC19:45
*** zul has joined #openstack19:45
*** mray has joined #openstack19:46
westmaasvishy: what exactly is missing from the v1 spec that you need?  we have a bunch of bug fixes in play for them already19:46
westmaasjustinsb: ^^19:46
*** hub_cap_ has joined #openstack19:47
westmaasI guess what I mean is, if v1.0 was fully functional, would that be sufficient?19:47
vishyjaypipes: I'm concerned about the stuff that justinsb's validation is picking up19:48
justinsbwestmaas, jaypipes: I think the bug fixes that the titan (?) team are working on for the 1.0 API are great19:48
justinsbThe OS v1.1 schema doesn't even load in lxml19:49
justinsbSo I think that's not worth pursuing in Cactus19:49
justinsbI think we should focus entirely on the OS v1.0 API, which should then be compatible with CloudServers19:49
justinsbI would like volumes, but I can probably figure that out separately19:50
justinsbMy biggest concern is areas where the v1.0 API isn't quite right19:50
justinsbMany of which are being addressed19:50
justinsbBut I know there are lingering issues vs the schema19:50
justinsbe.g. ...19:50
creiht[1]RickB17: looks like I told you wrong sorry :/19:51
justinsbOne of the tests uses RESIZE-CONFIRM.  I haven't narrowed it down yet.  To be honest, I'm just trying to get schemas in there.  Once the schemas are in there, the domain expert can easily look at the issue19:51
justinsb(RESIZE-CONFIRM is not a valid v1.0 state - VERIFY_RESIZE is though)19:52
*** hub_cap has quit IRC19:52
*** hub_cap_ is now known as hub_cap19:52
ttxvishy: pong19:52
creiht[1]RickB17: or maybe not, so what does your cluster url look like now?19:52
jaypipesvishy: can you point me to a link for that validation? is there a reason why issues can't be fixed as bugs in Cactus?19:53
creiht(from your proxy config)19:53
justinsbjaypipes: It's the check-xsd branch19:53
westmaasjustinsb: so that one in particular seems easy to fix19:53
jaypipesjustinsb: sorry, I thought vishy was referring to the 1.0 API? not the 1.1 API.19:53
vishyttx: we are just discussing the possiblity of delaying cactus slightly to make sure all of the 1.0 os api bugs are fixed19:53
westmaaswe have a few proposals in to tie off a few more of these thigns19:53
*** hub_cap has quit IRC19:54
justinsbwestmaas: Yes, you're squashing a lot of these bugs :-)19:54
vishywestmaas: what is your level of confidence in the 1.0 api for cactus19:54
creihtjaypipes: so back to the i18n stuff, I don't like that a.) we have to add yet another dependency to make things work, and it all seems kinda hackish to me19:54
justinsbjaypipes: I think we are talking about 1.019:54
justinsbjaypipes: The direct link is
jaypipescreiht: do you have an alternative solution?19:54
creihtguess there should have been a b.) in there19:54
ttxvishy: hrm. It needs more than a few days to be fixed. My curent estimate is ~2weeks19:54
jaypipesjustinsb: thx19:54
creihtjaypipes: that's what I'm trying to figure out :)19:54
*** _vinay has quit IRC19:54
ttxvishy: define "slightly"19:54
westmaasvishy: confidence in correctness?19:54
vishyttx: that is what I'm attempting to determine19:55
creihtnow that I have my head halfway wrapped around what the problem is, I can poke at it a bit more19:55
vishywestmaas: will we be compatible with client libraries using 1.0 api?19:55
jaypipescreiht: ok, I'll just let you do it then. note that it works perfectly fine for Nova...19:55
vishywestmaas: do we need more time to validate against schema and find remaining bugs? etc.19:55
creihtfor some measure of working ;)19:56
ttxvishy: this is not a light decision, since it affects the other projects as well.19:56
creihtI also don't like all the warning/error messages that get printed out19:56
jaypipescreiht: I was only trying to help Swift get i18n'd...19:56
justinsbvishy: I don't think we're very far away on v1.0, particularly if we're willing to prioritize it as we did with renaming computeFault -> cloudServersFault19:56
creihtjaypipes: oh I understand... not trying to criticize19:56
blamarvishy: doesn't the nova command line client use the 1.0 api?19:56
creihtto me it just seems like there should be a better way19:57
vishyttx: I understand, I'm just trying to determine what we are releasing here.19:57
creihtif there isn't then I guess this is what we have to do19:57
jaypipescreiht: if you find a better solution, pls do let us know. I'll port it over to Nova if it fixes the issues you have with it.19:57
vishyI haven't been using the os api beyond a few simple tests, so I'm depending on westmaas and justinsb for info on how close it is.19:57
westmaasvishy: you mean the Cloud Servers client libraries?  or the Open Stack libraries19:57
creihtjaypipes: of course :)19:57
*** hadrian has joined #openstack19:58
vishywestmaas: I was referring to the old libraries, but I'm interested in both.19:58
westmaasvishy: there are workign python and ruby bindings for openstack right now19:58
justinsbRemember also that just because the CS client library works, that doesn't mean that the app that uses it will work19:58
justinsbSo e.g. if the flavorId was supposed to be an integer, but was a string, the binding might tolerate it19:58
justinsbBut the app using that binding could still blow up19:59
ttxvishy: ok... from what I've seen, only a handful of devs care about it, and we keep finding bugs faster than we fix them.19:59
justinsbI have a Java binding that is close to working19:59
ttxvishy: I'm also interested in better data on how far we are19:59
vishyjustinsb, westmaas: if we can give users a client library that we know works, then I think we can focus on tightening up the discrepencies post release and put them in a point release.19:59
justinsbvishy: True, but it won't automatically be compatible with CloudServers20:00
justinsbI do think we're pretty close - once Dan's flavorId patch gets in, I might be able to create a server :-)20:00
westmaaswe are also finding places where bindings match cloudservers implementation...where cloudservers impl doesnt match the spec20:01
westmaaswhich is fun20:01
justinsbwestmaas: That is a joy.  Like what sort of thing?20:01
westmaasI'm not sure, I'm repeating what dprince says :D20:01
westmaasI think its small.20:01
vishyjustinsb: maybe the best approach is for you to just file any bugs you find while getting your validation / bindings working20:01
justinsbvishy: I have been doing that, and filing fixes as well20:02
creiht[1]RickB17: when running swauth-prep try dropping the /v1/ in the url20:02
justinsbvishy: Though we don't have a valid 1.1 schema, which is beyond annoying20:02
vishysounds like most of them are small, so we will just have to prioritize them post gamma?20:02
ttxwe can make them rc bugs, yes20:03
ttxhopefully some other people will help fixing them20:03
justinsbI think if we get dprince's fix for flavorId in there today, then we can probably make other fixes small enough20:03
westmaasvishy: ruby bindings will be fixed this sprint, hopefully before cactus is out20:03
westmaasvishy: python bindings are out and mostly working20:04
justinsbBut why not just import the schema, run the test and fix the problems?  We'll find lots of bugs quickly.  It just feels to me like we're making this overly hard for ourselves by not using the schema we have20:04
blamarfinding and fixing a lot of bugs does not a stable release make?20:04
[1]RickB17creiht:    this is what i'm using already         swauth-prep -K password! -A
justinsbblamar: I don't understand what you mean there20:05
creihttry swauth-prep -K password! -A
creiht(drop the v1/)20:06
[1]RickB17gotcha sorry, missread20:06
creihtsorry it has been a pain to setup20:06
[1]RickB17creiht: Auth subsystem prep failed: 500 Server Error20:06
creihtok, so now there should be an error in the proxy log20:06
blamarjustinsb: Fixing bugs is great for accuracy, but the more code you change closer to a deadline the more likely you're going to introduce other, potentially worse issues? Sometimes the devil you have is better than the devil you don't know.20:07
[1]RickB17posting now20:07
*** kakella has quit IRC20:07
vishyjustinsb: I agree with you in theory, but I think we should defer to the guys working on the bugfixes.  If it complicates their workflow then I think they have a right to request that it goes in after cactus.20:07
*** kakella has joined #openstack20:08
justinsbvishy: It won't even touch their workflow unless they enable the flags20:08
justinsbvishy: There are enough failures that I couldn't make it mandatory20:08
jaypipesjustinsb, vishy: just read through the add and check-xsd merge props. So, can I ask, is it possible to just focus on the API 1.0 schema for Cactus? I would definitely support adding the 1.0 CS schema file and any tests against that schema. Not sure I would support doing the same for the 1.1 schema I see added in the check-xsd branch. Thoughts?20:08
creiht[1]RickB17: looks like it got 503 errors on the storage servers trying to create the account, so now we need ot look at the logs on your storage server to figure out why it returned a 50320:09
[1]RickB17k i'll check those logs20:09
creiht[1]RickB17: did you say your system was working when you were using dev auth?20:09
justinsbjaypipes: I'm definitely fine with that.  The 1.1 schema doesn't even work20:09
[1]RickB17btorch: walked me throuhg changing the proxy config to use swauth20:09
*** reldan has joined #openstack20:09
[1]RickB17but we never got it functioning20:09
ttxvishy: I think we can still do gammafreeze and Gamma milestone tomorrow, and use the RC bugs mechanism to prioritize RC bugs.20:09
jaypipesjustinsb: shall I add that in a comment on the check-xsd MP?20:09
ttxvishy: if we need an extra day next week to get the fixes in, we cna decide that on Tuesday night meeting20:10
justinsbjaypipes: Sure!   Though if we're not even going to merge 1.0, there seems little point20:10
justinsbI added the v1.1 schema when I added the validation, so that it could load the correct schema20:11
jaypipesjustinsb: oh, I would support merging add-xsd. I don't agree with it being problematic for Cactus.20:11
justinsbLittle did I realize the 1.1 schema is borked20:11
[1]RickB17creiht: only entries on the storage node are
westmaasjustinsb: so far the only 1.0 the XSD has found is the VERIFY_RESIZE?20:11
ttxvishy: if we are onto an unlimited pile of bugs, I don't see the point of delaying, since that means we'll never release.20:11
*** pothos_ has joined #openstack20:12
*** fabiand__ has quit IRC20:12
vishyttx: sounds good20:12
westmaasjustinsb: the only 1.0 bug*20:12
justinsbwestmaas: I can try to get you a more complete list.  The problem is that it's not able to validate anything that hits 1.1, because the 1.1 schema is broken20:12
creiht[1]RickB17: it almost looks like the proxy never talks to the storage node20:12
justinsbwestmaas: And I'm glossing over the ones that are being worked on, of course!20:13
westmaasjustinsb: the 1.1 XSD?20:13
westmaashaha sure20:13
*** pothos has quit IRC20:13
[1]RickB17theres no firewall enabled on either of them20:13
creiht[1]RickB17: which server is running on ip .33?20:13
creihtyour proxy?20:13
justinsbwestmaas: Yes, the 1.1 XSD has problems... I'll find you the bug report20:13
[1]RickB17storage node20:14
[1]RickB17.31 is proxy20:14
[1]RickB17.32 was auth20:14
creihthrm, why are those requests showing up in the proxy log then?20:14
uvirtbotLaunchpad bug 752929 in nova "XML Schema for OS API 1.1 does not load" [Undecided,New]20:14
creihtnotice the top 3 lines?20:14
*** fabiand__ has joined #openstack20:14
*** pothos_ has quit IRC20:15
[1]RickB17to wich post?20:15
edayvishy, ttx: I could be missing something here, but I don't see why OS v1.0 API should delay anything. Folks using nova will still choose ec2 over OS 1.0 due to the large feature gap. The only folks who really care about 1.0 is rackspace for backwards compatibility for cloud servers (and they can run a 1.0 branch they finish after cactus if they are deploying before diablo release). It's still very important, but nothing that should postpone a community release.20:15
*** Ryan_Lane|food is now known as Ryan_Lane20:15
*** pothos has joined #openstack20:15
[1]RickB17 ?20:15
creiht[1]RickB17: yeah, in that paste the top 3 lines are PUTs that should have gone to the storage server20:15
creihtbut that makes it look like it was trying to send them to the proxy20:15
[1]RickB17.33 is the storage server20:15
vishyeday: I'm trying to get a usable os api.  Even if it is missing some features (although i think we have volumes which is most of it)20:16
* creiht scrolls to the right some more20:16
creihtmain point is the connection refused20:16
[1]RickB17mcould it have somehting to do with my rings?20:16
creihtit can't talk to the storage server20:16
[1]RickB17i was unclear on how the ports works20:16
[1]RickB17in the docs it isn't that clear20:16
[1]RickB17when setting up the zones20:17
*** hub_cap has joined #openstack20:17
creihtwell usually you set up everything on more than one server :)20:17
ttxeday: I agree with you. I think we can use the RC targeting to get the top bugs squashed.20:17
[1]RickB17i have not problem bringing up some addtional storage nodes if thats the case20:17
ttxeday: I really don't want to delay the release.20:17
[1]RickB17since i wa sin the start of testing here i wanted ot use as little reasoures as possible though20:17
creihtin teh saio docs, we have the service running on different ports just so that we can simulate several servers running20:18
justinsbeday: I'm starting a new project that works on nova, and I would be disappointed if I had to use the EC2 API.  But you're right, perhaps I should just suck it up...20:18
creiht[1]RickB17: yeah now worries, we can make it work this way20:18
edayvishy: well, that's talking about v1.1, no? I thought we already punted there for cactus due to things like keypairs, vpns, floating IPs, etc20:18
creihtin swift the zones are just a way of isolating failure20:18
creihtthey are very abstract20:18
edayjustinsb: or use os v1.1, and just run it against Nova trunk once it's ready?20:19
[1]RickB17what about when setting them up you specifiy the ip:port  how is the port defined on the storage node?20:19
*** fabiand__ has quit IRC20:19
edayjustinsb: is there any reason you need your project to use the cactus tarball?20:19
ttxvishy: let's resync on that at the very end of your day, when mine begins.20:20
creihtthe storage service is going to run on some port (by default 6000)20:20
vishyttx: ok20:20
justinsbblamar: I take your point about late fixes.  I don't think we really have a choice20:20
creihtat a bare minimum just to try things out, you will need at least 3 zones, with at least 1 partition in each of those zones20:20
ttxjustinsb: if we "released" more frequently, would you care as much ?20:20
justinsbeday: Just the normal reasons for preferring a stable release20:20
vishyI think at this point i'm targeting cactus.220:20
[1]RickB17thats the account.builder output20:21
justinsbttx: That's not an easy question.  I'm wearing a couple of hats here as well.20:21
[1]RickB17at first i had all the same ip:port, but i seen some docs that didn't match it they all had different ports, but same ip20:21
creihtok then on the storage node, do you have 3 separate object/account/container servers running?20:21
justinsbttx: As someone building a project on nova, I want a single stable release that has everything in it and works perfectly20:21
[1]RickB17i don't believe so20:21
justinsbttx: Which is obviously impossible20:22
[1]RickB17shoudl i put it back so that they are all the same port?20:22
justinsbttx: But if you're going to do a release, then I want it to work.20:22
blamarjustinsb: Right. So what we're trying to define is "work"20:22
*** jonkelly has joined #openstack20:22
justinsbttx: If the OS API doesn't work yet, I know it's going to be your primary API going forward20:22
creiht[1]RickB17: can you paste the output of "ps -ef | grep swift" on the storage node20:22
creiht[1]RickB17: how you set it up depends on what your goal is20:23
justinsbttx: So if I was doing this at arms length I'd still put Nova into the "too early" category20:23
*** omidhdl has joined #openstack20:23
justinsbttx: As a developer on nova, I just think it's silly that we're not validating our schemas20:23
creihtif this is just to try out and see how things works, then installing from the SAIO instructions on a single machine is the easiest20:23
justinsbblamar: Well, the goal is to have people using it, right?20:23
ttxjustinsb: I think a release is what we make of it. Most people didn't care about the OSAPI, leaving a handful of devs to care about it. One week to release is a bit late to realize it's buggy.20:24
[1]RickB17well i really want to show the ability to expand20:24
justinsbblamar: Would you tell your friends to use the EC2 or the OS API?20:24
[1]RickB17thats why i went with the multinode setup20:24
*** spectorclan has quit IRC20:24
blamarjustinsb: Friends? Aren't I a developer? ;)20:24
westmaasalso I don't allow my guys to go home20:24
creihteven though you have several servers, you really have one node :)20:24
westmaasso yeah20:24
justinsbblamar: :-)  I just mean someone that isn't working on nova!20:24
*** dirkx has joined #openstack20:25
* creiht looks at the last paste20:25
[1]RickB17but i'm intending to add a physical one to the mix20:25
[1]RickB17and i want to go through that process20:25
*** bkkrw has joined #openstack20:25
[1]RickB17i am workign on setting up a standalone one on the side of this jsut incase this is fubar20:25
ttxjustinsb: If we really cared (as a dev group) about OSAPI, we would have helped out Titan team way earlier.20:25
justinsbttx: We've put a huge amount of effort into the OS API, and I think we have the opportunity to have it work much better if we push things along20:25
creihtso for your ring then20:25
[1]RickB17but i'm hoping it's not, since i am learning alot by going through this process20:25
creihtI would change the ports20:26
creihtto the defaults20:26
[1]RickB17just remove the accoutn.builder file and recreate it?20:26
creiht6000 for object, 6001 for container and 6002 for account20:26
justinsbttx: Fair enough20:26
creiht[1]RickB17: yeah that is probably the easiest20:27
justinsbttx: Perhaps we should have a discussion at the design summit about this20:27
vishyjustinsb, westmaas, blamar: lets get as much fixed as we can in the next week.  We'll discuss progress on tuesday20:27
justinsbttx: At school we would call them "post-mortems", though I don't think it's that bad :-)20:27
westmaasvishy: sounds good20:28
justinsbblamar: Can you help me better understand why you're opposed to validating against the schemas?20:28
*** mahadev_ has joined #openstack20:28
westmaasbunch of things waiting for review already :D20:28
blamarjustinsb: Woah woah, not against validating XML.20:28
blamarjustinsb, I'm against adding a feature to Cactus in the Bug fix freeze.20:28
westmaasjustinsb: did you already file a bug for what you found?  I haven't gone to look.20:29
justinsbblamar: Though I can make it so that it is never triggered in production20:29
justinsbblamar: If that isn't the case already20:29
*** mahadev has quit IRC20:29
blamarjustinsb: And I'll say again, it's a feature IMO. Unnecessary for the release of Cactus.20:29
justinsbwestmaas: I did some MPs, and filed a bug on the 1.1 schema being broken20:30
[1]RickB17creiht: okay now swauth-prep worked.  atleast the command completed and there was not output20:30
westmaasjustinsb: thanks :)20:30
*** bkkrw has quit IRC20:30
[1]RickB17creiht: do you have any recommended reads for understanding how accounts work?20:30
[1]RickB17i really appreiate your help in this20:31
creiht[1]RickB17: something like this?
creihtand no problem :)20:31
*** dirkx has quit IRC20:32
[1]RickB17k thanks again. my ultimate could is to get a container mounted with cloudfuse, thats the original reason i switched to swauth.20:32
justinsbblamar: It doesn't add any functionality, other than things working better20:33
justinsbblamar: By virtue of helping us find more bugs20:33
*** dprince has quit IRC20:33
sorenvishy: Is it quick?20:33
blamarjustinsb: I would classify a commit which allows us to find more bugs a 'feature'.20:33
vishysoren: I was just pinging about the change20:34
vishysoren: lo works20:34
vishysoren: so changed the branch20:34
* vishy was surprised to find out that lo works.20:34
*** mahadev_ has quit IRC20:35
vishyjustinsb: want to flip your review here?
vishylooks like your concerns were addressed...20:35
justinsbvishy: Yes, just running it in production now.  Will flip back to Approve.20:36
*** kashyap has quit IRC20:36
*** mahadev has joined #openstack20:36
justinsbvishy: Thanks for the reminder20:36
sorenvishy: Cool.20:38
sorenvishy: Less stuff to configure ftw.20:38
*** rnirmal has quit IRC20:39
* ttx disappears, back in ~10h20:40
creihtmtaylor, jaypipes: ok so I have a proposal20:43
*** big_d has joined #openstack20:43
*** big_d has left #openstack20:44
[1]RickB17now that I'm using swauth what is a valid auth_url?  https://<PROXYIIP>:8080/auth/v1 ?20:44
creihtfrom the research that I have done, most projects expect that the translations already be in the package20:44
creiht[1]RickB17: yes20:44
[1]RickB17so no longer port 11000, correct?20:44
creihtso for a release, I would suggest we build the .mo files for the current translations through babel20:45
creiht[1]RickB17: yes20:45
creihtthen ensure that installs the po folder with the rest of the python code20:45
*** bkkrw has joined #openstack20:46
creihtand change our code that runs gettext.install to point the localedir to where the po files are20:46
*** Ryan_Lane has quit IRC20:47
*** Ryan_Lane has joined #openstack20:47
creihtjaypipes, mtaylor: thoughts?20:51
creihtThis will not "auto-build" the .mo files on an install, but will install the current .mo files in the archive20:52
creihtand should make the app use those installed .mo files20:52
*** mgoldmann has quit IRC20:54
*** mgoldmann has joined #openstack20:57
xtoddxcreiht: so when using swauth-cleanup-tokens as it goes through the .token_N urls, it will sometimes 404, is there something misconfigured?20:59
creihtgholt: -^?20:59
*** reldan has quit IRC20:59
*** Ryan_Lane1 has joined #openstack21:00
dsockwellhas there been any testing of openstack on FrameOS or Scientific Linux?21:01
uvirtbotNew bug: #754051 in nova "instance state polling finds compute instance not in DB" [Undecided,New]
*** mgoldmann_ has joined #openstack21:02
*** bcwaldon has quit IRC21:02
*** troytoman is now known as troytoman-away21:02
*** Ryan_Lane has quit IRC21:03
*** mgoldmann has quit IRC21:03
gholtxtoddx: I'm actually not sure. That part hasn't been well vetted; I guess all of Swauth hasn't been run very long for that matter.21:04
gholtxtoddx: If it goes to delete an old token and finds it already gone, that's probably okay.21:04
xtoddxgholt: ok, so what is the difference between all the .token_N urls?  I'm not very familiar with swauth yet.  The problem is the cleanup script bombing out as soon as it hits a 404, and I'm thinking it may need to continue to all the others?21:05
*** rnirmal has joined #openstack21:05
gholtxtoddx: Yes, definitely a bug in the cleanup. The .token_ containers are to keep each container lower on the number of entries for a really busy system. It's probably way overkill for most anyone.21:08
*** [1]RickB17 has quit IRC21:09
gholtxtoddx: Can you paste the actuall stack trace? I'm curious which thing it isn't finding and where it's looking.21:09
xtoddxgholt: sure thing21:09
*** icarus901 has quit IRC21:10
*** allsystemsarego has quit IRC21:14
openstackjenkinsProject nova build #784: SUCCESS in 2 min 31 sec:
openstackjenkinsTarmac: Update instances table to use instance_type_id instead of the old instance_type column which represented the name (ex: m1.small) of an instance type.21:14
openstackjenkinsThis allows the various subsystems to cleanly reference instance types by their IDs instead of using the name as the key and provides a clean way to do the outer join to the instance types table in several SQL queries.21:14
openstackjenkinsUpdated the instance_types module so that all methods now return instance_type dictionaries for consistency.21:14
openstackjenkinsUpdated the Openstack API so that flavor ID is returned correctly.21:14
*** h0cin has quit IRC21:14
xtoddxgholt: that is two different runs.  The first doesn't find a particular token, and the following (with -v) didn't list .token_021:15
*** vernhart has quit IRC21:16
xtoddxrunning out for a bit21:19
gholtxtoddx: The first is one thing but yeah that second is weird. You can run swauth-prep again and it should create that container.21:19
gholtxtoddx: That probably also means that any token generated that ended with 0 would've errored out.21:21
*** reldan has joined #openstack21:25
jaypipescreiht: sounds reasonable to me. I will let mtaylor weigh in on that.21:26
mtaylorcreiht: sounds reasonable... but can I ask what benefit it has over the distutils.extra approach used in nova? or what it is that you're trying to solve?21:28
mtaylorcreiht: I ask because it seems that while your way will work, it adds a potential manual step into a release that I don't understand the benefit of having21:28
creihtThe main thing being that I don't like the distutils.extra approach :)21:29
*** hadrian has quit IRC21:30
creihtthis approach seems to be the more standard way python apps do i18n21:30
creihtat least from what I have gathered today21:30
creihtThe thing that is a bit weird is the translation workflow in launchpad21:32
creihtyour .po files kinda appear automagically (at least from what I gather)21:32
mtaylorah, ok. well ...21:33
mtaylorI got the distutils.extra approach from talking with the launchpad folks about best practices :)21:33
edaymtaylor: could you please add a scp target in jenkins for burrow docs? :)21:33
creihtmtaylor: lol21:33
mtayloreday: shit. yes. sorry21:33
mtaylorcreiht: have I mentioned how much it pisses me off that there is _not_ actually _one_ comprehensive best practice for this in the python world?21:34
edaymtaylor: no reason to shit or apologize :)21:34
mtayloreday: no, I meant to do that yesterday21:34
mtaylorcreiht: or actually that _all_ of the things surrounding distutils/setuptools whatnot are only ever 75% of the way to being what you actually need21:35
* mtaylor stops ranting21:35
creihtmtaylor: well there seems to be a basic best practice, but launchpad screws that up with their magic, and then thus requires a different best practice :)21:35
creihthah... yeah that isn't ideal21:35
mtaylorcreiht: indeed... BUT - the basic best practice doesn't integrate with install without extra foo21:35
mtaylorwhich makes _NO_ sense to me21:35
creihtbut I think it is a difference that in a normal project, the i18n process looks more like the dev process21:35
creihtmtaylor: it does though21:35
mtaylorbut you have to run that babel command by hand21:36
creihtif you .mo files are installed with your app, and you point gettext at it, everything works21:36
mtaylorsure- it's the making/installing of the .mo files that I'm bitching about21:36
mtaylorI don't want someone to need to know that they should do that21:37
creihtthere are two different things21:37
edaymtaylor: couldn't you just hack to run the babel command for you before install?21:37
mtaylorbut I can get off my high horse- can you show me a branch and a list of commands to run in it that I can look at?21:37
mtayloreday: my beef is why should I need to do that since it's a standard thing that needs done21:37
creihtin a tarball release, the .mo files are up to date, so a normal install should install everything as needed21:37
*** masumotok has joined #openstack21:38
creihtin branch from trunk, it is likely that the .mo files will have to be updated21:38
edaymtaylor: agree, but might be an easy solution that doesn't require another pkg21:38
mtayloreday: well... either way requires a package- one requires bable, the other requires distutils-extra, so it's a wash as far as that goes21:39
creihtmtaylor: I don't have a branch yet, was just thinking out loud, and trying to decide if this is a reasonable solution21:39
edaymtaylor: ahh21:39
creihtmtaylor: well currently in nova, you have both babel and -extra21:39
masumotokvishy: sorry for annoying you about pep8 violation. regarding to pep8 violation for I just want to let you know I fixed it..21:39
mtaylorcreiht: oh, well, that's historical and is a bug, there is no need for babel21:39
mtaylorcreiht: so, in a perfect world, I want python install to do the same thing in a tarball as in a branch ... but I'm sort of an exremist21:40
vishymasumotok: thanks, reapproved21:40
*** ctennis has quit IRC21:40
*** MotoMilind has quit IRC21:41
creihtmtaylor: right now python install doesn't auto build the docs either, should it do that as well? :)21:41
mtaylorcreiht: as far as I'm concerned it should - although I thought that we put in a thing so that if you did not have sphinx installed it would silently do nothing21:41
mtaylorcreiht: but yeah - any process that requires someone to read a README to get a proper install is broken21:42
uvirtbotNew bug: #754086 in swift "swauth-cleanup-tokens "not found" error handling" [Medium,Confirmed]
creihtmtaylor: in python dev outside of lp, it seems that whenever a change to a .po is made, the .mo is generated, and both checked in, and thus installed when installed21:42
mtaylorso there's the fail21:42
mtaylorI am VERY against checking generated files into vc21:42
creihtyeah I know21:42
creihtI'm not all that high on it either21:42
vishyI have 3 bug branches that need love here:
vishy2 are pretty high priority21:43
creihtand no, install doesn't build docs, it does only if your run build_sphinx21:43
*** vernhart has joined #openstack21:43
*** dubsquared1 has joined #openstack21:43
dubsquared1jaypipes in the house?21:44
creihtmtaylor: most of this is a byproduct of how kinda works (or doesn't work)21:44
mtaylorcreiht: perhaps we should write _yet_ _another_ system ... called rax-po or something21:44
mtaylorthat is very true21:44 is, well, wonderful and butt-awful all at the same time21:44
creihtthere is supposedly a distutils2 coming that is supposed to fix all our problems, but it isn't read yet21:45
creihtmtaylor: is there any type of notification from lp when a .po is updated?21:45
mtaylorcreiht: not really21:45
mtaylorcreiht: launchpad just drops new revs into a branch21:45
devcamcarhey gang, starting my review day duties now21:46
*** imsplitbit has quit IRC21:46
devcamcargoing from thierry's most recent list and working backwards21:46
*** dendrobates is now known as dendro-afk21:46
devcamcarif anyone has any other high priorities reviews they'd like to see done today, please list them here for me21:47
devcamcarand i'll get to as many as i can21:47
creihtmtaylor: then what happens to that branch, does it get auto merged into trunk?21:47
*** ppetraki has quit IRC21:48
*** hub_cap has quit IRC21:48
*** hub_cap has joined #openstack21:49
mtaylorcreiht: I think we've been doing it by hand21:49
justinsbdevcamcar: I think this one needs a close look but is important:
creihtmtaylor: well at that step, would it make sense to have the .mo files generated?21:49
justinsbdevcamcar: I sort of forgot about it in all the strum und drang about schemas !21:50
*** reldan has quit IRC21:50
devcamcarjustinsb: with my feet on the ground and my code in the cloud21:50
ironcamel2devcamcar: it would be great if you could take a look at
creihtmtaylor: that way an install would always install the latest translations21:53
creihtI know it means binaries in the repo, but not sure there is a better way with how without the -extras hackery21:53
openstackjenkinsProject nova build #785: SUCCESS in 2 min 30 sec:
openstackjenkinsTarmac: This branch is a patch for fixing below issue.21:54
openstackjenkins> Bug #746821: live_migration failing due to network filter not found  Link a bug report21:54
uvirtbotLaunchpad bug 746821 in nova "live_migration failing due to network filter not found" [High,In progress]
*** _vinay has joined #openstack21:54
_vinayHi vishy21:55
_vinayI am outputting a simple print "START" in a test case21:55
_vinayand its not coming out on stdout ...21:55
_vinaywhile doing test_xenapi21:55
_vinayany reason?21:55
vishy_vinay: aye print does not work in tests21:55
creihtmtaylor: and as a side note, if anyone updates a .po in the code tree, does the translation stuff pick up on that? or can translations only be edited through launchpad?21:56
_vinayoh ok... then how do I output stuff on screen21:56
_vinayesp on stdout21:56
vishysomething to do with the way nose + run_tests handles the output21:56
_vinayyes I am looking at colorizer stuff21:56
_vinaypretty complicated21:56
_vinaybut anyways the logging thing didnt work for me21:57
_vinayI set the verbose flag to False in all the files21:57
_vinaybut still I get debug output from a lot of packages21:57
_vinaythe funny thing is that all that output goes to stderr21:57
_vinayand the test summary OK FAIL etc.. (green, red) they go on stdout21:57
*** troytoman-away is now known as troytoman21:58
_vinayso I am just trying to print something on stdout ... but cant seem to find a way to do it22:00
vishy_vinay: sys__stdout__.write should work22:01
*** miclorb has joined #openstack22:01
vishysys.__std.... that is22:01
*** ibatanov has quit IRC22:02
*** dendro-afk is now known as dendrobates22:04
*** coreym has quit IRC22:04
justinsbIs there any documentation on how to set up networking?  I want to set up a FlatNetwork, and I think I may have to edit the DB...22:05
vishyFlatNetwork is not fun22:05
vishynot recommended22:05
justinsbWhich should I be using?22:05
justinsbOK.  I wanted to keep the cloud VMs in their own network space, but I guess it doesn't really matter.  Thanks!22:06
creihtmtaylor: and as another thing, how many projects actually internationalize their log outputs? :)22:06
*** troytoman is now known as troytoman-away22:06
*** mgoldmann_ has quit IRC22:07
vishyjustinsb: you can put them in their own network space22:07
vishyFlatDHCP just actually works and Flat requires a ton of custom setup + network injection22:07
vishyor special images /etc22:07
justinsbvishy: I can?  OK - I'll read the docs on DHCP - thanks for the pointer!22:07
*** hggdh has quit IRC22:07
claygjustinsb: I added a couple of projects in my cluster, and I think I basically just successfully attached a volume from project a to an instance in project b - it doesn't seem like I should have been able to do that?22:08
claygbasically a quick nmap for the volume-manager (who's listening on 3260) and an iscsiadm discovery and I can attach22:09
claygis there blueprint to use like... idk, CHAP or something?22:09
vishyjustinsb: if you use, it is as easy as changing NET_MAN, although if you are doing it on multiple hosts you need a few other changes22:09
justinsbvishy: Thanks, I'll check out - I love that script!22:10
*** hggdh has joined #openstack22:10
justinsbclayg: There is some CHAP code22:10
*** gondoi has quit IRC22:10
justinsbclayg: But not on Linux OpeniSCSI targets yet22:10
vishyclayg: that sounds annoying22:10
claygthe HPSan stuff? or in the default ISCSIDriver?22:10
claygyeah gotcah22:10
justinsbclayg: Just on HP SANs, I think22:10
vishyclayg: although in our case vms don't have access to the host ip22:11
justinsbclayg: I don't believe that security is a consideration yet in the volume code22:11
justinsbclayg: Because we're relying on e.g. network isolation22:11
*** Ryan_Lane1 is now known as Ryan_Lane22:11
*** Ryan_Lane has joined #openstack22:11
*** rnirmal has quit IRC22:11
justinsbclayg: The HP SAN CHAP stuff is there as a side effect of a different implementation issue22:12
claygvishy: seperate interface on the host for iscsi, vms can't get to it... makes sense22:12
claygjustinsb, vishy: thanks for the clarification22:12
vishyyeah the network host doesn't route to the infrastructure ips22:12
justinsbclayg: Thank you!  We should probably document that in case anyone gets hit by that unexpectedly!22:14
*** dubsquared1 has quit IRC22:15
openstackjenkinsProject nova build #786: SUCCESS in 2 min 32 sec:
openstackjenkinsTarmac: Automatically add the metadata address to the network host.  This allows guests to ARP for the address properly.22:19
openstackjenkinsI also uncovered an issue with moving the gateway.  Apparently specifying route add doesn't actually work, you have use route add 'default'.  I also added a line to specifically delete the old gateway since it doesn't always automatically get deleted.22:19
*** ctennis has joined #openstack22:19
*** ctennis has joined #openstack22:19
*** nnahum has joined #openstack22:20
*** MotoMilind has joined #openstack22:21
*** Shentonfreude has quit IRC22:21
*** zigo-_- has quit IRC22:23
*** omidhdl has quit IRC22:23
*** nnahum has quit IRC22:26
*** dendrobates is now known as dendro-afk22:33
*** slyphon has quit IRC22:34
*** mray has quit IRC22:35
creihtmtaylor: after doing a bit more research, it seems that most python projects that support i18n include the .mo files in the source tree22:38
devcamcari'll approve when you remove the commented code blocks22:41
creihtmtaylor: or goes way out of there way to hack to build them22:41
ironcameldevcamcar: any chance you could look at today?22:43
creihtmtaylor: and if you don't use babel in nova, how will the .pot file be generated when changes are needed?22:43
devcamcarironcamel: its on my screen right now in fact :)22:43
creihtand does that happen automagically some how, or do we have to remember to do that each time changes are made to messages?22:43
devcamcarironcamel: can you help me understand how the code forces out this format? i'm confused how the proposed changes result in the correct format:22:44
devcamcar<server flavorId="1" hostId="" id="1" imageId="10" name="server1" status="BUILD">22:44
devcamcar    <addresses>22:44
devcamcar        <public>22:44
devcamcar            <ip addr=""/>22:44
devcamcar        </public>22:44
devcamcar        <private>22:44
devcamcar            <ip addr=""/>22:44
devcamcar        </private>22:44
devcamcar    </addresses>22:44
devcamcar    <metadata>22:44
devcamcar        <meta key="foo">bar</meta>22:44
devcamcar    </metadata>22:44
devcamcarthe proposed code is:22:44
devcamcar8+            "dict_collections": {22:44
devcamcar9+                "metadata": {"item_name": "meta", "item_key": "key"},22:44
devcamcar10+            },22:44
devcamcar11+            "list_collections": {22:44
devcamcar12+                "public": {"item_name": "ip", "item_key": "addr"},22:44
devcamcar13+                "private": {"item_name": "ip", "item_key": "addr"},22:44
devcamcar14+            },22:44
devcamcaractually! hah, nm22:45
ironcameldevcamcar: so there is the new concept called list_collections22:45
devcamcari misread, i thought it was ip_addr22:45
devcamcari see now22:45
ironcamelok good22:45
devcamcarironcamel: but yes, give me more info on list_collections22:45
devcamcarwould be helpful22:45
ironcameldevcamcar: so list_collections are for converting python lists to xml collections22:46
ironcameland dict_collections means convert python dicts to xml collections22:46
ironcamelif you look at the 1.0 spec and compare the json and xml responses for a server response, that would help you understand22:47
devcamcarironcamel: is this specific to this patch or are list_collections a standard practice in os api? i have more experience with ec2 api then os api at this point22:47
devcamcarironcamel: cool, will do22:47
devcamcarok yep, that makes sense to me22:48
devcamcarironcamel: approved for merge22:48
ironcameldevcamcar: the idea is that list_collections would be the way to go for osapi. a subsequent merge prop that depends on this uses list_collections for correctly supporting servers/1/ips, servers/i/ips/public, and servers/1/ips/private22:48
*** [1]RickB17 has joined #openstack22:48
devcamcaranyone else have high priority reviews they'd like done now?22:48
ironcameldevcamcar: thanks!22:48
devcamcargamma freeze is for tomorrow, so speak now! :)22:48
ironcameldevcamcar: if no one else speaks up, would be great if you could also look at
ironcamelit is very similar22:49
ironcamelwait, it has 2 approves. is that enough?22:50
devcamcarironcamel: yes but i'll give it another pass now as well22:50
*** carlp has joined #openstack22:50
devcamcarironcamel: just starting from the bug before i look at code:22:51
devcamcarTo recreate this bug, create a cloud server. Then try to view it's addresses via:22:51
devcamcar    GET /servers/1/ips22:51
devcamcarIt will return a 404.22:51
devcamcarwhat is expected result set?22:51
ironcameldevcamcar: what is expected is the ip addresses of that server22:52
devcamcarcool i see: Now you can GET22:52
devcamcar * /servers/1/ips22:52
devcamcar * /servers/1/ips/public22:52
devcamcar * /servers/1/ips/private22:52
jaypipescreiht: could you back up your claim about most python projects that do i18n put their .mo files in vcs with some specifics? which specific projects?22:52
ironcamelif you have the 1.0 spec open, it will show exactly what it should return22:52
*** rds__ has joined #openstack22:53
ironcameldevcamcar: /servers/1 will return an entity that contains an addresses entity inside it. /servers/1/ips is another way to return just the addresses entity.22:53
ironcamelnot sure if that is a great idea, but that's what the spec says :)22:54
devcamcarironcamel: self.builder.build_private_parts  huhuhuhuhuhuhuhuhuhuhhhh22:54
devcamcari'm so mature22:54
ironcameldevcamcar: glad you caught that :)22:54
ironcamela little subtle humour on my part22:55
devcamcarhaha it was not lost on me22:55
devcamcarironcamel: approved for merge22:56
ironcameldevcamcar: thanks!22:56
*** arun_ has quit IRC22:56
ironcameldevcamcar: does that mean you flipped the switch in hudson?22:56
devcamcaryea hudson should process it shortly22:56
creihtjaypipes: most places that you find i18n are web apps, so most django and pylons apps for example seem to do this22:58
creihta good example is the pinax group of django apps22:58
creihtdjango-cms does this as well23:01
*** dirkx has joined #openstack23:02
[1]RickB17creiht:  i was able to successfuly create accounts, but when i use st -A auth_url -U account:users -K password stat  i recieve "Account not found".  any ideas?  I read through that link your provided, but i'm unsure on how to proceed.23:03
[1]RickB17i did various swauth-list to verify the acctount info is there.23:03
openstackjenkinsProject nova build #787: SUCCESS in 2 min 34 sec:
openstackjenkinsTarmac: This fixes how the metadata and addresses collections are serialized in xml responses.23:04
*** kakella has left #openstack23:06
*** bkkrw has quit IRC23:06
*** Zangetsu_ has quit IRC23:06
creiht[1]RickB17: hrm23:11
creihtCan you paste the log output of the proxy for that request?23:11
openstackjenkinsProject nova build #788: SUCCESS in 2 min 29 sec:
openstackjenkinsTarmac: Added support for listing addresses of a server in the openstack api.23:11
openstackjenkinsNow you can GET23:11
openstackjenkins* /servers/1/ips23:11
openstackjenkins* /servers/1/ips/public23:11
openstackjenkins* /servers/1/ips/private23:11
openstackjenkinsSupports v1.0 json and xml.23:11
openstackjenkinsAdded corresponding tests.23:11
[1]RickB17that may be more info than needed, but it's what i know to collect23:12
[1]RickB17if there is anything else useful, let me know23:12
*** jonkelly has left #openstack23:13
[1]RickB17i used to be able to connect via cyberduck from a windows machine when using devauth.  I haven't been able to successfully do so since we got swauth working23:13
creiht[1]RickB17: try: st -A -U system:root -K testpass stat23:15
[1]RickB17that was it23:16
[1]RickB17what makes the difference in the authurl?23:16
[1]RickB17i'm unclear on when to use jsut /auth/ or /auth/v123:16
[1]RickB17or even /auth/v1.023:16
creihtyeah, sorry that is confusing23:16
creihtfor all auth, when acting as a user, your auth url should be /auth/v1.023:16
creihtand always should be v1.023:17
creihtwell until it upgrades :)23:17
*** dragondm has quit IRC23:17
[1]RickB17one last quick question :-D  how can you change a users password or other settings?  I had to delete and recreate it to set the password because i coudln't find how to do it any other way23:18
devcamcaranybody been looking at this one?
uvirtbotLaunchpad bug 712322 in nova "Issue in security group's revoke command" [High,Confirmed]23:18
[1]RickB17again i relaly appreciate your time and helpfulness23:18
creihtif you are directly interacting with swauth, then it should be just to the /auth/ path23:18
devcamcar"Issue in security group's revoke command"23:18
creihtthe storage_url that you access the cluster with to operate on your accounts/containers/objects ends with /v123:19
creihtsorry it is confusing, mostly due to legacy systems written by different people23:19
*** dirkx has quit IRC23:20
creiht[1]RickB17: if you re-add a user, it will change their information23:20
creiht(don't have to delete)23:20
[1]RickB17perfect, thanks23:20
*** aliguori has quit IRC23:20
[1]RickB17still having troubles mounting cloudfuse, "unable to authenticate." but i'll work on that another time.  time for some grub.23:20
[1]RickB17have a good one23:20
creihtyou as well23:20
justinsbdevcamcar: I think should be good now23:24
jaypipescreiht: so, no examples other than django/pylons (which both have hacks that rewrite parts of gettext themselves?)23:26
jaypipescreiht: just want to make sure that the reason you don't want to go down the route that Nova does is because it involves Launchpad...23:27
*** alekibango has quit IRC23:30
*** alekibango has joined #openstack23:31
edaycreiht, jaypipes: I really don't like the idea of checking in the generated files. If you go this route, I would say let the package builder run the generation command, not the thing that merges po branches.23:33
*** fysa has quit IRC23:33
jaypipeseday: Nova doesn't check in generated files. That's the whole point...23:36
edayjaypipes: I know, I'm saying if swift doesn't do what nova does, and is deciding to either check in .mo's vs run the extra command at pkg build time23:37
*** slyphon has joined #openstack23:37
*** jeffjapan has joined #openstack23:38
jaypipeseday: works quite well for nova IMHO. 14 languages so far, at varying levels of translation...23:39
*** rchavik has joined #openstack23:40
*** zaitcev has quit IRC23:42
edayjaypipes: I'm not saying it doesn't work, I know it does quite well. This was building off the discussion earlier with mtaylor and creiht :)23:42
*** Zangetsu has joined #openstack23:43
devcamcarjustinsb: approved where-has-my-instance-gone23:45
*** vernhart1 has joined #openstack23:45
*** vernhart has quit IRC23:46
justinsbdevcamar: Thanks!23:46
*** Zangetsu has left #openstack23:47
*** miclorb has quit IRC23:47
*** miclorb_ has joined #openstack23:47
jaypipeseday: I know.. I'm just trying to figure out how to move this forward and get Swift to the point that Nova is... getting translated.23:57
*** bcwaldon has joined #openstack23:58
*** fysa has joined #openstack23:58
openstackjenkinsProject nova build #789: SUCCESS in 2 min 30 sec:
openstackjenkinsTarmac: Keep guest instances when libvirt host restarts23:59

Generated by 2.14.0 by Marius Gedminas - find it at!