Wednesday, 2013-12-11

« Tuesday, 2013-12-10 Index Thursday, 2013-12-12 »
*** jcooley_ has quit IRC00:02
kuileaving . Have a good day! everyone00:02
*** jcooley_ has joined #openstack-ironic00:02
*** lexx has quit IRC00:03
*** jcooley_ has quit IRC00:07
*** epim has quit IRC00:07
NobodyCamhave a good night kui and thank you00:10
*** epim has joined #openstack-ironic00:12
*** jcooley_ has joined #openstack-ironic00:13
*** jcooley_ has quit IRC00:13
*** jcooley_ has joined #openstack-ironic00:13
openstackgerritdekehn proposed a change to openstack/ironic: Method to generate PXE options for Neutron ports  https://review.openstack.org/6090400:14
*** jcooley_ has quit IRC00:18
devanandaNobodyCam: now that I know how to push a release of ironic client, i'm goign to bump its version to 0.1.0. seem reasonable? think I should wait for any particular patches?00:30
*** kui has quit IRC00:34
NobodyCamhmmm00:35
NobodyCamnot in the client00:37
*** hemna has quit IRC00:37
NobodyCamdevananda: there will be a client side to https://review.openstack.org/#/c/6116000:41
NobodyCamAdd trigger to start provisioning the node00:41
devanandahm?00:41
devanandaahh yes00:42
NobodyCambut we can do another release. why do you want ot push a new version00:42
devanandaok. let's wait for that + the matching client00:42
NobodyCam:)00:42
devanandathen i'll do a 0.1.000:42
NobodyCamya that seems like a good time to do that00:43
NobodyCambrb quick walkies time00:43
lifelesshttps://review.openstack.org/#/c/60982/ needs another +200:45
lifelessfixes the sphinx thing00:45
*** epim has quit IRC00:46
NobodyCamlifeless: the multi line doc string thing?00:53
openstackgerritA change was merged to openstack/python-ironicclient: Updated from global requirements  https://review.openstack.org/6098200:53
lifelessno, gate was broken for a bunch o fprojects00:56
lifelesssphinx release fubared it all00:56
NobodyCamahh :)00:57
*** hemna has joined #openstack-ironic00:59
*** anniec has quit IRC01:03
*** hemna is now known as hemnafk01:05
*** rloo has quit IRC01:24
*** nosnos has joined #openstack-ironic01:29
*** dguitarbite has joined #openstack-ironic01:44
*** michchap_ has quit IRC01:46
*** michchap has joined #openstack-ironic01:47
*** kui has joined #openstack-ironic01:58
*** rongze has joined #openstack-ironic02:01
*** jcooley_ has joined #openstack-ironic02:04
*** rongze has quit IRC02:31
*** kushi_ has joined #openstack-ironic02:43
*** jcooley_ has quit IRC02:43
*** kui has quit IRC02:45
*** rongze has joined #openstack-ironic02:46
*** kushi_ has quit IRC02:48
*** prekarat has joined #openstack-ironic03:14
*** prekarat has quit IRC04:02
*** hemna_ has joined #openstack-ironic04:24
*** hemnafk has quit IRC04:24
*** hemna_ has quit IRC04:24
*** hemna_ has joined #openstack-ironic04:24
*** dguitarbite has quit IRC04:27
*** anniec has joined #openstack-ironic04:28
*** anniec_ has joined #openstack-ironic04:35
*** anniec has quit IRC04:35
*** anniec_ is now known as anniec04:35
*** rongze has quit IRC04:50
*** prekarat has joined #openstack-ironic05:12
*** prekarat1 has joined #openstack-ironic05:13
*** prekarat has quit IRC05:16
*** rongze has joined #openstack-ironic05:20
*** nosnos_ has joined #openstack-ironic05:31
*** jcooley_ has joined #openstack-ironic05:33
*** nosnos has quit IRC05:34
*** rongze has quit IRC05:35
*** anniec has quit IRC05:36
*** devananda has quit IRC05:57
*** davidlenwell_ has joined #openstack-ironic05:57
*** devananda has joined #openstack-ironic05:57
*** Ghe_HPDiscover has quit IRC05:57
*** davidlenwell has quit IRC05:58
*** NobodyCam has quit IRC05:58
openstackgerritJenkins proposed a change to openstack/ironic: Imported Translations from Transifex  https://review.openstack.org/5962706:00
*** rongze has joined #openstack-ironic06:03
*** nosnos_ has quit IRC06:04
*** nosnos has joined #openstack-ironic06:05
*** rongze has quit IRC06:07
*** vkozhukalov has quit IRC06:09
*** NobodyCam has joined #openstack-ironic06:14
*** rongze has joined #openstack-ironic06:15
*** vetalll has joined #openstack-ironic06:16
*** Ghe_HPDiscover has joined #openstack-ironic06:35
*** jcooley_ has quit IRC06:39
*** yfujioka has joined #openstack-ironic06:59
*** mdurnosvistov has joined #openstack-ironic07:02
*** lifeless has quit IRC07:05
*** mdurnosvistov has quit IRC07:06
*** jcooley_ has joined #openstack-ironic07:09
*** ifarkas has joined #openstack-ironic07:13
*** jcooley_ has quit IRC07:15
*** vetalll has quit IRC07:20
*** vetalll has joined #openstack-ironic07:23
*** vetalll has quit IRC07:27
*** urulama has joined #openstack-ironic07:31
*** lexx has joined #openstack-ironic07:32
*** vkozhukalov has joined #openstack-ironic07:57
*** jcooley_ has joined #openstack-ironic08:05
*** jcooley_ has quit IRC08:09
*** lexx has quit IRC08:15
*** rongze has quit IRC08:46
*** rongze has joined #openstack-ironic08:47
*** jistr has joined #openstack-ironic08:57
*** jcooley_ has joined #openstack-ironic09:00
*** romcheg has joined #openstack-ironic09:04
*** jcooley_ has quit IRC09:04
*** lucasagomes has joined #openstack-ironic09:14
*** max_lobur_afk is now known as max_lobur09:21
*** derekh has joined #openstack-ironic09:23
*** tatyana has joined #openstack-ironic09:28
*** mdurnosvistov has joined #openstack-ironic09:30
openstackgerritSushil Kumar proposed a change to openstack/ironic: Updated from global requirements  https://review.openstack.org/6135409:33
*** nosnos_ has joined #openstack-ironic09:36
*** nosnos has quit IRC09:40
*** lifeless has joined #openstack-ironic09:43
*** dguitarbite has joined #openstack-ironic10:12
openstackgerritAlexandra Fomina proposed a change to openstack/python-ironicclient: Import exceptions functionality from Oslo  https://review.openstack.org/6052810:14
openstackgerritAlexandra Fomina proposed a change to openstack/python-ironicclient: Import exceptions functionality from Oslo  https://review.openstack.org/6052810:16
*** prekarat1 has quit IRC10:18
*** homegrown has joined #openstack-ironic10:20
*** nosnos_ has quit IRC10:57
*** nosnos has joined #openstack-ironic10:57
openstackgerritLucas Alvares Gomes proposed a change to openstack/python-ironicclient: Rename nodes/<uuid>/state to nodes/<uuid>/states  https://review.openstack.org/6055810:58
*** anniec has joined #openstack-ironic11:03
*** ifarkas has quit IRC11:03
openstackgerritLucas Alvares Gomes proposed a change to openstack/python-ironicclient: Add set_provision_state to the client libs  https://review.openstack.org/6137911:16
*** rongze has quit IRC11:21
*** jcooley_ has joined #openstack-ironic11:43
*** anniec_ has joined #openstack-ironic11:43
*** rongze has joined #openstack-ironic11:44
*** anniec has quit IRC11:45
*** anniec_ is now known as anniec11:45
*** jcooley_ has quit IRC11:47
*** ifarkas has joined #openstack-ironic11:53
lucasagomeslifeless, https://review.openstack.org/#/c/60972/ will fix the problem with the docs http://logs.openstack.org/98/60798/1/gate/gate-ironic-docs/9353b3a/console.html (caused by sphinx >=1.2)12:04
lucasagomeslifeless, when you get some free time take a look at the patch and if that lgty: +2/A12:06
lucasagomeslifeless, thanks :)12:06
*** lucasagomes is now known as lucas-hungry12:06
*** yfujioka has quit IRC12:12
openstackgerritA change was merged to openstack/ironic: Updated from global requirements  https://review.openstack.org/6097212:19
devanandaHaomeng: are you still working on bug 1244541 ?12:21
HaomengMorning devananda, let me check12:22
Haomengdevananda: sorry, I missed this bug for some time12:23
Haomengdevananda: will go ahead to port your change, and commit the path this night12:23
devanandaHaomeng: it may not be a simple port-my-old-code. also, it's OK to unassign if you are busy with other things12:25
devanandaHaomeng: just trying to keep the bug list accurate12:25
Haomengdevananda: let me try to fix, and will ask you if I have questions12:27
Haomengdevananda: yes, we have a lot of defect in the list12:27
Haomengdevananda: yes, these days, I am focusing on some code review to get more understanding about our Ironic code12:32
Haomengdevananda: so will take back to fix defect if required12:33
devanandaHaomeng: both good things to do :)12:34
devanandaromcheg: ping12:34
Haomengdevananda: ok, I will do12:34
romchegMorning devananda12:34
devanandag'morning12:34
Haomengdevananda: for 1244541, I think we have to fix this both ironic api and ironic driver, an exception when overwriting the instance_uuid in ironic api, and implement a EAFP approach in nova ironic driver, right?12:42
devanandaGhe_HPDiscover: I imagine you're busy at discover -- just a quick question. are you going to be able to work on bugs assigned to you over next few months (such as 1236536) or should they be unassigned?12:43
devanandaHaomeng: that's correct. ironic/db/api needs a new method specifically for set/unset instance_uuid, which will raise an exception instead of overwriting it12:45
Haomengdevananda: ok, thank you12:45
devanandaHaomeng: then some changes eg. in the api to handle that exception12:46
devanandaHaomeng: welcome. thanks for taking a look at it!12:46
Haomengdevananda: yes12:46
Haomengdevananda: I like fix defect, that is challenge for me, can get more deep understand by fixing these bugs12:47
HaomengHaomeng: any time12:47
devanandalucas-hungry: hi! what do you think of https://bugs.launchpad.net/ironic/+bug/1238117 today? still a bug, or did it get fixed amidst the last 2 months of work?12:48
devanandaHaomeng: think you'll be able to fix that by end of january? if so, i'll target it to i2 milestone12:50
Haomengdevananda: that is ok, I will take this as high prority and ask you  core team members for any issue12:51
Haomengdevananda: I think that is enough time12:51
*** urulama has quit IRC12:53
ekarlso /win 3812:58
*** vkozhukalov has quit IRC13:00
*** urulama has joined #openstack-ironic13:13
devanandaNobodyCam: hi! just sent you some pull requests for tripleo ironic element13:28
*** max_lobur is now known as max_lobur_afk13:30
*** jcooley_ has joined #openstack-ironic13:30
*** sandeepr has joined #openstack-ironic13:31
*** jdob has joined #openstack-ironic13:35
*** jcooley_ has quit IRC13:35
*** prekarat has joined #openstack-ironic13:38
devanandaNobodyCam: also i think your undercloud-vm-ironic.yaml file needs to have13:39
devananda216         compute_driver: ironic.driver.IronicDriver13:39
devananda(or at least mine did)13:39
devanandaNobodyCam: and once I do that, I get teh following trace in nova-compute.log13:42
devanandahttp://paste.openstack.org/show/54824/13:43
*** rloo has joined #openstack-ironic13:47
* devananda goes afk again13:47
*** lucas-hungry is now known as lucasagomes14:01
lucasagomesdevananda, morning, I don't think the API is fixing that, I we still need to add some sanity check on that area, def fixed by end of jan14:06
*** nosnos has quit IRC14:06
*** dguitarbite has quit IRC14:11
openstackgerritA change was merged to openstack/ironic: Port StringType class from WSME trunk  https://review.openstack.org/6079814:16
HaomengIronic, nice day, I really want to go to bed, so tired:)14:18
*** michchap has quit IRC14:24
*** linggao has joined #openstack-ironic14:24
*** jcooley_ has joined #openstack-ironic14:25
*** jcooley_ has quit IRC14:29
*** matty_dubs|gone is now known as matty_dubs14:33
lucasagomesHaomeng, g'night!14:38
Haomenglucasagomes: nice day, Lucas:)14:39
lucasagomes:)14:39
*** prekarat has quit IRC14:42
*** max_lobur_afk is now known as max_lobur14:45
*** viktors has joined #openstack-ironic14:51
*** jdob has quit IRC15:00
*** jdob has joined #openstack-ironic15:06
*** vkozhukalov has joined #openstack-ironic15:07
*** jcooley_ has joined #openstack-ironic15:18
*** jcooley_ has quit IRC15:23
openstackgerritYuriy Zveryanskyy proposed a change to openstack/ironic: Check connectivity in SSH driver 'validate' method  https://review.openstack.org/6087615:35
NobodyCamgood morning Ironic15:42
NobodyCamdevananda: what are you doinf up at 5 am?15:47
*** ndipanov has quit IRC15:48
lucasagomesNobodyCam, morning15:49
NobodyCammorning lucasagomes :)15:51
NobodyCamsorry for the last start this morning15:51
NobodyCamhow are you doinf today15:51
lucasagomesnot too bad, urself?15:52
NobodyCamlol geeting a slow start this morning :-p15:53
*** ifarkas has quit IRC15:54
*** urulama has quit IRC15:54
lucasagomesNobodyCam, heh yeah, you need ur first cup of joe :)15:55
NobodyCam:) working on it ...lol15:56
lucasagomes:D15:56
*** anniec has quit IRC15:57
NobodyCambrb getting the coffee :)16:03
lucasagomes:) enjoy16:07
lucasagomesdevananda, ping16:11
*** hemna has joined #openstack-ironic16:13
*** jcooley_ has joined #openstack-ironic16:15
NobodyCamahh coffee16:15
*** jcooley_ has quit IRC16:20
devanandalucasagomes: pong16:24
NobodyCammornig devananda16:24
lucasagomesdevananda, thinking about ur review at 60194 re sanitize paths16:25
devanandaNobodyCam: g'morning! i see you already merged my pull reqs. hehe, thanks!16:25
lucasagomesas they are hardcoded (won't change) do we really need an sanitize method?16:25
* devananda updates the wiki accordingly16:26
devanandalucasagomes: nope16:26
lucasagomesdevananda, right :)16:26
devanandalucasagomes: that was my first thought. then i assumed you had put it there in case someone subclassed it and didn't follow the right pattern. but that's what unit tests are for, IMO.16:27
devanandai think it's worth a check somewhere. but not in object's runtime path :)16:27
lucasagomescool yea16:27
lucasagomesalso because we don't instantiate that class ourself, it's done by wsme16:27
lucasagomesso sanitizing on the constructor will have the same effect as sanitizing at the validate() method... the sanitize() will be called the same no of times16:28
lucasagomeswhich is, everytime a request comes in16:28
devanandaright16:29
devanandawhich seems very inefficient16:29
lucasagomesyup16:29
lucasagomeswsme instantiate that class and pass the object as a parameter to the validate() method16:29
lucasagomesanyway I will update the patch16:30
lucasagomesthanks for the reviews16:30
NobodyCam:)16:30
NobodyCamdevananda: Was that error just after startup with no nodes registered to ironic?16:31
devanandaNobodyCam: no. I've registered a node in ironic. it happens any time i restart nova-compute16:31
devanandalucasagomes: welcome!16:32
NobodyCamok16:32
*** hemna has quit IRC16:33
devanandalucasagomes: also, when you get a chance, you may want to skim https://review.openstack.org/#/c/59454/2/reference/incubation-integration-requirements for a baseline of what our requirements are16:33
*** jcooley_ has joined #openstack-ironic16:33
devanandalucasagomes: this came up at the TC meeting yesterday. I think it's something all our -core members should be at least aware of16:34
openstackgerritArtem Rozumenko proposed a change to openstack/ironic: ironic/common/exception.py wrap_exception method is never used  https://review.openstack.org/6147416:34
lucasagomesdevananda, right, will take a look at it. Thanks16:34
devanandalucasagomes: re bug 1238117, and in general about our power/provision state strings, what do you think of copying ironic.common.states into the client lib?16:36
*** hemna has joined #openstack-ironic16:37
lucasagomesdevananda, all the states or just the ones we are going to use?16:39
devanandalucasagomes: why not all?16:39
lucasagomesfor example "initializing" idk if that makes sense to exist on the client16:39
lucasagomesprobably won't be used at all16:40
lucasagomesI like the idea of having the states on the client16:40
lucasagomesand if copying the whole thing makes it easier to syncup later, I think it's a fair trade-off16:40
devanandalucasagomes: when I added rebooting to the client, it took some wrangling, because there is not a strict mapping of the states ENUM there16:40
devanandaideally, i think we should have the states defined in only one place16:41
devanandabut the client can't do "from ironic import states"16:41
devanandaand we also need the states in the nova driver16:41
lucasagomesdevananda, yea, I hear you... I when i was thinking about triggering the "unprovisioning" I was tihnking about which state to use on the client16:41
devanandayea, exactly...16:42
devanandaso syncing the file to the client seems the simplest approach. at least we can ensure that they stay in sync16:42
lucasagomesdevananda, fair I think it's a good idea16:42
devanandacool16:42
openstackgerritYuriy Zveryanskyy proposed a change to openstack/ironic: Check connectivity in SSH driver 'validate' method  https://review.openstack.org/6087616:42
lucasagomesdevananda, another q about the client that I saw today when was codying the provision thing16:43
lucasagomesshould the client itself be awere about which state is valid before sending the request to the API?16:43
lucasagomese.g16:43
lucasagomesfor states/power we accept power on, power off and rebooting16:43
lucasagomesbut the library doesnt actually check if the state passed to it is one of those16:44
lucasagomesit takes the state and submit to the api  and the API would return 400 (badrequest) in case that's a invalid state16:44
lucasagomesdo you think we should capture it before in the library and return a BadRequest before doing the request to the API?16:45
lucasagomeshttps://github.com/openstack/python-ironicclient/blob/master/ironicclient/v1/node.py#L86-L8716:46
devanandalucasagomes: nope. our client (both CLI and library) are, AFAIK, basically examples. Folks may (and will) write their own client libs (esp. in other languages)16:46
lucasagomesnote; the CLI does check it, but if you're using the libs it doesnt: https://github.com/openstack/python-ironicclient/blob/master/ironicclient/v1/node_shell.py#L14916:47
lucasagomesright16:47
devanandawell. so that doesn't make sense :)16:47
lucasagomesalso having the error returned from the API would be more informative (that's why I was confused)16:47
devanandayea16:47
lucasagomesyea so there's that incosistence there16:48
*** max_lobur has quit IRC16:48
devanandaneither should check it, and we should handle the API error. or the lib should check it and STILL handle the API error. because there could be version difference that changes the string or some such16:48
lucasagomes+116:48
lucasagomesI can file a bug about it later on16:49
devanandagreat16:49
devanandaoh, re: bugs, we (meaning, mostly, me) need to stay on top of bug triage and targeting bugs to milestones now16:49
lucasagomesi see, yea I should give you a hand on that too, the new bugs I open/or I'm working on I try to set the milestone16:50
NobodyCamlucasagomes: question on the deploy trigger stuff, are you also planning a destroy / undeploy in the client too?16:50
*** rongze_ has joined #openstack-ironic16:51
devanandaany bug assigned to someone should be updated by that person :)16:51
lucasagomesbut I haven't gone trough the list of bugs to verify the rest16:51
lucasagomesahh16:51
lucasagomesgotcha16:51
lucasagomesNobodyCam, yup16:51
NobodyCam:) sweet16:52
lucasagomesNobodyCam, it's architeted in a way that makes it easy to extend to add more functions16:52
*** rongze__ has joined #openstack-ironic16:52
*** rongze has quit IRC16:52
devanandathe rest (all the unassigned bugs) probably shouldn't have a milestone set16:53
lucasagomesbut I don't have anything in place that actually do/trigger any undeploy16:53
devanandalike, if no one is saying "i'm going to fix this" then why would we tell the community "we'll have it fixed by this $date" ...16:53
lucasagomesmakes sense16:54
devanandalucasagomes: i'm planning a 0.1.0 client release once the deploy/undeploy API is written & added to the client. sound good?16:55
lucasagomesbut at least I should go trough the list and verify some of the bugs to see whether they r valid or not16:55
openstackgerritJames E. Blair proposed a change to openstack/ironic: Add -U to pip install command in tox.ini  https://review.openstack.org/6148816:55
devanandalucasagomes: ++16:55
devanandawe have a lot of open bugs, many of which I suspect are stale. I closed a couple last night that weren't applicable any more16:55
lucasagomesdevananda, sounds good, the deploy/undeploy, has some dependencies and when they get merged I think that's a good time to make a release16:55
lucasagomesdevananda, yea, I'm haven't been paying much attention for the new bugs, I should I will try to look at them more often16:56
*** rongze_ has quit IRC16:56
lucasagomesshame on me16:56
*** max_lobur has joined #openstack-ironic16:58
*** yuriyz has quit IRC17:05
*** rloo has quit IRC17:06
devanandalucasagomes: i haven't been doing triage as often as I should -- so I'm getting on top of it now, but there's a bit of a backlog17:06
*** rloo has joined #openstack-ironic17:06
lucasagomesI see, yea I can imagine17:06
devanandafor now, most of the bug filing is still being done by us, so it should be easy for us to stay on top of -- if we just all follow the policy when we file 'em17:07
devanandaonce we get real users filing bugs, well, we'll see .... Nova's taken what I find to be a bit of a heavy approach, but then, they have a very high volume and needed to17:07
*** ifarkas has joined #openstack-ironic17:09
*** rongze__ has quit IRC17:11
lucasagomesyea, I believe we are going to need some approach to deal with that once ironic we have ironic ready to replace nova baremetal, but yea let's not complicate things for now17:13
*** matty_dubs is now known as matty_dubs|lunch17:16
openstackgerritJames E. Blair proposed a change to openstack/python-ironicclient: Have tox install via setup.py develop  https://review.openstack.org/6150617:17
*** rongze has joined #openstack-ironic17:17
*** tatyana has quit IRC17:17
NobodyCambbt...brb17:23
*** digambar has joined #openstack-ironic17:25
digambarI have cloned all the code of ironic17:27
digambarneed to test it17:27
digambarhow can I do it17:27
digambar?17:27
NobodyCamdigambar: have you seen our wiki17:28
digambaryes17:28
digambarI am following those steps17:29
digambaronce done get back to you17:29
NobodyCamok :)17:29
digambarThanks17:29
*** blamar has joined #openstack-ironic17:30
openstackgerritMikhail Durnosvistov proposed a change to openstack/ironic: Clean up duplicated change-building code in objects  https://review.openstack.org/6114317:33
lucasagomesdigambar, http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html17:33
lucasagomesah u following it already :)17:34
*** digambar has quit IRC17:34
*** digambar has joined #openstack-ironic17:39
digambarhi17:39
digambarAfter following setup steps for ironic17:39
digambarwhen I run17:39
digambarcreate virtualenv17:40
digambargot error17:40
digambarNo command 'create' found, did you mean:  Command 'mcreate' from package 'lustre-utils' (universe) create: command not found17:40
digambarI get above error17:40
NobodyCamdigambar: which are you doing local, devstack or DIB?17:41
digambarubuntu os17:41
digambarthis error is getting while installing ironic17:41
digambardevstack is local17:42
NobodyCamdigambar: https://wiki.openstack.org/wiki/Ironic17:42
NobodyCamI was reffering to the three ways we have up on hte wiki17:42
NobodyCamso you are doing the devstack version?17:42
devanandadigambar: you create teh virtualenv by running " tox -evenv -- echo 'done'17:43
devanandawoops, bad paste17:43
devanandadigambar: you create teh virtualenv by running: tox -evenv -- echo 'done'17:43
digambarokk17:43
digambartox -evenv -- echo 'done'17:54
digambarpython setup.py build_sphnix17:54
digambargetting error17:54
digambarerror: invalid command 'build_sphinx'17:55
NobodyCamdigambar: did you source the venv?  source .tox/venv/bin/activate17:55
digambaryes17:55
digambar(venv)root@openstack-cloud:~/ironic# python setup.py build_sphinx17:55
digambarusage: setup.py [global_opts] cmd1 [cmd1_opts] [cmd2 [cmd2_opts] ...]    or: setup.py --help [cmd1 cmd2 ...]    or: setup.py --help-commands    or: setup.py cmd --help  error: invalid command 'build_sphinx'17:56
digambarabove error it  throws17:56
*** derekh has quit IRC17:57
NobodyCamare you setting up ironic or building he docs?17:58
digambarsetting up th eironic18:00
digambarironic**18:00
digambarlet me try it out with devstack18:01
digambarI think that is easy way18:01
NobodyCamdigambar: you seem to be following the build the docs walk thru18:01
*** lexx has joined #openstack-ironic18:02
digambarok18:03
digambarhow to install the ironic ?18:03
digambarthat's the doc it is, right?18:03
NobodyCamhttps://wiki.openstack.org/wiki/Ironic#Try_it_on_Devstack18:03
NobodyCami use this one. https://wiki.openstack.org/wiki/Ironic#Using_Disk_Image_Builder18:03
*** matty_dubs|lunch is now known as matty_dubs18:03
devanandadigambar: for what purpose are you trying to install ironic?18:04
digambarI want to contribe to ironic ?18:04
NobodyCamya the two link I pasted will setup ironic18:04
digambarokk18:04
NobodyCamThe one on the top of the page is for building the docs18:05
digambaryep18:05
digambarThanks man18:06
NobodyCamsure :)18:06
digambarI'll setup up with devstack18:06
digambar:)18:06
*** hemna has quit IRC18:08
*** max_lobur is now known as max_lobur_afk18:09
*** jistr has quit IRC18:13
NobodyCambrb quick walkies18:13
*** mdurnosvistov has quit IRC18:24
*** digambar has quit IRC18:26
*** datajerk has joined #openstack-ironic18:28
openstackgerritA change was merged to openstack/ironic: Add -U to pip install command in tox.ini  https://review.openstack.org/6148818:28
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Add JsonPatchType class  https://review.openstack.org/6019418:33
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Rework patch validation on Ports  https://review.openstack.org/5698418:33
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Reworks Node validations  https://review.openstack.org/5749518:33
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Reworks Chassis validations  https://review.openstack.org/5768618:33
*** blamar has quit IRC18:41
*** blamar has joined #openstack-ironic18:41
*** lexx has quit IRC18:45
*** anniec has joined #openstack-ironic18:50
*** vetalll has joined #openstack-ironic18:55
*** jdob has quit IRC19:03
*** jdob has joined #openstack-ironic19:03
lucasagomesright I'm done for today have a good night everyone19:10
devanandag'night!19:11
*** lucasagomes has quit IRC19:11
NobodyCamgah he's quicl19:12
NobodyCamquick even19:12
*** rongze has quit IRC19:13
*** rongze has joined #openstack-ironic19:15
NobodyCamdevananda: with the computer driver it's set here19:21
NobodyCamhttps://github.com/openstack/tripleo-heat-templates/blob/master/undercloud-vm-source.yaml#L2019:21
NobodyCami'll update the walk thru19:22
*** rongze has quit IRC19:22
devanandaNobodyCam: yea. i think when you are make'ing the ironic undercloud yaml, that should be replaced with ironic.driver.IronicDriver19:22
NobodyCamyep19:23
*** mdurnosvistov has joined #openstack-ironic19:26
openstackgerritdekehn proposed a change to openstack/ironic: Method to generate PXE options for Neutron ports  https://review.openstack.org/6090419:29
*** lexx has joined #openstack-ironic19:30
*** ifarkas has quit IRC19:33
*** openstackgerrit has quit IRC19:34
*** openstackgerrit has joined #openstack-ironic19:34
NobodyCambrb19:41
devanandadkehn: reviewed ^ just now.20:06
devanandadkehn: any thoughts / progress on the actual "send this info to neutron" code? anything blocking that work that i can help unblock?20:07
*** jistr has joined #openstack-ironic20:08
*** jistr has quit IRC20:08
*** jistr_ has joined #openstack-ironic20:09
*** jistr_ has quit IRC20:09
*** jistr has joined #openstack-ironic20:10
*** jistr has quit IRC20:11
*** jistr has joined #openstack-ironic20:11
*** romcheg has left #openstack-ironic20:25
*** romcheg has joined #openstack-ironic20:25
*** romcheg1 has joined #openstack-ironic20:28
*** vkozhukalov has quit IRC20:29
*** romcheg has quit IRC20:30
NobodyCambrb again20:34
NobodyCam:-p20:34
*** romcheg has joined #openstack-ironic20:38
*** tatyana has joined #openstack-ironic20:39
*** romcheg1 has quit IRC20:40
*** tatyana has quit IRC20:42
romchegdevananda, vetalll: if you want to discuss the bug, I'm available for that.20:48
devanandaromcheg: which bug? I have been triaging a lot today :p20:48
romchegOh :)20:48
romchegThis one: https://bugs.launchpad.net/ironic/+bug/125188020:48
*** kui has joined #openstack-ironic20:49
*** lexx has quit IRC20:51
* devananda looks20:53
devanandaah!20:53
*** lucasagomes has joined #openstack-ironic20:54
lucasagomesdevananda, ping20:54
devanandalucasagomes: pong20:55
lucasagomesdevananda, I comment on the 60194 re staticmethod20:55
romchegdevananda: Does it make sense to add that example to the description20:55
lucasagomesmaybe that's not the best way to do it, so I want to discuss it here20:55
*** anniec has quit IRC20:55
devanandaromcheg: i have no problem with /v1.json or /v1.xml20:58
devanandaromcheg: without any auth20:58
devanandaromcheg: well, the root controller works. but the resource controllers don't20:59
devananda$ curl -H "Content-type: application/json" http://192.0.2.2:6385/v1.json/nodes/20:59
*** jistr has quit IRC20:59
devananda{"error_message": "{\"debuginfo\": null, \"faultcode\": \"Client\", \"faultstring\": \"Expected an uuid or int but received v1.json.\"}"}20:59
devanandalucasagomes: thinking ...20:59
romchegI think it should not work this way21:00
romchegJust compare http://192.0.2.2:6385/v1.json/nodes/ and http://192.0.2.2:6385/v1.json/nodes/21:00
romchegArgh!21:00
romchegJust compare http://192.0.2.2:6385/v1.json/nodes/ and http://192.0.2.2:6385/v1.json/nodes/21:00
devanandalucasagomes: i understand the need for an inherited class to access teh base classes internal_attr21:00
romchegWTF21:00
romchegJust compare http://192.0.2.2:6385/v1.json and http://192.0.2.2:6385/v121:01
devanandalucasagomes: but how is this different from mandatory_attr, except that the baseclass defines this as an empty list today?21:01
lucasagomesdevananda, no difference21:01
devanandaromcheg: there is no difference21:01
lucasagomesI mean21:01
lucasagomesthe base class does have any mandatory attributes by default21:02
romchegdevananda: Both do not require authentication, do they?21:02
devanandaromcheg: correct21:02
lucasagomesmaybe both should be staticmethod for consistence, or I should think in another way to access the attributes from the base class21:02
romchegThen the bug appears to be fixed21:03
romchegBy some brave but unknown hero :)21:03
devanandaromcheg: however. there may be something odd in my env, but i dont know how ...21:03
romchegI'm checking the sources21:03
devanandaromcheg: i am able to curl v1/nodes/{UUID} without passing any auth token in headers21:03
romchegAh, you have noauth in your settings file21:04
devanandaNobodyCam: do you have an environment spun up?21:04
devanandaromcheg: ahh. that may be21:04
devanandalucasagomes: i think both being staticmethod is fine21:04
romchegAfter changing that to keystone, you will notice the difference21:04
NobodyCam??? not atm was rebuilding my env21:04
lucasagomesdevananda, right, I will update the patches21:05
lucasagomescheers :)21:05
devanandalucasagomes: i pulled the following patch (that fixes up the port class) and changed it to staticmethod in both base and port modules, didn't have any problems21:05
devanandalucasagomes: cheers! I'll +2 it for realz this time :)21:05
lucasagomesdevananda, yea there's np, they get overwritten21:05
devanandaNobodyCam: nvm - think i figured it out21:05
lucasagomesdevananda, hehe np, if you spot another thing don't hesitate in -1 or question it21:05
* lucasagomes learns a lot from the reviews21:06
devanandaromcheg: wait, where are you seeing noauth in a setting file?21:06
romchegIf you are able to access /v1/nodes without submitting an auth token, than means you have noauth21:07
romchegWhich is default21:07
devanandaahhh got it21:08
romcheg# Method to use for auth: noauth or keystone. (string value)21:08
romcheg#auth_strategy=noauth21:08
devanandabingo21:08
devanandaNobodyCam: your template needs to include "auth_strategy=keystone"21:08
devanandaactually21:08
devanandawe just need to change the default21:09
devanandaNobodyCam: nvm again -- i'm just going to patch it21:09
devanandano auth is not a sane default21:09
NobodyCamlol sure21:09
NobodyCamnova-ironic adds all the auth info to nova conf for ironic21:09
romchegWe do not have keystone options in the sample config file21:09
NobodyCamthere was a line on the agenda for that21:10
romchegI think I know how to fix that21:10
openstackgerritDevananda van der Veen proposed a change to openstack/ironic: Change default API auth to keystone-based  https://review.openstack.org/6156821:11
devanandaromcheg: ^21:11
devananda:)21:11
*** epim has joined #openstack-ironic21:11
romchegI definitely have problems with reviews.21:11
romchegShame on me21:11
romchegDo other projects use keystone as the default value for that?21:12
devanandaromcheg: ok, NOW i see the problem you're talking about in the bug21:12
romchegIn nova they use noauth: https://github.com/openstack/nova/blob/master/etc/nova/nova.conf.sample#L32521:13
*** anniec has joined #openstack-ironic21:14
devanandalifeless: i suspect i know your opinion on using auth_strategy=noauth as the default. Mine is that it is not a #sanedefault21:14
romchegIn neutron they use keystone21:16
devanandahere's another oddity21:17
devanandachanging header to -H "Content-type: text/xml" does not change response type21:17
romchegdevananda: you should specify Accept21:19
romchegContent-type defines the format of the request but does not specifies the expected responce21:19
romchegresponse21:19
romchegDamn word, I will never remember how to write it :)21:19
devanandahmm21:20
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Add JsonPatchType class  https://review.openstack.org/6019421:20
devanandaah. duh, of course21:20
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Rework patch validation on Ports  https://review.openstack.org/5698421:20
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Reworks Node validations  https://review.openstack.org/5749521:20
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Reworks Chassis validations  https://review.openstack.org/5768621:21
devanandaromcheg: right. and that works :)21:21
devanandaromcheg: curl -H "Accept: text/xml" ...21:21
devanandaromcheg: also taht works without auth. so only the URI suffix is getting blocked21:22
*** epim has quit IRC21:25
*** lucasagomes is now known as lucas-dinner21:29
*** epim has joined #openstack-ironic21:32
*** anniec has quit IRC21:33
*** jcooley_ has quit IRC21:34
lifelessdevananda: I agree :)21:35
*** epim has quit IRC21:36
openstackgerritArtem Rozumenko proposed a change to openstack/ironic: removed wrap_exception method from ironic/common/exception.py  https://review.openstack.org/6147421:36
*** jcooley_ has joined #openstack-ironic21:39
*** anniec has joined #openstack-ironic21:42
*** lexx has joined #openstack-ironic21:48
NobodyCamdevananda: do we need to poke any one about this review? https://review.openstack.org/#/c/6063821:49
devanandamaybe. lemme see21:51
*** vetalll has quit IRC21:51
devanandaI've added a few requirements-core to the review. and we can poke when it becomes more pressing21:54
devanandaI'll tag it with a bp to raise the review ranking21:54
NobodyCam:)21:54
devanandawell, nvm. i'm not sure that works across projects21:55
devanandathe bp is in nova....21:55
*** kushi_ has joined #openstack-ironic21:57
*** kui has quit IRC21:58
SpamapShm21:58
SpamapSwill Ironic have something to mark a baremetal node as "don't deploy new things to this" ?21:59
SpamapSOh and hey, y'all should write some Heat resource plugins.21:59
*** anniec has quit IRC22:00
devanandaheya SpamapS22:01
devanandawe don't currently have that. it sounds like a bug (want to file it?)22:01
devananda'cause we totally should expose that in our API22:01
SpamapSdevananda: yes.. unless you have a better answer for "how do I prevent new things from going onto hardware right after I delete the instance deployed on it?"22:02
devanandaSpamapS: as far as heat resource plugins, I'm not up on that, but NobodyCam has been working on adding Ironic to tripleo / dib / etc. Perhaps he has something?22:02
devanandaSpamapS: well, as it is today, *you* dont do that. Ironic will not allow new requests on that node until it is finished doing what ever teardown it needed to22:03
SpamapSNobodyCam: we haven't talked about Heat plugins.. but it would make a lot of sense to be able to express inventory as a heat stack. :)22:03
*** jdob has quit IRC22:03
devanandaSpamapS: in what situation would you want to explicitly delete an instance and then stop any activity on the node that instance was previously on? that seems like conflating two things to me22:03
lifelessSpamapS: maybe22:03
devanandabut IMBMS22:03
SpamapSdevananda: So I'm responding to a laundry list of DC OPS / TechOps situations from almost a year ago.. and the "how do I replace a hard drive that requires power off / power on" story is tricky without that.22:04
lifelessSpamapS: but you can't deploy to a thing you can't match hardware descriptions against :)22:04
*** linggao has quit IRC22:04
NobodyCamSpamapS: ya22:04
devanandaSpamapS: the "put this node into maintenance mode" thing is _totaly_ something we need to expose22:05
lifelessSpamapS: I know the list of which you speak, and mmm the use cases are good but it was shoehorned into specific products on a marketing basis.22:05
devanandaSpamapS: the relationship between "delete an instance" and "put that node into maintenance" doesn't seem like something Ironic cares about, but we should give you enough info to do it22:05
SpamapSlifeless: Right, I've been asked to resurrect it for a deep dive w/ some HP peeps next week.22:05
lifelessputting a node into maintenance seems like 'evacuate this hypervisor' to me.22:05
devanandalifeless: yep22:06
devanandasimilar22:06
lifelessSpamapS: face to face or virt? May I hang with?22:06
SpamapSIt is evacuate yes.22:06
SpamapSlifeless: in person in Sunnyvale. Would be happy to G+ you in :)22:06
lifelessso, if you have evacuate with say kvm and no live migrate, how should heat handle that... is how we should frame it, I think.22:06
SpamapSlifeless: In this case, I want to delete the instance.22:07
devanandaSpamapS: sounds like something I should be pulled into as well. and sunnyvale isn't that far for me ....22:07
SpamapSlifeless: I don't want that IP to be up anymore. Heat has already let go of the machine.22:07
lifelessto whit, you want to a) stop scheduling more work to the hypervisor, b) signal out somewhere that the instances on it are about to be deleted, c) wait for a timeout, d) delete.22:07
SpamapSdevananda: The more the merrier! and I would feel quite a bit better about things if I had an ironic expert there with me.22:07
lifelessSpamapS: How did Heat know to let go ?22:07
SpamapSlifeless: so the box is already down logically. Disk is dead.22:08
devanandaSpamapS: pls send me the details (assuming someone hasn't already ... i'm a bit behind on corp mail right now)22:08
SpamapSno raid, or raid double-degraded.. disk problems.. box is not serving traffic22:08
lifelessSpamapS: then heat will have deleted the instance already, nothing to see here22:08
lifelessSpamapS: 'I need to delete the instance' only applies if heat doesn't yet know there is a problem22:08
SpamapSlifeless: right, and then ironic starts trying to put new things on it... wasting new things' time.22:08
lifelessSpamapS: so thats a different scenario now - I agree they relate22:09
lifelessbut 'I have a box that can't be deployed to and it shouldn't make everything super slow'22:09
lifelessvs 'I have a box that I want to take out usage which currently has a workload on it'22:09
SpamapSthe progression is that between box down and delete instance I want to make the hardware unavailable.22:10
NobodyCamsorry guy I have to step away for just a bit... be back22:10
lifelessSpamapS: why?22:10
SpamapSlifeless: because if it gets reclaimed, how do I then un-claim it?22:10
SpamapSI can't delete an in-use baremetal node.22:11
devanandaSpamapS: a) sure you can22:11
SpamapSAssuming here, that Ironic also has a similar requirement.22:11
lifelessSpamapS: so this is why I'm challenging you to make these really crystal clear scenarios22:11
lifelessSpamapS: because if the box is *down*, it won't get anything new on it.22:11
lifelessSpamapS: because it will fail to deploy.22:11
SpamapSit will get tried22:11
lifelessSpamapS: which is the 'I have a box that can't be deployed to and it shouldn't make everything super22:11
lifeless                  slow'22:11
lifelessscenario that happens independently22:11
lifelessso we have to solve that22:12
devanandalifeless: i think the technical solution to both scenarios is identical22:12
lifelessbut why does heat care that Ironic will try a bad box ? -> Because it makes things slow. So we need to fix *that*22:12
lifelessdevananda: could be!22:12
devanandalifeless: though i agree they are different scenarios22:12
SpamapSseems relatively simple and I'm surprised you are trying to flush things out with something that is pretty obvious to me: boxes will be taken out of the pool of available hardware.22:12
lifelessSpamapS: I'm trying to avoid layer confusion22:12
* devananda likes layer cakes22:13
SpamapSHeat, btw, does not care that ironic will try a bad box.22:13
SpamapSI just mentioned Heat because it came to mind.22:13
lifelessSpamapS: hey, we were going to do face time about service enablement22:13
devanandaso, here is the technical side of solving this22:13
SpamapSSo ignore Heat. Heat will not care. :)22:13
SpamapSlifeless: we should yes22:13
devanandamark the node as in maintenance mode first -- this should work whether the physical hardware is functioning or not22:14
devanandait won't kill any existing workload, but it should stop ironic from putting new workload on it AND remove its availablility from nova-scheduler (at the next periodic-task cycle)22:14
devanandathen delete the instance -- this should also work whether or not the physical hardware is functioning, though i'm not sure it will, today22:15
lifelessdevananda: seems like we should be able to tell the scheduler immediately22:15
lifelessdevananda: if you cannot, please file a bug?22:15
devanandalifeless: AFAIK, nova-scheduler only processes updates to the list of compute hosts once a minute22:15
lifelessdevananda: we're fixing that :)22:15
devanandaooh! great22:16
devanandathen nvm :)22:16
lifelesspoint is that I think there may not be an 'unregister' call today22:16
lifelessso we need to add that22:16
devanandaright22:16
devanandain both the baremetal and ironic drivers in nova22:16
lifelessfor which a bug would be useful22:16
devanandaSpamapS: did you file a bug for ^^ in ironic?22:17
SpamapSno because I am still wrapping head around what is actually needed :)22:17
SpamapSand as you all know, my head is rather thick and inflexible ;)22:17
lifeless:)22:17
* devananda files it, then takes a break22:17
lifelessso, I think at the ironic layer maintenance + instance deletion being decoupled is super useful22:17
lifelessbut22:17
lifelessI want to point out that the folk running the overcloud22:18
lifelessand the folk provisioning the hardware may be separate22:18
lifelessand not share creds22:18
lifelessso this is why I'm talking layer confusion.22:18
lifelessLong term I think we want to be able to say 'node X is in trouble, it is to be removed' and have that propogate gracefully out to $whereever22:18
lifelessheat, chef, Fred in the corner with a laptop - I don't care22:18
lifelessto let them gracefully [assuming it's not a hard failure] get their instance off the node22:19
lifelessAnd secondly I think we have to make sure we don't bake in an assumption like 'the Heat that deploys the overcloud will be able to do admin operations on Ironic'.22:19
lifelessNote that I am not arguing that such ops team separation is *good*, I am just arguing that it *is*.22:20
lifelessSpamapS: so - throw me a meeting invite for the deep dive22:21
lifelessSpamapS: and g+ now?22:21
*** anniec has joined #openstack-ironic22:21
SpamapSlifeless: G+ in a minute, just captured neutron/nova in a race that takes a few minutes to resolve hoping to get more data into a bug22:22
lifelessok22:22
devanandalifeless: I agree on the unassailable fact that there *will* be privilege separation at that layer, whether or not we agree with it22:23
devanandaalso, here's the bug: https://bugs.launchpad.net/ironic/+bug/126009922:23
lifelesscool, thanks!22:25
lifelessdevananda: you say 'should not adversely affect management of...'22:25
lifelessdevananda: I'd like to propose adding 'or operation' before 'of' :)22:25
lifelesse.g. it shouldn't power the node off :)22:26
devanandalifeless: ++22:26
lifelessI'll edit the description22:26
* devananda did22:26
*** kushi_ has quit IRC22:26
*** kushi_ has joined #openstack-ironic22:27
lifelessheh, inflight - cool22:27
SpamapSoh hmmmm22:28
SpamapSno agents for compute node22:28
*** matty_dubs is now known as matty_dubs|gone22:29
devanandalifeless: https://bugs.launchpad.net/nova/+bug/126010222:31
*** michchap has joined #openstack-ironic22:31
devanandalifeless: i need to run. please edit that if you are so inclined :)22:31
lifelessdevananda: I will triage it for you22:31
lifelessdevananda: tsk!22:31
lifelessdevananda: and ciao!22:32
* devananda steps out for a while22:32
openstackgerritdekehn proposed a change to openstack/ironic: Method to generate PXE options for Neutron ports  https://review.openstack.org/6090422:35
*** anniec has quit IRC22:45
*** anniec has joined #openstack-ironic22:47
*** mdurnosvistov has quit IRC22:48
*** kushi_ has quit IRC22:51
*** kushi_ has joined #openstack-ironic22:52
*** romcheg has quit IRC22:54
*** datajerk has quit IRC22:55
*** anniec has quit IRC23:01
*** kushi_ has quit IRC23:05
*** kushi__ has joined #openstack-ironic23:05
*** anniec has joined #openstack-ironic23:07
*** anniec has quit IRC23:10
*** lucas-dinner has quit IRC23:33
*** openstackgerrit has quit IRC23:35
*** openstackgerrit has joined #openstack-ironic23:36
*** blamar has quit IRC23:39
*** anniec has joined #openstack-ironic23:40
*** kushi__ has quit IRC23:43
*** openstackgerrit has quit IRC23:56
*** openstackgerrit has joined #openstack-ironic23:56
*** blamar has joined #openstack-ironic23:56
« Tuesday, 2013-12-10 Index Thursday, 2013-12-12 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!