*** aginwala has quit IRC | 00:01 | |
openstackgerrit | Sinval Vieira Mendes Neto proposed openstack/ironic: Improving capabilities_to_dict and adding helpful tests https://review.openstack.org/254972 | 00:03 |
---|---|---|
*** alexpilotti has joined #openstack-ironic | 00:03 | |
openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add documentation for Swiftless deploy https://review.openstack.org/241893 | 00:05 |
NobodyCam | yes | 00:06 |
*** Sukhdev_ has joined #openstack-ironic | 00:06 | |
*** aginwala has joined #openstack-ironic | 00:06 | |
*** ChubYann has quit IRC | 00:07 | |
*** naohirot has joined #openstack-ironic | 00:07 | |
*** chlong has quit IRC | 00:09 | |
NobodyCam | I like #1 contrib dir :) | 00:11 |
NobodyCam | I feel it help keep the community as a whole in touch with how their work is being used, | 00:13 |
NobodyCam | helps* | 00:13 |
*** ijw has joined #openstack-ironic | 00:14 | |
devananda | jroll: a comment on lucas-dinner' "indexable JSON fields" spec just made me think of something scheduler / claims related | 00:15 |
*** rajathagasthya has quit IRC | 00:15 | |
devananda | jroll: scheduling based on compute capabilities filter | 00:16 |
devananda | have I already brought this up? | 00:16 |
NobodyCam | ++++ | 00:16 |
jroll | devananda: I believe so? | 00:16 |
jroll | we're planning on allowing capabilities | 00:17 |
jroll | we even talked abuot making that filter a library in tokyo | 00:17 |
devananda | jroll: making it a library implies using it across projects | 00:20 |
jroll | well, the point was the incomprehensible code | 00:20 |
*** ChubYann has joined #openstack-ironic | 00:21 | |
devananda | jroll: https://github.com/openstack/nova/blob/master/nova/scheduler/filters/compute_capabilities_filter.py#L33 | 00:22 |
devananda | it's incomprehensible -- but its inputs are not things that Ironic has | 00:23 |
jroll | devananda: right, we make it take arguments and pass in what we need | 00:24 |
devananda | jroll: except no one wants to change that code in nova | 00:25 |
jroll | devananda: surely it's refactorable | 00:26 |
devananda | i mean, no one is willing to take the risk of changing it | 00:26 |
devananda | nope | 00:26 |
devananda | I'm not | 00:26 |
devananda | I'm not going to try refactoring it again | 00:26 |
jroll | sigh | 00:27 |
openstackgerrit | Yuiko Takada proposed openstack/ironic: Migrate Tempest tests into Ironic tree https://review.openstack.org/253982 | 00:31 |
jroll | devananda: so what are you trying to say then? | 00:31 |
devananda | jroll: dunno. that it's cider-o'clock? | 00:39 |
jroll | s/cider/whiskey | 00:41 |
jroll | :) | 00:41 |
devananda | ++ | 00:41 |
openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add documentation for user inputs as HTTPS URLs https://review.openstack.org/242771 | 00:42 |
*** rajathagasthya has joined #openstack-ironic | 00:53 | |
*** aginwala has quit IRC | 00:56 | |
*** garthb has quit IRC | 00:57 | |
*** sinval has quit IRC | 01:01 | |
*** chlong has joined #openstack-ironic | 01:05 | |
*** aginwala has joined #openstack-ironic | 01:06 | |
JayF | Writing some downstream hardware manager tests when I discover ipa.extensions.get_clean_steps doesn't always return clean steps in a consistent order | 01:08 |
JayF | :( | 01:08 |
JayF | is that something we'd consider a bug and want to fix? | 01:08 |
*** yuanying_ has joined #openstack-ironic | 01:09 | |
*** yuanying_ has quit IRC | 01:09 | |
*** yuanying has quit IRC | 01:11 | |
*** zhenguo has joined #openstack-ironic | 01:13 | |
*** Sukhdev_ has quit IRC | 01:15 | |
*** rajathagasthya has quit IRC | 01:17 | |
*** chlong has quit IRC | 01:23 | |
*** mtanino has quit IRC | 01:23 | |
*** yuanying has joined #openstack-ironic | 01:23 | |
*** rajathagasthya has joined #openstack-ironic | 01:24 | |
*** rajathagasthya has quit IRC | 01:28 | |
*** Sukhdev has joined #openstack-ironic | 01:30 | |
*** chenke__ has quit IRC | 01:40 | |
*** chenke__ has joined #openstack-ironic | 01:41 | |
*** smoriya has joined #openstack-ironic | 01:42 | |
*** alexpilotti has quit IRC | 01:47 | |
*** jamielennox|away is now known as jamielennox | 01:48 | |
*** aginwala has quit IRC | 01:48 | |
*** zhenguo has quit IRC | 01:54 | |
*** boris-42_ has quit IRC | 01:54 | |
*** dolphm has quit IRC | 01:54 | |
*** johnthetubaguy has quit IRC | 01:54 | |
*** dolphm has joined #openstack-ironic | 01:54 | |
*** zhenguo has joined #openstack-ironic | 01:55 | |
*** boris-42_ has joined #openstack-ironic | 01:56 | |
*** rajathagasthya has joined #openstack-ironic | 01:56 | |
*** johnthetubaguy has joined #openstack-ironic | 01:57 | |
*** Sukhdev has quit IRC | 01:58 | |
*** rcernin has quit IRC | 02:01 | |
*** aginwala has joined #openstack-ironic | 02:03 | |
*** davidlenwell has quit IRC | 02:10 | |
*** jdk has joined #openstack-ironic | 02:13 | |
openstackgerrit | Ruby Loo proposed openstack/ironic: Add FSM.is_stable() method https://review.openstack.org/255640 | 02:13 |
*** davidlenwell has joined #openstack-ironic | 02:13 | |
*** rloo has quit IRC | 02:13 | |
jdk | TheJulia: I found this on ipxe.org http://ipxe.org/howto/chainloading#uefi I think that explains why chainloading wasn't working today in UEFI mode. | 02:14 |
jdk | TheJulia: I suggest that you make some mention of this in the readme | 02:14 |
openstackgerrit | Ruby Loo proposed openstack/ironic: Add FSM.is_stable() method https://review.openstack.org/255640 | 02:18 |
*** jdk has left #openstack-ironic | 02:19 | |
*** hoangcx has joined #openstack-ironic | 02:24 | |
*** links has joined #openstack-ironic | 02:33 | |
*** thanhnt-z has joined #openstack-ironic | 02:37 | |
*** aginwala has quit IRC | 02:38 | |
*** baoli has joined #openstack-ironic | 02:43 | |
*** chlong has joined #openstack-ironic | 02:44 | |
*** aginwala has joined #openstack-ironic | 02:44 | |
openstackgerrit | Zhenguo Niu proposed openstack/ironic-specs: Console ports allocation management https://review.openstack.org/249876 | 02:46 |
*** praneshp has quit IRC | 02:55 | |
*** aginwala has quit IRC | 03:02 | |
*** rajathagasthya has quit IRC | 03:03 | |
*** boris-42_ has quit IRC | 03:03 | |
*** aginwala has joined #openstack-ironic | 03:06 | |
*** aginwala has quit IRC | 03:10 | |
*** ijw has quit IRC | 03:13 | |
jlvillal | naohirot, I just commented in the patch | 03:15 |
jlvillal | naohirot, Basically asking if your development environment might be named 'localhost' | 03:16 |
*** yuanying has quit IRC | 03:19 | |
*** Haomeng|2 has quit IRC | 03:28 | |
*** Haomeng|2 has joined #openstack-ironic | 03:28 | |
*** Sukhdev has joined #openstack-ironic | 03:31 | |
*** yuanying has joined #openstack-ironic | 03:37 | |
*** baoli has quit IRC | 03:39 | |
*** praneshp has joined #openstack-ironic | 03:50 | |
*** praneshp has quit IRC | 03:57 | |
*** praneshp has joined #openstack-ironic | 03:58 | |
*** kkoski has quit IRC | 03:59 | |
*** yuanying has quit IRC | 04:02 | |
*** changzhi has joined #openstack-ironic | 04:03 | |
*** davidlenwell has quit IRC | 04:06 | |
*** yuanying has joined #openstack-ironic | 04:06 | |
naohirot | jlvillal: Hi, I was out for lunch. | 04:07 |
*** praneshp_ has joined #openstack-ironic | 04:08 | |
*** davidlenwell has joined #openstack-ironic | 04:09 | |
*** praneshp has quit IRC | 04:11 | |
*** praneshp_ is now known as praneshp | 04:11 | |
jlvillal | naohirot, Hey there | 04:13 |
jlvillal | naohirot, I hope your lunch was good :) | 04:13 |
naohirot | jlvillal: Yeah, it was good, and started sleepy noon :) | 04:14 |
jlvillal | naohirot, :) | 04:14 |
jlvillal | naohirot, Did my comments make sense? | 04:14 |
jlvillal | naohirot, Is your development environment named 'localhost'? | 04:14 |
naohirot | jlvillal: Yes, I'm just writing reply to you and tan. | 04:15 |
jlvillal | Okay, thanks | 04:15 |
jlvillal | naohirot, I'm going to go to sleep. Thanks for working on this! | 04:15 |
naohirot | jlvillal: My env, hostname is ubuntu and fqdn is localhost | 04:15 |
jlvillal | naohirot, That would cause the issue :) | 04:16 |
jlvillal | I think my suggested fix will work. | 04:16 |
naohirot | jlvillal: My dev is in the private network, so I don't use dns. | 04:16 |
* jlvillal thinks that for all of his suggested fixes :) | 04:16 | |
jlvillal | naohirot, Makes sense. | 04:16 |
jlvillal | naohirot, I think this is not unusual. So I think a fix would be good too. | 04:17 |
naohirot | jlvillal: did you already submit a patch? | 04:17 |
jlvillal | naohirot, If you do the fix for 'localhost' please update the commit message. | 04:17 |
jlvillal | naohirot, No. I just think my proposed fix will work. | 04:17 |
jlvillal | naohirot, I'll let you test it, if you like :) | 04:17 |
jlvillal | naohirot, Proposed as in I put a comment with an idea in your patch. | 04:18 |
*** praneshp_ has joined #openstack-ironic | 04:18 | |
jlvillal | naohirot, Have a good day. | 04:18 |
naohirot | jlvillal: Oh, I just read your comment in generator.py | 04:18 |
jlvillal | naohirot, That's the one! :) | 04:19 |
* jlvillal gets ready for bed. | 04:19 | |
jlvillal | naohirot, Good night | 04:19 |
naohirot | jlvillal: I think your suggestion certainly fix so as not to generate 'ironic'. | 04:19 |
*** praneshp has quit IRC | 04:19 | |
*** praneshp_ is now known as praneshp | 04:19 | |
jlvillal | naohirot, Great. I think using 'localhost' is not that unusual with development environments. So I think a good thing to fix | 04:20 |
jlvillal | naohirot, bye | 04:20 |
naohirot | jlvillal: wait, but do you have any idea why the current code generate 'ironic'? | 04:20 |
jlvillal | naohirot, Yes | 04:20 |
jlvillal | naohirot, Because your system is named 'localhost' | 04:20 |
naohirot | jlvillal: No no | 04:21 |
jlvillal | And the default value is 'localhost' | 04:21 |
jlvillal | fqdn is localhost | 04:21 |
jlvillal | That also works | 04:21 |
jlvillal | If either is localhost then it matches | 04:21 |
naohirot | jlvillal: why does the current specification of this generator need to set 'ironic' as a host name? | 04:21 |
jlvillal | naohirot, It is trying to fix things if the user had set it to their current system name. | 04:22 |
jlvillal | naohirot, I wouldn't change that part of the logic. | 04:22 |
naohirot | jlvillal: why does the author of this code choose 'ironic' as a host name? | 04:22 |
jlvillal | naohirot, I don't know the history. | 04:22 |
jlvillal | naohirot, I was going with the simplest solution. | 04:23 |
jlvillal | naohirot, I have to go now. | 04:23 |
naohirot | jlvillal: that's the reason I hesitate to change 'ironic' to just 'localhost' | 04:23 |
jlvillal | naohirot, I need sleep.... :) | 04:23 |
jlvillal | jlvillal, That is not what I suggested | 04:23 |
jlvillal | naohirot, I suggested adding an 'elif' statement | 04:23 |
naohirot | jlvillal: Yeah, thanks!! have a good night, I'll reply to you in the gerrit :) | 04:24 |
jlvillal | naohirot, So it would perform the same on a system with a real FQDN/hostname as it does if the system has either a FQDN/hostname of 'localhost' | 04:24 |
jlvillal | Ciao | 04:25 |
naohirot | jlvillal: gn :-) | 04:25 |
*** amotoki has quit IRC | 04:25 | |
changzhi | hi, all. I boot a vm by using "nova boot --image xx --flavor xx --nic net-id=xx --key-name new test3". when the machine boot up, i want to login the machine. I use the command "ssh -i ~/.ssh/id_rsa.pub root@10.0.0.205" why it let me input password?? | 04:32 |
changzhi | the nova keypair named "new" is generated by "nova keypair-add --pub-key ~/.ssh/id_rsa.pub new" | 04:33 |
*** kozhukalov_ has joined #openstack-ironic | 04:37 | |
*** hoangcx has quit IRC | 04:43 | |
openstackgerrit | vinay kumar muddu proposed openstack/proliantutils: Made changes to avoid creating objects every time a RIBCL/RIBS operation is requested https://review.openstack.org/255670 | 04:44 |
openstackgerrit | vinay kumar muddu proposed openstack/proliantutils: Made changes to avoid creating objects every time a RIBCL/RIBS operation is requested https://review.openstack.org/255671 | 04:44 |
*** hoangcx has joined #openstack-ironic | 04:45 | |
*** kkoski has joined #openstack-ironic | 05:00 | |
*** ijw has joined #openstack-ironic | 05:01 | |
*** ijw_ has joined #openstack-ironic | 05:03 | |
*** ijw has quit IRC | 05:03 | |
*** kkoski has quit IRC | 05:04 | |
*** kozhukalov_ has quit IRC | 05:28 | |
*** amotoki has joined #openstack-ironic | 05:30 | |
openstackgerrit | Merged openstack/bifrost: Correct DHCP configuration examples https://review.openstack.org/255524 | 05:42 |
openstackgerrit | Syed Ismail Faizan Barmawer proposed openstack/ironic: Refactor disk partitioner code from ironic and use ironic-lib. https://review.openstack.org/184443 | 05:42 |
*** ijw_ has quit IRC | 05:51 | |
*** yonglihe has quit IRC | 05:51 | |
*** saripurigopi has joined #openstack-ironic | 05:57 | |
*** greghayn1 has joined #openstack-ironic | 05:57 | |
saripurigopi | morning all | 05:57 |
*** greghaynes has quit IRC | 05:58 | |
*** greghayn1 is now known as greghaynes | 05:58 | |
*** greghaynes has quit IRC | 05:59 | |
*** greghaynes has joined #openstack-ironic | 06:00 | |
*** aginwala has joined #openstack-ironic | 06:00 | |
*** links has quit IRC | 06:01 | |
*** links has joined #openstack-ironic | 06:01 | |
*** Sukhdev has quit IRC | 06:04 | |
*** greghaynes has quit IRC | 06:05 | |
*** greghaynes has joined #openstack-ironic | 06:06 | |
*** ChubYann has quit IRC | 06:21 | |
*** HimanshuGarg has joined #openstack-ironic | 06:23 | |
openstackgerrit | Tan Lin proposed openstack/python-ironicclient: Add --uuid to port-create https://review.openstack.org/252762 | 06:27 |
*** chlong has quit IRC | 06:28 | |
openstackgerrit | Tan Lin proposed openstack/python-ironicclient: Add --uuid to chassis-create https://review.openstack.org/252751 | 06:33 |
*** links has quit IRC | 06:49 | |
*** chlong has joined #openstack-ironic | 06:49 | |
*** praneshp_ has joined #openstack-ironic | 06:54 | |
*** Fdaisuke_ has joined #openstack-ironic | 06:55 | |
*** Fdaisuke has quit IRC | 06:57 | |
*** praneshp has quit IRC | 06:58 | |
*** praneshp_ is now known as praneshp | 06:58 | |
openstackgerrit | Tan Lin proposed openstack/ironic: Follow-up Patch on 1ab3c9217edc8ebed7643f8ded7b16c39dd4bc70 https://review.openstack.org/252883 | 07:01 |
*** chlong has quit IRC | 07:04 | |
*** links has joined #openstack-ironic | 07:05 | |
*** aginwala has quit IRC | 07:05 | |
*** ukalifon has joined #openstack-ironic | 07:10 | |
*** rcernin has joined #openstack-ironic | 07:11 | |
*** ukalifon has joined #openstack-ironic | 07:19 | |
*** chlong has joined #openstack-ironic | 07:20 | |
*** dims_ has joined #openstack-ironic | 07:27 | |
*** kozhukalov_ has joined #openstack-ironic | 07:32 | |
openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: support multi-nodes operation for node-set-power-state command https://review.openstack.org/228301 | 07:34 |
*** aginwala has joined #openstack-ironic | 07:38 | |
*** Fdaisuke_ has quit IRC | 07:42 | |
*** kozhukalov_ has quit IRC | 07:48 | |
*** Fdaisuke_ has joined #openstack-ironic | 07:48 | |
*** wshao has joined #openstack-ironic | 07:52 | |
*** Haomeng|2 has quit IRC | 07:54 | |
*** Haomeng|2 has joined #openstack-ironic | 07:54 | |
*** wshao has quit IRC | 07:57 | |
*** mkovacik has joined #openstack-ironic | 07:58 | |
*** hoangcx has quit IRC | 08:00 | |
*** kkoski has joined #openstack-ironic | 08:01 | |
*** dansmith has quit IRC | 08:05 | |
*** chlong has quit IRC | 08:05 | |
*** sirushti has quit IRC | 08:05 | |
*** kkoski has quit IRC | 08:06 | |
*** thanhnt-z has quit IRC | 08:06 | |
*** praneshp has quit IRC | 08:06 | |
*** dansmith has joined #openstack-ironic | 08:07 | |
*** dansmith is now known as Guest17779 | 08:07 | |
*** praneshp has joined #openstack-ironic | 08:08 | |
*** sirushti has joined #openstack-ironic | 08:11 | |
*** aginwala has quit IRC | 08:13 | |
*** links has quit IRC | 08:14 | |
*** ionutbalutoiu has joined #openstack-ironic | 08:17 | |
*** athomas has quit IRC | 08:22 | |
*** mbound_ has joined #openstack-ironic | 08:23 | |
*** athomas has joined #openstack-ironic | 08:23 | |
openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector-specs: High Availability for Ironic Inspector https://review.openstack.org/253675 | 08:27 |
*** praneshp has quit IRC | 08:27 | |
*** smoriya_afk has quit IRC | 08:27 | |
*** MattMan has quit IRC | 08:31 | |
*** MattMan has joined #openstack-ironic | 08:31 | |
*** ionutbalutoiu has quit IRC | 08:32 | |
*** e0ne has joined #openstack-ironic | 08:32 | |
openstackgerrit | Tan Lin proposed openstack/ironic: Add a config option [upgrade_levels]/conductor https://review.openstack.org/253355 | 08:33 |
*** ionutbalutoiu has joined #openstack-ironic | 08:33 | |
openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add oslo_config.Opt support in Ironic config generator https://review.openstack.org/253498 | 08:34 |
*** ifarkas has joined #openstack-ironic | 08:39 | |
*** ifarkas has quit IRC | 08:39 | |
*** ifarkas has joined #openstack-ironic | 08:40 | |
openstackgerrit | Tan Lin proposed openstack/ironic: Remove extra space in 'host' config comment https://review.openstack.org/255733 | 08:41 |
*** Marga_ has quit IRC | 08:44 | |
*** smoriya_afk has joined #openstack-ironic | 08:47 | |
openstackgerrit | Naohiro Tamura proposed openstack/ironic: Update ironic.conf.sample for oslo.messaging (3.1.0) https://review.openstack.org/255738 | 08:49 |
*** e0ne has quit IRC | 08:52 | |
*** links has joined #openstack-ironic | 08:52 | |
*** Marga_ has joined #openstack-ironic | 09:03 | |
*** Marga_ has quit IRC | 09:03 | |
*** Marga_ has joined #openstack-ironic | 09:03 | |
*** Marga_ has quit IRC | 09:04 | |
*** Marga_ has joined #openstack-ironic | 09:05 | |
*** jamielennox is now known as jamielennox|away | 09:09 | |
*** yuanying has quit IRC | 09:10 | |
*** ndipanov has joined #openstack-ironic | 09:14 | |
*** Fdaisuke_ has quit IRC | 09:14 | |
openstackgerrit | Naohiro Tamura proposed openstack/ironic: Update ironic.conf.sample for oslo.messaging (3.1.0) https://review.openstack.org/255738 | 09:18 |
*** ndipanov has quit IRC | 09:20 | |
aarefiev | morning! | 09:21 |
*** openstackgerrit has quit IRC | 09:32 | |
*** openstackgerrit has joined #openstack-ironic | 09:33 | |
*** ndipanov has joined #openstack-ironic | 09:33 | |
*** links has quit IRC | 09:42 | |
*** _degorenko|afk is now known as degorenko | 09:43 | |
*** athomas has quit IRC | 09:51 | |
*** jcoufal has joined #openstack-ironic | 09:51 | |
*** e0ne has joined #openstack-ironic | 09:51 | |
*** naohirot has quit IRC | 09:54 | |
*** links has joined #openstack-ironic | 09:55 | |
*** athomas has joined #openstack-ironic | 09:56 | |
*** kozhukalov_ has joined #openstack-ironic | 09:57 | |
*** igordcard_ has quit IRC | 09:58 | |
*** igordcard_ has joined #openstack-ironic | 09:58 | |
*** jistr has joined #openstack-ironic | 09:59 | |
*** igordcard_ has quit IRC | 09:59 | |
*** links has quit IRC | 10:04 | |
*** zhenguo has quit IRC | 10:06 | |
*** mkovacik has quit IRC | 10:12 | |
*** saripurigopi has quit IRC | 10:16 | |
*** links has joined #openstack-ironic | 10:17 | |
*** electrofelix has joined #openstack-ironic | 10:19 | |
*** kozhukalov_ has quit IRC | 10:20 | |
*** sambetts-afk is now known as sambetts | 10:27 | |
*** dtantsur|afk is now known as dtantsur | 10:30 | |
dtantsur | Morning Ironic | 10:30 |
sambetts | o/ dtantsur | 10:30 |
dtantsur | o/ | 10:30 |
*** kozhukalov_ has joined #openstack-ironic | 10:33 | |
*** rkharya has joined #openstack-ironic | 10:33 | |
*** Fdaisuke has joined #openstack-ironic | 10:37 | |
*** alexpilotti has joined #openstack-ironic | 10:48 | |
*** igordcar1_ has joined #openstack-ironic | 10:50 | |
*** igordcar1_ has quit IRC | 10:50 | |
*** igordcar1_ has joined #openstack-ironic | 10:50 | |
*** igordcar1_ has quit IRC | 10:50 | |
openstackgerrit | Anton Arefiev proposed openstack/ironic-inspector-specs: Ironic node auto-discovery https://review.openstack.org/252946 | 10:51 |
*** chenke_ has joined #openstack-ironic | 10:52 | |
aarefiev | dtantsur, sambetts : morning, I've updated spec ^^, please take a look way you have free min | 10:52 |
aarefiev | s/way/when | 10:52 |
*** alexpilotti has quit IRC | 10:53 | |
dtantsur | sure, will try | 10:53 |
*** chenke__ has quit IRC | 10:55 | |
aarefiev | dtantsur: btw, what do you think about `Ironic python agent impact` in spec? | 10:56 |
dtantsur | aarefiev, good idea | 10:57 |
aarefiev | dtantsur: ok, I'll propose it :) | 10:58 |
dtantsur | thanks | 10:58 |
*** changzhi has quit IRC | 11:01 | |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Explicitly depend on WebTest>=2.0 https://review.openstack.org/255793 | 11:01 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Explicitly depend on WebTest>=2.0 https://review.openstack.org/255793 | 11:02 |
*** dims_ has quit IRC | 11:04 | |
*** alexpilotti has joined #openstack-ironic | 11:06 | |
*** kozhukalov_ has quit IRC | 11:07 | |
dtantsur | aarefiev, see on the ML, we've decided to switch from blueprints to bugs, please adjust your spec accordingly | 11:13 |
aarefiev | dtantsur: sure | 11:13 |
*** igordcard_ has joined #openstack-ironic | 11:20 | |
*** mkovacik has joined #openstack-ironic | 11:21 | |
*** kozhukalov_ has joined #openstack-ironic | 11:22 | |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Use launchpad bugs instead of blueprints https://review.openstack.org/255807 | 11:27 |
dtantsur | jroll, ^^ | 11:27 |
*** lucas-dinner is now known as lucasagomes | 11:28 | |
lucasagomes | morning all | 11:29 |
dtantsur | morning lucasagomes | 11:31 |
openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic-specs: Allow specifying bug ref instead of BP https://review.openstack.org/255811 | 11:32 |
openstackgerrit | Anton Arefiev proposed openstack/ironic-inspector-specs: Add Ironic python agent impact to template https://review.openstack.org/255813 | 11:35 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Driver composition reform spec https://review.openstack.org/188370 | 11:37 |
dtantsur | jroll, how it looks for a spec ^^ | 11:37 |
*** Marga_ has quit IRC | 11:38 | |
*** Marga_ has joined #openstack-ironic | 11:38 | |
*** sinval-afk is now known as sinval | 11:40 | |
sinval | good morning everyone | 11:40 |
dtantsur | morning sinval | 11:42 |
sinval | hey dtantsur, please take a look when you have time: https://review.openstack.org/#/c/255591/, it's the follow up patch about exception message assurance | 11:43 |
dtantsur | sure-sure, lemme finish some urgencies first :) | 11:44 |
sinval | dtantsur, no problem | 11:47 |
*** mgoddard has joined #openstack-ironic | 11:49 | |
openstackgerrit | Vasyl Saienko proposed openstack/ironic: Introduce libvirt dirver https://review.openstack.org/253096 | 11:49 |
jroll | dtantsur: nice, thanks | 11:54 |
sambetts | o/ HenryG | 11:55 |
sambetts | o/ jroll | 11:55 |
sambetts | sigh I can't type today... | 11:55 |
jroll | morning sambetts :P | 11:55 |
sinval | morning jroll sambetts | 11:56 |
sambetts | hey sinval | 11:56 |
sambetts | jroll: I don't know if I | 11:56 |
sambetts | 've hit a weird race condition on the context.auth_token | 11:57 |
sambetts | really randomly I get an error from neutronclient saying that auth_strategy = None is not valid, but the only way for that to happen seem to be if the auth token is set initally and then unset after calling out to neutronclient | 11:58 |
openstackgerrit | Sinval Vieira Mendes Neto proposed openstack/ironic: Improving capabilities_to_dict and adding helpful tests https://review.openstack.org/254972 | 12:00 |
*** nicodemos has joined #openstack-ironic | 12:00 | |
jroll | sambetts: ooo, I feel like I've seen this before | 12:01 |
jroll | is that in dhcp code? | 12:02 |
sambetts | I've had it happen in there, and also in my network provider I'm working on | 12:02 |
sambetts | I've now stoped passing the token into get_client to avoid this in my network provider | 12:03 |
*** kkoski has joined #openstack-ironic | 12:03 | |
openstackgerrit | Anton Arefiev proposed openstack/ironic-inspector-specs: Ironic node auto-discovery https://review.openstack.org/252946 | 12:03 |
jroll | sambetts: https://bugs.launchpad.net/ironic/+bug/1323783 | 12:03 |
openstack | Launchpad bug 1323783 in Ironic "Neutron client doesn't properly honor noauth" [Medium,Fix released] | 12:03 |
jroll | but that's long fixed | 12:04 |
*** mgoddard_ has joined #openstack-ironic | 12:05 | |
jroll | you sure the token is initially set? | 12:06 |
jroll | oh actually, I see the bug | 12:06 |
sambetts | yeah, because otherwise auth_token would equal "keystone" | 12:06 |
sambetts | oh? | 12:06 |
jroll | if auth_strategy is keystone and token is not None: https://github.com/openstack/ironic/blob/master/ironic/dhcp/neutron.py#L90 | 12:06 |
jroll | so the auth_strategy we're passing in there needs to change | 12:07 |
sambetts | in neutron client they have if not token: authenticate() and its running authenticate though | 12:07 |
*** kkoski has quit IRC | 12:08 | |
sambetts | thats were the error is coming from | 12:08 |
*** caiobo has joined #openstack-ironic | 12:08 | |
*** mgoddard has quit IRC | 12:08 | |
sambetts | http://git.openstack.org/cgit/openstack/python-neutronclient/tree/neutronclient/client.py#n121 | 12:09 |
sambetts | it shouldn't run authenticate if token exists | 12:09 |
jroll | O_o | 12:11 |
jroll | unless the token expires and it tries to re-auth for you maybe? | 12:11 |
sambetts | isn't token a string? | 12:11 |
jroll | sure | 12:12 |
jroll | so I'm thinking | 12:12 |
jroll | he tries to do a request with your token, and it fails, so it re-auths | 12:13 |
jroll | he/it | 12:13 |
jroll | but that can't be it because it doesn't have creds :| | 12:13 |
sambetts | yeah :/ | 12:13 |
jroll | although, it doesn't need creds to try | 12:13 |
jroll | https://github.com/openstack/python-neutronclient/blob/master/neutronclient/client.py#L173 | 12:13 |
jroll | BAM | 12:13 |
jroll | I bet that's it | 12:14 |
jroll | check out neutron logs, I bet you see a 404 right before it happens | 12:14 |
jroll | er, 401 | 12:14 |
*** kozhukalov_ has quit IRC | 12:14 | |
smoriya | morning ironic | 12:16 |
smoriya | dtantsur: If you have a change, would you take a look at https://review.openstack.org/#/c/200496/ ? | 12:16 |
smoriya | dtantsur: I answered your comments and updated the spec | 12:16 |
dtantsur | yeah, sure | 12:17 |
smoriya | dtantsur: thank you :) | 12:17 |
dtantsur | smoriya, btw, linking DeployInterface to a specific BMC is unacceptable IMO. so if you do that (as you said in the comment) - please don't :) | 12:17 |
smoriya | dtantsur: Hmm.... it seems that I couldn't explain well what I thought... | 12:19 |
openstackgerrit | John L. Villalovos proposed openstack/ironic: genconfig: Fix issue when hostname/FQDN is 'localhost' https://review.openstack.org/255834 | 12:20 |
sambetts | jroll: http://paste.openstack.org/show/481469/ spot on | 12:21 |
jroll | fun! | 12:22 |
*** raildo-afk is now known as raildo | 12:22 | |
sambetts | yup... so we should be providing the token, and all the auth creds etc at the same time | 12:22 |
sambetts | and let neutronclient pick which one ? | 12:23 |
jroll | I guess so? | 12:23 |
jroll | seems like that's the client's intent | 12:23 |
*** kozhukalov_ has joined #openstack-ironic | 12:27 | |
sambetts | jroll: https://bugs.launchpad.net/ironic/+bug/1524779 | 12:28 |
openstack | Launchpad bug 1524779 in Ironic "Neutron client will fail to re-authenticate if token expires because no credentials are provided" [Undecided,New] | 12:28 |
jroll | sambetts: cool, ty | 12:28 |
sambetts | jroll: I can put a patch up for it too, just ditching the else and else if, and just passing the token and the all the creds together, but I'm wondering if there is a reason behind having them separated? | 12:30 |
jroll | sambetts: I'm not aware of any reason, probably was just "only pass creds if we aren't auth'd yet" | 12:30 |
jroll | but I don't know everything so take that with a grain of salt | 12:31 |
jroll | :) | 12:31 |
sambetts | :) well I'm sure it'd get picked up in the patch comments if there is something like that | 12:31 |
*** rkharya has quit IRC | 12:32 | |
amotoki | sambetts: regarding bug 1524779, is there a way to renew a token without credential? | 12:32 |
openstack | bug 1524779 in Ironic "Neutron client will fail to re-authenticate if token expires because no credentials are provided" [Undecided,New] https://launchpad.net/bugs/1524779 - Assigned to Sam Betts (sambetts) | 12:32 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Enable ramdisk log collection in devstack plugin https://review.openstack.org/255842 | 12:33 |
sambetts | amotoki: in the neutron client it doesn't appear that it will be able to do it | 12:35 |
sambetts | https://github.com/openstack/python-neutronclient/blob/master/neutronclient/client.py#L173 | 12:35 |
*** thrash|g0ne is now known as thrash | 12:35 | |
sambetts | https://github.com/openstack/python-neutronclient/blob/master/neutronclient/client.py#L194 | 12:35 |
amotoki | sambetts: I know that. btw, is there a way to do so in general? trust? | 12:36 |
lucasagomes | jroll, ping re 3rd party CI, what you think about the options devananda put in the review? | 12:37 |
lucasagomes | jroll, since the spec is merged, the 2# seems the most viable IMO | 12:37 |
amotoki | sambetts: I think nova does re-authenticate when a token expires and neutronclient raises Unauthorized exception. | 12:37 |
*** kozhukalov_ has quit IRC | 12:37 | |
lucasagomes | if we agree in having another project under ironic umbrealla for the oot drivers | 12:37 |
jroll | lucasagomes: they're good options, I'm not terribly opinionated | 12:37 |
jroll | lucasagomes: *but* I don't want to move all drivers we drop to a contrib thing | 12:38 |
lucasagomes | cause that's better because at least all (or most) drivers will be there. And people don't have to go to many sites to find them all | 12:38 |
jroll | just the dev ones and such we care about as a community | 12:38 |
jroll | like, seamicro should be on its own imo | 12:38 |
lucasagomes | jroll, right | 12:38 |
lucasagomes | jroll, contrib thing you mean in ironic tree? | 12:38 |
lucasagomes | or that applies also for a separated repo? | 12:38 |
jroll | lucasagomes: contrib thing meaning #1 or #2 | 12:39 |
lucasagomes | right | 12:39 |
jroll | whichever we end up with | 12:39 |
lucasagomes | fair enuff | 12:39 |
pas-ha | sambetts, jroll https://review.openstack.org/#/c/236982/ re how we should do Keystone stuff | 12:39 |
jroll | I think I do prefer #2 but I haven't had much coffee | 12:39 |
lucasagomes | jroll, yeah no worries we have time | 12:40 |
jroll | pas-ha: awesome! | 12:40 |
*** sambetts has quit IRC | 12:40 | |
pas-ha | not sure if it solves the bug in question right away though. will take a look | 12:40 |
jroll | pas-ha: oh, no, this is authing to neutron client, I think it's unrelated | 12:40 |
jroll | lucasagomes: yeah, maybe we start working on the driver facade thing we talked about at the beginning of N, and get it to a usable point by middle of cycle, and start working on moving the drivers we care about | 12:41 |
pas-ha | well, all the client will be using new session stuff, so it might still be related | 12:41 |
jroll | ah | 12:41 |
*** sambetts has joined #openstack-ironic | 12:41 | |
jroll | might be worth a try | 12:41 |
lucasagomes | jroll, I have to read again what we've agreed there | 12:42 |
jroll | sambetts: 12:39:31 pas-ha | sambetts, jroll https://review.openstack.org/#/c/236982/ re how we should do Keystone stuff | 12:42 |
* lucasagomes short memory | 12:42 | |
jroll | lucasagomes: line 30-40 or so https://etherpad.openstack.org/p/summit-mitaka-ironic-driver-api | 12:42 |
lucasagomes | will do | 12:42 |
openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Add power and thermal limits support https://review.openstack.org/251957 | 12:45 |
lucasagomes | jroll, ++ | 12:45 |
*** mbound_ has quit IRC | 12:45 | |
*** baoli has joined #openstack-ironic | 12:45 | |
sambetts | jroll: Thanks for the link, I'm looking at https://review.openstack.org/#/c/236982/14/ironic/dhcp/neutron.py L97 | 12:47 |
*** baoli_ has joined #openstack-ironic | 12:47 | |
sambetts | and won't this suffer from the same issue? | 12:47 |
sambetts | pas-ha: ^ | 12:47 |
*** openstackgerrit has quit IRC | 12:47 | |
pas-ha | sambetts, left some comments in the bug | 12:47 |
jroll | sambetts: yeah, looks like it | 12:48 |
*** openstackgerrit has joined #openstack-ironic | 12:48 | |
jroll | probably need to pass the session in there | 12:48 |
pas-ha | this might be related to Keystone V2 vs v3 | 12:48 |
jroll | but really... | 12:48 |
jroll | we can just combine those two cases right? | 12:48 |
jroll | if it's keystone, pass in token + creds | 12:48 |
jroll | neutronclient will handle if token is None or not | 12:48 |
pas-ha | we have keystone_authtoken configured to use Keystone v2 explicitly, and reusing its options for clients | 12:49 |
pas-ha | so when Keystone v3 token comes in, i suspect there might be some problems | 12:50 |
*** baoli has quit IRC | 12:50 | |
dtantsur | lucasagomes, jroll, no, like seriously: every time we fix a crash in conductor we bump RPC API version? | 12:51 |
jroll | wat | 12:51 |
dtantsur | I suspect we landed plenty of bug fixes without doing that.. | 12:51 |
dtantsur | jroll, https://review.openstack.org/#/c/253365/3/ironic/conductor/rpcapi.py | 12:52 |
lucasagomes | dtantsur, if it affects the consumer of the API we should bump AFAIUI | 12:52 |
lucasagomes | changing the exception raised by the methods is one of those cases | 12:52 |
jroll | dtantsur: is NotImplemented an exception that exists? | 12:53 |
dtantsur | no | 12:53 |
lucasagomes | it will raise a TypeError | 12:53 |
dtantsur | raising NotImplemented gives TypeError | 12:53 |
*** kozhukalov_ has joined #openstack-ironic | 12:53 | |
jroll | heh | 12:53 |
lucasagomes | I'm not sure either if we should bump or not | 12:53 |
lucasagomes | but hmm | 12:53 |
jroll | we *should* bump it, by the rules | 12:53 |
jroll | however, I don't think anything expects that error | 12:54 |
dtantsur | it's like whether we bump API version when we fix internal server error (HTTP 500) | 12:54 |
dtantsur | I think we should be more clear on it | 12:54 |
jroll | well, it IS an rpc api change | 12:54 |
jroll | we should bump it | 12:55 |
jroll | I don't think not bumping it will break anything | 12:55 |
jroll | but I don't see any good reason not to bump it | 12:55 |
dtantsur | then we should do it when fixing a failure in any sync method in conductor, right? | 12:57 |
dtantsur | it's not immediately obvious, I think we should communicate it better | 12:57 |
jroll | it's not about the failure | 12:57 |
lucasagomes | it's about API | 12:57 |
jroll | the exception raised is part of the api | 12:57 |
jroll | whether intentional or not | 12:57 |
jroll | we do need to document it here, because it's not: http://docs.openstack.org/developer/ironic/dev/code-contribution-guide.html | 12:57 |
dtantsur | jroll, failure == unexpected exception | 12:57 |
*** crinkle has quit IRC | 12:57 | |
jroll | ah, true | 12:57 |
* lucasagomes can argue both sides... | 12:57 | |
dtantsur | so if we have a sync method in driver, and we fix some unexpected exception (TypeError, etc), we bump API version in conductor, right? | 12:57 |
lucasagomes | the change is safe, whether bumping or not | 12:57 |
lucasagomes | but yeah would be good to clarify | 12:57 |
* dtantsur also does not have a strong opinion, but wants to be clear on it | 12:57 | |
jroll | wait | 12:57 |
jroll | this is the client side, no? | 12:58 |
*** crinkle has joined #openstack-ironic | 12:58 | |
jroll | we don't bump for changes in the client side | 12:58 |
lucasagomes | jroll, that's a good point | 12:58 |
dtantsur | oh yeah, you're right in this case | 12:58 |
lucasagomes | maybe only for manager.py | 12:58 |
dtantsur | RPC is not involved at all here | 12:58 |
jroll | right | 12:58 |
jroll | carry on! | 12:58 |
lucasagomes | cool | 12:58 |
dtantsur | but still it's a good question to raise | 12:58 |
* lucasagomes retract his vote | 12:58 | |
lucasagomes | dtantsur, ++ | 12:59 |
*** smoriya has quit IRC | 12:59 | |
*** smoriya_afk is now known as smoriya | 12:59 | |
jroll | dtantsur: totally | 12:59 |
* dtantsur gets more coffee | 12:59 | |
*** mgoddard__ has joined #openstack-ironic | 13:01 | |
openstackgerrit | Andre Aranha proposed openstack/ironic: Just a test https://review.openstack.org/255859 | 13:01 |
*** mgoddard has joined #openstack-ironic | 13:02 | |
*** mgoddard_ has quit IRC | 13:03 | |
*** mgoddard__ has quit IRC | 13:05 | |
*** mgould has joined #openstack-ironic | 13:05 | |
lucasagomes | devananda, re fact table... it seems possible to re-use the same fact tables to more than one field | 13:07 |
lucasagomes | s/fact tables/fact table* | 13:07 |
sambetts | damn it who added pillow to the requirements for running ironic tests, now I have to go an install libjpeg libjpeg-dev | 13:09 |
*** dprince has joined #openstack-ironic | 13:21 | |
jroll | sambetts: our state machine generator did :/ | 13:21 |
jroll | it was actually fine until some version of pillow started requiring libjpeg instead of an optional dep | 13:22 |
*** trown|outttypeww is now known as trown | 13:24 | |
openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Intel Node Manager policies support https://review.openstack.org/214657 | 13:24 |
*** mkovacik_ has joined #openstack-ironic | 13:27 | |
*** mkovacik has quit IRC | 13:29 | |
*** kozhukalov_ has quit IRC | 13:37 | |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Explicitly depend on WebTest>=2.0 https://review.openstack.org/255793 | 13:43 |
*** dims has joined #openstack-ironic | 13:43 | |
*** derekh has joined #openstack-ironic | 13:45 | |
*** [1]cdearborn has joined #openstack-ironic | 13:47 | |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Copy devstack code to ironic tree https://review.openstack.org/255879 | 13:47 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add devstack plugin https://review.openstack.org/255880 | 13:47 |
*** kozhukalov_ has joined #openstack-ironic | 13:48 | |
*** dims has quit IRC | 13:48 | |
*** rloo has joined #openstack-ironic | 13:51 | |
*** kozhukalov_ has quit IRC | 13:54 | |
*** dims_ has joined #openstack-ironic | 13:57 | |
*** openstackstatus has quit IRC | 13:57 | |
*** mariojv has quit IRC | 13:57 | |
*** links has quit IRC | 13:57 | |
*** sc68cal has quit IRC | 13:58 | |
openstackgerrit | Sam Betts proposed openstack/ironic: Always pass keystone credentials to neutronclient https://review.openstack.org/255885 | 13:58 |
*** mariojv has joined #openstack-ironic | 14:00 | |
*** sc68cal has joined #openstack-ironic | 14:01 | |
*** sc68cal has quit IRC | 14:01 | |
*** sc68cal has joined #openstack-ironic | 14:01 | |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add devstack plugin https://review.openstack.org/255880 | 14:04 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Enable ramdisk log collection in devstack plugin https://review.openstack.org/255842 | 14:05 |
openstackgerrit | Julia Kreger proposed openstack/bifrost: Add VM creation support on Debian jessie https://review.openstack.org/253171 | 14:11 |
jroll | this devstack plugin actually works O_o | 14:11 |
lucasagomes | w00t | 14:12 |
openstackgerrit | Julia Kreger proposed openstack/bifrost: Adjust primary bifrost settings for Debian jessie support https://review.openstack.org/255331 | 14:12 |
*** [1]cdearborn has quit IRC | 14:12 | |
openstackgerrit | Julia Kreger proposed openstack/bifrost: Support to wait for node deployment to complete https://review.openstack.org/254245 | 14:13 |
jroll | yeah. that totally just worked. | 14:15 |
* jroll is shocked | 14:15 | |
* jroll finishes the patch series | 14:15 | |
dtantsur | jroll, it's an awesome thing, we've been using it for a while for inspector | 14:17 |
dtantsur | the best feature is gating on devstack changes | 14:17 |
rloo | hi dtantsur, lucasagomes, jroll, TheJulia and everyone else this fine Thursday morning | 14:19 |
lucasagomes | rloo, hi there! | 14:19 |
dtantsur | morning rloo | 14:19 |
mgould | rloo, morning! | 14:19 |
rloo | if anyone has some time (not a quick one), I'm finally happy with the ironic-lib refactoring: https://review.openstack.org/#/c/184443/ | 14:19 |
dtantsur | \o/ | 14:19 |
jroll | rloo: good morning, woot! | 14:19 |
jroll | dtantsur: wait, gating on devstack changes? | 14:19 |
dtantsur | jroll, well, if you have devstack plugin in tree, then any change to it will have to pass all project gates, right? | 14:20 |
jroll | dtantsur: oh yeah, that, I thought you meant devstack core changes :) | 14:20 |
jroll | but yeah ++ | 14:20 |
dtantsur | no, sorry for confusion :) | 14:20 |
dtantsur | jroll, wdyt about pas-ha's comment on https://review.openstack.org/#/c/255807/1/tests/test_titles.py ? | 14:21 |
jroll | dtantsur: it's a good question | 14:22 |
dtantsur | yeah | 14:22 |
jroll | but I can't imagine a spec for either of those projects, really | 14:22 |
jroll | or at least one that doesn't require major ironic changes as well | 14:22 |
dtantsur | I think all specs more or less target ironic itself.. e.g. if we change something in ironic-lib, we probably expect something to change in ironic | 14:22 |
jroll | yep | 14:22 |
jroll | I think it's fine as is | 14:22 |
TheJulia | good morning everyone | 14:24 |
mgould | TheJulia, morning! | 14:24 |
jroll | morning TheJulia, mgould :) | 14:25 |
dtantsur | morning TheJulia | 14:25 |
mgould | jroll, morning :-) | 14:27 |
rloo | morning mgould | 14:33 |
*** jcoufal has quit IRC | 14:34 | |
*** jcoufal has joined #openstack-ironic | 14:34 | |
lucasagomes | morning all :-) | 14:35 |
*** lucasagomes is now known as lucas-hungry | 14:35 | |
* lucas-hungry brb go grab some lunch | 14:35 | |
*** Guest17779 is now known as dansmith | 14:36 | |
*** dansmith is now known as Guest18124 | 14:37 | |
*** Guest18124 is now known as dansmith | 14:37 | |
NobodyCam | good morning Iornicers | 14:38 |
jroll | hey, morning NobodyCam :) | 14:38 |
*** cdearborn has joined #openstack-ironic | 14:38 | |
dtantsur | morning NobodyCam | 14:39 |
NobodyCam | morning dtantsur lucas-hungry mgould rloo sambetts jroll jlvillal sinval devananda and all others not listed here | 14:39 |
NobodyCam | :) | 14:39 |
mgould | morning NobodyCam dtantsur lucas-hungry sambetts jlvillal sinval devananda and everyone else I've missed :-) | 14:40 |
dtantsur | o/ | 14:42 |
rloo | hi, morning NobodyCam | 14:43 |
*** garthb has joined #openstack-ironic | 14:46 | |
*** mgoddard has quit IRC | 14:46 | |
NobodyCam | :) | 14:46 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add devstack plugin https://review.openstack.org/255880 | 14:46 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Remove override-defaults https://review.openstack.org/255914 | 14:46 |
*** garthb_ has joined #openstack-ironic | 14:47 | |
*** mtanino has joined #openstack-ironic | 14:48 | |
*** garthb has quit IRC | 14:50 | |
sinval | morning NobodyCam and mgould | 14:54 |
*** mgoddard has joined #openstack-ironic | 14:55 | |
*** sinval is now known as sinval-afk | 14:56 | |
jroll | dtantsur: oh, I have the enable_service things in local.conf heh | 14:57 |
jroll | I think devstack-gate currently does that for us though | 14:58 |
dtantsur | jroll, maybe it's right, dunno... | 14:58 |
*** boris-42_ has joined #openstack-ironic | 14:59 | |
*** amotoki has quit IRC | 15:02 | |
*** kozhukalov_ has joined #openstack-ironic | 15:04 | |
*** wuhg has joined #openstack-ironic | 15:04 | |
jroll | vsaienko: thanks for the idea on is_ironic_hardware, I was wondering what to do about that :) | 15:04 |
*** kozhukalov_ has quit IRC | 15:08 | |
*** yonglihe has joined #openstack-ironic | 15:08 | |
aarefiev | sambetts: have a min? | 15:10 |
*** mgould has quit IRC | 15:10 | |
dtantsur | which user is the default on the coreos ramdisk? aka how to I ssh to it? | 15:20 |
*** UForgotten has quit IRC | 15:21 | |
*** UForgotten has joined #openstack-ironic | 15:21 | |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add devstack plugin https://review.openstack.org/255880 | 15:23 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Copy devstack code to ironic tree https://review.openstack.org/255879 | 15:23 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Clean up variables in plugin.sh https://review.openstack.org/255936 | 15:23 |
NobodyCam | :: | 15:23 |
NobodyCam | gah | 15:23 |
NobodyCam | :) | 15:23 |
jroll | dtantsur: I promise I'm done now | 15:23 |
*** HimanshuGarg has quit IRC | 15:23 | |
jroll | dtantsur: it's 'core' | 15:23 |
* NobodyCam need lots of coffee this morning | 15:23 | |
*** trown is now known as trown|brb | 15:23 | |
openstackgerrit | Ruby Loo proposed openstack/ironic: Add FSM.is_stable() method https://review.openstack.org/255640 | 15:23 |
dtantsur | jroll, thanks, I would not guess :) | 15:24 |
jroll | yeah, it's kinda annoying | 15:24 |
dtantsur | wow, the coreos ramdisk builder just stripped away 2 setuptools entry points | 15:24 |
dtantsur | Oo | 15:24 |
dtantsur | oO | 15:24 |
dtantsur | O___O | 15:25 |
jroll | x.x | 15:25 |
jroll | wat, how | 15:25 |
dtantsur | is it really not Monday today? | 15:25 |
dtantsur | no idea.. | 15:25 |
jroll | must be friday :P | 15:25 |
*** mgould has joined #openstack-ironic | 15:25 | |
NobodyCam | I wish it was friday | 15:25 |
dtantsur | hmmm, how came I got some ancient IPA there... | 15:26 |
dtantsur | it's like liberty version or even older | 15:26 |
dtantsur | what. the. hell. is. going. on. | 15:26 |
*** kkoski has joined #openstack-ironic | 15:27 | |
*** trown|brb is now known as trown | 15:27 | |
dtantsur | 0.1.0.dev763 is in the tar.gz, but the output says it installed 1.1.0.dev64. WUT?? | 15:30 |
dtantsur | JayF, hi! could you build IPA with https://review.openstack.org/#/c/254896/ and post somewhere for me to grab? something mad is going on for me and I want to check it.. | 15:33 |
* dtantsur rebuilds everything | 15:37 | |
*** cdearborn has quit IRC | 15:37 | |
*** HimanshuGarg has joined #openstack-ironic | 15:37 | |
sambetts | aarefiev: hi | 15:38 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add devstack plugin https://review.openstack.org/255880 | 15:38 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Copy devstack code to ironic tree https://review.openstack.org/255879 | 15:38 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Clean up variables in plugin.sh https://review.openstack.org/255936 | 15:38 |
jroll | ok NOW I am done | 15:38 |
dtantsur | lol | 15:38 |
jroll | so | 15:38 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Enable ramdisk log collection in devstack plugin https://review.openstack.org/255842 | 15:38 |
jroll | since plugin gets turned on with project-config | 15:38 |
aarefiev | sambetts: hi, thanks for review, I'll update spec due your suggestions, about new action | 15:38 |
jroll | and that has to be applied to jenkins before it can be used | 15:39 |
aarefiev | The idea was to fetch node info in case of different credentials, etc, instead of having multiple rules. | 15:39 |
jroll | there's no way to ensure this works before landing that change | 15:39 |
jroll | so, gate may break when that happens | 15:39 |
dtantsur | jroll, we can start with one non-voting job, check, then switch all jobs. if infra agrees to fast-track our patches | 15:39 |
jroll | we could, it could still break things :D | 15:40 |
jroll | and the way we template them, it may be painful | 15:40 |
openstackgerrit | Julia Kreger proposed openstack/bifrost: Support to wait for node deployment to complete https://review.openstack.org/254245 | 15:41 |
dtantsur | jroll, yeah, but chances of breakage are much lower then.. | 15:41 |
jroll | right | 15:42 |
*** kkoski has quit IRC | 15:44 | |
*** cdearborn has joined #openstack-ironic | 15:45 | |
*** yarkot has joined #openstack-ironic | 15:45 | |
jroll | dtantsur: other option is run devstack-gate locally and trust | 15:46 |
jroll | run the exact script from project config that is | 15:46 |
jroll | stepping away for a while and then I'll hack on that | 15:46 |
dtantsur | jroll, I would do both actually :) | 15:46 |
jroll | dtantsur: that template makes it hard-ish, but yeah | 15:46 |
jroll | will do | 15:47 |
dtantsur | maybe, dunno.. we can hard code "if ipxe, then use plugin" | 15:47 |
dtantsur | (ipxe job is not voting yet, right?) | 15:47 |
*** lucas-hungry is now known as lucasagomes | 15:51 | |
lucasagomes | dtantsur, I think it's now, it was approved yesterday | 15:51 |
lucasagomes | the patch to make it voting | 15:51 |
* lucasagomes checks | 15:51 | |
dtantsur | well, damn.. | 15:51 |
dtantsur | nononononoNONONONONOOOOOOO...... I've rebuilt everything from scratch and I still get ancient IPA embedded >_< | 15:52 |
dtantsur | what am I doing so terribly wrong? | 15:52 |
lucasagomes | dtantsur, yeah ipxe is not voting | 15:55 |
lucasagomes | dtantsur, rebuilt the IPA image? coreos? dib? | 15:55 |
dtantsur | lucasagomes, coreos.. something terrible stale is somewhere on my system | 15:55 |
sambetts | aarefiev: if you've already got a mapping of addresses to credentials, why don't you just enroll the ironic nodes using a bash script and then trigger bulk inspection on them, the way you've defined it you'll have to manually power on a node you already have the remote power credentially for? | 15:55 |
* dtantsur kills everything in docker with fire | 15:56 | |
* sambetts *coughs* tiny *cough* ipa | 15:56 | |
sambetts | ;) | 15:56 |
*** genago has joined #openstack-ironic | 16:00 | |
zer0c00l | ping rloo | 16:01 |
rloo | zer0c00l: ? | 16:01 |
dtantsur | sambetts++ | 16:01 |
zer0c00l | rloo: was about to ask whether you are joining | 16:01 |
rloo | zer0c00l: :) | 16:02 |
openstackgerrit | Sinval Vieira Mendes Neto proposed openstack/ironic: Improving capabilities_to_dict and adding helpful tests https://review.openstack.org/254972 | 16:02 |
NobodyCam | sambetts: :) | 16:02 |
sambetts | aarefiev: perhaps making the value field of set-attribute more flexible, so it can retreive data from different places that would leave it open for both possiblity | 16:03 |
sambetts | s | 16:03 |
aarefiev | sambetts: in your case we don't need discovery at all:) , yeah I can extend set-attributte | 16:05 |
openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector-specs: High Availability for Ironic Inspector https://review.openstack.org/253675 | 16:06 |
*** dtantsur is now known as dtantsur|brb | 16:08 | |
*** Haomeng|2 has quit IRC | 16:14 | |
dtantsur|brb | jroll, JayF, only 'make docker_clean' actually forced the coreos builder to pick new IPA code.. I'd say it's a bug | 16:14 |
*** devananda has quit IRC | 16:15 | |
aarefiev | could somebody take a look, please: https://review.openstack.org/#/c/229286/ | 16:17 |
*** dprince has quit IRC | 16:19 | |
*** e0ne has quit IRC | 16:20 | |
*** devananda has joined #openstack-ironic | 16:21 | |
*** dprince has joined #openstack-ironic | 16:21 | |
jroll | dtantsur|brb: sounds like docker layer caching... this is upstream docker file? | 16:25 |
*** Marga_ has quit IRC | 16:27 | |
*** genago has quit IRC | 16:32 | |
*** kkoski has joined #openstack-ironic | 16:34 | |
*** dprince has quit IRC | 16:41 | |
*** rcernin has quit IRC | 16:42 | |
*** pas-ha has quit IRC | 16:44 | |
openstackgerrit | Anton Arefiev proposed openstack/ironic-inspector-specs: Ironic node auto-discovery https://review.openstack.org/252946 | 16:45 |
JayF | dtantsur|brb: there's lots and lots of caching built into the build process. IDK if that's a bug or not | 16:46 |
JayF | I certainly don't feel like it's a bug when I know which pieces of cache to invalidate and it makes my testing go faster :) | 16:47 |
jroll | dtantsur|brb: also, https://review.openstack.org/#/c/255981/ | 16:50 |
lucasagomes | cache invalidation is hard | 16:50 |
lucasagomes | JayF, btw I saw ur question this morning re the order of clean_steps | 16:50 |
lucasagomes | I think that's fine because in python dicts are not ordered anyway | 16:51 |
lucasagomes | that's why we have the "priority" field right? | 16:51 |
JayF | lucasagomes: well basically, we have multiple extensive downstream hardware managers | 16:51 |
JayF | and I'm trying to write a giant unit test that loads all the managers, mocks out evaluate_hw_support, and tests the sets of steps are correct | 16:52 |
JayF | so lack of ordering in the steps was a bit ... frustrating in that context :) | 16:52 |
*** yarkot has quit IRC | 16:52 | |
jroll | https://rackhd.readthedocs.org/en/latest/ | 16:53 |
jroll | ^ another bare metal provisioning system | 16:53 |
*** cemason has joined #openstack-ironic | 16:53 | |
*** david-lyle has quit IRC | 16:53 | |
NobodyCam | https://github.com/RackHD/RackHD | 16:55 |
NobodyCam | saw that and thought it was reackerlab at first... but I've only have one cup of coffee | 16:57 |
jroll | right? | 16:58 |
*** dtantsur|brb is now known as dtantsur | 17:01 | |
*** dprince has joined #openstack-ironic | 17:02 | |
lucasagomes | NobodyCam, nice | 17:04 |
sambetts | jroll: https://review.openstack.org/#/c/255885 | 17:04 |
lucasagomes | tho it's node.js x.x | 17:04 |
sambetts | oh god JS | 17:04 |
* sambetts runs and hides | 17:04 | |
lucasagomes | heh | 17:05 |
sambetts | I did my final project at Uni in node.js ... never again ... *shudders* | 17:05 |
*** dims_ has quit IRC | 17:06 | |
openstackgerrit | Merged openstack/ironic: Correct NotImplemented to NotImplementedError in rpcapi.py https://review.openstack.org/253365 | 17:09 |
lucasagomes | JayF, oh yeah for testing it sucks... you may want to use assertEqual(sorted(<expected_steps>), sorted(<steps>)) when testing it | 17:10 |
*** ig0r_ has joined #openstack-ironic | 17:11 | |
*** dims has joined #openstack-ironic | 17:12 | |
JayF | lucasagomes: well the data model for that whole command result from the agent is {' | 17:13 |
JayF | {'clean_steps': {'some_hw_mgr':[list of dicts containing steps]}} | 17:14 |
JayF | so gotta sort it deep inside | 17:14 |
lucasagomes | JayF, oh right, yeah that can be a bit more tricky | 17:16 |
*** dims has quit IRC | 17:16 | |
JayF | yeah, I think I have it figured out | 17:16 |
lucasagomes | like using sorted(... key=lambda) | 17:16 |
JayF | for manager, steps in expected['clean_steps'].items(): | 17:16 |
JayF | steps.sort(key=lambda x: x['priority']) | 17:16 |
JayF | doing that on both the expected and the results has it working now | 17:16 |
* JayF credits jroll for the PM python tutoring | 17:16 | |
JayF | I was falling back into my bad habit of writing half python half ruby which is zero working | 17:17 |
JayF | lol | 17:17 |
*** dims has joined #openstack-ironic | 17:17 | |
lucasagomes | :-) | 17:17 |
lucasagomes | w00t! | 17:17 |
*** Sukhdev has joined #openstack-ironic | 17:17 | |
dtantsur | JayF, jroll, I thought it's a bug because of this comment: https://github.com/openstack/ironic-python-agent/blob/master/Dockerfile#L3-L5 | 17:20 |
dtantsur | maybe I misunderstood it though | 17:20 |
dtantsur | JayF, +2 on the chroot patch | 17:21 |
JayF | dtantsur: hmm. | 17:22 |
JayF | dtantsur: that builder is a bit different upstream than down, I'll try to take a look today if I have time | 17:22 |
*** dprince has quit IRC | 17:22 | |
dtantsur | thanks | 17:22 |
JayF | dtantsur: the chroot change is one of a few patches needed to upstream all our changes to it. Am going to try to get the rest up | 17:22 |
*** vishwanathj has quit IRC | 17:23 | |
dtantsur | cool | 17:23 |
*** vishwanathj has joined #openstack-ironic | 17:23 | |
*** Sukhdev has quit IRC | 17:23 | |
*** [1]cdearborn has joined #openstack-ironic | 17:23 | |
dtantsur | I expect the chroot patch to fix quite a few things | 17:24 |
*** dims has quit IRC | 17:24 | |
jroll | dtantsur: yeah, so I wonder if docker now caches ADD commands | 17:25 |
dtantsur | jroll, https://github.com/docker/docker/issues/880 ? | 17:26 |
jroll | dtantsur: yep, good eye https://github.com/docker/docker/pull/2809 | 17:28 |
*** wuhg has quit IRC | 17:28 | |
jroll | that's fun | 17:29 |
*** Walts has joined #openstack-ironic | 17:30 | |
Walts | HI all, just wondering if someone could help me with a quick question about Ironic | 17:30 |
TheJulia | Walts: we gladly answer questions :) | 17:31 |
Walts | I have it installed and its provisioning servers nicely | 17:31 |
Walts | I was just wondering about the tftp directory | 17:31 |
Walts | do I need to back it up | 17:31 |
Walts | if I lost the data in there would ironic recreate it | 17:32 |
Walts | it creates folders for each instance | 17:32 |
Walts | with booting configuration | 17:32 |
Walts | its this stuff I am worried about | 17:32 |
jroll | if ironic fails over to a different ironic-conductor host, it will re-create it, but I don't believe it will recreate it if it just disappears | 17:32 |
jroll | I could be wrong though :) | 17:32 |
TheJulia | Ideally you should back it up, but it will get re-created as-needed | 17:32 |
TheJulia | what jroll said | 17:32 |
TheJulia | the odd case is if your netbooting, you have no other conductor, and you rebooted your machine | 17:33 |
TheJulia | hence why its handy to have the tftpboot folder contents | 17:33 |
Walts | jroll: TheJulia: thanks thats good to know | 17:33 |
*** mkovacik_ has quit IRC | 17:33 | |
devananda | I just realized that the state machine on_[enter,exit] logging hooks are useless because they do not indicate which Node is changing state | 17:34 |
TheJulia | doh | 17:34 |
jroll | nice :/ | 17:34 |
*** Marga_ has joined #openstack-ironic | 17:35 | |
lucasagomes | Walts, yeah back it up... A possible workaround you can do in case you lose it | 17:35 |
lucasagomes | is to bring down the ironic-conductor you're currently running | 17:35 |
lucasagomes | and start another one with a diff hostname | 17:35 |
devananda | jroll: we should probably test that the conductor correctly rebuilds that directory | 17:36 |
lucasagomes | that will take over the nodes and it will reconstruct the files (tho it's a workaround) | 17:36 |
devananda | lucasagomes: what is it a workaround for? | 17:36 |
lucasagomes | devananda, in case someone lose the PXE configuration files from the TFTP dir | 17:36 |
*** openstackstatus has joined #openstack-ironic | 17:37 | |
*** ChanServ sets mode: +v openstackstatus | 17:37 | |
*** athomas has quit IRC | 17:37 | |
* sambetts is calling it a day | 17:37 | |
lucasagomes | devananda, take_over() will reconstruct that | 17:37 |
NobodyCam | night sambetts | 17:37 |
*** Nisha has joined #openstack-ironic | 17:37 | |
TheJulia | goodnight sambetts | 17:37 |
NobodyCam | morning devananda | 17:37 |
jroll | oh, this is going to be fun to unwind: 2015-12-10 16:53:39.386 | /opt/stack/new/grenade/upgrade-tempest: line 54: /opt/stack/new/devstack/lib/ironic: No such file or directory | 17:37 |
TheJulia | good morning Nisha | 17:37 |
sambetts | night all o/ | 17:37 |
*** ukalifon has quit IRC | 17:37 | |
jroll | grenade depends on ironic in devstack's tree | 17:37 |
jroll | wheeeeeeeeeee | 17:37 |
Nisha | good morning TheJulia :) | 17:37 |
lucasagomes | jroll, x.x | 17:37 |
devananda | lucasagomes: oh, gotcha. if it's the same hostname, it wont trigger take_over() | 17:37 |
devananda | jroll: ugh | 17:37 |
lucasagomes | devananda, exactly! | 17:38 |
jroll | I didn't want to move grenade yet :| | 17:38 |
devananda | jroll: depends-on | 17:38 |
*** cdearborn has quit IRC | 17:38 | |
devananda | but yea, now there's a 3-way (or is it 4-way) dependency? | 17:38 |
jroll | devananda: I've already got like 4 depends-on in this tree, I'm well aware | 17:38 |
jroll | :P | 17:38 |
lucasagomes | lol | 17:38 |
devananda | haha | 17:38 |
jroll | context https://review.openstack.org/#/q/status:open+branch:master+topic:ironic-devstack-plugin,n,z | 17:38 |
Walts | lucasagomes: thanks for the tip | 17:38 |
lucasagomes | devananda, btw, not sure you saw the last ping... but having the same fact table for multiple join tables works | 17:39 |
devananda | lucasagomes: awesome | 17:39 |
lucasagomes | devananda, hard is to find a name for it now | 17:39 |
lucasagomes | devananda, node_keys ? | 17:39 |
lucasagomes | node_facts? | 17:39 |
devananda | lucasagomes: names are hard. how about 'labels' or 'property_names' | 17:40 |
lucasagomes | devananda, labels seems good | 17:41 |
devananda | it's a table of names of things that are referenced in other tables | 17:41 |
lucasagomes | property_names not much cause we may confuse that it's only for properties (but it will be used for other stuff too, driver_info and future maybe capabilities) | 17:41 |
*** dprince has joined #openstack-ironic | 17:41 | |
lucasagomes | devananda, yeah | 17:41 |
devananda | it could be used by tags, too | 17:42 |
lucasagomes | oh yeah tags too! | 17:42 |
lucasagomes | labels seems good | 17:42 |
lucasagomes | it's generic | 17:42 |
devananda | yea | 17:42 |
Walts | actually I had not considered running more than one conductor node, can I run them HA or does it have to be a failover? | 17:42 |
devananda | Walts: active-active | 17:42 |
jroll | Walts: you can run as many as you like at the same time, ironic will shard node resources between them | 17:42 |
lucasagomes | Walts, it has HA embeeded in the ironic-conductor so you don't actually need to care about it | 17:43 |
devananda | Walts: all Nodes will get distributed between the active conductors, and as conductors join/leave the cluster, they rebalance | 17:43 |
devananda | :) | 17:43 |
lucasagomes | yeah active-active as deva said | 17:43 |
Walts | nice! | 17:43 |
*** cemason has quit IRC | 17:44 | |
openstackgerrit | Merged openstack/bifrost: Add VM creation support on Debian jessie https://review.openstack.org/253171 | 17:44 |
openstackgerrit | Merged openstack/bifrost: Adjust primary bifrost settings for Debian jessie support https://review.openstack.org/255331 | 17:44 |
*** dims has joined #openstack-ironic | 17:47 | |
*** garthb_ has quit IRC | 17:47 | |
* TheJulia sits back and watches the merges | 17:47 | |
*** trown is now known as trown|lunch | 17:49 | |
*** degorenko is now known as _degorenko|afk | 17:49 | |
*** Sukhdev has joined #openstack-ironic | 17:49 | |
Walts | the tftpboot server address gets passed to the neutron port of each bare-metal server, how is this handled when there are multiple conductors each with a tftp server. Or is this the wrond way to do it | 17:50 |
Walts | *wrong | 17:50 |
devananda | Walts: each conductor knows its IP, and passes that to neutron when deploying to a node | 17:50 |
jroll | (or when taking over a node from a conductor that goes down) | 17:51 |
Walts | ah i see so if that node fails and a bare-metal server has to re-boot is this not a issue as the conductor node that takes over will update neutron | 17:52 |
devananda | Walts: each Node is associated to a specific conductor, though that association can change over time (and when it does, it updates Neutron and /tftp/, etc) | 17:52 |
Walts | devananda: jroll: thanks you so much this has been very helpful | 17:53 |
jroll | sure thing :) | 17:53 |
*** dims has quit IRC | 17:53 | |
devananda | Walts: np! | 17:54 |
*** praneshp has joined #openstack-ironic | 17:54 | |
openstackgerrit | Merged openstack/bifrost: Follow up to 251527 https://review.openstack.org/252695 | 17:56 |
openstackgerrit | Merged openstack/bifrost: Added jessie under Debian versions https://review.openstack.org/253007 | 17:56 |
openstackgerrit | Merged openstack/bifrost: Added jessie under Debian versions https://review.openstack.org/253021 | 17:56 |
openstackgerrit | Merged openstack/bifrost: Added jessie under Debian versions https://review.openstack.org/253026 | 17:56 |
lucasagomes | devananda, you know if alembic allow me to use the models to migrate the data? | 17:56 |
NobodyCam | TheJulia: bam ^^^ | 17:57 |
devananda | lucasagomes: I think so? | 17:57 |
lucasagomes | devananda, somehow it would map already the stuff to the right tables? Or will have to do the logic in python? | 17:57 |
jroll | lucasagomes: we don't want to do the migration in alembic, right? | 17:57 |
devananda | lucasagomes: we did that with sqla-migrate back in the day | 17:57 |
jroll | because that will force it at upgrade time, rather than "when the deployer wants to" | 17:57 |
devananda | lucasagomes: no, it'll do it in python | 17:57 |
*** dims has joined #openstack-ironic | 17:57 | |
lucasagomes | jroll, idk... | 17:58 |
*** athomas has joined #openstack-ironic | 17:58 | |
* lucasagomes is not familiar with those bits | 17:58 | |
devananda | if you use alembic to transform the data, then it does it in python IIRC | 17:58 |
jroll | yeah, alembic is all python | 17:58 |
jroll | but iirc we didn't want to force this huge migration at upgrade time | 17:58 |
devananda | if we do ^ then we can do it on the fly, when nodes are accessed | 17:58 |
devananda | otoh, if we did an UPDATE, we would have to do it at migration time | 17:58 |
lucasagomes | jroll, devananda right... you have any example of how we want it to work then | 17:59 |
lucasagomes | I was thinking migration time... but seems we don't wanna do that | 17:59 |
devananda | lucasagomes: I think nova hasn't even done this before, but it's a good way to do no-downtime data migrations | 17:59 |
jroll | lucasagomes: on save() / load(), transform as needed in the objects and/or model layer | 17:59 |
jroll | nova has done this, with their big flavor migration | 17:59 |
*** ifarkas has quit IRC | 18:00 | |
devananda | I put details in the scheduler spec draft that I posted a while back | 18:00 |
lucasagomes | jroll, but I will have to drop the column from the properties table no? | 18:00 |
lucasagomes | the properties column from the node table | 18:00 |
devananda | jroll: you removed the db specifics fromthat spec, and I was supposed to recreate it, right? | 18:00 |
devananda | lucasagomes: don't drop the column -- use a placeholder value to indicate that a record has been converted | 18:00 |
jroll | lucasagomes: then we provide a script that forces everything to migrate *now*, that ops need to run before the next cycle | 18:00 |
jroll | and drop columns the next cycle | 18:00 |
devananda | jroll: oh, right. with flavors. | 18:01 |
jroll | devananda: I did, lucas has that spec going | 18:01 |
devananda | ugh. too many specs ... | 18:01 |
*** maratoid has quit IRC | 18:02 | |
jroll | lucasagomes: so you set node.properties = None (or even {}) when you migrate a node (at save() time). at load() time, if node.properties == None, load them from the new table. otherwise load from old table. | 18:02 |
jroll | does that make sense? | 18:02 |
lucasagomes | jroll, trying to grasp | 18:04 |
lucasagomes | and what about downgrade ? | 18:04 |
jroll | heh, so | 18:05 |
jroll | just don't | 18:05 |
lucasagomes | so when we save() we then migrate to the new table and set the properties to {} | 18:05 |
jroll | funny enough, I saw this recently http://specs.openstack.org/openstack/openstack-specs/specs/no-downward-sql-migration.html | 18:05 |
lucasagomes | on load we check it to know where we should load the data from | 18:05 |
jroll | which we haven't done the work for but we can not support downgrade here | 18:05 |
jroll | yep, that's exactly right | 18:06 |
lucasagomes | ok | 18:06 |
lucasagomes | then the filter api will only come next cycle? | 18:07 |
lucasagomes | because until there we have no guarantees that all the data is now indexable | 18:07 |
lucasagomes | well maybe it can be documented... | 18:07 |
jroll | so my plan is to have an experimental one this cycle | 18:07 |
*** cemason has joined #openstack-ironic | 18:07 | |
jroll | tell people it's slow until they run that migration | 18:07 |
jroll | or something like that | 18:07 |
jroll | that's a good point | 18:07 |
lucasagomes | right | 18:08 |
lucasagomes | or to run that script to migrate all data if they want to use that api | 18:08 |
dtantsur | going now, see you tomorrow! | 18:08 |
lucasagomes | dtantsur, g'night | 18:08 |
jroll | yeah, I'm struggling to figure out how we decide if the migration is done or not | 18:08 |
jroll | night dtantsur | 18:08 |
*** dtantsur is now known as dtantsur|afk | 18:08 | |
jroll | but maybe we just only filter on the new table, and nodes not in it end up excluded | 18:08 |
lucasagomes | right | 18:10 |
lucasagomes | yeah well this will come later right? | 18:10 |
lucasagomes | what we can do is to check if the node id is in the node_properties table | 18:11 |
lucasagomes | that means it migrated | 18:11 |
lucasagomes | and properties is {}/None | 18:11 |
*** dims has quit IRC | 18:12 | |
devananda | jroll: so if we tie it into object load(), we'll have a stampeding herd in the first 60 seconds after conductors are upgraded | 18:13 |
devananda | periodic power state check | 18:13 |
lucasagomes | x.x | 18:13 |
lucasagomes | true | 18:13 |
devananda | every node not in maintenance mode will get migrated right away | 18:13 |
NobodyCam | night dtantsur|afk | 18:14 |
devananda | dtantsur|afk: g'night! | 18:14 |
lucasagomes | devananda, even a heartbeat can migrate the node | 18:14 |
jroll | devananda: no, don't migrate data in load() | 18:14 |
jroll | load() just loads differently depending on if data is migrated | 18:14 |
devananda | migrate on save()? | 18:14 |
lucasagomes | jroll, but save's same no ? We update the node quite often | 18:14 |
*** amit213 has quit IRC | 18:14 | |
lucasagomes | well not that often, but maintenance, the power state sync etc | 18:15 |
*** amit213 has joined #openstack-ironic | 18:15 | |
jroll | we don't maint often | 18:15 |
jroll | power state sync only saves on change | 18:15 |
jroll | heartbeat in most deployments is only during deploy/clean | 18:15 |
jroll | I think save() shouldn't be too bad | 18:15 |
openstackgerrit | Merged openstack/bifrost: Fix dhcp range for static inventory https://review.openstack.org/255182 | 18:17 |
*** nicodemos has quit IRC | 18:19 | |
lucasagomes | jroll, ack I will try to figure how it's going to work | 18:20 |
*** aginwala has joined #openstack-ironic | 18:20 | |
jroll | lucasagomes: cool, dansmith gave me some of those ideas so he may be able to help out too | 18:21 |
lucasagomes | ack | 18:21 |
lucasagomes | jroll, last thing, so in the spec I will mention that downgrade won't be done | 18:22 |
lucasagomes | won't be possible* | 18:22 |
jroll | yep | 18:22 |
lucasagomes | fair enuff | 18:22 |
jroll | well, we shouldn't be doing them anymore now anyway | 18:22 |
jroll | because http://specs.openstack.org/openstack/openstack-specs/specs/no-downward-sql-migration.html | 18:22 |
lucasagomes | ack | 18:22 |
lucasagomes | thanks | 18:22 |
jroll | np, thank you for working on this | 18:23 |
*** trown|lunch is now known as trown | 18:24 | |
*** garthb has joined #openstack-ironic | 18:28 | |
*** rajathagasthya has joined #openstack-ironic | 18:28 | |
openstackgerrit | Chris Krelle proposed openstack/bifrost: add check for genisoimage https://review.openstack.org/256030 | 18:32 |
*** dims has joined #openstack-ironic | 18:36 | |
*** jdk has joined #openstack-ironic | 18:37 | |
*** jistr has quit IRC | 18:38 | |
jdk | hi TheJulia | 18:38 |
*** ionutbalutoiu has quit IRC | 18:39 | |
jdk | TheJulia: switching to legacy mode worked, now the ipa ram disk loads but it is getting an error trying to post back to ironic saying network is unreachable | 18:39 |
*** aginwala has quit IRC | 18:39 | |
jdk | TheJulia: looking at troubleshooting.rst , it says to add ssh keys in order to login to coreos ipa ramdisk | 18:40 |
jdk | that is what i'm doing now | 18:40 |
*** aginwala has joined #openstack-ironic | 18:43 | |
rajathagasthya | lucasagomes: Your suggestion of using CoreOS upstream ramdisk worked partially yesterday. I don't see "failed to initialize RDMA" error. But I see a weird issue where I see login prompt for CoreOS and not my user image Ubuntu. And Ironic status is all messed up. It shows power off and deploying, and after a while reports deploy failed. | 18:43 |
jdk | TheJulia: network is unreachable, is there a way to login to coreos ipa ramdisk from console? | 18:45 |
lucasagomes | rajathagasthya, you will see the prompt while deploying right? | 18:46 |
lucasagomes | does the node even goes to active? | 18:46 |
mtanino | dtantsur|afk: Sorry for my interrupt, if you have a chance, could you revisit smoriya's spec? https://review.openstack.org/#/c/200496/ To get Nova's exception, this needs to be merged within tomorrow. | 18:46 |
*** Marga_ has quit IRC | 18:46 | |
*** aginwala has quit IRC | 18:47 | |
rajathagasthya | lucasagomes: Yeah, I see the prompt during deploy. Node never goes to active. | 18:47 |
lucasagomes | rajathagasthya, ok so the image didn't get deployed | 18:47 |
*** Marga_ has joined #openstack-ironic | 18:47 | |
lucasagomes | rajathagasthya, in the prompt you only see the console? You don't actually see any errors ? | 18:47 |
rajathagasthya | lucasagomes: Right, no errors. | 18:48 |
rajathagasthya | lucasagomes: I built a new Ubuntu deploy ramdisk and I now see a Ubuntu 14.04 prompt, but Ironic still shows deploy failed and power off. | 18:48 |
lucasagomes | right, you may not be seeming errors because we are not redirecting it to the console tho | 18:48 |
lucasagomes | rajathagasthya, can you edit the pxe_append_params config opt in Ironic and add (/me one sec lemme find the systemd option for that) | 18:49 |
lucasagomes | rajathagasthya, "systemd.journald.forward_to_console=yes" | 18:49 |
rajathagasthya | lucasagomes: Is it possible that the pxe_ucs driver is not reporting correct power state state? | 18:49 |
lucasagomes | rajathagasthya, right probably the deployment timeout due some error | 18:49 |
rajathagasthya | lucasagomes: Okay, will add that. | 18:49 |
lucasagomes | rajathagasthya, right, it's possible but... Ironic would fail straight away if it can set the power state to on when the deployment starts | 18:50 |
*** cemason has quit IRC | 18:50 | |
*** mgoddard has quit IRC | 18:51 | |
rajathagasthya | lucasagomes: Okay, will try a deploy and let you know how it goes. | 18:51 |
lucasagomes | ack | 18:51 |
*** mgoddard has joined #openstack-ironic | 18:52 | |
*** aginwala has joined #openstack-ironic | 18:56 | |
jlvillal | Sukhdev, jroll: When getting the service url from keystone. Shouldn't the 'service_type' be 'network' and not 'neutron'? | 18:56 |
*** harlowja has quit IRC | 18:56 | |
* jlvillal learned about keystone recently... | 18:56 | |
*** harlowja has joined #openstack-ironic | 18:56 | |
jroll | sounds right | 18:56 |
jlvillal | jroll, https://review.openstack.org/#/c/255885/1/ironic/dhcp/neutron.py | 18:57 |
jlvillal | jroll, That is why I asked. My comment | 18:57 |
jroll | jlvillal: heh | 18:57 |
jlvillal | jroll, If you think it is wrong, I can spin a patch. | 18:57 |
jlvillal | After my meeting.... | 18:57 |
jroll | jlvillal: I have no clue, honestly | 18:58 |
jroll | it's been there for at least a year... | 18:58 |
jroll | so don't panic :P | 18:58 |
*** jcoufal has quit IRC | 18:58 | |
jlvillal | :) Yeah, I think it almost always use the CONF value, I think. | 18:58 |
jlvillal | I will put it on my TODO list. Thanks! | 18:58 |
jroll | thank you :) | 18:58 |
*** dprince has quit IRC | 18:58 | |
* jlvillal goes to team brown-bag session. | 18:58 | |
*** kkoski has quit IRC | 18:59 | |
* jroll takes a break for a while | 18:59 | |
jroll | jlvillal: oh btw, I'm working on the devstack and grenade plugin stuff | 18:59 |
*** dprince has joined #openstack-ironic | 18:59 | |
*** Sukhdev has quit IRC | 19:00 | |
*** kkoski has joined #openstack-ironic | 19:00 | |
*** Sukhdev has joined #openstack-ironic | 19:00 | |
rajathagasthya | lucasagomes: Okay, I see "cannot create /var/lib/cloud/data/no-net" and "/bin/sh: 1: cloud-init not found". Then "booting system without full network configuration" after which I see Ubuntu 14.04 login prompt. | 19:00 |
rajathagasthya | lucasagomes: Ironic says power off and deploying, Nova says spawning. | 19:01 |
lucasagomes | rajathagasthya, that's after the deployment? What's the provision state of the ne? | 19:01 |
lucasagomes | oh | 19:01 |
lucasagomes | rajathagasthya, are you sure you are not using the tenant image as the deploy ramdisk | 19:02 |
lucasagomes | rajathagasthya, if you look at the node's driver_info/deploy_ramdisk=<glance uuid> | 19:02 |
lucasagomes | that glance uuid is pointing to which image? | 19:02 |
rajathagasthya | lucasagomes: I built a new ramdisk using "disk-image-create ironic-agent ubuntu -o ironic-deploy". | 19:03 |
lucasagomes | oh right so ur not using the upstream one anymore? | 19:03 |
rajathagasthya | lucasagomes: deploy_ramdisk points to ubuntu-deploy-initrd. | 19:03 |
rajathagasthya | lucasagomes: No, but I still got the same error for upstream ramdisk. | 19:04 |
lucasagomes | rajathagasthya, same error re cloud-init not found? | 19:05 |
lucasagomes | I thought you couldn't see any error with the upstream ramdisk, that's why we added that systemd line there | 19:05 |
*** jaybeale has joined #openstack-ironic | 19:05 | |
lucasagomes | <rajathagasthya> lucasagomes: Right, no errors. | 19:05 |
rajathagasthya | lucasagomes:Same error as in Ironic reporting deploying and power off. | 19:05 |
lucasagomes | ah | 19:05 |
rajathagasthya | lucasagomes: You want me to try upstream ramdisk once? | 19:05 |
lucasagomes | rajathagasthya, yes please | 19:06 |
rajathagasthya | lucasagomes: Cool, will do now. | 19:06 |
lucasagomes | still the power off/deploying is very weird | 19:06 |
lucasagomes | haven't seem that before | 19:06 |
SpamapS | lucasagomes: hey didn't you recently add an element to dib to help debug things on ironic agent? | 19:07 |
devananda | lucasagomes: I noticed today that I see (poweroff, cleaning) + (poweron, clean wait) + (poweroff, cleaning) + (poweroff, available) | 19:07 |
rajathagasthya | lucasagomes: That's wierd and it seems to be in sync with "Desired power state" in UCS Manager, not the actual power state. Hence I suspect that it's UCS driver problem. | 19:07 |
lucasagomes | SpamapS, yup! Good point we can use that | 19:07 |
SpamapS | jdk: ^^ | 19:07 |
jdk | SpamapS: hi | 19:07 |
SpamapS | lucasagomes: jdk is also looking for some debug support in IPA, what's it called again? | 19:07 |
lucasagomes | rajathagasthya, when building a ramdisk with DIB, you can use a dynamic-login element. That will allow you to log-in the node | 19:08 |
rajathagasthya | BTW, I'm using Kilo. Very sorry if I haven't mentioned this before. | 19:08 |
lucasagomes | SpamapS, dynamic-login | 19:08 |
rajathagasthya | lucasagomes: No, I didn't use dynamic-login. How do I use that? Is there an example? | 19:08 |
*** ukalifon1 has joined #openstack-ironic | 19:09 | |
lucasagomes | rajathagasthya, https://github.com/openstack/diskimage-builder/tree/master/elements/dynamic-login | 19:09 |
rajathagasthya | lucasagomes: Thanks, that's helpful. | 19:09 |
lucasagomes | devananda, right... hmm it kinda makes sense tho... cause cleaning == conductor executing clean steps, cleanwait == clean steps being executed in the node | 19:10 |
lucasagomes | so if it's cleaning and powered off it should be fine | 19:10 |
lucasagomes | probably it's a oob clean step | 19:10 |
devananda | jroll, Sukhdev: do you know of any commodity switches that will support the ML2 integration work, like, for testing at home? | 19:11 |
devananda | lucasagomes: yea, I agree that that is what happens. I think it's poor UX though | 19:11 |
lucasagomes | indeed | 19:11 |
lucasagomes | those distinctions are not very user friendly | 19:11 |
lucasagomes | at least it's consistent with deploying/deploywait | 19:12 |
Sukhdev | devananda: nothing comes to mind | 19:12 |
*** e0ne has joined #openstack-ironic | 19:13 | |
Sukhdev | devananda : mirantis guys wrote a generic switch implementation for Ironic_neutron testing - perhaps you can play with it - let me look for the link | 19:13 |
*** maratoid has joined #openstack-ironic | 19:13 | |
devananda | lucasagomes: indeed. "clean wait" is displayed during the time that the disks are being shredded ... | 19:14 |
devananda | same for "deploy wait" while the image is being written | 19:14 |
Sukhdev | devananda : check out this - https://github.com/jumpojoy/generic_switch | 19:15 |
lucasagomes | devananda, yeah, and we allow abort/delete on the *wait states | 19:15 |
*** Marga_ has quit IRC | 19:15 | |
* mgould -> home; good night! | 19:15 | |
*** mgould has quit IRC | 19:15 | |
*** ukalifon1 has quit IRC | 19:16 | |
*** Nisha has quit IRC | 19:18 | |
*** yarkot has joined #openstack-ironic | 19:19 | |
*** aginwala has quit IRC | 19:19 | |
*** ijw has joined #openstack-ironic | 19:20 | |
*** Sukhdev has quit IRC | 19:20 | |
*** ijw has quit IRC | 19:24 | |
*** ijw has joined #openstack-ironic | 19:24 | |
*** aginwala has joined #openstack-ironic | 19:24 | |
*** Marga_ has joined #openstack-ironic | 19:25 | |
*** ukalifon has joined #openstack-ironic | 19:26 | |
*** r-daneel has joined #openstack-ironic | 19:27 | |
rajathagasthya | lucasagomes: I used upstream deploy ramdisk and see "Error heartbeating to agent API" errors. | 19:27 |
NobodyCam | night mgould | 19:29 |
openstackgerrit | Merged openstack/ironic: Always pass keystone credentials to neutronclient https://review.openstack.org/255885 | 19:35 |
*** electrofelix has quit IRC | 19:39 | |
lucasagomes | rajathagasthya, hmm you know if the node you're booting can access the ironic api? | 19:41 |
lucasagomes | rajathagasthya, what you can do it is | 19:41 |
rajathagasthya | lucasagomes: Oh yeah, it can. | 19:41 |
lucasagomes | rajathagasthya, in the pxe parameters add "coreos.autologin" | 19:41 |
lucasagomes | that will give a prompt in the upstream coreos ramdisk | 19:41 |
lucasagomes | so you can troubleshoot it, ping, check the network etc | 19:42 |
JayF | lucasagomes: turns out the bug I found ... was in an unmerged patch that we carried downstream | 19:42 |
rajathagasthya | lucasagomes: Ah, okay. But wouldn't I get the same heartbeat error again? Or does allow me to login with that option? | 19:43 |
JayF | lucasagomes: so I'm fixing it in the patch and getting it back up | 19:43 |
JayF | after playing about 30 minutes of "where's that code?" | 19:43 |
lucasagomes | rajathagasthya, you will get the same eror again, but right now you can't log in the ramdisk to troubleshoot it can you? | 19:43 |
lucasagomes | JayF, oh boomer! At least you found it, so it's good :D | 19:44 |
JayF | well it's code that should be merged in the agent | 19:44 |
JayF | really almost a requirement for cleaning | 19:44 |
JayF | https://review.openstack.org/#/c/189969 I'm about to put up a fresh patchset of it | 19:44 |
* lucasagomes looks | 19:45 | |
lucasagomes | oh right | 19:45 |
rajathagasthya | lucasagomes: I can't now, after I enabled the systemd.journald.forward_to_console option. I don't see the login prompt at all. But let me see what happens with auto login. | 19:46 |
lucasagomes | rajathagasthya, ack, you can add both | 19:47 |
lucasagomes | this will give you a prompt | 19:47 |
lucasagomes | and redirect errors to the console | 19:47 |
rajathagasthya | lucasagomes: So 'ack' is another option? | 19:47 |
lucasagomes | rajathagasthya, no sorry... ack == understood | 19:48 |
lucasagomes | it's a slang | 19:48 |
rajathagasthya | lucasagomes: Lol, sorry. What was I thinking? ;) | 19:48 |
lucasagomes | like SYN-ACK from tcp | 19:48 |
rajathagasthya | lucasagomes: Yes, yes, yes. | 19:48 |
jdk | SpamapS: the problem yesterday was that undionly.kpxe is for legacy mode; you need to use ipxe.efi for uefi mode | 19:50 |
*** Marga_ has quit IRC | 19:50 | |
SpamapS | jdk: OH, so I was dead wrong about UEFI not mattering. Thanks. :) | 19:50 |
SpamapS | jdk: (I told you this isn't really my area. ;) | 19:50 |
JayF | Can someone confirm for me that tests on IPA master are not broken? | 19:51 |
JayF | Locally, I see a very leaky test that's failing trying to mess with my local /dev/sdb | 19:52 |
*** jaybeale has quit IRC | 19:52 | |
JayF | https://gist.github.com/jayofdoom/aa510056f73aa8b64c1c is the error I'm getting in both py27 and py34 | 19:52 |
*** baoli_ has quit IRC | 19:53 | |
*** rajathagasthya has quit IRC | 19:55 | |
*** kkoski has quit IRC | 19:56 | |
lucasagomes | jroll, seems I got some migration going on by changing update_node, and the get_node_by_X in the sqlalchemy/api.py | 19:57 |
lucasagomes | jroll, I'm only migrating when properties is updated | 19:57 |
*** rajathagasthya has joined #openstack-ironic | 19:57 | |
lucasagomes | so other save()'s update()'s wont' interfere | 19:58 |
*** ChubYann has joined #openstack-ironic | 19:58 | |
JayF | dtantsur|afk: lucasagomes: I'm like 90% sure b569e37d065910cb25e4817951fa75937b7ef357 broke tests as noted ^ above | 19:59 |
JayF | tests will only pass on machines that have a /dev/sdb device actually on them | 19:59 |
JayF | which my VM does not | 19:59 |
JayF | leaky test = scary test when we're talking about a rest API to erase a computer :P | 19:59 |
lucasagomes | JayF, lemme clone the repo and create the test env | 19:59 |
lucasagomes | from scratch and see | 19:59 |
lucasagomes | JayF, py27 and 34? | 20:00 |
JayF | lucasagomes: yep, but make sure your test box doesn't have a /dev/sdb at all | 20:00 |
JayF | lucasagomes: if it does the bad test will pass, which is why it passes in the gate | 20:00 |
lucasagomes | JayF, only sda | 20:00 |
JayF | just looking at the test, it only mocks out utils.execute | 20:01 |
JayF | when it uses pyudev now to do a lot of that stuff | 20:01 |
lucasagomes | will tell u in a bit... (internet in the country side of brazil is kinda poor, so takes me a while to build stuff) | 20:01 |
JayF | the pyudev stuff is not mocked out at all | 20:01 |
*** rcernin has joined #openstack-ironic | 20:01 | |
JayF | https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/hardware.py#L94 will have to be mocked out to make that test stop failing | 20:02 |
JayF | wait | 20:02 |
JayF | ... | 20:02 |
JayF | is this rebase fail? | 20:02 |
* JayF wonders suddenly why his hardware.py isn't hte same as master | 20:03 | |
lucasagomes | JayF, http://paste.openstack.org/show/481549/ | 20:03 |
* lucasagomes looks at the patch | 20:03 | |
*** mgoddard_ has joined #openstack-ironic | 20:03 | |
*** mgoddard has quit IRC | 20:03 | |
JayF | lucasagomes: it's all my fault | 20:04 |
JayF | lucasagomes: I was syncing up and rebasing on our very not-up-to-date master | 20:05 |
JayF | instead of upstream | 20:05 |
lucasagomes | oh | 20:05 |
lucasagomes | hah, happens | 20:05 |
JayF | that's what happens when you read code in the browser | 20:05 |
JayF | and assume it's the same as on disk | 20:05 |
NobodyCam | :p | 20:05 |
JayF | that test /is/ still leaky though, I think | 20:06 |
JayF | because pyudev is actually checking on /dev/sdb in my local sysrtem | 20:06 |
lucasagomes | yeah we should mock it for sure | 20:06 |
*** ionutbalutoiu has joined #openstack-ironic | 20:06 | |
openstackgerrit | Chris Krelle proposed openstack/bifrost: add check for genisoimage https://review.openstack.org/256030 | 20:07 |
*** kkoski has joined #openstack-ironic | 20:07 | |
openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Allow hardware managers to override clean step priority https://review.openstack.org/189969 | 20:09 |
JayF | rloo: ^ You put comments on that months ago. I just came across a bug in it downstream to discover it was never pushed through. I'd love a re-review on it and I'll make sure it gets completed | 20:10 |
rloo | JayF: oh. will try to get to it next week. i have been dragging my feet on the OSC CLI stuff and i really want to look into that one today. | 20:11 |
*** Marga_ has joined #openstack-ironic | 20:12 | |
jroll | lucasagomes: ++ that sounds good | 20:14 |
lucasagomes | jroll, I will continue tomorrow and then update the spec | 20:14 |
lucasagomes | but so far so good, seems quite simple | 20:14 |
lucasagomes | I just need refactor some of the get_node_by_* because so they use the same function to pull the properties from the right place | 20:14 |
jroll | awesome, good stuff :) | 20:15 |
jroll | yep | 20:15 |
openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Fix iPXE template for whole disk image https://review.openstack.org/255362 | 20:18 |
lucasagomes | rloo, ^ updated the release note with ur suggestion | 20:18 |
lucasagomes | if folks have a time please take a look at ^, it's a simple patch | 20:18 |
lucasagomes | and with that I will call it a day | 20:18 |
lucasagomes | have a great night everyone! | 20:18 |
*** lucasagomes is now known as lucas-beer | 20:19 | |
NobodyCam | night lucas-beer | 20:20 |
rloo | thx lucas-beer | 20:20 |
jroll | \o lucas | 20:20 |
* rloo had a thought about raffling off a +2. | 20:20 | |
* rloo goes back to work | 20:20 | |
*** rajathagasthya has quit IRC | 20:22 | |
jroll | rloo: I have things you can +2! | 20:22 |
jlvillal | jroll, The devstack plugin stuff sounds great! Thanks | 20:22 |
jroll | https://review.openstack.org/#/c/255879/ | 20:22 |
jroll | https://review.openstack.org/#/c/255880/ | 20:22 |
*** Sukhdev has joined #openstack-ironic | 20:22 | |
jroll | ^ need eyes on those, dtantsur|afk had +2s earlier | 20:22 |
rloo | jroll: is that something we urgently want or can it wait til next week? | 20:23 |
jroll | jlvillal: np. it needs to get done. I was going to save grenade for later, but it's a blocker to finishing the devstack plugin stuff | 20:23 |
jroll | rloo: I'd prefer sooner than later as it's just the beginning of the work (and it's pretty much copy paste), but it isn't urgent | 20:23 |
jlvillal | jroll, Sorry that all this training came up. I had no idea it was going to be so long.... | 20:23 |
jroll | rloo: it's devstack plugin so that we can move faster on devstack changes | 20:23 |
jroll | jlvillal: it's fine | 20:24 |
jroll | nothing you could do about it | 20:24 |
jroll | hope you learned a thing or 5 :D | 20:24 |
*** rajathagasthya has joined #openstack-ironic | 20:24 | |
jlvillal | jroll, I did! A lot more about OpenStack for sure. | 20:24 |
rloo | jroll: I feel bad about this spec: https://review.openstack.org/#/c/247539/. was going to look/think about it today (and am procrastinating about it) | 20:24 |
jroll | awesome, then it's worth it | 20:25 |
jroll | rloo: yeah, that's been lagging, go for it | 20:25 |
jroll | rloo: hit my patches when you need a mental break, because it's copy/paste and doesn't actually change anything until we make gate changes :) | 20:25 |
jroll | so if it doesn't work yet it's ok | 20:25 |
rloo | jroll: oh. ok. might not be so bad then. will see... | 20:26 |
jroll | jlvillal: if you get curious, I'm using this topic for everything https://review.openstack.org/#/q/status:open++branch:master+topic:ironic-devstack-plugin,n,z | 20:26 |
rloo | what's NobodyCam up to? Maybe he has a few minutes? ^^ | 20:26 |
jlvillal | jroll, Looking at the patches now... | 20:26 |
jlvillal | starting with https://review.openstack.org/#/c/255879/ | 20:26 |
jroll | cool | 20:27 |
jroll | thanks! | 20:27 |
*** rajathagasthya has quit IRC | 20:27 | |
NobodyCam | I was making soup :0 | 20:28 |
rloo | NobodyCam: no worries. If you have time, jroll has those devstadck plugin patches. | 20:29 |
NobodyCam | lookin now() | 20:30 |
*** alex_xu has quit IRC | 20:30 | |
*** athomas has quit IRC | 20:32 | |
jlvillal | jroll, +2 on https://review.openstack.org/#/c/255879/ I compared every file added to what is in devstack and perfect match :) | 20:32 |
*** yarkot has quit IRC | 20:32 | |
jroll | cool, thanks man | 20:32 |
*** alex_xu has joined #openstack-ironic | 20:33 | |
*** lucas-beer has quit IRC | 20:34 | |
jroll | jlvillal: oh, and as a note, plugin.sh in the next one is basically the same as extras.d/50-ironic | 20:34 |
jroll | one minor change | 20:34 |
*** dims has quit IRC | 20:35 | |
*** dims has joined #openstack-ironic | 20:35 | |
jlvillal | Okay | 20:36 |
*** dprince has quit IRC | 20:37 | |
jlvillal | jroll, I assume the doc files were left behind intentionally. | 20:39 |
jlvillal | In 1st patch | 20:39 |
*** lucasagomes has joined #openstack-ironic | 20:40 | |
jroll | jlvillal: there weren't any ironic-specific doc files in devstack, just bits in existing docs | 20:40 |
jroll | we should totally add a readme | 20:40 |
jroll | but I haven't done it yet | 20:40 |
*** e0ne has quit IRC | 20:41 | |
jlvillal | jroll, I was thinking these docs: http://paste.openstack.org/show/481552/ | 20:41 |
jlvillal | docs/* | 20:41 |
jlvillal | s/docs/doc/ | 20:42 |
NobodyCam | jroll: have you seen this: http://logs.openstack.org/98/255898/5/check/gate-grenade-dsvm/5950b6f/logs/grenade.sh.txt.gz#_2015-12-10_16_53_39_386 | 20:42 |
jlvillal | Looks like at most it would have been one doc file: ./doc/build/html/lib/ironic.html | 20:43 |
jroll | NobodyCam: yeah, that's further down the road right? | 20:43 |
jroll | where we remove the devstack code | 20:43 |
jroll | NobodyCam: yeah, so that one needs some work yet - need to finagle grenade | 20:44 |
jroll | it's on my list | 20:44 |
NobodyCam | :) | 20:44 |
* jroll needs to go get a sandwich, bbiab | 20:45 | |
NobodyCam | :) I have a all hands meeting in 15 too | 20:45 |
*** yarkot has joined #openstack-ironic | 20:46 | |
jlvillal | jroll, Ignore my doc comment. Those aren't in git. | 20:47 |
*** cdearborn has joined #openstack-ironic | 20:49 | |
*** jaybeale has joined #openstack-ironic | 20:58 | |
*** Sukhdev has quit IRC | 21:01 | |
*** cemason has joined #openstack-ironic | 21:02 | |
*** Walts has quit IRC | 21:03 | |
*** [1]cdearborn has quit IRC | 21:03 | |
*** cemason has quit IRC | 21:07 | |
*** raildo is now known as raildo-afk | 21:07 | |
*** baoli has joined #openstack-ironic | 21:09 | |
jdk | SpamapS: which distribution should I use for the IPA ramdisk? it has to support the x710 intel nics with a recent i40e driver | 21:10 |
*** rajathagasthya has joined #openstack-ironic | 21:10 | |
*** mgoddard has joined #openstack-ironic | 21:11 | |
*** mgoddard_ has quit IRC | 21:12 | |
*** ukalifon has quit IRC | 21:13 | |
*** rajathagasthya has quit IRC | 21:14 | |
devananda | harlowja: just saw your reply on the ironic claims spec | 21:17 |
harlowja | devananda :-P | 21:17 |
harlowja | ya, not trying to start a fire, just trying to understand a few thigns | 21:17 |
devananda | harlowja: i'm writing a lenghthy email now ... but ... i want to clarify / get clarification on a few things | 21:17 |
harlowja | uh ohs | 21:17 |
* harlowja runs away | 21:17 | |
harlowja | lol | 21:17 |
harlowja | :) | 21:17 |
devananda | harlowja: it sounds like you're proposing that nova and ironic both write to the same db table | 21:18 |
devananda | but that's crazy talk :) | 21:18 |
harlowja | lets not call it a db table | 21:18 |
harlowja | let's call it something else | 21:18 |
devananda | harlowja: will the nova team accept such a change? | 21:18 |
*** Sukhdev has joined #openstack-ironic | 21:18 | |
harlowja | db table has to many connotations | 21:18 |
harlowja | will they | 21:18 |
harlowja | that is the question | 21:19 |
harlowja | the million dollar question, ha | 21:19 |
devananda | well | 21:19 |
*** Sukhdev has quit IRC | 21:19 | |
devananda | all of my past discussions/interactions with them leads me very much to avoid that | 21:19 |
devananda | but hey, let's ask them :) | 21:20 |
harlowja | ;) | 21:20 |
*** Sukhdev has joined #openstack-ironic | 21:20 | |
*** baoli has quit IRC | 21:20 | |
harlowja | depends on how much flake u feel like taking around xmas, ha | 21:21 |
*** baoli has joined #openstack-ironic | 21:21 | |
devananda | harlowja: zero | 21:21 |
harlowja | :) | 21:21 |
harlowja | *i guess its flak, not flake, lol | 21:27 |
harlowja | whatever | 21:27 |
rloo | harlowja: lots of snowflakes would be good! | 21:27 |
harlowja | ;) | 21:27 |
mordred | TheJulia: you might want to double-check this patch: https://review.openstack.org/#/c/254399/ | 21:28 |
mordred | TheJulia: it changes how the None auth type works, which I know you use in bifrost | 21:28 |
mordred | TheJulia: it SHOULD work for you | 21:28 |
mordred | TheJulia: (it turns out that we've been making that much too hard this whole time, and an auth plugin value of None does what you want all this time) | 21:29 |
BadCub | hey mordred long time no see! | 21:30 |
NobodyCam | hey hey mordred :) | 21:31 |
*** rajathagasthya has joined #openstack-ironic | 21:31 | |
*** baoli has quit IRC | 21:32 | |
*** baoli has joined #openstack-ironic | 21:33 | |
*** kkoski has quit IRC | 21:33 | |
*** vishwanathj has quit IRC | 21:34 | |
*** vishwanathj has joined #openstack-ironic | 21:34 | |
*** caiobo has quit IRC | 21:34 | |
BadCub | mordred: TheJulia asked us to relay that she is away for a while at an appointment | 21:37 |
*** baoli has quit IRC | 21:44 | |
*** baoli has joined #openstack-ironic | 21:45 | |
jlvillal | jroll, Darn, you stole my thunder. I thought I found a valid issue ;) | 21:48 |
*** baoli has quit IRC | 21:52 | |
*** baoli has joined #openstack-ironic | 21:53 | |
*** Sukhdev has quit IRC | 21:54 | |
jroll | :P | 21:56 |
*** Sukhdev has joined #openstack-ironic | 21:56 | |
JayF | jroll.append_to_review_queue(https://review.openstack.org/#/c/189969/, https://review.openstack.org/#/c/254896/) :P | 21:59 |
*** baoli has quit IRC | 22:00 | |
jroll | AttributeError: 'NoneType' object has no attribute 'append_to_review_queue' | 22:00 |
*** baoli has joined #openstack-ironic | 22:01 | |
JayF | jroll: Hey, apparently I didn't do this right but I'd love a review on 189969 and 254896 </exception_handler> | 22:01 |
JayF | lol | 22:01 |
jroll | heh | 22:01 |
* jroll appends to queue | 22:01 | |
*** ionutbalutoiu has quit IRC | 22:04 | |
*** jdk has quit IRC | 22:04 | |
*** aginwala has quit IRC | 22:04 | |
*** alex_xu has quit IRC | 22:07 | |
*** derekh has quit IRC | 22:07 | |
*** rajathagasthya has quit IRC | 22:08 | |
openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Add command metrics to IPA API https://review.openstack.org/119981 | 22:08 |
*** alex_xu has joined #openstack-ironic | 22:09 | |
*** praneshp has quit IRC | 22:11 | |
*** shakamunyi has quit IRC | 22:12 | |
*** penick has joined #openstack-ironic | 22:12 | |
*** yarkot has quit IRC | 22:12 | |
*** spandhe has joined #openstack-ironic | 22:12 | |
*** baoli has quit IRC | 22:13 | |
*** baoli has joined #openstack-ironic | 22:13 | |
*** yarkot has joined #openstack-ironic | 22:13 | |
*** aginwala has joined #openstack-ironic | 22:14 | |
*** trown is now known as trown|outttypeww | 22:14 | |
*** shakamunyi has joined #openstack-ironic | 22:15 | |
penick | hey, so I was chatting with melwitt and she pointed out some problems with my plan to drop resource tracker and go straight to the ironic DB. Really one problem now, but I could see it growing. If I drop the RT and go direct to ironic nodes, then we’ll lose out on host aggregates | 22:15 |
*** praneshp has joined #openstack-ironic | 22:15 | |
penick | so, new plan is to look into whether or not we can either a: make RT smarter and event driven for everything, or B: see if at least we can make a smarter ironic RT | 22:16 |
harlowja | devananda ^ | 22:16 |
penick | Oops, truncated that on accident. Thanks JOsh | 22:16 |
penick | s/O/o | 22:16 |
harlowja | np | 22:16 |
*** yarkot has quit IRC | 22:17 | |
*** david-lyle has joined #openstack-ironic | 22:20 | |
*** jamielennox|away is now known as jamielennox | 22:22 | |
*** derekh has joined #openstack-ironic | 22:24 | |
*** baoli has quit IRC | 22:24 | |
*** baoli has joined #openstack-ironic | 22:25 | |
*** kkoski has joined #openstack-ironic | 22:26 | |
devananda | penick: completely dropping the RT seems like you'll lose a lot more than that | 22:28 |
openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Add command metrics to IPA API https://review.openstack.org/119981 | 22:28 |
devananda | quotas? | 22:28 |
devananda | penick: how about RPC? which compute host will the scheduler send a given job to? | 22:29 |
*** caiobo has joined #openstack-ironic | 22:33 | |
*** maratoid has quit IRC | 22:36 | |
*** baoli has quit IRC | 22:36 | |
*** baoli has joined #openstack-ironic | 22:37 | |
penick | Quota would be fine, and selecting a compute host could be done by round robining through the hosts, if they’re no longer needed for any scheduling or “ownership” constraints. | 22:37 |
*** thrash is now known as thrash|g0ne | 22:38 | |
penick | But, the primary point being that RT probably needs to be kept, but improved and that moving away from the nova scheduler is the wrong thing to do | 22:38 |
penick | devananda: ^ | 22:41 |
devananda | penick: has there been a discussion of this with the nova-core team yet? | 22:42 |
penick | Not yet, that’s next. I was going to wait for the midcycle in jan, since that’s coming up | 22:42 |
devananda | why wait? | 22:43 |
*** alexpilotti has quit IRC | 22:43 | |
penick | devananda: It’s a little over a month from the midcycle, so I was going to wait until then. | 22:43 |
devananda | yes, well, the current solution was discussed at the last midcycle and approved in october | 22:44 |
penick | devananda: higher bandwidth discussion seemed better for this | 22:44 |
devananda | so waiting two more moths is kinda annoying | 22:44 |
penick | Ok, well, I can fire off an email to the list and we can try to start something sooner | 22:44 |
devananda | penick: see PM. I already have an email drafted. | 22:45 |
*** rcernin has quit IRC | 22:45 | |
*** baoli has quit IRC | 22:48 | |
*** baoli has joined #openstack-ironic | 22:49 | |
*** alexpilotti has joined #openstack-ironic | 22:49 | |
*** ndipanov has quit IRC | 22:52 | |
openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Allow hardware managers to override clean step priority https://review.openstack.org/189969 | 22:54 |
*** david-lyle has quit IRC | 22:55 | |
openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Allow hardware managers to override clean step priority https://review.openstack.org/189969 | 22:57 |
*** baoli has quit IRC | 23:01 | |
*** baoli has joined #openstack-ironic | 23:01 | |
*** openstackstatus has quit IRC | 23:01 | |
*** baoli has quit IRC | 23:07 | |
*** baoli has joined #openstack-ironic | 23:11 | |
openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Allow hardware managers to override clean step priority https://review.openstack.org/189969 | 23:13 |
*** baoli has quit IRC | 23:17 | |
*** yuanying has joined #openstack-ironic | 23:18 | |
*** aginwala has quit IRC | 23:21 | |
mordred | BadCub: coolio. and thanks! and hiya! | 23:24 |
*** mgoddard has quit IRC | 23:24 | |
BadCub | mordred: you are welcome sir! | 23:26 |
NobodyCam | hey hey mordred :) hope all is good and well | 23:26 |
*** aginwala has joined #openstack-ironic | 23:26 | |
mordred | NobodyCam: you know it! | 23:26 |
NobodyCam | :) | 23:26 |
NobodyCam | wow havent seen this name in along while: Creates baremetal poseur nodes for ironic testing purposes | 23:31 |
*** Sukhdev has quit IRC | 23:32 | |
*** rajathagasthya has joined #openstack-ironic | 23:35 | |
*** derekh has quit IRC | 23:35 | |
*** cdearborn has quit IRC | 23:45 | |
NobodyCam | jroll: any benifet to removing the devstack review dependicy on 255879. so we can start making changes? its got two +2's already :p | 23:48 |
jroll | NobodyCam: I'd rather have it there to be sure | 23:50 |
jroll | we can approve it, it just won't merge until the devstack change lands | 23:50 |
NobodyCam | ok :) | 23:50 |
jroll | thanks though! | 23:51 |
NobodyCam | great stuff :) | 23:51 |
*** Haomeng has joined #openstack-ironic | 23:56 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!