| devananda | I'd love the docs to be done and in the release -- but it matters more that they're online soon | 00:00 |
|---|---|---|
| rloo | devananda: ok, since no one can really use RAID now anyway, I won't consider those a high priority then. | 00:02 |
| devananda | and I don't think we should even be considering a feature-release of the client at this point in Liberty just to add RAID support there | 00:02 |
| devananda | well, and there's that ... | 00:02 |
| rloo | devananda: oh, too late to release a client | 00:02 |
| devananda | if I understand it (and I'm not sure that I do) -- I would call the feature incomplete | 00:03 |
| rloo | devananda: yeah, if doc + client are part of the feature. | 00:03 |
| devananda | even though the REST interface is there, it's not really functional (because of zapping), the client doesn't support it, and it's not documented??? | 00:04 |
| devananda | yea, it's not done | 00:04 |
| rloo | so maybe you/jroll can change the milestone on that BP | 00:04 |
| *** tsekiyam_ has joined #openstack-ironic | 00:04 | |
| devananda | jroll: would you agree with my assesment? | 00:05 |
| *** jerrygb_ has joined #openstack-ironic | 00:05 | |
| *** mtanino has quit IRC | 00:07 | |
| *** tsekiyama has quit IRC | 00:08 | |
| *** jerrygb_ has quit IRC | 00:08 | |
| *** tsekiyam_ has quit IRC | 00:09 | |
| *** jerrygb has quit IRC | 00:09 | |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-python-agent: Allow hardware managers to override clean step priority https://review.openstack.org/189969 | 00:09 |
| *** naohirot has joined #openstack-ironic | 00:13 | |
| jroll | devananda: "generic raid interface" seems done to me, and needs a followup to be useful | 00:16 |
| *** Guest18166 is now known as dims_ | 00:17 | |
| jroll | the weird thing is there isn't RAID-specific code to be done to expose this through zapping | 00:17 |
| jroll | also rloo: ^ | 00:19 |
| rloo | jroll: there is. just not generic raid. | 00:19 |
| rloo | jroll: oh wait. zapping. thinking... | 00:19 |
| devananda | so the spec lists zapping as a dependency | 00:19 |
| devananda | is it _not_ ? | 00:20 |
| jroll | rloo: I think the 'manual clean steps' are there, just need a way to trigger them | 00:20 |
| rloo | devananda, jroll: it is a dependency because yeah, as jroll says | 00:20 |
| devananda | jroll: right, so, the *interface* is done. but it's not usable | 00:20 |
| jroll | yar | 00:21 |
| rloo | right. which is why i wanted to try to get it into automated cleaning but we nixed that today. | 00:21 |
| jroll | it's a weird situation | 00:22 |
| jroll | we basically decided to re-architect both of the ways it could be triggered | 00:22 |
| jroll | this week. | 00:22 |
| jroll | so it should probably be untargeted, marked unimplemented or whatever | 00:22 |
| rloo | well... not sure we decided to rearchitect the automated cleaning part; we just couldn't resolve some issues in a timely fashion. | 00:22 |
| *** shadower has quit IRC | 00:23 | |
| jroll | and when zapping is marked done, it is done | 00:23 |
| rloo | unless we decide to rearchitect cleaning which was one of the proposals... | 00:23 |
| *** shadower has joined #openstack-ironic | 00:23 | |
| jroll | lol | 00:23 |
| jroll | I'm not sure what we decided, other than defer to later | 00:23 |
| jroll | :) | 00:23 |
| rloo | jroll: oh, i think it is going to be a duel between you and lucas. i'm taking bets now :) | 00:23 |
| jroll | hah | 00:24 |
| jroll | devananda: it is untargeted, marked as 'good progress' | 00:25 |
| * jroll pours one out for RAID | 00:25 | |
| devananda | cool | 00:25 |
| devananda | also sad | 00:25 |
| devananda | that is a cool feature :( | 00:25 |
| rloo | jroll: maybe i should go through all the targetted features and punch holes in them to untarget them from L* | 00:25 |
| jroll | devananda: also, zapping will just add the ability to trigger it, we still need drivers and/or agents to support it :D | 00:25 |
| jroll | rloo: :| | 00:26 |
| rloo | I'm so sad too. all that time reviewing the code ... (and I'm sure ramesh is even sadder) | 00:26 |
| devananda | jroll: yup yup. and docs and CLI support | 00:26 |
| *** saripurigopi has joined #openstack-ironic | 00:26 | |
| devananda | and functional tests o.0 | 00:26 |
| jroll | hehe | 00:26 |
| devananda | jroll: were you preparing a patch to move completed specs to liberty/ ? | 00:27 |
| jroll | devananda: it's on my todo list, if that counts | 00:27 |
| devananda | np. I got it | 00:27 |
| jroll | \o/ | 00:27 |
| jroll | I kinda wanna land the rest of the boot interface | 00:27 |
| jroll | (I say that without looking at the code first) | 00:28 |
| *** ijw has quit IRC | 00:28 | |
| *** lucas-dinner has quit IRC | 00:28 | |
| devananda | I kinda do too | 00:28 |
| devananda | except those patches are huge | 00:28 |
| rloo | jroll: I looked earlier today. there were 4 patches. i couldn't do it. | 00:28 |
| devananda | two patches are ~1k lines each :( | 00:29 |
| rloo | I'm reviewed out wrt large patches. Need a break. | 00:29 |
| jroll | ouch | 00:29 |
| jroll | maybe not, then | 00:29 |
| *** garthb has quit IRC | 00:30 | |
| devananda | huh .. wasn't this implemented before 4.1 ? http://specs.openstack.org/openstack/ironic-specs/specs/approved/enroll-node-state.html | 00:30 |
| jroll | yes | 00:31 |
| devananda | k. i'll bump that too, separate patch | 00:31 |
| jroll | I don't love this moving to liberty/ thing because it's super manual and we miss things like that | 00:31 |
| jroll | :( | 00:31 |
| devananda | look at what nova's doing. I like this more, fwiw | 00:31 |
| devananda | (or dislike it less?) | 00:32 |
| devananda | it's manual, which means it's prone to error | 00:32 |
| jroll | well, it looks like they do roughly the same thing, but at end-of-cycle? | 00:32 |
| jroll | I like their priorities docs in the specs repo | 00:33 |
| jroll | also, our summit proposal list is still pretty short | 00:33 |
| jroll | https://etherpad.openstack.org/p/mitaka-ironic-design-summit-ideas | 00:33 |
| jroll | can't decide if people aren't proposing things or we just don't have much to talk about or? | 00:33 |
| jroll | exactly 2.5 of these are user facing features, curious what that's indicitive of | 00:34 |
| jroll | (I totally butchered that word) | 00:34 |
| jroll | (maybe) | 00:34 |
| devananda | jroll: it's been light every previous cycle too | 00:34 |
| devananda | maybe raise it at the meeting on monday | 00:35 |
| jroll | I feel like not this light, but idk | 00:35 |
| jroll | yeah | 00:35 |
| devananda | most of that 'pad was done live during the meeting last week, I think | 00:35 |
| rloo | i don't think we run out of things to talk about at the summit | 00:35 |
| jroll | indeed | 00:36 |
| jroll | rloo: of course, it just feels less feature-heavy | 00:36 |
| devananda | rloo: no, but I don't htink we've done a great job planing / organizing them very well ahead of time | 00:36 |
| rloo | devananda: ++ | 00:36 |
| devananda | or perhaps the priorities spreadsheet is actually accurate and we all just have a bunch of work to do? :) | 00:36 |
| jroll | heh | 00:36 |
| jroll | brb | 00:37 |
| *** davideagnello has quit IRC | 00:37 | |
| devananda | rloo: given the discussion around RAID this week, perhaps we should have a session on that | 00:37 |
| rloo | I feel like we didn't even accomplish what we discussed at last summit. i want to review those notes. | 00:37 |
| devananda | rloo: ++ | 00:37 |
| devananda | I know we didn't accomplish a lot of it | 00:37 |
| rloo | re-architecting cleaning might be a topic if we can't resolve it before then. | 00:37 |
| *** davideagnello has joined #openstack-ironic | 00:37 | |
| devananda | I think the midcycles have been more helpful than the summit for our design discussions for a while now | 00:37 |
| rloo | i wonder what the other projects think. summit vs mid-cycles. how to get the most out of summits, etc... | 00:40 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic-specs: Update spec file locations for 4.2 / Liberty https://review.openstack.org/226606 | 00:41 |
| *** achanda has quit IRC | 00:41 | |
| devananda | jroll: oh weird! that publish job is not deleting old spec files?? | 00:43 |
| devananda | both of these exist | 00:43 |
| devananda | http://specs.openstack.org/openstack/ironic-specs/specs/liberty-implemented/enroll-node-state.html | 00:43 |
| devananda | http://specs.openstack.org/openstack/ironic-specs/specs/approved/enroll-node-state.html | 00:43 |
| devananda | git shows the correct things -- it was moved already -- http://git.openstack.org/cgit/openstack/ironic-specs/tree/specs/approved | 00:44 |
| jroll | devananda: I have no clue, great question | 00:44 |
| * devananda files a bug to track it, since it's late and I need to make dinner soon | 00:45 | |
| jroll | I should also drop for the night | 00:48 |
| *** ijw has joined #openstack-ironic | 00:48 | |
| jroll | see you tomorrow devananda, rloo | 00:48 |
| devananda | g'night! | 00:48 |
| rloo | jroll: night | 00:48 |
| *** davideagnello has quit IRC | 00:49 | |
| *** davideagnello has joined #openstack-ironic | 00:49 | |
| *** Sukhdev has quit IRC | 00:50 | |
| devananda | https://bugs.launchpad.net/ironic/+bug/1498708 | 00:52 |
| openstack | Launchpad bug 1498708 in Ironic "Implemented specs are moved but an old copy remains" [Low,Triaged] | 00:52 |
| devananda | ok - i'm out for the night as well... see ya tmw! | 00:52 |
| *** ijw has quit IRC | 01:03 | |
| rloo | bye devananda. I'm going too. ciao everyone else :) | 01:04 |
| *** rloo has left #openstack-ironic | 01:04 | |
| *** priteau has joined #openstack-ironic | 01:09 | |
| *** praneshp has quit IRC | 01:10 | |
| *** achanda has joined #openstack-ironic | 01:11 | |
| *** priteau has quit IRC | 01:14 | |
| *** puranamr has joined #openstack-ironic | 01:15 | |
| *** [1]cdearborn has quit IRC | 01:17 | |
| *** chenglch has joined #openstack-ironic | 01:32 | |
| *** puranamr has quit IRC | 01:40 | |
| *** jerrygb has joined #openstack-ironic | 01:44 | |
| *** jerrygb has quit IRC | 01:45 | |
| *** saripurigopi has quit IRC | 01:45 | |
| *** jerrygb has joined #openstack-ironic | 01:45 | |
| *** harshs has quit IRC | 01:53 | |
| *** dims_ has quit IRC | 01:57 | |
| *** baoli has joined #openstack-ironic | 01:57 | |
| *** dims_ has joined #openstack-ironic | 01:57 | |
| *** dims_ has quit IRC | 02:10 | |
| *** dims_ has joined #openstack-ironic | 02:13 | |
| *** dims__ has joined #openstack-ironic | 02:18 | |
| *** dims_ has quit IRC | 02:18 | |
| *** dims__ has quit IRC | 02:23 | |
| *** bizarrochristy has quit IRC | 02:29 | |
| openstackgerrit | Merged openstack/ironic: Make end-points discoverable via Ironic API https://review.openstack.org/205895 | 02:36 |
| *** baoli has quit IRC | 02:37 | |
| *** baoli has joined #openstack-ironic | 02:38 | |
| *** early has quit IRC | 02:47 | |
| *** harshs has joined #openstack-ironic | 02:51 | |
| *** early has joined #openstack-ironic | 02:55 | |
| *** bizarrochristy has joined #openstack-ironic | 03:06 | |
| *** naohirot has quit IRC | 03:06 | |
| *** bizarrochristy has quit IRC | 03:07 | |
| *** boris-42 has quit IRC | 03:09 | |
| *** david-lyle has joined #openstack-ironic | 03:13 | |
| *** baoli has quit IRC | 03:32 | |
| *** olaph has quit IRC | 03:36 | |
| *** olaph has joined #openstack-ironic | 03:43 | |
| *** ijw has joined #openstack-ironic | 03:44 | |
| *** Marga__ has quit IRC | 03:46 | |
| *** Marga_ has joined #openstack-ironic | 03:46 | |
| *** dims_ has joined #openstack-ironic | 03:47 | |
| *** vinbs has joined #openstack-ironic | 03:48 | |
| *** ijw has quit IRC | 03:49 | |
| *** Marga_ has quit IRC | 03:50 | |
| *** shuquan has joined #openstack-ironic | 03:53 | |
| *** dims_ has quit IRC | 03:54 | |
| *** jcrubio has joined #openstack-ironic | 03:56 | |
| *** jerrygb has quit IRC | 03:56 | |
| *** garthb has joined #openstack-ironic | 03:57 | |
| *** shuquan has quit IRC | 03:58 | |
| *** jcrubio has quit IRC | 04:02 | |
| *** Marga_ has joined #openstack-ironic | 04:02 | |
| *** jcrubio has joined #openstack-ironic | 04:02 | |
| *** jcrubio has left #openstack-ironic | 04:03 | |
| *** naohirot has joined #openstack-ironic | 04:04 | |
| *** ijw has joined #openstack-ironic | 04:09 | |
| *** ijw has quit IRC | 04:09 | |
| *** ijw has joined #openstack-ironic | 04:09 | |
| *** rameshg87 has joined #openstack-ironic | 04:10 | |
| *** achanda has quit IRC | 04:17 | |
| *** jlk has joined #openstack-ironic | 04:20 | |
| *** jlk has quit IRC | 04:20 | |
| *** jlk has joined #openstack-ironic | 04:20 | |
| *** harshs has quit IRC | 04:26 | |
| *** Haomeng|2 has quit IRC | 04:34 | |
| *** Marga_ has quit IRC | 04:36 | |
| *** Marga_ has joined #openstack-ironic | 04:37 | |
| *** achanda has joined #openstack-ironic | 04:48 | |
| *** garthb has quit IRC | 04:55 | |
| *** praneshp has joined #openstack-ironic | 04:56 | |
| *** harshs has joined #openstack-ironic | 04:59 | |
| *** praneshp_ has joined #openstack-ironic | 05:06 | |
| *** praneshp has quit IRC | 05:09 | |
| *** praneshp_ is now known as praneshp | 05:09 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-python-agent: Fix default ironic api port number https://review.openstack.org/226626 | 05:11 |
| *** Sukhdev has joined #openstack-ironic | 05:11 | |
| *** Sukhdev has quit IRC | 05:17 | |
| *** harshs_ has joined #openstack-ironic | 05:29 | |
| *** Haomeng has joined #openstack-ironic | 05:29 | |
| *** harshs has quit IRC | 05:30 | |
| *** harshs_ is now known as harshs | 05:30 | |
| openstackgerrit | Ramakrishnan G proposed stackforge/proliantutils: Add logging for RIS and RIBCL modules https://review.openstack.org/219636 | 05:44 |
| *** harshs_ has joined #openstack-ironic | 05:47 | |
| *** boris-42 has joined #openstack-ironic | 05:47 | |
| *** Sukhdev has joined #openstack-ironic | 05:48 | |
| *** harshs has quit IRC | 05:49 | |
| *** harshs_ is now known as harshs | 05:49 | |
| *** dims_ has joined #openstack-ironic | 05:50 | |
| *** harshs has quit IRC | 05:53 | |
| *** dims_ has quit IRC | 05:55 | |
| *** jerrygb has joined #openstack-ironic | 06:08 | |
| *** jerrygb has quit IRC | 06:12 | |
| *** Sukhdev has quit IRC | 06:28 | |
| *** romainh has joined #openstack-ironic | 06:40 | |
| *** yog_ has joined #openstack-ironic | 06:52 | |
| *** karimb has joined #openstack-ironic | 06:54 | |
| openstackgerrit | Ramakrishnan G proposed stackforge/proliantutils: Add logging for RIS and RIBCL modules https://review.openstack.org/219636 | 07:02 |
| *** praneshp has quit IRC | 07:09 | |
| *** achanda has quit IRC | 07:14 | |
| *** karimb has quit IRC | 07:15 | |
| *** karimb has joined #openstack-ironic | 07:16 | |
| *** karimb has quit IRC | 07:27 | |
| openstackgerrit | Merged stackforge/proliantutils: Add logging for RIS and RIBCL modules https://review.openstack.org/219636 | 07:31 |
| *** karimb has joined #openstack-ironic | 07:32 | |
| *** ifarkas has joined #openstack-ironic | 07:32 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic-python-agent: Fixes config drive creation failure in UEFI boot mode https://review.openstack.org/215097 | 07:36 |
| *** dims_ has joined #openstack-ironic | 07:52 | |
| *** getvasanth has joined #openstack-ironic | 07:53 | |
| *** dims_ has quit IRC | 07:59 | |
| *** athomas has joined #openstack-ironic | 07:59 | |
| *** dtantsur|afk is now known as dtantsur | 07:59 | |
| dtantsur | Morning Ironic | 07:59 |
| getvasanth | Good Monring dtantsur :) | 08:03 |
| *** BobBall_AWOL is now known as BobBall | 08:04 | |
| *** derekh has joined #openstack-ironic | 08:13 | |
| *** lazy_prince has joined #openstack-ironic | 08:14 | |
| *** MattMan has joined #openstack-ironic | 08:14 | |
| *** lucas-dinner has joined #openstack-ironic | 08:19 | |
| *** lucasagomes has joined #openstack-ironic | 08:19 | |
| *** stendulker has joined #openstack-ironic | 08:20 | |
| *** priteau has joined #openstack-ironic | 08:22 | |
| *** romcheg has joined #openstack-ironic | 08:38 | |
| *** romcheg has quit IRC | 08:43 | |
| vdrok | morning ironic | 08:44 |
| dtantsur | vdrok, o/ | 08:44 |
| vdrok | morning dtantsur and getvasanth :) | 08:44 |
| getvasanth | vdrok: good morning | 08:45 |
| *** pelix has joined #openstack-ironic | 08:46 | |
| *** athomas has quit IRC | 08:52 | |
| lucasagomes | vdrok, dtantsur getvasanth good morning | 08:57 |
| *** athomas has joined #openstack-ironic | 08:58 | |
| dtantsur | lucasagomes, o/ | 08:59 |
| *** romcheg has joined #openstack-ironic | 09:05 | |
| vdrok | morning lucas-dinner | 09:06 |
| vdrok | or lucasagomes :) | 09:07 |
| dtantsur | unfortunately, it's not time for dinner :) | 09:07 |
| lucasagomes | oh hah not dinnering anymore | 09:07 |
| lucasagomes | lol | 09:07 |
| *** lucas-dinner has quit IRC | 09:07 | |
| lucasagomes | boom | 09:07 |
| lucasagomes | that's my brother | 09:07 |
| vdrok | :D | 09:08 |
| *** jerrygb has joined #openstack-ironic | 09:08 | |
| *** getvasanth has quit IRC | 09:10 | |
| openstackgerrit | Merged openstack/python-ironicclient: Added unit test cases for command-line shell https://review.openstack.org/213654 | 09:10 |
| *** dims_ has joined #openstack-ironic | 09:11 | |
| dtantsur | ifarkas, morning! could you review https://review.openstack.org/#/c/225173 please? it's required to close our last bp for 2.2.0 | 09:13 |
| *** jerrygb has quit IRC | 09:13 | |
| ifarkas | dtantsur, sure, will do! | 09:13 |
| *** e0ne has joined #openstack-ironic | 09:15 | |
| mrda | hey all - I've put a patch up for the Ironic review dashboard so that it includes all ironic projects. If you care about such things, here's the review: https://review.openstack.org/#/c/226592/ (just because you won't know about it unless I promote it in channel :) | 09:17 |
| rameshg87 | mrda: so will all these projects start getting listed in the same ironic inbox ? | 09:21 |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 09:21 |
| xek | dansmith, jroll, lucasagomes, ^ | 09:22 |
| lucasagomes | thanks | 09:22 |
| lucasagomes | mrda, a-ha thanks for that! | 09:24 |
| *** getvasanth has joined #openstack-ironic | 09:26 | |
| *** alexpilotti has joined #openstack-ironic | 09:32 | |
| *** vinbs has quit IRC | 09:33 | |
| sambetts | Morning everyone o/ | 09:36 |
| *** vinbs has joined #openstack-ironic | 09:37 | |
| dtantsur | morning sambetts! | 09:38 |
| lucasagomes | sambetts, hey there, morning | 09:44 |
| *** dims_ has quit IRC | 09:44 | |
| sambetts | o/ dtantsur, lucasagomes hows it going? | 09:45 |
| lucasagomes | sambetts, all is well! Few days away from holidays o/ | 09:46 |
| lucasagomes | sambetts, how are you? | 09:46 |
| dtantsur | pretty good, getting closer to the release | 09:46 |
| sambetts | lucasagomes: Nice! You going anywhere? | 09:47 |
| mrda | rameshg87: yeah, they will all be grouped together (specs handled differently though) | 09:47 |
| sambetts | lucasagomes: I'm pretty good thanks :) | 09:47 |
| lucasagomes | sambetts, yeah, I'm going to Iceland | 09:47 |
| lucasagomes | going to drive around there and be away from technology for a couple of days | 09:47 |
| rameshg87 | mrda: but will not that bundle all the project together | 09:47 |
| sambetts | dtantsur: Awesome :D | 09:48 |
| sambetts | lucasagomes: That sounds sweet! I've never been to iceland before | 09:48 |
| *** saripurigopi has joined #openstack-ironic | 09:48 | |
| dtantsur | lucasagomes, oh Iceland, sounds really interesting | 09:48 |
| rameshg87 | mrda: idk if it's only me, but personally I don't prefer that :) | 09:48 |
| mrda | rameshg87: Sure. It does this already for ironic server, client, and ipa | 09:48 |
| lucasagomes | sambetts, dtantsur yea, I've never been there before either. But man, there's so many things to do | 09:48 |
| mrda | rameshg87: this just adds ironic-lib, inspector, webclient and bifrost | 09:49 |
| lucasagomes | really 10 days is not enough, I don't know how many are required if you really want to see everything there | 09:49 |
| rameshg87 | mrda: ironic server, client, ipa are part of core bucket. but webclient and birforst seem too distant | 09:49 |
| *** e0ne has quit IRC | 09:49 | |
| sambetts | lucasagomes: you going to drive out to that huge volcano ? I think thats in Iceland | 09:49 |
| mrda | rameshg87: my motivation was to get ironic-lib in there, but the question remains for inspector, webclient and bifrost | 09:49 |
| rameshg87 | mrda: yeah ironic-lib too I agree | 09:50 |
| mrda | rameshg87: so review comments welcomed :) | 09:50 |
| saripurigopi | Hello Ironic | 09:50 |
| *** e0ne has joined #openstack-ironic | 09:50 | |
| lucasagomes | sambetts, I'm going to do the "route 1" there, it pass some vulcanos but not sure about the biggest there | 09:50 |
| rameshg87 | mrda: but inspector, webclient and bifrost in that list seems not good for me .. | 09:50 |
| * lucasagomes will check | 09:50 | |
| rameshg87 | mrda: anyway I will add comments | 09:50 |
| lucasagomes | sambetts, https://en.wikipedia.org/wiki/Route_1_%28Iceland%29 | 09:50 |
| mrda | rameshg87: thanks | 09:50 |
| rameshg87 | it's a personal choice anyway :) | 09:50 |
| lucasagomes | if possible I wanna do the golden circle thing as well (may take 1 full day) | 09:51 |
| sambetts | lucasagomes: Oh man that looks so cool! | 09:51 |
| lucasagomes | yeah, looking fwd to it! | 09:52 |
| * lucasagomes thinks we should have a midcycle in iceland one day :-) | 09:52 | |
| lucasagomes | imagine that! | 09:52 |
| sambetts | lucasagomes: Would certainly be different :-P | 09:53 |
| lucasagomes | yeah | 09:53 |
| dtantsur | :) | 09:55 |
| *** naohirot has quit IRC | 10:01 | |
| *** dims_ has joined #openstack-ironic | 10:09 | |
| *** ndipanov has quit IRC | 10:09 | |
| *** ndipanov has joined #openstack-ironic | 10:11 | |
| *** degorenko has quit IRC | 10:13 | |
| stendulker | lucasagomes: Good morning ! Did you get a chance to look at the hybrid image thing? | 10:18 |
| *** trown|outttypeww is now known as trown | 10:19 | |
| *** degorenko has joined #openstack-ironic | 10:19 | |
| lucasagomes | stendulker, oh #@!% I haven't, will look at it now | 10:20 |
| lucasagomes | I forgot it | 10:20 |
| stendulker | lucasagomes: oh ok.. | 10:20 |
| lucasagomes | stendulker, I will do some tests on it now, sorry for that | 10:20 |
| lucasagomes | it's the core os images right? | 10:21 |
| stendulker | lucasagomes: yes | 10:21 |
| *** chenglch has quit IRC | 10:21 | |
| stendulker | lucasagomes: It worked in uefi mode with gdisk | 10:22 |
| lucasagomes | stendulker, oh, so we can replace parted with gdisk I suppose? | 10:22 |
| stendulker | lucasagomes: OpenSUSE does not... | 10:22 |
| lucasagomes | we already have an interface for image partitioning | 10:22 |
| lucasagomes | ouch | 10:22 |
| stendulker | lucasagomes: Yes. I did that. | 10:22 |
| lucasagomes | ok... | 10:22 |
| stendulker | lucasagomes: Have raised review withgdisk implementation. But can't claim its universal solution. | 10:23 |
| lucasagomes | I see, yeah parted doesn't seem to be either | 10:23 |
| lucasagomes | that kinda sucks :-/ | 10:24 |
| stendulker | lucasagomes: Interestingly, openSUSE had worked with parted and does not work with gdisk | 10:24 |
| lucasagomes | it makes also me think about the propose we had before about having the configdrive in a loop device | 10:24 |
| lucasagomes | I still think it's a valid thing | 10:24 |
| stendulker | coreos works with gdisk but not with parted | 10:24 |
| lucasagomes | tho not all people may agree with me | 10:24 |
| lucasagomes | fwiw https://review.openstack.org/#/c/173142/ | 10:24 |
| stendulker | lucasagomes: looking into spec... | 10:25 |
| lucasagomes | stendulker, yeah most cores didn't agree with it | 10:26 |
| * trown is super jealous | 10:26 | |
| lucasagomes | I don't like the idea of injecting/editing stuff into user's images. But IMHO it's a valid case | 10:26 |
| trown | of the iceland trip that is...just realized I was lost in the scrollback | 10:26 |
| lucasagomes | as optional | 10:26 |
| lucasagomes | trown, heh yeah, but don't be man. You should plan one trip to there as well | 10:27 |
| trown | lucasagomes: there is no such thing as a vacation with kids :P | 10:27 |
| stendulker | lucasagomes: your proposal looks good... it would solve all these issues where different distros would have used their own partitioning tools | 10:27 |
| trown | at least not little ones | 10:28 |
| stendulker | lucasagomes: which may not be compatible with the deploy ramdisk distro partition tools | 10:28 |
| lucasagomes | stendulker, it's not actually mine | 10:28 |
| lucasagomes | but yeah I think it's a fair solution for certain situations where partition is not an option | 10:28 |
| stendulker | lucasagomes: oh yes.. | 10:28 |
| lucasagomes | which can exist, like this one you're facing right now | 10:28 |
| lucasagomes | or deploying an image with MBR with 4 primary partitions already created | 10:29 |
| lucasagomes | such things are valid situations where creating another partition is either impossible or not ideal | 10:29 |
| lucasagomes | IMHO | 10:29 |
| stendulker | lucasagomes: May be it should be an option | 10:29 |
| stendulker | lucasagomes: For images where partitioning fails, user can opt for loop device | 10:29 |
| lucasagomes | trown, oh, yeah you gotta wait them grow a little more | 10:29 |
| lucasagomes | trown, how old are they? | 10:30 |
| lucasagomes | stendulker, yeah, we can bring it up again with the rest of the people | 10:30 |
| lucasagomes | if you think that's a valid solution | 10:30 |
| lucasagomes | JayF, devananda ^ | 10:30 |
| trown | lucasagomes: one is 1.5 and there is not another ... yet | 10:31 |
| stendulker | lucasagomes: loop device is a valid option to get around such situations. Moreover with UEFI support increasing we may see more hybrid images | 10:31 |
| lucasagomes | stendulker, yup, tho the UEFI case surprises me a bit | 10:32 |
| stendulker | lucasagomes: HYbrid images makes sense for operators as well wherein they need not main tain uefi only or bios only images | 10:32 |
| lucasagomes | if it was only UEFI with GPT I don't see much problems there | 10:32 |
| lucasagomes | the hybrid seems complicated | 10:32 |
| lucasagomes | in theory GPT is hybrid but reality is diff AFAICT | 10:32 |
| stendulker | lucasagomes: UEFI only images works fine with parted/gdisk.. tested using ubuntu uefi whole disk image | 10:32 |
| lucasagomes | some BIOS simple doesn't work with the private MBR of gpt | 10:32 |
| lucasagomes | but hey... | 10:32 |
| lucasagomes | trown, nice one! | 10:33 |
| *** trown is now known as trown|relocating | 10:34 | |
| lucasagomes | stendulker, btw this is a good read about it http://www.rodsbooks.com/gdisk/bios.html | 10:34 |
| stendulker | lucasagomes: sure | 10:36 |
| lucasagomes | stendulker, so yeah I think I would suggest you to comment on that spec let the author knows you are also good with the change | 10:41 |
| lucasagomes | and it would solve some of ur problems | 10:41 |
| lucasagomes | and bring up the discussion again with the rest of the community | 10:42 |
| *** vinbs has quit IRC | 10:42 | |
| *** vinbs has joined #openstack-ironic | 10:44 | |
| *** Nisha has joined #openstack-ironic | 10:46 | |
| *** lucasagomes has quit IRC | 10:48 | |
| openstackgerrit | Merged openstack/ironic-inspector: Add missing plugins for introspection rules https://review.openstack.org/225173 | 10:51 |
| stendulker | lucasagomes: sure, will do | 10:53 |
| *** lucasagomes has joined #openstack-ironic | 10:53 | |
| stendulker | lucasagomes: thank you | 10:54 |
| lucasagomes | np | 10:54 |
| * lucasagomes connection just dropped :-( | 10:54 | |
| *** lucasagomes_ has joined #openstack-ironic | 11:00 | |
| *** lucasagomes_ has quit IRC | 11:01 | |
| openstackgerrit | Merged openstack/ironic-inspector: Allow empty lookup attributes if node_not_found_hook is provided https://review.openstack.org/225668 | 11:02 |
| dtantsur | merge, merge, MERGE!! | 11:02 |
| dtantsur | :D | 11:02 |
| *** ukalifon1 has joined #openstack-ironic | 11:04 | |
| dtantsur | ifarkas, sambetts, looking at https://launchpad.net/ironic-inspector/+milestone/2.2.0 I think we have good chances of finishing today, except for rootwrap patch. I think I'll update it myself, if I don't hear from Yuiko in a couple of hours, wdyt? | 11:05 |
| ifarkas | dtantsur, sounds good | 11:05 |
| sambetts | dtantsur: Yup sounds good to me :D | 11:06 |
| dtantsur | agreed :) I'd really like to see it all green by today evening | 11:07 |
| sambetts | dtantsur: What do you think about getting the alembic patch in for 2.2.0? | 11:07 |
| sambetts | I think it'll make people upgrading from 2.1.0 to 2.2.0 easier, because they'll need to add the rules tables, or will current model system just add rules anyway? | 11:08 |
| dtantsur | sambetts, I think we're on track with it, just need 2nd +2. Some people will probably hate us for complicating upgrade just before Liberty final, but I guess it's worth it | 11:09 |
| dtantsur | sambetts, with the current model, it will create missing table. However, as most people will anyway have to upgrade from discoverd, which is anyway painful.... | 11:09 |
| sambetts | dtantsur: ah yes :/ I guess all peoples tooling will be written for the old binaries | 11:10 |
| dtantsur | I wonder if we need some kind of upgrade docs to cover it. At least cover discoverd-kilo to inspector-liberty | 11:11 |
| *** lucasagomes is now known as lucas-hungry | 11:15 | |
| sambetts | dtantsur: hmmm, its not dramically different in the way its run is it, so its most likely just ironic-discoverd is now ironic-inspector right? I guess theres the deprecated configs and stuff too :/ | 11:15 |
| dtantsur | sambetts, and client got split away, and plugin interface was changed, and deprecated API call was dropped | 11:16 |
| sambetts | dtantsur: XD I lose track of how many changes we made | 11:17 |
| * dtantsur too | 11:17 | |
| sambetts | Can we generate a change log between 1.1.0 and 2.2.0 from launchpad?? hmmmm ... | 11:18 |
| *** baoli has joined #openstack-ironic | 11:23 | |
| *** ramineni_ has joined #openstack-ironic | 11:24 | |
| *** stendulker has quit IRC | 11:25 | |
| *** shakamunyi has quit IRC | 11:25 | |
| *** ramineni_ has left #openstack-ironic | 11:25 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Use rootwrap to execute iptables instead of requiring root https://review.openstack.org/223930 | 11:28 |
| * rameshg87 goes home | 11:29 | |
| *** rameshg87 has quit IRC | 11:29 | |
| dtantsur | sambetts, ifarkas, the last patch is up ^^ | 11:29 |
| sambetts | Nice! | 11:29 |
| dtantsur | trown|relocating, have a look at that change ^^ please, it might be important for packaging | 11:29 |
| ifarkas | dtantsur, excellent! I will do reviews later on | 11:30 |
| dtantsur | thanks | 11:30 |
| *** saripurigopi has quit IRC | 11:32 | |
| *** baoli_ has joined #openstack-ironic | 11:33 | |
| *** jerrygb has joined #openstack-ironic | 11:35 | |
| *** baoli has quit IRC | 11:36 | |
| *** jerrygb has quit IRC | 11:39 | |
| *** jerrygb has joined #openstack-ironic | 11:48 | |
| *** getvasanth has quit IRC | 11:49 | |
| *** krtaylor has quit IRC | 12:01 | |
| *** achanda has joined #openstack-ironic | 12:03 | |
| *** ukalifon1 has quit IRC | 12:04 | |
| *** trown|relocating is now known as trown | 12:12 | |
| *** krtaylor has joined #openstack-ironic | 12:13 | |
| *** e0ne has quit IRC | 12:13 | |
| *** achanda has quit IRC | 12:18 | |
| *** lucas-hungry is now known as lucasagomes | 12:19 | |
| *** ukalifon has joined #openstack-ironic | 12:22 | |
| *** ionutbalutoiu has joined #openstack-ironic | 12:23 | |
| *** zhenguo has quit IRC | 12:24 | |
| *** caiobo has quit IRC | 12:29 | |
| ionutbalutoiu | Hello, guys! I'm doing some work with Ironic, focusing on the deployment of the service in standalone mode. I have a question: How does Ironic scale out ? Let's say you have 100 nodes each having their own services (ironic-api, ironic-conductor, dhcp server and tftp server). All nodes communicate with a central database. Are there any concerns I should be are of ? | 12:29 |
| dtantsur | ionutbalutoiu, hi! you probably want to wait until jroll and JayF come online, they're operating ironic on scale | 12:35 |
| ionutbalutoiu | Sure. | 12:36 |
| dtantsur | ionutbalutoiu, I personally didn't quite get the question: do you want to have one ironic conductor per one node it manages? | 12:36 |
| ionutbalutoiu | I was planning to do that. I'm not sure how does ironic-conductor scales out. | 12:36 |
| dtantsur | ionutbalutoiu, highly depends on a driver you're planning to use. even with the simplest pxe_ipmitool driver, one conductor can manage dozens of nodes | 12:37 |
| dtantsur | ionutbalutoiu, using agent_ipmitool will reduce load on conductor (somewhat increasing network traffic). | 12:38 |
| *** uggla_ has joined #openstack-ironic | 12:38 | |
| uggla_ | Hello | 12:38 |
| dtantsur | ionutbalutoiu, if you have hardware supporting virtual media (only HP iLO and Fujitsu IIRC now), that will allow you to scale even better | 12:38 |
| dtantsur | uggla_, o/ | 12:39 |
| dtantsur | ionutbalutoiu, but one conductor per one baremetal node is not what you probably want in any case | 12:39 |
| lucasagomes | ionutbalutoiu, scaling out the conductor should be simple, Ironic will already distribute the workload across the conductors dynamically | 12:40 |
| lucasagomes | ionutbalutoiu, for api can put it behind a HA proxy | 12:40 |
| dtantsur | ionutbalutoiu, also if you don't use virtual media, use iPXE, not PXE | 12:40 |
| lucasagomes | ionutbalutoiu, and yeah, as dtantsur says. One conductor is capable of managing many nodes at the same time with different drivers and all | 12:41 |
| ionutbalutoiu | lucasagomes, tantsur, if the node with the single conductor fails, I will probably have problems. As far as I understand, I can have multiple conductors managing the same pool of nodes and Ironic distributes the workload. lucasagomes, do you know how Ironic does this distribution ? | 12:44 |
| lucasagomes | ionutbalutoiu, yes we use a distributed hashing algorithm. So you can have multiple conductors able to manage the same types of nodes (based on the driver they use) | 12:45 |
| lucasagomes | but there's always 1 conductor managing each node and it's dynamic | 12:45 |
| lucasagomes | if that conductor dies, ironic will rebalance the ring and another conductor able to manage nodes with that driver will take over that node | 12:45 |
| lucasagomes | conductors can join and leave the cluster and Ironic will dynamically balance ir | 12:46 |
| lucasagomes | it* | 12:46 |
| ionutbalutoiu | When a conductor comes up, it is registered in the database and Ironic is aware of it. | 12:48 |
| lucasagomes | ionutbalutoiu, the module is this one if you are interested https://github.com/openstack/ironic/blob/297f14469d23dd793de625251366380cc648fdb2/ironic/common/hash_ring.py | 12:48 |
| lucasagomes | ionutbalutoiu, yes, a conductor comes up and register itself in the database along with the drivers it have enabled | 12:49 |
| lucasagomes | so when a request comes to the API to a specific node the ironic-api service will know how to route that message to a conductor that is able to manage that node | 12:49 |
| lucasagomes | it uses the node uuid as a hash | 12:50 |
| lucasagomes | and there's a hash ring per driver type | 12:50 |
| *** rloo has joined #openstack-ironic | 12:51 | |
| *** albertoffb has joined #openstack-ironic | 12:53 | |
| ionutbalutoiu | That's pretty cool. So, I shouldn't be worried about high availability. From what I understand, Ironic already takes care of it in a scenario with multiple ironic-conductors and a central database. | 12:54 |
| *** Nisha has quit IRC | 12:54 | |
| lucasagomes | ionutbalutoiu, yes, HA for the ironic conductor is embedded in it | 12:54 |
| lucasagomes | ionutbalutoiu, you should care about the API, but that's pretty standard | 12:54 |
| lucasagomes | put it behind a HA proxy or something | 12:54 |
| ionutbalutoiu | Yep. That was clear already. Thank-you guys. :) | 12:54 |
| lucasagomes | yw | 12:55 |
| lucasagomes | ionutbalutoiu, there's a couple of videos on youtube from previous summits where devananda (and maybe others) have talked about how it works | 12:55 |
| lucasagomes | you may want to check | 12:55 |
| lucasagomes | ironic was designed that way so this feature is been around since the beginning | 12:56 |
| lucasagomes | ionutbalutoiu, here https://www.youtube.com/watch?v=vjGPuBfZ87A#t=1031 | 12:57 |
| ionutbalutoiu | lucasagomes, Awesome! I'll check them out for sure. Thanks | 12:57 |
| lucasagomes | cool np | 12:57 |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 12:58 |
| ionutbalutoiu | I do have another question if you don't mind guys. | 12:58 |
| lucasagomes | ionutbalutoiu, ask away | 12:58 |
| *** rbudden has joined #openstack-ironic | 12:58 | |
| ionutbalutoiu | So, we have only neutron as dhcp_provider for Ironic, right? But if using I'm using it in standalone, I have to take care of the dhcp reservation and such. | 12:59 |
| lucasagomes | ionutbalutoiu, yes, dhcp providers (as we call it) is pluggable. But currently we only have 2 plugins in tree "neutron" and "none" (which is no-op) | 13:00 |
| ionutbalutoiu | I'm trying to write a dhcp_provider using dhcpd from Linux to be used by Ironic in standalone. | 13:00 |
| lucasagomes | ionutbalutoiu, so if you use "none" you may be looking at using a static PXE configuration file that will boot the deploy images onto the node | 13:01 |
| lucasagomes | but you can also implement an interface for another dhcp server you may have | 13:01 |
| lucasagomes | ionutbalutoiu, right | 13:01 |
| lucasagomes | ionutbalutoiu, https://github.com/openstack/ironic/blob/master/ironic/dhcp/base.py | 13:01 |
| lucasagomes | I'm not sure how mature is that interface, because we only have 1 implementation right now | 13:01 |
| lucasagomes | so it would be good to have more :-) | 13:02 |
| ionutbalutoiu | Cool, life's good. But I don't get any trigger for any of the methods when a node is deleted. | 13:02 |
| ionutbalutoiu | Everytime I create an Ironic node, I make an a dhcp reservation for it and reload the dhcpd service. | 13:03 |
| * lucasagomes checks that bit | 13:04 | |
| ionutbalutoiu | I would like to delete the dhcp reservation when the node is deleted and reload the dhcpd service again. But unfortunately I don't get any trigger on delete. | 13:05 |
| trown | dtantsur: no more packaging changes after the alembic patch for a bit ok :P | 13:05 |
| dtantsur | trown, rootwrap change is pretty trivial: adds one more executable and one more configuration file. after it we no longer need to run inspector as root, which is a good thing IMO | 13:07 |
| trown | dtantsur: ya rootwrap one is not so bad, alembic requires more puppet though | 13:08 |
| trown | but both are good changes | 13:08 |
| dtantsur | trown, https://launchpad.net/ironic-inspector/+milestone/2.2.0 is all we want to land before 2.2. doesn't look like we have anything else concerning | 13:08 |
| lucasagomes | ionutbalutoiu, that sucks :-( yeah I skimmed a bit here it seems that we just call update_dhcp() in prepare_instance() or prepare_ramdisk() | 13:09 |
| lucasagomes | ionutbalutoiu, that may because nova will delete the neutron network for us (when used with neutron) | 13:09 |
| lucasagomes | and none well... it doesn't do anything | 13:09 |
| *** [1]cdearborn has joined #openstack-ironic | 13:09 | |
| ionutbalutoiu | yep, neutron takes care of deleting the dhcp reservation when an instance is deleted | 13:09 |
| lucasagomes | ionutbalutoiu, perhaps that's something that needs to be added to that interface, at the clean_up() methods we should call some sort of delete | 13:09 |
| lucasagomes | ionutbalutoiu, yeah... see, that's why we need more dhcp implementations | 13:10 |
| ionutbalutoiu | lucasagomes, yep I agree, that would be good. | 13:10 |
| lucasagomes | ionutbalutoiu, mind opening a bug about it please? | 13:10 |
| trown | dtantsur: cool, pretty close on all of those | 13:10 |
| ionutbalutoiu | lucasagomes, not at all. | 13:10 |
| lucasagomes | ionutbalutoiu, thanks much! | 13:10 |
| lucasagomes | ionutbalutoiu, the fix should be straight forward, the same way we have prepare_{instance, ramdisk}() methods for the boot interfaces such as pxe | 13:11 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic-specs: Using multicast for image transfer https://review.openstack.org/225570 | 13:11 |
| lucasagomes | we also have the clean_up_{instance, ramdisk}() methods where the delete hooks should be called | 13:11 |
| lucasagomes | they both get the "task" object as parameter which gives you access to the whole node and all it's ports objects | 13:12 |
| ionutbalutoiu | lucasagomes, I agree. At the moment, I have the implementation so when a node is created I do the clean_up(). Hackish, but does the job for the moment. | 13:13 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic-specs: Using multicast for image transfer https://review.openstack.org/225570 | 13:14 |
| ionutbalutoiu | lucasagomes, still I will have the dhcp reservations until a node is created, no matter how many nodes are deleted meanwhile. | 13:14 |
| lucasagomes | ionutbalutoiu, cool! That would be great if you could polish it a bit and submit upstream :-) | 13:14 |
| * lucasagomes will gladly review it | 13:14 | |
| ionutbalutoiu | lucasagomes, Sure, I need to finish the implementation when the clean_up() method is merged. :) | 13:16 |
| lucasagomes | ++ thanks again | 13:16 |
| *** thiagop has joined #openstack-ironic | 13:16 | |
| ionutbalutoiu | lucasagomes, I'm going to open the bug. Thank-you for all your help. | 13:16 |
| thiagop | Morning folks | 13:17 |
| lucasagomes | ionutbalutoiu, yvw, cheers for looking into it | 13:17 |
| lucasagomes | thiagop, morning | 13:17 |
| thiagop | lucasagomes: o/ | 13:18 |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 13:18 |
| *** dims_ has quit IRC | 13:21 | |
| *** dims_ has joined #openstack-ironic | 13:21 | |
| *** alexpilotti has quit IRC | 13:22 | |
| openstackgerrit | Merged openstack/ironic-inspector: Use rootwrap to execute iptables instead of requiring root https://review.openstack.org/223930 | 13:22 |
| dtantsur | \o/ 3 patches to go | 13:23 |
| *** olaph has quit IRC | 13:28 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/python-ironicclient: Allow 'abort' verb for node-set-provision-state https://review.openstack.org/202184 | 13:28 |
| sambetts | dtantsur: w00p! | 13:30 |
| *** vinbs has quit IRC | 13:31 | |
| *** bradjones has quit IRC | 13:35 | |
| ifarkas | dtantsur, 3? not just 2? | 13:38 |
| dtantsur | ifarkas, now only 2 left: https://review.openstack.org/225658 and https://review.openstack.org/223665 | 13:39 |
| *** ukalifon has quit IRC | 13:39 | |
| dtantsur | :) | 13:39 |
| ifarkas | dtantsur, same here :-) | 13:39 |
| ionutbalutoiu | lucasagomes, Bug submitted: https://bugs.launchpad.net/ironic/+bug/1498923 | 13:39 |
| openstack | Launchpad bug 1498923 in Ironic "No clean up method for the dhcp provider base class" [Undecided,New] | 13:39 |
| lucasagomes | ionutbalutoiu, thanks! | 13:40 |
| *** olaph has joined #openstack-ironic | 13:41 | |
| *** lsmola has quit IRC | 13:41 | |
| dtantsur | sambetts, ooops! alembic patch needs rebasing | 13:42 |
| sambetts | dtantsur: :-P i can guess its because of the setup.cfg changes made my rootwrap | 13:42 |
| dtantsur | yeah, I suppose so | 13:42 |
| sambetts | dtantsur: just toxing then I'll git review | 13:45 |
| dtantsur | thanks | 13:45 |
| openstackgerrit | Sam Betts proposed openstack/ironic-inspector: Add alembic migrations for the inspector database https://review.openstack.org/223665 | 13:46 |
| dtantsur | awesome! I'll wait for Jenkins before re +2 | 13:46 |
| sambetts | :D | 13:49 |
| *** lsmola has joined #openstack-ironic | 13:50 | |
| *** e0ne has joined #openstack-ironic | 13:51 | |
| *** romcheg has quit IRC | 13:55 | |
| liliars | hi ionutbalutoiu, lucasagomes :) so if I understood the situation correctly, I may have had a similar issue, but not using ironic standalone; sometimes when deleting an instance in error state (if I recall correctly), the neutron-port wouldn't be deleted as expected, and in a next deployment I would get "port already in use" or something | 14:00 |
| liliars | had to delete the neutron-port manually so I could still use the same node-port for deployment; it's not a regular behaviour so I don't know if it was any network failure from my side (?) or a bug on the code (nova or ironic); not sure if other folks have experienced this too | 14:00 |
| lucasagomes | liliars, oh, and you're using nova as well? | 14:01 |
| sambetts | lucasagomes: I think one of my colleages saw that too at one point recently | 14:02 |
| lucasagomes | I haven't seem that particularly, but it wouldn't surprise me if that happens when deleting a node in error state | 14:03 |
| *** jerrygb has quit IRC | 14:03 | |
| *** afaranha has quit IRC | 14:03 | |
| *** romcheg has joined #openstack-ironic | 14:03 | |
| liliars | lucasagomes, yes using nova | 14:04 |
| lucasagomes | sambetts, right, and where did it fail? At deployment time ("deploy error") or instance tear down (I think it's just "error") ? | 14:04 |
| lucasagomes | liliars, ^ | 14:04 |
| lucasagomes | you remember? | 14:04 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic-inspector: Updated from global requirements https://review.openstack.org/226805 | 14:05 |
| openstackgerrit | Merged openstack/python-ironicclient: Set a default endpoint value of None https://review.openstack.org/226475 | 14:06 |
| liliars | lucasagomes, deployment time; when I tried a new deployment with the same node/node-port then I got the "port already in use" error; and when I checked the neutron ports list, the port was really still there | 14:06 |
| *** jerrygb has joined #openstack-ironic | 14:07 | |
| lucasagomes | liliars, right, yeah odd looking at the nova code it seems to call _unprovision() for nodes in error state https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L909-L910 | 14:08 |
| lucasagomes | liliars, sambetts yeah I will have to investigate more see if I can reproduce that | 14:08 |
| lucasagomes | I dunno off the top of my head | 14:08 |
| lucasagomes | but it's good to know it's possible | 14:08 |
| * lucasagomes would be good to have a bug about it too, but will need some more info for that | 14:08 | |
| lucasagomes | thinks it would be good* | 14:09 |
| xek | dansmith, jroll, lucasagomes, jenkins just finished on https://review.openstack.org/224079 | 14:10 |
| sambetts | Yeah, I didn't get much more info, other than he asked me if I had ever seen it happen before, I put it down to his env (because he was having other weird issues anyway), so I didn't dig much deeper | 14:10 |
| lucasagomes | yeah that's fair enuff, and good that at least there's a workaround for that (deleting the port) | 14:12 |
| openstackgerrit | Merged openstack/ironic-inspector: Convert eDeploy data so that rules can process it https://review.openstack.org/225168 | 14:13 |
| sambetts | lucasagomes: Yeah, its a fairly simple fix :) | 14:13 |
| *** mtanino has joined #openstack-ironic | 14:13 | |
| rloo | hi iroKnickers | 14:13 |
| rloo | anyone know if yuiko is around these days? | 14:14 |
| rloo | dtantsur: ^^ ? | 14:14 |
| dansmith | xek: I guess I never saw a decision from jroll | 14:15 |
| liliars | lucasagomes, yeah I'm sure about the neutron-port not being deleted at some point, but not really sure about the exact scenario; I'm gonna try to reproduce it too (: thanks | 14:15 |
| thiagop | morning rloo | 14:16 |
| *** romcheg has quit IRC | 14:16 | |
| liliars | morning rloo | 14:16 |
| dtantsur | rloo, hi! haven't seen you for some time, and she wasn't reviewing code as well | 14:16 |
| rloo | hi thiagop, liliars | 14:16 |
| *** afaranha has joined #openstack-ironic | 14:16 | |
| rloo | dtantsur: am wondering if she's on vaca. ok, there's a bug/patch she's working on that we'd like in, so maybe i'll finish it. | 14:17 |
| sambetts | rloo: Last thursday was the last time I saw her on here | 14:17 |
| rloo | sambetts: thx. | 14:18 |
| TheJulia | devananda: Woud you be awesome and add me to the bifrost-release group? | 14:18 |
| TheJulia | s/Woud/Would | 14:18 |
| *** harshs has joined #openstack-ironic | 14:18 | |
| lucasagomes | liliars, thank you for looking into it | 14:19 |
| lucasagomes | rloo, TheJulia morning | 14:19 |
| rloo | hiya lucasagomes, morning TheJulia. (Could deva be more awesome than he already is?) | 14:20 |
| TheJulia | good morning lucasagomes, good morning rloo | 14:20 |
| sambetts | o/ TheJulia | 14:21 |
| TheJulia | (That is quite a question.... I suspect there is always room for increased awesomeness, then again there are certain curses that go along with being awesome.....) | 14:21 |
| thiagop | xek: wow, I'm really trying to review this indirection_api thing, but it's complex | 14:22 |
| *** romcheg has joined #openstack-ironic | 14:25 | |
| jroll | ionutbalutoiu: some scale notes: at rackspace onmetal we run around 250 nodes per conductor, if that's a helpful number. we also manage dhcp/tftp outside of ironic with static configs, and serve the ramdisk over http rather than glance | 14:26 |
| jroll | morning everyone | 14:26 |
| thiagop | morning jroll | 14:27 |
| dtantsur | jroll, morning | 14:27 |
| rloo | TheJulia: :D | 14:27 |
| ionutbalutoiu | jroll, morning. Cool stuff. :) | 14:27 |
| xek | jroll, afternoon ;) | 14:27 |
| rloo | morning jroll | 14:27 |
| jroll | so active today \o/ | 14:28 |
| jroll | xek: this channel is UGT :D http://www.total-knowledge.com/~ilya/mips/ugt.html | 14:28 |
| sambetts | o/ jroll | 14:28 |
| lucasagomes | jroll, morning | 14:28 |
| jroll | dansmith: I think I am going to land that, would love a quick review from you if you have bandwidth | 14:29 |
| * jroll also reviews | 14:29 | |
| xek | jroll, :D | 14:30 |
| rloo | jroll: question about the 'topic' parameter. | 14:30 |
| dansmith | jroll: okay cool | 14:30 |
| jroll | rloo: mhmmm? | 14:30 |
| rloo | jroll: wrt xek's patch. I'm fine deleting that, or not. but if not, the comment doesn't say much to me. | 14:30 |
| rloo | jroll: topic isn't used. so what? should it be used, should it be deleted? | 14:31 |
| lucasagomes | topic for RPC? that's how we route messages from API to a specific conductor | 14:32 |
| rloo | jroll: 'FIXME(xek): The topic argument is actually never used.'. What's the fix? | 14:32 |
| lucasagomes | why it's not being used? | 14:32 |
| * lucasagomes checks the patch | 14:32 | |
| rloo | lucasagomes: https://review.openstack.org/#/c/224079/17/ironic/conductor/rpcapi.py | 14:32 |
| jroll | lucasagomes: because these don't need to be a specific conductor | 14:32 |
| lucasagomes | jroll, so it will default to the generic topic? | 14:33 |
| jroll | lucasagomes: yes, and any conductor will pick it up AIUI | 14:33 |
| rloo | lucasagomes: right. | 14:33 |
| lucasagomes | (the one not using the cond. hostname) | 14:33 |
| lucasagomes | ok | 14:33 |
| rloo | lucasagomes, jroll: the FIXME comment doesn't make sense to me. | 14:33 |
| jroll | what I'm wondering is if some conductors are on version X, and some are on Y, where Y > X, if there's a case to be had there | 14:33 |
| jroll | I don't think there is | 14:34 |
| rloo | lucasagomes, jroll: it may not be used 'now' cuz it is None, but it IS used, just that it is None. If it wasn't None, it'd be used. | 14:34 |
| lucasagomes | cctxt = self.client.prepare(topic=topic or self.topic, version='1.31') | 14:34 |
| lucasagomes | seems it's being used? | 14:34 |
| lucasagomes | at least within that function scope | 14:34 |
| jroll | right, so the FIXME is it could be prepare(self.topic...) | 14:34 |
| rloo | lucasagomes: exactly. that's why the comment doesn't make sense. | 14:34 |
| jroll | it's not used by any callers | 14:34 |
| rloo | jroll: that's a dangerous thing to say cuz code changes and a caller could use it. | 14:35 |
| lucasagomes | well that's outside that execution body | 14:35 |
| lucasagomes | in that method POV it's used | 14:35 |
| rloo | if we don't want to use it, as dan sez, we should delete it | 14:35 |
| jroll | rloo: right, it's more that a caller shouldn't be using it | 14:36 |
| rloo | if we think we want to use it, we should leave it and update the comment. | 14:36 |
| rloo | jroll: i don't think we can say 'we have this argument but please don't use it' | 14:36 |
| jroll | yeah, it's silly | 14:36 |
| jroll | here's what I'm wondering: | 14:36 |
| lucasagomes | rloo, bad interfaces | 14:36 |
| jroll | if API makes a call to a specific conductor for a node | 14:36 |
| jroll | and then needs to translate that response | 14:37 |
| lucasagomes | but, if we have to pass topic and we don't want it to be used | 14:37 |
| jroll | should it go to any conductor or that specific one? | 14:37 |
| jroll | I think any conductor is fine in all cases | 14:37 |
| rloo | jroll: why does it matter which conductor? | 14:37 |
| lucasagomes | we perhaps could put a fix me there and error out if "topic is not None" | 14:37 |
| jroll | I'm wondering if there's a case where it's not fine | 14:37 |
| jroll | where conductors have different versions | 14:37 |
| jroll | but maybe we just say different versions of conductors isn't supported | 14:37 |
| lucasagomes | jroll, hmm yeah, well when an upgrade is going on | 14:37 |
| lucasagomes | you may have different versions of conductors at the same time | 14:38 |
| jroll | right | 14:38 |
| rloo | jroll: that's an interesting point. | 14:39 |
| lucasagomes | also if topic should not be used, perhaps it should be indicated in the docstring | 14:39 |
| lucasagomes | "not used by this method" | 14:39 |
| jroll | I guess it's more like, I don't want to change this method signature later if we decide we need it | 14:39 |
| jroll | but I'm 99% sure we don't need it | 14:39 |
| devananda | g'morning, all | 14:39 |
| rloo | my suggestion then is to leave topic in (cuz it doesn't hurt) but change the comment | 14:39 |
| sambetts | Morning devananda | 14:40 |
| rloo | mornin' devananda | 14:40 |
| thiagop | good morning devananda | 14:40 |
| xek | rloo, maybe to something like "FIXME: currently no caller of this method is using the topic argument" | 14:40 |
| TheJulia | good morning devananda | 14:40 |
| lucasagomes | rloo, ++ | 14:41 |
| lucasagomes | xek, I think it's fine to just delete that comment to be honest | 14:41 |
| lucasagomes | devananda, morning | 14:41 |
| rloo | jroll, lucasagomes: :param topic: RPC topic. Defaults to self.topic. In most cases, the default should be used since any conductor should be able to handle this. | 14:41 |
| jroll | lucasagomes: I agree, no need for a FIXME | 14:41 |
| rloo | xek ^^ | 14:41 |
| jroll | rloo++ | 14:41 |
| xek | lucasagomes, I already run tox on the code without the topic, so it's fine either way | 14:42 |
| lucasagomes | xek, right, I mean it has a use | 14:42 |
| lucasagomes | even if the current callers are not using it | 14:42 |
| rloo | xek: the point is that the FIXME comment is confusing cuz there isn't anything to fix. | 14:42 |
| lucasagomes | so the FIXME makes no sense there | 14:42 |
| thiagop | xek: If you'll send another patch, I'd vote for better exception messages | 14:43 |
| lucasagomes | but code wise it seems good | 14:43 |
| devananda | jroll: what if different conductors have different drivers loaded? | 14:43 |
| xek | lucasagomes, rloo, jroll, ok, I'm adding the above param description and deleting the fixme comment | 14:43 |
| thiagop | xek: let me know if you'll send another one | 14:43 |
| *** r-daneel has joined #openstack-ironic | 14:43 | |
| lucasagomes | xek, thanks | 14:43 |
| jroll | devananda: that shouldn't matter for object version translation, right? | 14:43 |
| rloo | thx xek! | 14:43 |
| *** aarefiev has quit IRC | 14:43 | |
| jroll | devananda: maybe later, if we object-ify driver things, idk | 14:44 |
| devananda | jroll: right | 14:44 |
| devananda | the subresources like driver-info, driver-internal-info | 14:44 |
| devananda | architecturally speaking, I think it could matter, though it may not today | 14:44 |
| jroll | yeah, so I think leaving it is a good idea | 14:44 |
| jroll | our conductor isn't just an orchestration thing like nova's | 14:45 |
| devananda | right | 14:45 |
| dansmith | jroll: xek: why are we still adding this method with a topic argument? | 14:45 |
| devananda | nova's conductor is more of a database proxy | 14:45 |
| devananda | whereas our API service accesses the DB directly when it can | 14:45 |
| jroll | dansmith: because conductors can be tied to a given node, and may have different drivers loaded | 14:46 |
| dansmith | devananda: our api accesses the db directly as well, and our conductor does more than db proxy, it also runs longer tasks | 14:46 |
| jroll | dansmith: no topic argument is probably ok now but I don't want to change the method signature later if we need it | 14:46 |
| dansmith | jroll: but we never pass anything but None | 14:46 |
| xek | dansmith, It would be actually pretty cool to support the scenario when conductors run with different versions | 14:46 |
| jroll | dansmith: *today*, yes | 14:46 |
| devananda | dansmith: in some situations, there may be only one conductor which can evaluate certain methods for a node, such as validating the driver_info | 14:46 |
| dtantsur | morning, devananda | 14:46 |
| dansmith | jroll: this isn't on-the-wire api, so I'm not sure why we wouldn't just add it once it's required, if that day comes, but okay | 14:47 |
| jroll | dansmith: true | 14:48 |
| rloo | dansmith: there are pros/cons to both approaches. One other thing is that all the other RPC calls (I think) have that topic argument, and if we remove it, i can just see someone submitting a patch to add it in. | 14:48 |
| *** harshs_ has joined #openstack-ironic | 14:48 | |
| dansmith | rloo: well, that patch would be silly, in the absence of some use of it :) | 14:48 |
| dansmith | the other thing is, | 14:48 |
| dansmith | I'm not really sure how you're going to go from an arbitrary object to the topic that you need to send the request to | 14:49 |
| rloo | dansmith: yeah. but folks have a short memory. whatever we do, we need to comment. | 14:49 |
| dansmith | so I think there's likely more machinery around this required to make it work anyway | 14:49 |
| *** aarefiev has joined #openstack-ironic | 14:49 | |
| jroll | dansmith: yeah that was my next question :) | 14:49 |
| dansmith | rloo: I meant, I would hope that nobody would +2 a patch to add a thing that is not used :) | 14:49 |
| rloo | dansmith: so would I but really, have you never seen a +2 for something that maybe shouldn't? | 14:49 |
| *** harshs has quit IRC | 14:50 | |
| *** harshs_ is now known as harshs | 14:50 | |
| rloo | dansmith: besides, why add all that extra overhead of someone submitting a patch and reviewers looking at it. nip it in the bud. | 14:50 |
| *** garthb has joined #openstack-ironic | 14:50 | |
| dansmith | rloo: of course, accidents happen, but I can't say I've seen something approved where someone added an argument to a function for no reason.. but, whatever, I'm not opposed to comments of course :) | 14:50 |
| jroll | dansmith: look at you all persuasive, I think I agree with you now | 14:52 |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 14:52 |
| jroll | the main point being we'll need more infra to sanely pass a topic | 14:52 |
| dansmith | yep | 14:52 |
| rloo | dansmith: I have seen +2, and either it wasn't documented and the others knew or didn't. but i had to ask why and the initial answer was cuz it was inconsistent. the next was 'it could be used', and the following was 'i'd like to use it in case blah'. I'm not quite sure if that case blah has shown up yet but i think the patch was approved. | 14:52 |
| dansmith | rloo: heh, fair enough.. | 14:53 |
| dansmith | anyway, I don't really have a voice here, I'm just throwing in my two cents.. happy to +1 it as is if that's the consensus | 14:53 |
| rloo | dansmith: well, you do have a voice and we are listening. I'm happy either way, just want it to be clear in the code/comments. | 14:55 |
| xek | rloo, I addressed your other nits | 14:55 |
| rloo | thx xek, looking... | 14:55 |
| dansmith | rloo: I just meant "I ain't tryin' to tell ya'll how to run yer project" :) | 14:56 |
| *** mgoddard has quit IRC | 14:56 | |
| rloo | dansmith: OH. yeah, that's probably a good thing. Although I'd listen to that too, if it is an improvement :) | 14:56 |
| jroll | dansmith: get off our lawn :D | 14:57 |
| dansmith | :) | 14:57 |
| * xek pushed ctrl+C when jroll seemed to change his mind but didn't make it | 14:57 | |
| jroll | lol | 14:57 |
| jroll | xek: I think we probably should remove it for now, and add it when we need it | 14:57 |
| rloo | xek, ha ha. you copied dan's 'oslo.messaging fu' ! | 14:57 |
| rloo | jroll: ++ | 14:58 |
| dansmith | rloo: you told him to and he followed directions :) | 14:58 |
| jroll | xek: can you do that so I don't have to waste a -1 on that patchset? | 14:59 |
| jroll | sorry for the run-around :( | 14:59 |
| xek | jroll, ok | 14:59 |
| rloo | dansmith: guess i wasn't explicit enough. at least it was only 'fu'. | 14:59 |
| jroll | I'm running low on -1s for this cycle | 14:59 |
| jroll | need to find someone to borrow from | 14:59 |
| rloo | jroll: you serious? you want to remove topic? | 14:59 |
| thiagop | I'd like to read more about how this rolling update feature is supposed to work. Any reading suggestions? | 14:59 |
| jroll | rloo: yeah, to be useful it's going to take quite a bit of work | 15:00 |
| *** mgoddard has joined #openstack-ironic | 15:00 | |
| rloo | jroll: ok. dansmith gets his wish. | 15:00 |
| jroll | heh | 15:00 |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 15:01 |
| openstackgerrit | Merged openstack/ironic-python-agent: Fix default ironic api port number https://review.openstack.org/226626 | 15:04 |
| rloo | xek: I commented. I did mention that I wanted a comment either way. | 15:06 |
| *** romcheg has quit IRC | 15:08 | |
| *** bizarrochristy has joined #openstack-ironic | 15:09 | |
| *** karimb has quit IRC | 15:09 | |
| *** karimb has joined #openstack-ironic | 15:09 | |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 15:10 |
| xek | rloo, rloo ^ | 15:11 |
| rloo | xek, you are fast. i just commented on 'fu', but that isn't a big deal cuz it is associated with dan. | 15:11 |
| *** romcheg has joined #openstack-ironic | 15:12 | |
| *** penick has joined #openstack-ironic | 15:14 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic-inspector: Updated from global requirements https://review.openstack.org/226805 | 15:14 |
| dansmith | rloo: heh | 15:15 |
| rloo | dansmith: I feel like I have only so many 'nit' credits ;) | 15:16 |
| dansmith | heh | 15:16 |
| *** uggla_ has quit IRC | 15:17 | |
| xek | http://www.urbandictionary.com/define.php?term=google-fu - oslo.messaging fu should be right up there | 15:17 |
| *** amotoki has quit IRC | 15:17 | |
| rloo | xek: wow, didn't realize that existed. | 15:18 |
| rloo | xek: ahh, for kung fu. Ok, that's probably what dansmith meant then, not the other fu. It is OK then! | 15:19 |
| dansmith | rloo: oh, yeah, sorry.. a lot of people mess it up by using "foo", but ${thing}-fu means "skills at $thing" | 15:19 |
| *** garthb has quit IRC | 15:20 | |
| *** garthb has joined #openstack-ironic | 15:20 | |
| rloo | dansmith: i'm not 'up' on american slang. | 15:20 |
| dansmith | heh | 15:20 |
| *** amotoki has joined #openstack-ironic | 15:20 | |
| *** lazy_prince has quit IRC | 15:22 | |
| *** tsekiyama has joined #openstack-ironic | 15:24 | |
| *** harshs has quit IRC | 15:26 | |
| *** puranamr has joined #openstack-ironic | 15:35 | |
| *** Sukhdev has joined #openstack-ironic | 15:38 | |
| *** yog_ has quit IRC | 15:46 | |
| *** penick has quit IRC | 15:46 | |
| *** e0ne has quit IRC | 15:47 | |
| *** puranamr has quit IRC | 15:47 | |
| *** e0ne has joined #openstack-ironic | 15:48 | |
| *** Nisha has joined #openstack-ironic | 15:48 | |
| *** puranamr has joined #openstack-ironic | 15:50 | |
| *** bizarrochristy has quit IRC | 15:52 | |
| *** ijw has quit IRC | 15:56 | |
| *** yuriyz has quit IRC | 15:57 | |
| *** sdake has joined #openstack-ironic | 15:57 | |
| *** garthb_ has joined #openstack-ironic | 16:00 | |
| *** trown is now known as trown|lunch | 16:00 | |
| *** ukalifon1 has joined #openstack-ironic | 16:00 | |
| *** ukalifon1 has left #openstack-ironic | 16:00 | |
| *** yuriyz has joined #openstack-ironic | 16:01 | |
| *** lazy_prince has joined #openstack-ironic | 16:01 | |
| openstackgerrit | Kurt Taylor proposed openstack/ironic: Unit test environment setup clarification https://review.openstack.org/226445 | 16:02 |
| *** mgoddard has quit IRC | 16:02 | |
| *** garthb has quit IRC | 16:02 | |
| *** mgoddard has joined #openstack-ironic | 16:03 | |
| *** sdake has quit IRC | 16:04 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Use standard locale when executing 'dd' command https://review.openstack.org/216478 | 16:06 |
| *** sdake_ has joined #openstack-ironic | 16:07 | |
| devananda | TheJulia: added you to bifrost-release. anyone else I should include? | 16:08 |
| TheJulia | devananda: not at this time, thank you! | 16:10 |
| *** karimb has quit IRC | 16:10 | |
| *** sdake has joined #openstack-ironic | 16:11 | |
| *** bizarrochristy has joined #openstack-ironic | 16:12 | |
| xek | jroll, lucasagomes, rloo, looks like tempest is timing out on boot_instance() https://jenkins07.openstack.org/job/gate-tempest-dsvm-ironic-pxe_ssh/2016/console | 16:13 |
| dtantsur | oh, so ironic-inspector is not the only broken? | 16:13 |
| jroll | xek: that's an occassional failure we have, unless something just broke | 16:13 |
| dtantsur | 2 of our patches just timed out when booting a node | 16:13 |
| jroll | ogod | 16:13 |
| jroll | no | 16:13 |
| jroll | pls | 16:13 |
| dtantsur | I first thought we broke something in inspector, but everything before nova boot looks sane | 16:14 |
| dtantsur | anyone knows how to decode http://logs.openstack.org/05/226805/2/check/gate-ironic-inspector-dsvm/31ff507/logs/ironic-bm-logs/baremetalbrbm_0_console.txt.gz | 16:14 |
| dtantsur | ? | 16:14 |
| *** sdake_ has quit IRC | 16:14 | |
| lucasagomes | :-( | 16:14 |
| lucasagomes | damn we had many gate breakages recently :-/ | 16:15 |
| jroll | dtantsur: I'm not sure it gives any useful info right now, I've been meaning to fix that | 16:15 |
| jroll | dtantsur: that basically means it didn't boot up right, see normal output: http://logs.openstack.org/01/226201/3/check/gate-tempest-dsvm-ironic-agent_ssh/6774602/logs/ironic-bm-logs/baremetalbrbm_0_console.txt.gz | 16:16 |
| jroll | I think it just hangs when it fails | 16:16 |
| jroll | same junk, and then it boots | 16:17 |
| *** amit213 has quit IRC | 16:17 | |
| *** amit213 has joined #openstack-ironic | 16:17 | |
| rloo | jroll, lucasagomes, xek, dtantsur: it's been timing out 'randomly'. i have a note in the etherpad about it. | 16:18 |
| jroll | yeah | 16:18 |
| lucasagomes | right | 16:18 |
| dtantsur | ok, I'll check locally if latest inspector has some problems... | 16:18 |
| jroll | dtantsur: I don't know what inspector changes you could make that would make ramdisk boot hang like that :P | 16:19 |
| jroll | and also affect normal pxe_ssh job | 16:20 |
| dtantsur | we definitely couldn't affect normal pxe_ssh job :) we could break something around our firewall logic, but I don't see any errors in logs... | 16:20 |
| *** romainh has left #openstack-ironic | 16:20 | |
| sambetts | noo :( | 16:21 |
| dtantsur | but all changes passed the gate, so I really doubt... | 16:21 |
| *** Sukhdev has quit IRC | 16:29 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: devstack: allow nodes to be already MANAGEABLE in exercise https://review.openstack.org/226873 | 16:29 |
| *** garthb_ has quit IRC | 16:30 | |
| *** garthb has joined #openstack-ironic | 16:30 | |
| *** Sukhdev has joined #openstack-ironic | 16:31 | |
| dtantsur | hmm, I've updated inspector on my devstack, and nova boots nodes correctly | 16:35 |
| dtantsur | I'll figure out tomorrow, if it does not self-heal | 16:35 |
| dtantsur | c u | 16:35 |
| *** dtantsur is now known as dtantsur|afk | 16:35 | |
| *** praneshp has joined #openstack-ironic | 16:39 | |
| sambetts | dtantsur|afk: cya later | 16:40 |
| *** nicodemos has joined #openstack-ironic | 16:41 | |
| *** Sukhdev has quit IRC | 16:43 | |
| *** Sukhdev has joined #openstack-ironic | 16:46 | |
| *** Sukhdev has quit IRC | 16:47 | |
| *** puranamr has quit IRC | 16:47 | |
| *** derekh has quit IRC | 16:49 | |
| devananda | jroll: any new thoughts today on the last 4 patches for the boot iface refactoring? | 16:53 |
| *** Sukhdev has joined #openstack-ironic | 16:53 | |
| jroll | devananda: I haven't looked, given the size I'm inclined to continue to do so | 16:53 |
| devananda | I started reviewing 216538 but ran out of steam a few hundred lines in | 16:53 |
| jroll | precisely | 16:53 |
| devananda | +1711, -2094 is, frankly, too much to be landing the day before a release -- even if it is almost all just driver code | 16:54 |
| *** Sukhdev has quit IRC | 16:54 | |
| rloo | jroll, devananda: I agree. Should defer it. | 16:54 |
| jroll | yep | 16:54 |
| devananda | call the boot interface done, and indicate which drivers were updated in the release notes -- and that it's fully backwards compatible at this stage? | 16:55 |
| *** romcheg has quit IRC | 16:56 | |
| jroll | devananda: I'm good with that | 16:56 |
| * devananda annotates the BP | 16:58 | |
| jroll | woo, 3 bugs left | 16:58 |
| lucasagomes | folks I'm going to call it a day | 16:58 |
| *** harshs has joined #openstack-ironic | 16:58 | |
| lucasagomes | have a good night everyone | 16:59 |
| JoshNang | o/ | 16:59 |
| *** penick has joined #openstack-ironic | 17:00 | |
| *** lucasagomes is now known as lucas-dinner | 17:00 | |
| thiagop | night lucas-dinner | 17:00 |
| thiagop | night dtantsur|afk | 17:00 |
| devananda | g'night, lucas-dinner | 17:00 |
| jroll | see ya lucas-dinner \o | 17:03 |
| *** trown|lunch is now known as trown | 17:04 | |
| jroll | rloo: one note on the dd patch, tests are likely going to fail there | 17:07 |
| rloo | jroll: ohhhhhhhhhhhhhhhhhh | 17:07 |
| rloo | jroll: why? | 17:07 |
| rloo | jroll: do we run the tests on OSs with diff lang? | 17:08 |
| jroll | rloo: see the comment, mock cares about argument order | 17:08 |
| rloo | jroll: I wondered about that but it passed on my box. in 2.7. maybe 3.4 won't. | 17:08 |
| jroll | really? weird | 17:09 |
| *** achanda has joined #openstack-ironic | 17:10 | |
| rloo | jroll: maybe it is smarter than we think, and know that key=value args can be in any order | 17:10 |
| jroll | rloo: maybe, if it does I think that's new :) | 17:11 |
| *** jxiaobin has quit IRC | 17:11 | |
| thiagop | are you talking about 216478? | 17:12 |
| jroll | yes | 17:12 |
| thiagop | I was looking into that bug and it seems that the problem is after the dd (just by the log, couldn't reproduce that) | 17:12 |
| jroll | really? | 17:14 |
| jroll | hrm | 17:14 |
| thiagop | the thing does do logout and delete from iscsiadm | 17:14 |
| rloo | thiagop: you mean the actual bug? not the patch? | 17:14 |
| thiagop | rloo: yep | 17:15 |
| *** ijw has joined #openstack-ironic | 17:15 | |
| *** e0ne has quit IRC | 17:16 | |
| jroll | thiagop: damn, you're right | 17:16 |
| thiagop | I'm guessing this error is because python is trying to output something that isn't ascii on the logs | 17:16 |
| thiagop | don't know where, though | 17:16 |
| jroll | which is why making everything use standard locale fixed it for yuiko | 17:17 |
| rloo | thiagop: $#@#% If I have to read every bug in detail... :-( | 17:17 |
| thiagop | I'm suspecting of get_disk_identifier | 17:17 |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 17:17 |
| thiagop | just gut feeling | 17:17 |
| rloo | thiagop, jroll: i think the bug should be punted then. | 17:18 |
| thiagop | rloo: that's why we're here :) | 17:18 |
| jroll | yeah probably | 17:18 |
| thiagop | to read what you folks are too tired to read | 17:18 |
| jroll | I'm not putting my machine into japanese to test it :D | 17:18 |
| rloo | thiagop: :) I'd want someone to reproduce/test. | 17:18 |
| *** praneshp has quit IRC | 17:18 | |
| rloo | thiagop: and thank you very much. I would love more reviewers. | 17:19 |
| *** ijw has quit IRC | 17:20 | |
| jroll | man, I wonder if we do have gate problems, this patch also failed all the tempest tests | 17:20 |
| devananda | jroll: I saw some gate failures for tempest timeouts yesterday | 17:20 |
| devananda | things taking >5min to spawn | 17:20 |
| jroll | orly | 17:21 |
| jroll | what I've seen so far looks like VMs hanging at boot | 17:21 |
| devananda | ok - BP updated and two new ones created for tracking the followup in Mitaka for ilo / irmc | 17:21 |
| jroll | \o/ | 17:21 |
| devananda | jroll: oh? | 17:21 |
| jroll | devananda: yeah, it's a random failure we have | 17:22 |
| devananda | awesome | 17:22 |
| jroll | it's well known | 17:22 |
| *** lazy_prince has quit IRC | 17:25 | |
| *** garthb has quit IRC | 17:26 | |
| rloo | I have to admit though, I've NEVER seen it fail so much. Usually 1, sometimes 2 tests fail. But all of them? https://review.openstack.org/#/c/216478/ | 17:26 |
| *** garthb has joined #openstack-ironic | 17:26 | |
| *** ijw has joined #openstack-ironic | 17:26 | |
| jroll | rloo: right, that's what worries me | 17:26 |
| * devananda updates the title of https://bugs.launchpad.net/ironic/+bug/1488289 | 17:26 | |
| openstack | Launchpad bug 1488289 in Ironic "Ironic fails to deploy instance with pxe_ipmitool driver on system with Japanese locale" [High,In progress] - Assigned to Yuiko Takada (takada-yuiko) | 17:26 |
| rloo | thx devananda | 17:27 |
| devananda | wow, no kidding | 17:27 |
| thiagop | I'll try to lauch an instance here to test it | 17:27 |
| rloo | jroll: wondering if some other changes have slowed things down enough... | 17:27 |
| thiagop | rloo: ^ | 17:27 |
| rloo | thx thiagop | 17:27 |
| thiagop | not sure if my cloud will help | 17:27 |
| thiagop | it's damn slow these days | 17:28 |
| thiagop | (not sure if I'll need to read japanese) | 17:31 |
| devananda | judging by the nearly empty baremetalbrbm_0_console.txt file -- yea, the VM hung and didn't start in each of those failures | 17:32 |
| *** afaranha has left #openstack-ironic | 17:32 | |
| jroll | devananda: yeah, exactly | 17:33 |
| *** Nisha has quit IRC | 17:35 | |
| *** Big-Boss has joined #openstack-ironic | 17:35 | |
| jroll | FYI, removed the milestone from that locale bug | 17:35 |
| jroll | so 2 left. | 17:35 |
| *** Big-Boss is now known as bigboss | 17:35 | |
| *** bigboss has left #openstack-ironic | 17:36 | |
| thiagop | ^ dahell?! | 17:36 |
| thiagop | lol | 17:36 |
| jroll | heh | 17:37 |
| *** pelix has quit IRC | 17:38 | |
| *** sinval has quit IRC | 17:39 | |
| rloo | jroll: one of those bugs left is IPA and it has two +2. I left it for you. | 17:39 |
| JayF | rloo: which one? do you know | 17:39 |
| jroll | JayF: https://review.openstack.org/#/c/215097/ | 17:39 |
| jroll | which I would love your eyes on | 17:39 |
| JayF | oh it's /that one/ | 17:39 |
| jroll | :) | 17:39 |
| JayF | the scary partitioning one | 17:39 |
| *** solid-sinval has joined #openstack-ironic | 17:39 | |
| JayF | that is going to break the world in a thousand edge cases | 17:40 |
| rloo | JayF: https://bugs.launchpad.net/ironic/+bug/1486887 | 17:40 |
| openstack | Launchpad bug 1486887 in Ironic "Config drive creation fails in UEFI boot mode for whole disk image " [High,In progress] - Assigned to Shivanand Tendulker (stendulker) | 17:40 |
| devananda | jroll: ack. it sounds like a very serious bug. I'd support backporting if the fix turns out to be simple | 17:40 |
| jroll | devananda: +1 | 17:40 |
| jroll | devananda: I think the fix will be simple, getting to the fix and verifying it will be the hard part | 17:40 |
| devananda | right | 17:41 |
| devananda | getting others to validate both the bug and the fix | 17:41 |
| devananda | jroll: on https://bugs.launchpad.net/ironic/+bug/1493816 it sounded from your conversation with dansmith like even the work in progress here won't close the bug | 17:41 |
| openstack | Launchpad bug 1493816 in Ironic "Rolling-upgrade: lack of indirection API" [High,In progress] - Assigned to Grzegorz Grasza (xek) | 17:41 |
| dansmith | devananda: it will solve the lack of an indirection api | 17:42 |
| dansmith | devananda: which from the subject of the bug seems reasonable | 17:42 |
| jroll | +1 | 17:42 |
| dansmith | devananda: yeah, I think the bug is closed by this patch | 17:42 |
| devananda | dansmith: perhaps I misunderstood then. does this fix give ironic rolling upgrade capabilities? | 17:42 |
| dansmith | all of rolling upgrades isn't, but the bug doesn't address the whole scope | 17:42 |
| dansmith | devananda: no, it's necessary but not sufficient | 17:43 |
| devananda | ok. so what exactly is not working as desired now which this bug corrects? | 17:43 |
| * devananda rereads it | 17:43 | |
| devananda | ahh. it will allow a rolling upgrade *to* mitaka | 17:43 |
| devananda | gotcha | 17:43 |
| dansmith | it won | 17:43 |
| dansmith | won't do that all by itself | 17:44 |
| dansmith | but it's a necessary step.. it removes one kind of failure | 17:44 |
| dansmith | you remove this, you get a little farther along the path so you can see what the next issue is that you need to fix | 17:44 |
| devananda | gotcha, thanks | 17:45 |
| * devananda updates bug wording slightly to clarify | 17:45 | |
| *** e0ne has joined #openstack-ironic | 17:56 | |
| thiagop | This bug on the UEFI boot, is it targeted to IPA instead of Ironic? | 17:56 |
| thiagop | the patch is for IPA, the bug is targeted to Ironic | 17:57 |
| thiagop | weird | 17:57 |
| rloo | thiagop: the ascii codec error? | 17:57 |
| thiagop | https://bugs.launchpad.net/ironic/+bug/1486887 | 17:57 |
| openstack | Launchpad bug 1486887 in Ironic "Config drive creation fails in UEFI boot mode for whole disk image " [High,In progress] - Assigned to Shivanand Tendulker (stendulker) | 17:57 |
| rloo | thiagop: oh. yeah, that's ipa. | 17:57 |
| thiagop | wrt ascii error, stacking... | 17:57 |
| rloo | thiagop: we don't (didn't) have ability to target bugs to IPA. | 17:58 |
| *** krtaylor has quit IRC | 17:58 | |
| jroll | thiagop: ipa shares a launchpad with ironic (for now) | 17:58 |
| thiagop | aw, ok. Thanks for clarification. | 17:58 |
| rloo | thiagop: but I think jroll or someone else is going to fix that :) | 17:58 |
| jroll | yeah, soon | 17:59 |
| jroll | next week maybe | 17:59 |
| thiagop | xek: I can send the patch with the i18n thing if you're too busy | 18:00 |
| *** garthb has quit IRC | 18:03 | |
| *** garthb has joined #openstack-ironic | 18:03 | |
| *** praneshp has joined #openstack-ironic | 18:03 | |
| *** olaph has quit IRC | 18:04 | |
| *** e0ne has quit IRC | 18:06 | |
| jroll | stepping away for a while | 18:06 |
| *** Nisha has joined #openstack-ironic | 18:08 | |
| *** olaph has joined #openstack-ironic | 18:09 | |
| *** krtaylor has joined #openstack-ironic | 18:11 | |
| *** e0ne has joined #openstack-ironic | 18:13 | |
| *** olaph has quit IRC | 18:23 | |
| *** achanda has quit IRC | 18:26 | |
| * betherly is heading out for supper - laters all | 18:26 | |
| TheJulia | goodnight betherly | 18:26 |
| thiagop | night betherly | 18:27 |
| *** dims__ has joined #openstack-ironic | 18:28 | |
| *** vishwanathj has joined #openstack-ironic | 18:29 | |
| *** achanda has joined #openstack-ironic | 18:30 | |
| *** dims_ has quit IRC | 18:31 | |
| JayF | jroll: https://review.openstack.org/#/c/215097/10/ironic_python_agent/shell/copy_configdrive_to_disk.sh,cm line 64 | 18:32 |
| JayF | jroll: partprobe issues in gate make me go hrm as to how well this will be tested | 18:32 |
| *** Sukhdev has joined #openstack-ironic | 18:34 | |
| *** amotoki has quit IRC | 18:35 | |
| *** romcheg has joined #openstack-ironic | 18:39 | |
| jroll | JayF: we don't have UEFI in gate, so yeah it won't be tested :( | 18:46 |
| JayF | yeah I just landed it | 18:46 |
| JayF | I feel very mehhh about it, we def need to test the hell outta that agent downstream before deploying it | 18:47 |
| JayF | but just because I think code is spooky doesn't mean it's wrong or shouldn't land | 18:47 |
| JayF | probably means more I'm paranoid about breaking things ;) | 18:47 |
| jroll | heh, yeah | 18:48 |
| jroll | though paranoia is welcome a day before a release :) | 18:48 |
| jroll | paranoia without an alternative, though, is always hard | 18:48 |
| *** ionutbalutoiu has quit IRC | 18:48 | |
| *** e0ne has quit IRC | 18:49 | |
| JayF | yeah exactly why I landed it | 18:49 |
| JayF | I don't have factual reasions it shouldn't hand | 18:49 |
| JayF | land | 18:49 |
| JayF | 3 people I trust say it should (author & 2x+2s) | 18:50 |
| *** penick has quit IRC | 18:51 | |
| jroll | so... just the indirection api thing | 18:52 |
| jroll | woo | 18:52 |
| rloo | jroll: I think we have problems at the gate: https://review.openstack.org/#/c/224079/ | 18:54 |
| jroll | yep. | 18:54 |
| jroll | now how the %^&*% do we debug this | 18:54 |
| rloo | jroll, devananda: gate isn't failing randomly any more | 18:54 |
| rloo | jroll: i was hoping you'd know. | 18:55 |
| jroll | I'm going to poke around | 18:55 |
| jroll | it isn't a DIB problem, we can rule that out | 18:55 |
| *** penick has joined #openstack-ironic | 19:06 | |
| *** Marga_ has quit IRC | 19:06 | |
| *** Marga_ has joined #openstack-ironic | 19:08 | |
| *** e0ne has joined #openstack-ironic | 19:12 | |
| *** Marga_ has quit IRC | 19:13 | |
| *** dims__ has quit IRC | 19:24 | |
| *** e0ne has quit IRC | 19:24 | |
| *** dims_ has joined #openstack-ironic | 19:25 | |
| *** Nisha has quit IRC | 19:26 | |
| jroll | blahhhhhhhhhhhhh | 19:26 |
| jroll | trying to reproduce this locally now | 19:27 |
| jroll | don't see anything super suspicious | 19:27 |
| rcarrillocruz | hmm, folks , do we do any kind of schema validation for node uuid ? | 19:28 |
| rcarrillocruz | i'm about to enroll a node | 19:28 |
| rcarrillocruz | and turns out the GUID is all set to zeroes | 19:29 |
| rcarrillocruz | :/ | 19:29 |
| rcarrillocruz | dunno what to put , in case i should stick to some convention db wise... | 19:29 |
| TheJulia | rcarrillocruz: it can be blank | 19:30 |
| TheJulia | ironic will assign a uuid then | 19:30 |
| rcarrillocruz | oh | 19:30 |
| rcarrillocruz | excellent | 19:30 |
| rcarrillocruz | thx :-) | 19:30 |
| TheJulia | np | 19:31 |
| *** penick has quit IRC | 19:31 | |
| TheJulia | rcarrillocruz: did you ever figure out why your node names were not getting populated? | 19:31 |
| rcarrillocruz | no, i have to come back to this , probably on friday | 19:32 |
| rcarrillocruz | i just finished a 35 blades provision + bootstrap | 19:32 |
| thiagop | wow | 19:33 |
| rcarrillocruz | but fwiw | 19:33 |
| rcarrillocruz | http://paste.openstack.org/show/473799/ | 19:33 |
| rcarrillocruz | and as i told you, i do have 'name' populated on baremetal.json | 19:33 |
| rcarrillocruz | dunno, i have my bifrost pinned to a commit | 19:33 |
| rcarrillocruz | will put that in the launchpad bug report | 19:33 |
| rcarrillocruz | bleh | 19:34 |
| rcarrillocruz | TheJulia: | 19:34 |
| rcarrillocruz | i just got 'failed: [lolserver.domain.lol -> localhost] => {"failed": true} | 19:34 |
| rcarrillocruz | msg: A uuid or name value must be defined to use this module' | 19:34 |
| rcarrillocruz | i set the uuid to null | 19:34 |
| rcarrillocruz | should i set to blank | 19:34 |
| rcarrillocruz | ? | 19:34 |
| * rcarrillocruz goes try | 19:35 | |
| TheJulia | yeah, you shouldn't have any reference if it is to use name | 19:35 |
| *** derekh has joined #openstack-ironic | 19:36 | |
| TheJulia | oddly enough, if your not getting name through the inventory, then that might be problematic for you :) | 19:36 |
| rcarrillocruz | yeah, i'm having lots of fun :-) | 19:36 |
| rcarrillocruz | greping uuid to get hostname from baremetal.json | 19:36 |
| rcarrillocruz | *not* fun | 19:36 |
| *** Sukhdev has quit IRC | 19:37 | |
| TheJulia | :( | 19:37 |
| *** Marga_ has joined #openstack-ironic | 19:37 | |
| rcarrillocruz | doh | 19:38 |
| rcarrillocruz | so | 19:38 |
| rcarrillocruz | blank it complains | 19:38 |
| rcarrillocruz | "" complains | 19:38 |
| rcarrillocruz | null complains | 19:38 |
| rcarrillocruz | not putting uuid field for the node complains about not having uuid | 19:38 |
| rcarrillocruz | so it seems i really need to have a uuid | 19:39 |
| rcarrillocruz | even if it's made up | 19:39 |
| TheJulia | that doesn't seem right, filing a bug for that | 19:40 |
| jroll | WELP http://ipxe.org/err/040ee1 | 19:41 |
| *** penick has joined #openstack-ironic | 19:41 | |
| *** achanda has quit IRC | 19:41 | |
| TheJulia | rcarrillocruz: you know what, I think I may have just reproduced your lacking name issue as well | 19:41 |
| TheJulia | jroll: *pours a whisky and passes it over* | 19:42 |
| rcarrillocruz | :S | 19:42 |
| rcarrillocruz | ok | 19:42 |
| rcarrillocruz | uuidgen to the rescue | 19:43 |
| * rcarrillocruz googled | 19:43 | |
| jroll | TheJulia: thanks, much needed | 19:44 |
| TheJulia | rcarrillocruz: I'm going ahead and creating a bug for both since they are likely rooted in same/similar places, and there has to be one or the other realistically | 19:46 |
| rcarrillocruz | sweet, thanks | 19:46 |
| trown | jroll: that looks like we did not get DHCP | 19:46 |
| trown | just going by the error name anyways EADDRNOTAVAIL_CONFIG | 19:47 |
| jroll | yeah | 19:47 |
| jroll | there's a dhcp-related neutron change, going to revert it and restack | 19:47 |
| trown | suppose if I read the additional notes...it says it right there | 19:48 |
| jroll | so either neutron broke us | 19:49 |
| jroll | or there's suddenly a new package supplying our pxe rom or ipxe rom | 19:49 |
| *** Marga_ has quit IRC | 19:50 | |
| trown | the second seems less likely | 19:50 |
| jroll | yeah | 19:50 |
| jroll | but the former is a v6-only change, it seems | 19:51 |
| jroll | anyway, we'll know shortly | 19:51 |
| jroll | testing with a revert now | 19:51 |
| jroll | wtb 'ironic boot-me-an-agent $uuid' | 19:51 |
| jroll | JoshNang: ^ last feature before you leave? :D | 19:51 |
| jroll | most of the infra is there | 19:52 |
| *** athomas has quit IRC | 19:53 | |
| jroll | this would be terrible if it's a neutron bug, that was the last commit before they cut stable/liberty, lol | 19:53 |
| *** olaph has joined #openstack-ironic | 19:55 | |
| trown | jroll: just checked dpkg from working(http://logs.openstack.org/36/210436/3/check/gate-tempest-dsvm-ironic-pxe_ssh-dib/85a6e53/logs/) and non-working(http://logs.openstack.org/36/210436/3/check/gate-tempest-dsvm-ironic-pxe_ssh-dib/1ba2ec1/logs/) | 19:58 |
| trown | both have the same version of ipxe | 19:58 |
| jroll | ohey, it is neutron, joy | 19:58 |
| *** ukalifon1 has joined #openstack-ironic | 20:00 | |
| rloo | jroll: neutron folks must be happy that we're testing their stuff | 20:03 |
| jroll | >.> | 20:03 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 20:04 |
| jroll | ^ added depends-on for extra verification | 20:04 |
| rloo | jroll: we don't know for sure though? | 20:05 |
| jroll | rloo: I mean, I reproduced locally, then reverted the neutron patch as such https://review.openstack.org/#/c/226969/ | 20:06 |
| jroll | and post-revert it worked fine | 20:06 |
| TheJulia | rcarrillocruz: bug 1499057, if you want to provide any additional detail/context, it might be helpful but I think i got the essentials | 20:06 |
| openstack | bug 1499057 in bifrost "Bifrost inventory parsing issues" [High,Triaged] https://launchpad.net/bugs/1499057 | 20:06 |
| *** ukalifon1 has left #openstack-ironic | 20:06 | |
| rloo | jroll: oh, that's good to know! | 20:06 |
| rcarrillocruz | ++ | 20:06 |
| jroll | :) | 20:06 |
| trown | jroll: cool I will put that on the DIB patch that I was using to troubleshoot too | 20:06 |
| trown | that gives two datapoints | 20:07 |
| jroll | cool | 20:07 |
| jroll | brb | 20:07 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: WIP: Ansible 2.0 compatibility **DO NOT MERGE** https://review.openstack.org/196832 | 20:08 |
| *** derekh has quit IRC | 20:16 | |
| *** sdake_ has joined #openstack-ironic | 20:23 | |
| *** nicodemos has quit IRC | 20:23 | |
| *** sdake has quit IRC | 20:25 | |
| *** trown is now known as trown|mtg | 20:27 | |
| *** albertoffb has quit IRC | 20:34 | |
| *** achanda has joined #openstack-ironic | 20:42 | |
| *** MattMan has left #openstack-ironic | 20:42 | |
| *** Sukhdev has joined #openstack-ironic | 20:43 | |
| *** liliars has quit IRC | 20:45 | |
| *** lucas-dinner has quit IRC | 20:46 | |
| *** achanda has quit IRC | 20:47 | |
| *** achanda has joined #openstack-ironic | 20:49 | |
| *** achanda has quit IRC | 20:55 | |
| *** eglute has quit IRC | 20:55 | |
| *** achanda has joined #openstack-ironic | 20:55 | |
| *** eglute has joined #openstack-ironic | 20:55 | |
| *** puranamr has joined #openstack-ironic | 20:57 | |
| thiagop | I'm calling it a day | 20:57 |
| thiagop | good night folks | 20:57 |
| thiagop | see you tomorrow | 20:57 |
| *** thiagop has quit IRC | 20:57 | |
| *** harshs has quit IRC | 21:01 | |
| *** penick has quit IRC | 21:01 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Ansible 2.0 compatibility https://review.openstack.org/196832 | 21:03 |
| *** penick has joined #openstack-ironic | 21:04 | |
| *** harshs has joined #openstack-ironic | 21:04 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 21:05 |
| jroll | ^ depends-on for a fix rather than a revert | 21:05 |
| jroll | the revert did pass tempest jobs | 21:06 |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Add 'abort' to state machine diagram https://review.openstack.org/226997 | 21:07 |
| *** trown|mtg is now known as trown | 21:08 | |
| trown | jroll: ya it passed here too https://review.openstack.org/#/c/210436/ | 21:08 |
| jroll | yeah | 21:08 |
| jroll | neutron folks are cool with landing the revert if needed, as they don't want to hold our release up | 21:09 |
| *** penick has quit IRC | 21:09 | |
| *** trown is now known as trown|outttypeww | 21:10 | |
| rloo | jroll: seems like they should land the revert. that change could have broken things for someone else? | 21:12 |
| *** puranamr has quit IRC | 21:12 | |
| rloo | jroll: or is it a bug on our side? | 21:13 |
| jroll | rloo: well, they're fixing the actual bug | 21:13 |
| jroll | nah it's their side | 21:13 |
| rloo | jroll: yeah, assuming no-dhcp == not allocated or whatever. | 21:13 |
| jroll | something like that | 21:13 |
| jroll | kevinbenton explained it to me in -neutron, I don't fully understand it | 21:14 |
| rloo | jroll: as long as they understand it :) | 21:14 |
| jroll | that's the goal | 21:14 |
| jroll | :P | 21:14 |
| *** Marga_ has joined #openstack-ironic | 21:21 | |
| *** ndipanov has quit IRC | 21:25 | |
| rloo | jroll: I didn't -1 (and I don't care right now): https://review.openstack.org/#/c/224079/21..23/ironic/conductor/rpcapi.py | 21:25 |
| jroll | rloo: oh, good one | 21:25 |
| jroll | rloo: I -1'd, I'll come back to it after I see how tests do | 21:27 |
| rloo | jroll: ok. guess i should have done that then. | 21:28 |
| jroll | nah it's alright | 21:28 |
| jroll | I feel your pain | 21:28 |
| rloo | jroll: :) I hate being the bad 'guy' | 21:28 |
| jroll | heh | 21:28 |
| *** achanda has quit IRC | 21:32 | |
| mrda | Morning Ironic | 21:35 |
| *** achanda has joined #openstack-ironic | 21:36 | |
| rloo | morning mrda | 21:38 |
| mrda | hi rloo | 21:38 |
| jroll | mornin' mrda | 21:39 |
| mrda | hey jroll | 21:39 |
| *** harshs has quit IRC | 21:51 | |
| *** harshs has joined #openstack-ironic | 21:54 | |
| *** achanda has quit IRC | 21:57 | |
| *** sdake_ has quit IRC | 21:59 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Update install guide to reflect 4.2.0 release https://review.openstack.org/227029 | 22:00 |
| *** achanda has joined #openstack-ironic | 22:00 | |
| rloo | jroll, devananda: ^^ that should be updated before the release is cut. Not quite sure what it should say. | 22:01 |
| jroll | rloo: good catch, thanks. left a -1 for you :) | 22:02 |
| rloo | jroll: I don't know either. I think there was a wiki or something about it but I didn't pay attention. sigh, let me see if i can find it. | 22:02 |
| jroll | rloo: every project has their own thing now | 22:03 |
| jroll | nova is 12.0.0, glance is maybe 10.0.0 or something | 22:03 |
| jroll | there's no "number" for liberty as a whole | 22:03 |
| rloo | jroll: so every project changed? I didn't know that! | 22:03 |
| jroll | yep :) | 22:03 |
| jroll | rloo: though we're somewhat unique in that we don't bump the major version each cycle | 22:03 |
| *** alexpilotti has joined #openstack-ironic | 22:04 | |
| *** vdrok has quit IRC | 22:04 | |
| *** vdrok has joined #openstack-ironic | 22:05 | |
| rloo | jroll: http://superuser.openstack.org/articles/confused-about-the-openstack-release-cycle-check-out-this-handy-chart | 22:05 |
| jroll | rloo: huh. I'll ask the release team | 22:06 |
| rloo | jroll: so '4.2.0 (Liberty release)' I think | 22:06 |
| jroll | I think so too | 22:06 |
| jroll | though that notes 2015.2 | 22:06 |
| rloo | jroll: i think that was a copy/paste from something that wasn't updated. this article doesn't use 2015.2. | 22:07 |
| *** [1]cdearborn has quit IRC | 22:08 | |
| jroll | rloo: yeah, indeed | 22:08 |
| jroll | I agree | 22:08 |
| *** garthb has quit IRC | 22:10 | |
| *** penick has joined #openstack-ironic | 22:10 | |
| *** garthb has joined #openstack-ironic | 22:10 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Update install guide to reflect 4.2.0 release https://review.openstack.org/227029 | 22:11 |
| rloo | jroll: gotta go. Doesn't quite seem right still ^^ but better. | 22:12 |
| jroll | rloo: cool, I have release notes patch coming later. have a good night :) | 22:12 |
| *** openstackgerrit has quit IRC | 22:16 | |
| *** openstackgerrit has joined #openstack-ironic | 22:17 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add 4.2.0 release notes https://review.openstack.org/227033 | 22:17 |
| jroll | devananda: rloo ^ | 22:17 |
| jroll | devananda: are you still working a patch to move specs around? | 22:17 |
| *** priteau has quit IRC | 22:19 | |
| *** mestery has quit IRC | 22:21 | |
| *** mestery has joined #openstack-ironic | 22:22 | |
| jroll | well, I figured out how to set a topic if we're using a remotable method | 22:29 |
| jroll | this is going to be fun | 22:29 |
| jroll | heh | 22:29 |
| jroll | dansmith: is deleting an RPC method something that should ever be done? just curious | 22:30 |
| jroll | like, eventually we want to transition the do_node_update rpc call to node.update() | 22:31 |
| dansmith | jroll: you can delete it if you bump the major version | 22:31 |
| *** bizarrochristy has quit IRC | 22:31 | |
| jroll | eww | 22:31 |
| dansmith | bumping the major comes with caveats of course | 22:31 |
| jroll | yeah | 22:31 |
| dansmith | we try to bump our "busy" rpc apis once every couple releases at least | 22:31 |
| dansmith | so we can drop stuff like that | 22:31 |
| jroll | I'm curious what caveats other than we'd need to deprecate for a couple cycles | 22:31 |
| dansmith | that, basically | 22:32 |
| dansmith | so, the way this works is, | 22:32 |
| jroll | but we'll never support like, L API and N conductor | 22:32 |
| dansmith | you support both versions on the server side across the release boundary, but always send the newer version | 22:32 |
| dansmith | then after the release you can drop support for the older one once you know that you won't have a two-cycles-old client talking to you (since that's the rule) | 22:32 |
| dansmith | not a huge deal, but it can be confusing the first time you do it | 22:33 |
| jroll | yep | 22:33 |
| jroll | makes sense | 22:33 |
| jroll | thanks | 22:33 |
| devananda | jroll: I proposed and WIPd it | 22:34 |
| jroll | devananda: ok cool, thanks | 22:35 |
| *** jamielennox is now known as jamielennox|away | 22:35 | |
| jroll | dansmith: I'm also curious where this object manifest comes from, but maybe that's a lesson for another day | 22:35 |
| jroll | I feel like there may be lots of magic involved | 22:35 |
| devananda | jroll: re: release notes, mind if I update with the list of in tree drivers updated? | 22:36 |
| devananda | I'd rather be precise than say "most" | 22:36 |
| jroll | devananda: sure, I never mind people picking up my patches :) | 22:37 |
| devananda | dansmith: two-cycle-boundary rule makes sense for nova, but less so for swift / ironic | 22:37 |
| devananda | jroll: :) | 22:37 |
| jroll | devananda: an 'except' list may be easier | 22:37 |
| jroll | devananda: the way I see it, a two-cycle-boundary is always more than enough for us | 22:37 |
| jroll | "playing it safe" | 22:37 |
| jroll | as you always remind me, there's people that will only run stable branch | 22:38 |
| devananda | I do not necessarily agree | 22:38 |
| devananda | but I agree for now :) | 22:38 |
| jroll | devananda: you think we may need *longer*? | 22:38 |
| devananda | in my ideal world, the distros would maintain LTS-style releases, ie, need 2 year support for upgrades | 22:38 |
| devananda | but none of openstack is there yet | 22:39 |
| jroll | oh god | 22:39 |
| jroll | yeah | 22:39 |
| jroll | I think we'd need the rest of openstack doing it | 22:39 |
| devananda | so -- two-cycle boundary is good for now. I just bring this up cause you said "always" | 22:39 |
| jroll | heh | 22:39 |
| jroll | "today" | 22:39 |
| devananda | :) | 22:39 |
| dansmith | devananda: we do one cycle boundary | 22:40 |
| dansmith | devananda: like L->M only | 22:40 |
| dansmith | jroll: there is a method in o.vo that walks a given object, figures out which objects are related, gathers the versions and that's the manifest | 22:40 |
| jroll | aha | 22:41 |
| dansmith | the rule has always been N->N+1 and no more | 22:41 |
| dansmith | i.e. no skipping releases | 22:41 |
| dansmith | and honestly, I can't imagine skipping releases is on the horizon for anyone in openstack | 22:41 |
| jroll | heh, we don't have obj_make_compatible in ironic | 22:42 |
| jroll | so much infra missing for this | 22:42 |
| dansmith | yep | 22:42 |
| dansmith | it's been a lot of effort in nova to get to where we are | 22:42 |
| jroll | totally | 22:42 |
| jroll | why can't we just subclass o.vo and have everything work, cmon | 22:43 |
| jroll | :) | 22:43 |
| *** dims__ has joined #openstack-ironic | 22:43 | |
| * jroll learning way too much about objects this week | 22:43 | |
| dansmith | yeah, sorry :) | 22:43 |
| *** w_ has joined #openstack-ironic | 22:43 | |
| jroll | dansmith: I'd love to hear a 'what we would do if we started over' talk on this, btw | 22:43 |
| JayF | jroll: "use go" | 22:44 |
| jroll | so we don't copy/paste a bunch of cruft etc | 22:44 |
| * JayF ducks | 22:44 | |
| jroll | >.> | 22:44 |
| *** alexpilotti has quit IRC | 22:44 | |
| jroll | JayF: legitimate argument for a completely different problem set :) | 22:44 |
| JayF | jroll: "I trade this problem set in for another problem set, please" | 22:44 |
| dansmith | jroll: for ironic or objects? | 22:45 |
| jroll | JayF: sure, we can trade, please review this objects code | 22:45 |
| jroll | dansmith: ouch | 22:45 |
| jroll | objects. | 22:45 |
| dansmith | jroll: no, I mean.. for ironic objects or o.vo? | 22:45 |
| *** dims_ has quit IRC | 22:46 | |
| jroll | dansmith: for the 'user' of o.vo | 22:46 |
| *** olaph has quit IRC | 22:46 | |
| dansmith | well, I got to correct several mistakes when we cut o.vo out of nova | 22:46 |
| dansmith | fundamental things we should have done differently from the beginning but just didn't know | 22:46 |
| dansmith | so that was good | 22:46 |
| dansmith | ironic pulled the stuff out of nova and ran with it, | 22:46 |
| jroll | dansmith: so like, for microversions, y'all have this cool decorator thing rather than if statements that makes life easy | 22:46 |
| jroll | which I still want to switch ironic over to | 22:47 |
| dansmith | and didn't really benefit from the same level of experience we were gaining because ironic had other things on its mind | 22:47 |
| jroll | thinking more code things like that, you know? | 22:47 |
| dansmith | so, the fact that you used objects from the start is a huge win over things like cinder who are retrofitting | 22:47 |
| jroll | nod | 22:47 |
| JayF | Did we use objects from the start? I thought comstud converted that stuff over in Icehouse | 22:47 |
| jroll | bearhands got all ragey every time he saw us using the dict stuff | 22:47 |
| JayF | although I guess Icehouse is probably " | 22:47 |
| dansmith | but yeah, going through the code as little as I have, I definitely have a list of things to work on | 22:47 |
| jroll | JayF: it was like half there when he started | 22:47 |
| JayF | jroll: aha | 22:48 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Allow vendor drivers to acquire shared tasks https://review.openstack.org/226058 | 22:48 |
| dansmith | jroll: well, very early at least | 22:48 |
| jroll | yep | 22:48 |
| *** jerrygb has quit IRC | 22:55 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Implement indirection_api https://review.openstack.org/224079 | 22:57 |
| jroll | ^ translated some things, going to +2 | 22:57 |
| *** trown|outttypeww is now known as trown | 23:01 | |
| trown | jroll: the path forward for the neutron issue is the fix? | 23:01 |
| trown | as opposed to the revert? | 23:01 |
| jroll | trown: I believe so, it has a +2 | 23:01 |
| jroll | other folks are looking at it | 23:02 |
| trown | ok, thanks I will switch the DIB patch depends | 23:02 |
| jroll | cool | 23:02 |
| *** Sukhdev has quit IRC | 23:03 | |
| *** olaph has joined #openstack-ironic | 23:04 | |
| *** w_ has quit IRC | 23:06 | |
| *** olaph has quit IRC | 23:09 | |
| *** garthb has quit IRC | 23:13 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-python-agent: Fixes config drive creation failure in UEFI boot mode https://review.openstack.org/215097 | 23:14 |
| jroll | ^ added depends-on and re-approved | 23:15 |
| *** penick has quit IRC | 23:21 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add 4.2.0 release notes https://review.openstack.org/227033 | 23:21 |
| devananda | oops. forgot to hit review ... | 23:21 |
| devananda | jroll: ^ also re-ordered the notes so that deprecation warning is at the top | 23:22 |
| jroll | cool thanks | 23:22 |
| devananda | shutting down for a bit, gotta catch a bus | 23:23 |
| jroll | devananda: also, curious what you think is best here https://review.openstack.org/227029 | 23:23 |
| jroll | heh ok, later | 23:23 |
| * devananda looks | 23:23 | |
| devananda | heh | 23:24 |
| devananda | well | 23:24 |
| devananda | perhaps we should remove any mention of release name | 23:24 |
| jroll | I'm fine with that too | 23:24 |
| jroll | though | 23:24 |
| jroll | many users just know it by name, not version | 23:24 |
| jroll | (I assume) | 23:24 |
| devananda | and just say "is contualy updated and reflects the latest available code" | 23:25 |
| devananda | and then work with the docs team to publish the docs as they were at each release point | 23:25 |
| jroll | then people ask where the docs for liberty are :) | 23:25 |
| jroll | yeah | 23:25 |
| devananda | right, well, yea | 23:25 |
| devananda | at the moment, that is neither the docs for kilo nor liberty | 23:25 |
| devananda | so it's wrong now, and it's wrong with that patch | 23:25 |
| jroll | well, it isn't wrong | 23:25 |
| jroll | it says it should be accurate for kilo | 23:25 |
| devananda | but it's not | 23:26 |
| devananda | it's kilo + delta | 23:26 |
| openstackgerrit | Stephanie Miller proposed openstack/bifrost: gcc can also be missing on deb distros https://review.openstack.org/227047 | 23:26 |
| jroll | I guess | 23:26 |
| jroll | I tend to think the setup guide would work for Kilo | 23:26 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add 4.2.0 release notes https://review.openstack.org/227033 | 23:27 |
| jroll | devananda: ^ one quick update | 23:27 |
| jroll | feel free to look later though | 23:27 |
| devananda | jroll: eg, http://docs.openstack.org/developer/ironic/webapi/v1.html#api-versions-history | 23:27 |
| jroll | yeah | 23:27 |
| *** penick has joined #openstack-ironic | 23:28 | |
| devananda | ++ on the doc link | 23:28 |
| devananda | ok, gotta run | 23:28 |
| jroll | see ya | 23:28 |
| *** achanda has quit IRC | 23:33 | |
| *** romcheg has quit IRC | 23:35 | |
| *** bizarrochristy has joined #openstack-ironic | 23:44 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Update install guide to reflect 4.2 release https://review.openstack.org/227029 | 23:51 |
| *** baoli_ has quit IRC | 23:51 | |
| jroll | rloo: did you see the discussion above? | 23:51 |
| *** jamielennox|away is now known as jamielennox | 23:52 | |
| rloo | jroll: oh, no. looking. | 23:54 |
| jroll | heh | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!