*** ChuckC has joined #openstack-ironic | 00:07 | |
*** praneshp has quit IRC | 00:09 | |
*** chuckC_ has joined #openstack-ironic | 00:09 | |
*** praneshp has joined #openstack-ironic | 00:13 | |
*** todd_dsm has joined #openstack-ironic | 00:35 | |
*** todd_dsm has quit IRC | 00:45 | |
*** ryanpetrello_ has joined #openstack-ironic | 00:57 | |
*** ryanpetrello_ has quit IRC | 00:59 | |
*** nosnos has joined #openstack-ironic | 01:01 | |
*** ryanpetrello_ has joined #openstack-ironic | 01:09 | |
*** r-daneel has quit IRC | 01:10 | |
*** ryanpetrello_ has quit IRC | 01:10 | |
*** ryanpetrello_ has joined #openstack-ironic | 01:16 | |
*** ryanpetrello_ has quit IRC | 01:16 | |
*** todd_dsm has joined #openstack-ironic | 01:19 | |
*** todd_dsm has quit IRC | 01:21 | |
*** ryanpetrello_ has joined #openstack-ironic | 01:21 | |
*** ryanpetrello_ has quit IRC | 01:26 | |
*** takadayuiko has joined #openstack-ironic | 01:27 | |
openstackgerrit | Mark Atwood proposed a change to openstack/ironic: Cleans up some Sphinx rST warnings in Ironic https://review.openstack.org/125273 | 01:28 |
---|---|---|
*** ryanpetrello_ has joined #openstack-ironic | 01:30 | |
*** marcoemorais has quit IRC | 01:31 | |
*** ryanpetr- has joined #openstack-ironic | 01:34 | |
*** ryanpetr- has quit IRC | 01:34 | |
*** ryanpetrello_ has quit IRC | 01:36 | |
*** ryanpetrello_ has joined #openstack-ironic | 01:40 | |
*** spandhe has quit IRC | 01:43 | |
*** Kui has joined #openstack-ironic | 01:47 | |
*** ryanpetrello_ has quit IRC | 01:48 | |
*** ryanpetrello_ has joined #openstack-ironic | 01:51 | |
*** ryanpetrello_ has quit IRC | 01:54 | |
*** Poornima has joined #openstack-ironic | 01:54 | |
*** todd_dsm has joined #openstack-ironic | 01:56 | |
*** ryanpetrello_ has joined #openstack-ironic | 01:59 | |
*** ryanpetrello_ has quit IRC | 02:01 | |
*** ryanpetrello_ has joined #openstack-ironic | 02:06 | |
*** rloo has quit IRC | 02:07 | |
*** todd_dsm has quit IRC | 02:10 | |
*** Haomeng has joined #openstack-ironic | 02:12 | |
*** Haomeng|2 has quit IRC | 02:12 | |
*** ryanpetrello_ has quit IRC | 02:12 | |
*** dlaube has quit IRC | 02:12 | |
*** todd_dsm has joined #openstack-ironic | 02:17 | |
*** ryanpetrello_ has joined #openstack-ironic | 02:18 | |
*** pcrews has quit IRC | 02:19 | |
*** praneshp has quit IRC | 02:20 | |
*** ryanpetrello_ has quit IRC | 02:21 | |
*** ryanpetrello_ has joined #openstack-ironic | 02:25 | |
*** praneshp has joined #openstack-ironic | 02:26 | |
*** ryanpetrello_ has quit IRC | 02:27 | |
*** zigo has quit IRC | 02:27 | |
*** zigo has joined #openstack-ironic | 02:29 | |
*** ryanpetrello_ has joined #openstack-ironic | 02:31 | |
*** ryanpetrello_ has quit IRC | 02:32 | |
*** todd_dsm has quit IRC | 02:35 | |
*** ryanpetrello_ has joined #openstack-ironic | 02:36 | |
*** ryanpetrello_ has quit IRC | 02:42 | |
*** ryanpetrello_ has joined #openstack-ironic | 02:46 | |
*** ryanpetrello_ has quit IRC | 02:46 | |
*** notsogentle is now known as annegentle | 02:47 | |
*** ryanpetrello_ has joined #openstack-ironic | 02:50 | |
*** praneshp has quit IRC | 02:51 | |
*** praneshp has joined #openstack-ironic | 02:51 | |
*** ryanpetrello has quit IRC | 02:57 | |
*** ryanpetrello_ is now known as ryanpetrello | 02:57 | |
*** praneshp has quit IRC | 02:58 | |
*** nosnos has quit IRC | 03:24 | |
*** praneshp has joined #openstack-ironic | 03:29 | |
*** praneshp_ has joined #openstack-ironic | 03:32 | |
*** praneshp has quit IRC | 03:35 | |
*** praneshp_ is now known as praneshp | 03:35 | |
*** rushiagr_away is now known as rushiagr | 03:38 | |
*** ramineni has joined #openstack-ironic | 04:01 | |
*** rushiagr is now known as rushiagr_away | 04:09 | |
*** nosnos has joined #openstack-ironic | 04:10 | |
*** pensu has joined #openstack-ironic | 04:11 | |
*** Poornima has quit IRC | 04:15 | |
*** pensu has quit IRC | 04:19 | |
*** Poornima has joined #openstack-ironic | 04:44 | |
*** lazy_prince2 has joined #openstack-ironic | 05:10 | |
*** lazy_prince2 has quit IRC | 05:12 | |
*** rakesh_hs has joined #openstack-ironic | 05:17 | |
*** loki184 has joined #openstack-ironic | 05:21 | |
*** rushiagr_away is now known as rushiagr | 05:23 | |
*** coolsvap|afk is now known as coolsvap | 05:24 | |
*** harlowja is now known as harlowja_away | 05:26 | |
*** nosnos has quit IRC | 05:37 | |
*** nosnos has joined #openstack-ironic | 05:38 | |
*** amitpp has joined #openstack-ironic | 05:40 | |
*** amitpp has quit IRC | 05:41 | |
*** nosnos has quit IRC | 05:42 | |
*** nosnos has joined #openstack-ironic | 05:43 | |
*** pensu has joined #openstack-ironic | 06:03 | |
GheRivero | morning all | 06:49 |
*** jcoufal has joined #openstack-ironic | 07:09 | |
*** arif-ali has quit IRC | 07:11 | |
*** arif-ali has joined #openstack-ironic | 07:11 | |
*** ifarkas has joined #openstack-ironic | 07:21 | |
*** jistr has joined #openstack-ironic | 07:46 | |
*** lazy_prince has quit IRC | 07:48 | |
*** lazy_prince has joined #openstack-ironic | 07:49 | |
*** lazy_prince2 has joined #openstack-ironic | 07:51 | |
*** lazy_prince is now known as killer_prince | 07:53 | |
*** killer_prince is now known as lazy_prince | 07:56 | |
*** derekh has joined #openstack-ironic | 07:57 | |
*** lazy_prince2 has quit IRC | 07:58 | |
*** foexle has joined #openstack-ironic | 08:00 | |
*** lucasagomes has joined #openstack-ironic | 08:11 | |
*** openstackgerrit has quit IRC | 08:11 | |
*** praneshp has quit IRC | 08:20 | |
*** nosnos has quit IRC | 08:23 | |
*** athomas has joined #openstack-ironic | 08:31 | |
*** nosnos has joined #openstack-ironic | 08:40 | |
*** dtantsur|afk is now known as dtantsur | 09:20 | |
dtantsur | g'afternoon Ironic | 09:20 |
*** loki184 has quit IRC | 09:26 | |
*** lazy_prince has quit IRC | 09:29 | |
lucasagomes | dtantsur, afternoon :) | 09:36 |
*** kincl_ has joined #openstack-ironic | 09:43 | |
*** kincl has quit IRC | 09:47 | |
*** GheRivero has quit IRC | 09:47 | |
*** GheRivero has joined #openstack-ironic | 09:47 | |
*** loki184 has joined #openstack-ironic | 09:48 | |
*** pelix has joined #openstack-ironic | 09:51 | |
*** tatyana has joined #openstack-ironic | 10:09 | |
*** romcheg has joined #openstack-ironic | 10:12 | |
*** lazy_prince has joined #openstack-ironic | 10:34 | |
*** lucasagomes_ has joined #openstack-ironic | 10:48 | |
*** lucasagomes has quit IRC | 10:54 | |
*** yuanying has quit IRC | 10:54 | |
*** agordeev has quit IRC | 10:54 | |
*** dhellmann has quit IRC | 10:54 | |
*** mordred has quit IRC | 10:54 | |
*** ramineni has quit IRC | 10:55 | |
*** dhellmann has joined #openstack-ironic | 10:58 | |
*** loki184 has quit IRC | 11:02 | |
*** tatyana has quit IRC | 11:04 | |
*** yuanying has joined #openstack-ironic | 11:07 | |
*** agordeev has joined #openstack-ironic | 11:07 | |
*** mordred has joined #openstack-ironic | 11:07 | |
*** romcheg has quit IRC | 11:07 | |
*** foexle has quit IRC | 11:09 | |
*** foexle has joined #openstack-ironic | 11:11 | |
*** lazy_prince has quit IRC | 11:23 | |
*** lazy_prince has joined #openstack-ironic | 11:23 | |
*** lucasagomes_ is now known as lucasagomes | 11:28 | |
*** romcheg has joined #openstack-ironic | 11:29 | |
*** lazy_prince has quit IRC | 11:30 | |
*** jistr has quit IRC | 11:32 | |
*** yuanying has quit IRC | 11:32 | |
*** yuanying has joined #openstack-ironic | 11:33 | |
*** takadayuiko has quit IRC | 11:35 | |
*** takadayuiko has joined #openstack-ironic | 11:37 | |
*** yuanying has quit IRC | 11:37 | |
takadayuiko | lucasagomes, are you here?? | 11:38 |
lucasagomes | takadayuiko, hello there, yes I am | 11:39 |
takadayuiko | lucasagomes, Hi! I have a question about your patch, can I ask? | 11:40 |
*** lucasagomes has quit IRC | 11:44 | |
*** Poornima has quit IRC | 11:48 | |
*** jistr has joined #openstack-ironic | 11:52 | |
*** jistr is now known as jistr|english | 11:54 | |
*** lucasagomes has joined #openstack-ironic | 11:54 | |
*** lucasagomes is now known as lucas-hungry | 11:54 | |
lucas-hungry | takadayuiko, my connection is terrible today! please leave the comment on the patch... I will take a look at it soon after lunch | 11:54 |
lucas-hungry | or here | 11:56 |
takadayuiko | mmm, in fact it's not about your patch... about some source comments which you wrote, OK, I'll send an e-mail to you! | 11:56 |
lucas-hungry | takadayuiko, oh, is it? ok... well u can ask here too | 11:57 |
lucas-hungry | I think my conn is back now | 11:57 |
lucas-hungry | back to normal (I hope) | 11:57 |
takadayuiko | OK OK | 11:57 |
takadayuiko | https://github.com/openstack-dev/devstack/blob/master/lib/ironic#L500 | 11:57 |
takadayuiko | Can we delete L.500-503 which you wrote? | 11:57 |
*** ndipanov has joined #openstack-ironic | 11:58 | |
takadayuiko | I've checked the BP(spec) https://blueprints.launchpad.net/ironic/+spec/add-node-instance-info but some patches related to this BP have been abandoned. | 11:58 |
lucas-hungry | takadayuiko, oh good one! yeah i think we are good now, we should include the pxe_deploy_kernel and pxe_deploy_ramdisk as part of the driver_info of the node directly | 11:59 |
lucas-hungry | instead of using the flavor | 11:59 |
*** dprince has joined #openstack-ironic | 11:59 | |
lucas-hungry | takadayuiko, yeah the one that touches that bit is this one here https://review.openstack.org/#/c/101864/4/ironic/nova/virt/ironic/patcher.py | 12:00 |
lucas-hungry | that's basically a backwards compat thing, if the flavor contains the deploy kernel and deploy ramdisk id we use it | 12:01 |
lucas-hungry | if not we assume the node have it in the driver_info | 12:01 |
* lucas-hungry brb 1 sec | 12:01 | |
lucas-hungry | I mean couple of minutes (is cooking) | 12:02 |
takadayuiko | hmm | 12:02 |
*** yuanying has joined #openstack-ironic | 12:03 | |
takadayuiko | thank you, I'll read codes and patches again! | 12:06 |
*** coolsvap is now known as coolsvap|afk | 12:07 | |
takadayuiko | I'm too hungry to think about... | 12:08 |
takadayuiko | I'm sorry to have kept you. | 12:09 |
*** linggao has joined #openstack-ironic | 12:30 | |
lucas-hungry | takadayuiko, no worries :) | 12:30 |
*** tatyana has joined #openstack-ironic | 12:30 | |
lucas-hungry | yeah so on devstack we can delete that lines and make it add the kernel/ramdisk to driver_info directly :) | 12:30 |
*** tatyana has joined #openstack-ironic | 12:30 | |
*** r-mibu has joined #openstack-ironic | 12:31 | |
*** r-mibu has left #openstack-ironic | 12:31 | |
*** tatyana has quit IRC | 12:31 | |
* NobodyCam wishes he was ot awake :-p | 12:38 | |
NobodyCam | s/ot/not :-p | 12:39 |
*** lucas-hungry is now known as lucasagomes | 12:44 | |
lucasagomes | NobodyCam, morning | 12:46 |
*** openstackgerrit has joined #openstack-ironic | 12:49 | |
NobodyCam | morning lucasagomes :) | 12:50 |
*** viktors|afk is now known as viktors | 12:50 | |
dtantsur | NobodyCam, morning, today is a hard day to wake up :( | 12:54 |
NobodyCam | :) morning dtantsur :) was easy to get up.. hard to sleep ..lol :-p | 12:55 |
dtantsur | oh, that doesn't sound good either | 12:55 |
NobodyCam | hehehehe | 12:56 |
*** pensu has quit IRC | 12:56 | |
vdrok | afternoon ironic! | 13:01 |
vdrok | there are some questions about https://review.openstack.org/#/c/125361/ | 13:02 |
vdrok | how do you think tests should be split up? | 13:02 |
vdrok | lucasagomes suggested including context in base TestCase, and db fixture and dbapi in DbTestCase | 13:05 |
*** hyakuhei has joined #openstack-ironic | 13:06 | |
lucasagomes | yup, well I think it make things easier... The context is lighweight afaict... so on tests you just need to decide between do I need access to db or not? and inherit from TestCase or DbTestCase | 13:06 |
*** jistr|english is now known as jistr | 13:07 | |
lucasagomes | instead of having 4 different classes: TestCase (no context, no db), ContextTestCase, DbTestCase and DbAndContextTestCase | 13:07 |
vdrok | lucasagomes, yup, I think it's a good idea, and context is separated from db fixture | 13:07 |
lucasagomes | yup | 13:07 |
*** ndipanov is now known as ndipanov_gone | 13:08 | |
hyakuhei | Hey all, I'm working to updates to our security notes, we have one about Nova Baremetal, currently it suggests that baremetal is only viable for test systems: http://lists.openstack.org/pipermail/openstack-dev/2013-July/011213.html | 13:10 |
hyakuhei | That seems a little stale... | 13:10 |
dtantsur | vdrok, ++ for lucasagomes | 13:12 |
hyakuhei | Can I get a very quick gerrit review to update it please? https://review.openstack.org/#/c/126203/1/security-notes/OSSN-0031 just point out anything that looks wrong and I'll make the changes. | 13:13 |
dtantsur | hyakuhei, hi. the problem is still in place and will probably be solved only in K | 13:13 |
openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Fix exceptions names and messages for Keystone errors https://review.openstack.org/126500 | 13:13 |
hyakuhei | Thanks dtantsur I'll update the OSSN. | 13:14 |
vdrok | lucasagomes, dtantsur, ok, thanks, will do it that way for now, and will wait for comments if someone comes up with something different | 13:14 |
*** bradjones has quit IRC | 13:15 | |
*** dprince has quit IRC | 13:15 | |
*** dprince has joined #openstack-ironic | 13:16 | |
lucasagomes | hyakuhei, dtantsur this are notes to Nova baremetal right? not ironic | 13:16 |
lucasagomes | so it won't be fixed in K | 13:16 |
hyakuhei | Yeah, sorry for the ambiguity. | 13:17 |
lucasagomes | the problem affects ironic as well, so in Ironic yes it may be fixed in K | 13:17 |
dtantsur | hyakuhei, yeah right, I was assuming you're talking about Ironic | 13:17 |
dtantsur | nova-bm will never be fixed | 13:17 |
hyakuhei | I'll modify the note. Mind if I ping you a review link once that's done? | 13:17 |
lucasagomes | hyakuhei, that's all good :) | 13:17 |
lucasagomes | feel free to ping me | 13:18 |
hyakuhei | Cheers | 13:19 |
*** igordcard has joined #openstack-ironic | 13:21 | |
*** r-daneel has joined #openstack-ironic | 13:23 | |
*** nosnos has quit IRC | 13:24 | |
*** zz_jgrimm is now known as jgrimm | 13:29 | |
openstackgerrit | A change was merged to openstack/ironic: Cleans up some Sphinx rST warnings in Ironic https://review.openstack.org/125273 | 13:31 |
*** kincl_ is now known as kincl | 13:35 | |
*** kincl has quit IRC | 13:35 | |
*** kincl has joined #openstack-ironic | 13:35 | |
*** Kui has quit IRC | 13:38 | |
openstackgerrit | A change was merged to openstack/ironic: Validate the power interface before deployment https://review.openstack.org/126523 | 13:39 |
*** rloo has joined #openstack-ironic | 13:47 | |
*** takadayuiko_ has joined #openstack-ironic | 13:47 | |
NobodyCam | brb | 13:48 |
*** killer_prince has joined #openstack-ironic | 13:48 | |
*** killer_prince is now known as lazy_prince | 13:48 | |
jroll | morning ironic :) | 14:04 |
jroll | are we going to try to get the whole verifiable boot chain in kilo? | 14:04 |
*** foexle has quit IRC | 14:05 | |
NobodyCam | good morning jroll | 14:13 |
jroll | heya NobodyCam :) | 14:13 |
NobodyCam | :) | 14:13 |
NobodyCam | jroll: verifiable boot chain == trusted boot? | 14:14 |
jroll | ye | 14:14 |
NobodyCam | I hope so :) | 14:14 |
jroll | nice | 14:14 |
lucasagomes | jroll, motning | 14:16 |
lucasagomes | morning* | 14:16 |
lucasagomes | yeah hope so :) | 14:16 |
* lucasagomes gotta finish the boot interface soon cause works are going to conflict | 14:17 | |
NobodyCam | worth puting it up on the spread sheet | 14:17 |
jroll | morning lucas :) | 14:17 |
*** bradjones has joined #openstack-ironic | 14:20 | |
* NobodyCam added to spread sheet fwtw :-p | 14:23 | |
dtantsur | jroll, morning | 14:25 |
jroll | morning :) | 14:25 |
dtantsur | brb | 14:25 |
rloo | morning, afternoon everyone! | 14:29 |
rloo | jroll: someone was asking me about IPA documentation. ?? | 14:30 |
jroll | hiya, rloo :) | 14:30 |
NobodyCam | good morning rloo :) | 14:30 |
* jroll hides | 14:30 | |
jroll | yeah, we need to get on that :/ | 14:30 |
jroll | rloo: for IPA itself or the agent driver in ironic? | 14:30 |
rloo | jroll: ok, do it soon, your fans are waiting. | 14:30 |
rloo | jroll: for IPA | 14:30 |
jroll | ok | 14:31 |
rloo | jroll: at least for the agent driver, there is the spec :-) | 14:31 |
jroll | we started on docstrings | 14:31 |
jroll | heh | 14:31 |
GheRivero | morning all | 14:31 |
rloo | hi GheRivero | 14:31 |
jroll | rloo: when is juno final? | 14:31 |
NobodyCam | ieek: http://www.bbc.com/news/world-europe-29514920 | 14:31 |
NobodyCam | thats a little close to the summit | 14:32 |
NobodyCam | morning GheRivero :) | 14:32 |
rloo | jroll: in 2 weeks I think? Oh, Oct 16. That's soon. | 14:32 |
jroll | D: | 14:32 |
jroll | yeah | 14:32 |
* jroll cries | 14:32 | |
rloo | jroll: about the juno final, or the ebola, or both? | 14:33 |
jroll | juno :P | 14:33 |
*** rakesh_hs has quit IRC | 14:33 | |
GheRivero | NobodyCam: nothing to worry.... yet | 14:34 |
NobodyCam | :) | 14:34 |
NobodyCam | phew | 14:34 |
* NobodyCam trusts GheRivero :) | 14:34 | |
*** dprince has quit IRC | 14:35 | |
*** jjohnson2 has joined #openstack-ironic | 14:35 | |
*** pcrews has joined #openstack-ironic | 14:36 | |
*** coolsvap|afk is now known as coolsvap | 14:37 | |
NobodyCam | ok heres a question for you operations type folks, if you were getting metrics for, oh say Ironic, would you want the output as statsd like or snmp like? | 14:38 |
lucasagomes | GheRivero, rloo morning | 14:38 |
rloo | hi lucasagomes | 14:38 |
jroll | NobodyCam: we use statsd in production :) | 14:38 |
NobodyCam | with some log store backend like log stash? | 14:40 |
NobodyCam | or stright into cati, graphite , ect... | 14:41 |
jroll | yeah, into graphite | 14:41 |
jroll | right now it's just API call timings | 14:42 |
jroll | but we want to add more | 14:42 |
jroll | there's a spec for this... | 14:42 |
jroll | https://review.openstack.org/#/c/100729/ | 14:43 |
NobodyCam | question comes from : https://review.openstack.org/#/c/100729/4/specs/juno/add-metrics.rst | 14:43 |
jroll | needs rebase though | 14:43 |
jroll | aha, yeah | 14:43 |
jroll | NobodyCam: aweeks is on our team, fwiw :) | 14:44 |
NobodyCam | side question: have we talked about what to do with the old specs? | 14:44 |
rloo | NobodyCam: which old specs? | 14:45 |
jroll | they need to be rebased and match the new template | 14:45 |
jroll | or do you mean approved specs or? | 14:45 |
NobodyCam | rloo: https://review.openstack.org/#/q/status:open+project:openstack/ironic-specs,n,z | 14:45 |
rloo | speaking of specs, I just came across this one, can't figure out what the status is: https://review.openstack.org/#/c/95025/ | 14:46 |
lucasagomes | NobodyCam, owners should adapt and submit it again for kilo | 14:46 |
lucasagomes | adapt == adhere to the new kilo spec | 14:46 |
jroll | yep | 14:46 |
jroll | or abandon | 14:46 |
rloo | NobodyCam: what lucasagomes said. | 14:46 |
NobodyCam | lucasagomes: can / should we move the old juno specs to a new folder or such | 14:46 |
lucasagomes | NobodyCam, hmm I don't see why, the old ones are in the "juno" folder | 14:47 |
lucasagomes | new ones will go to the "kilo" folder | 14:47 |
lucasagomes | so it's fine to leave as-is I think | 14:47 |
rloo | hmm. Whoever -2'd those spec patches (JayF?) needs to un -2 them. | 14:47 |
rloo | dtantsur, JayF: I randomly looked at a couple of spec patches, and you -2'd them. Kilo is open now, so whichever ones you -2'd, I think you can remove. | 14:49 |
lucasagomes | we also need to remember that for specs that still not clear whether it belongs to ironic or is aligned to the project goals we need to submit a small first, and then once we have an agreement the owner should submit the full one | 14:50 |
jroll | right | 14:50 |
jroll | so, people don't *need* to do that, but they can to avoid extra work | 14:50 |
NobodyCam | rloo: ohh | 14:50 |
lucasagomes | jroll, yeah | 14:50 |
rloo | NobodyCam: ohh? | 14:51 |
jroll | lucasagomes: I think we need to make that very clear to folks | 14:51 |
NobodyCam | that spec from nova | 14:51 |
lucasagomes | jroll, yeah, true | 14:51 |
lucasagomes | jroll, +1 | 14:51 |
rloo | NobodyCam: yeah. Not sure what that spec from nova means... | 14:51 |
*** chuckC_ has quit IRC | 14:52 | |
lucasagomes | rloo, can't fully remember what was agreed too... but seems that with the merge of the ironic driver nova bm is already "deprecated" | 14:53 |
lucasagomes | meaning that no new features can be added to it, only bug fixes | 14:53 |
lucasagomes | and soon should be removed from the nova tree | 14:54 |
* lucasagomes might be wrong tho | 14:54 | |
lucasagomes | soon = next cycle | 14:54 |
rloo | lucasagomes: yeah. do not sure what to do with that proposed spec. Hopefully devananda knows ;) | 14:55 |
lucasagomes | yeah | 14:55 |
lucasagomes | rloo, probably we gotta get that merged this cycle | 14:56 |
NobodyCam | gah all kinds of competing calls this morning :-p | 14:57 |
* lucasagomes just got the first machine booted with the new boot interface! | 14:58 | |
lucasagomes | \o/ | 14:58 |
rloo | lucasagomes: nice! | 14:59 |
lucasagomes | rloo, ... so the way I'm thinking about it ^ is | 14:59 |
jroll | wow, nice one | 14:59 |
lucasagomes | the deploy interface should have control over the boot interface right? | 14:59 |
rloo | wrt 126500, can we just rename an exception? Does that break backwards compatibility? | 15:00 |
lucasagomes | booting and controlling power, is part of the deploy | 15:00 |
lucasagomes | so the invokation for boot() a machine comes from a deploy interface | 15:00 |
lucasagomes | does this logic sounds correct to u guys? | 15:00 |
lucasagomes | jroll, NobodyCam dtantsur ^ | 15:00 |
lucasagomes | everyone really ^ | 15:01 |
rloo | lucasagomes: deploy shouldn't be the only place where power is controlled, but I'm not sure that's what you meant. | 15:01 |
jroll | lucasagomes: yeah, that sounds reasonable... | 15:01 |
lucasagomes | rloo, oh yeah no... for e.g for PXE the boot() cause the node to power | 15:02 |
jroll | lucasagomes: though should allow for boot() and shutdown() | 15:02 |
*** chuckC_ has joined #openstack-ironic | 15:02 | |
lucasagomes | power on* | 15:02 |
devananda | morning, all | 15:02 |
jroll | morning deva :) | 15:02 |
lucasagomes | devananda, morning | 15:02 |
rloo | morning devananda | 15:02 |
*** pensu has joined #openstack-ironic | 15:02 | |
NobodyCam | lucasagomes: I can agree with that logic | 15:02 |
NobodyCam | logic | 15:03 |
jroll | lucasagomes: and how do you see pxe/disk working with this? | 15:03 |
NobodyCam | morning devananda | 15:03 |
lucasagomes | devananda, ^ ur input re boot interface there too please | 15:03 |
lucasagomes | jroll, disk ? iscsi? | 15:03 |
jroll | lucasagomes: like... boot from disk | 15:03 |
jroll | whole disk images | 15:03 |
jroll | agent driver, etc | 15:03 |
devananda | lucasagomes: re: nva.virt.baremetal, afaik it will be deleted soon. I'll check. | 15:03 |
*** pensu has quit IRC | 15:03 | |
*** rushiagr is now known as rushiagr_away | 15:04 | |
Shrews | rloo: re: 126500... i was looking at that too. i don't see how it would break backward compatibility. | 15:04 |
devananda | re: boot, power, deploy ... these interfaces are able to call to each other through the driver interface (but not directly) | 15:04 |
rloo | Shrews: only if the CatalogExc* or whatever was exposed, and someone coded for it. | 15:05 |
devananda | hypothetical example: code in drivers.modules.pxe.PXEBoot should not call directly to drivers.modules.iscsi.iSCSIDeploy -- instead it should use task.node.driver.deploy | 15:05 |
lucasagomes | jroll, ah right... so I don't see much where it changes... the deploy() method for the agent just update some ports in neutron, set the boot device and then power on the node | 15:06 |
lucasagomes | this is what the boot interface does for pxe | 15:06 |
lucasagomes | so in the deploy() of the agent you simply would call driver.boot.boot() | 15:06 |
lucasagomes | because doing all that is related to the way it boots not the way it deploys, the deploy works after you boot the node | 15:06 |
jroll | lucasagomes: right... but with whole disk images, you don't need to update ports in neutron, and boot device is set differently | 15:07 |
jroll | but you still might pxe boot the ramdisk | 15:07 |
jroll | so deploy boot and instance boot are different, might need different drivers? | 15:07 |
lucasagomes | devananda, yes! | 15:07 |
devananda | jroll: or same driver but state awareness | 15:07 |
Shrews | rloo: those aren't used beyond the changed files, that i've seen | 15:07 |
jroll | devananda: right, or that | 15:07 |
rloo | Shrews: so eg what if someone has their own driver that isn't in the ironic tree. | 15:08 |
lucasagomes | devananda, as well the PXEDeploy (which should maybe be called iSCSIDeploy) should not call pxe related functions but driver.boot... | 15:08 |
devananda | we really need to decouple (disk structure=[partition|whole]) and (boot device=[disk|pxe|other]) this cycle | 15:08 |
devananda | which seems like it is a consideration in how you're refactoring deploy interface | 15:08 |
devananda | lucasagomes: correct | 15:09 |
Shrews | rloo: IMO, exceptions aren't an agreed to interface, like an API. | 15:09 |
jroll | devananda: right... I'm less asking how this will work and more making sure it's considered :) | 15:09 |
lucasagomes | jroll, I haven't start playing with the agent driver to decouple it yet... I started with the PXE one which I'm more familiar with | 15:10 |
rloo | Shrews: so our drivers try to catch specific exceptions from 3rd party libraries. If those exceptions changed, that'd break our code. I don't want to do a similar thing to our users. | 15:10 |
lucasagomes | I have looked the code of the other drivers ofc to see if I'm not missing anything | 15:10 |
lucasagomes | I will play with it soonish | 15:11 |
jroll | lucasagomes: right, I'm just saying think about it as you design :) | 15:11 |
lucasagomes | jroll, will do :) | 15:11 |
Shrews | rloo: valid point, though not sure how much we need to guard against that. what other drivers are outside our tree? | 15:11 |
rloo | Shrews: how would I know? It ain't in our tree. Whether we need to worry about them or not, is the question. | 15:12 |
jroll | Shrews: there could be thousands for all we know :P | 15:12 |
rloo | Shrews: we can alias the exceptions. | 15:12 |
devananda | Shrews: exceptions are an expected return code from a method. So they are part of the API. | 15:12 |
Shrews | rloo: ooh, i like the alias idea | 15:12 |
devananda | Shrews: at least our code documents what exceptions it expects to raise ... | 15:12 |
jroll | as someone who used to run ironic with an out of tree driver, I think we shouldn't break them :) | 15:13 |
lucasagomes | Shrews, hmm I think we should try to not break the drivers out of the tree | 15:13 |
rloo | Shrews, devananda, jroll: ok, we'll alias the exceptions then. thx. (something to remember for future patches). | 15:13 |
*** kylestev has joined #openstack-ironic | 15:14 | |
devananda | rloo: we need to be considerate of out-of-tree drivers | 15:14 |
lucasagomes | this is something I will have to deal with this boot interface too... tho I want it to be a *core* interface we may want to consider having it as an optional interface for this cycle | 15:14 |
lucasagomes | and promote it later to give people time to adjust their drivers | 15:14 |
rloo | devananda: do you know of a list of things to look out for, wrt maintaining backwards compatibility? | 15:15 |
lucasagomes | devananda, ^ sounds correct? | 15:15 |
devananda | lucasagomes: I think it's worth creating a stubbed-out driver that implements the current interfaces which we /dont/ update, just to test ourselves | 15:15 |
*** takadayuiko_ has quit IRC | 15:15 | |
lucasagomes | devananda, yeah... well we can get the let's say old pxe one... name it different and enable it and see if that still works | 15:17 |
devananda | lucasagomes: so, this also kinda supports something else I started thinking about recently | 15:17 |
devananda | within the big-tent discussion and nova splitting out their drivers | 15:17 |
devananda | rather than take our drivers out of tree (there are a lot, and they're small, and I don't think we need to yet) | 15:18 |
devananda | what if we took the driver base class and made that a separate package | 15:18 |
devananda | then imported it | 15:18 |
devananda | so we could properly version that interface | 15:18 |
lucasagomes | hmm that's not a bad idea | 15:18 |
lucasagomes | (and just to say, I'm a bit against separating the drivers from the tree) | 15:18 |
* devananda shamelessly borrows mordred's idea | 15:18 | |
lucasagomes | I just think that nova does more than it should | 15:18 |
lucasagomes | IMO nova should splitt all that creating instances, manage flavors etc in another project | 15:19 |
russell_h | does versioning the base class actually solve a problem? | 15:19 |
lucasagomes | and be something like ironic is but for hypervisors | 15:19 |
devananda | this would let us, for example, test with both juno/stable version of the driver interface and kilo/master | 15:19 |
lucasagomes | that would allow them to keep drivers in tree cause it's now not a gigantic code | 15:19 |
lucasagomes | devananda, I like the idea, gotta think more about it | 15:19 |
devananda | russell_h: I think it would solve the "oops, we didn't realize we broke backwards compat" problem | 15:19 |
lucasagomes | but it sounds good | 15:19 |
russell_h | ah I see what you mean | 15:20 |
lucasagomes | kinda like the RPC objects interface we try to maintain | 15:20 |
russell_h | more explicitly version the base class | 15:20 |
devananda | lucasagomes: right. that's all unrelated to the actual abstraction of the hypervisor. | 15:20 |
lucasagomes | s/try/do maintain | 15:20 |
jroll | devananda: but if the code in ironic that interacts with the drivers breaks backward compat, then what? | 15:20 |
lucasagomes | devananda, exactly... well but it's outside the conv topic | 15:20 |
devananda | lucasagomes: indeed | 15:20 |
jroll | devananda: presumably if you change a method signature in the base class, you're going to call that differently from conductor.manager or whatever, no? | 15:21 |
devananda | jroll: sure. but we can easily unit test conductor/manager against both current version of the base class package and any prior release of it | 15:21 |
devananda | jroll: we can't do that when they're in the same tree, tho | 15:22 |
devananda | s/we can easily/we could, if we released base class as a package, easily/ | 15:22 |
jroll | hmm | 15:22 |
russell_h | is the objective to make a really stable driver interface, or to build a system where you can rev it all the time really easily? | 15:24 |
russell_h | assuming the former, you could write some unit tests that:1 | 15:24 |
devananda | russell_h: objective is two-fold: assure taht we are backwards compatible, give out-of-tree driver maintainers a stable interface to code against | 15:25 |
russell_h | 1) validate the heck out of ConductorManager's use of the driver interface | 15:25 |
devananda | there's a third objective too: set an example for other projects (eg, Nova) | 15:25 |
russell_h | 2) validate an individual driver's implementation of the interface | 15:25 |
russell_h | right | 15:26 |
devananda | as all of openstack moves to a less tightly-coupled gate system | 15:26 |
devananda | russell_h: even using unit tests to enforce that, it wouldn't give us versioning | 15:26 |
russell_h | right, you'd basically need to copy/paste the unit tests every time you rev it | 15:27 |
russell_h | or try to factor things nicely to avoid that | 15:27 |
devananda | eg, if I'm maintaining an out of tree driver (which I'm not, fwiw) I could write it against v1.x of the base class and know that it'll work with Ironic as long as Ironic doesn't require >=2.x of the base class | 15:27 |
lucasagomes | the way I see it, is very similar to the RPC objects interface we have, with versioning and all... the base class lives in a diff project (right now nova, but once oslofied will be in oslo) | 15:28 |
russell_h | mm, right | 15:28 |
russell_h | there is also the reverse problem that we dealt with for a while | 15:28 |
russell_h | of how you test an out of tree driver | 15:28 |
devananda | lucasagomes: not really. the RPC interface is micro versioned for each method call,a nd can discover based on producer/consumer what methods are supported where | 15:29 |
devananda | lucasagomes: it's designed to handle a rolling upgrade of a large system where different RPC producers/consumers might be at different versions | 15:29 |
devananda | lucasagomes: we're not importing a SemVer'd class for the RPC code. On the API and Conductor side, we define what version of the API they each implement, and they negotiate which methods can be called | 15:30 |
lucasagomes | devananda, right and with a versioned base driver class, how it would work? let's say the version the driver written was based on is diff than the present base class? so we would say that the driver can't be enabled because the versions differ? | 15:31 |
russell_h | I guess the general approach I'm advocating is to try to think of the driver interface as a public API to Ironic, very much like the REST API | 15:31 |
*** Hefeweizen has joined #openstack-ironic | 15:31 | |
russell_h | write tests to validate that iterface, and if you _ever_ have to change those tests you probably need to rev the API | 15:31 |
russell_h | (or at least give a lot of scrutiny to why those tests are changing) | 15:31 |
russell_h | not that revving REST APIs is something people are especially good at | 15:32 |
* NobodyCam watches the ALL hands meeting | 15:32 | |
*** k4n0 has quit IRC | 15:33 | |
mordred | russell_h, devananda: I think if you did it right, you could get BOTH a ensuring backwards compat and facilitation of reving the interface all the time | 15:34 |
mordred | because as long as you rev the interface in a backwards compatible manner, it's fine | 15:34 |
russell_h | right | 15:35 |
russell_h | I was thinking more of REST APIs where we suck at semver | 15:36 |
mordred | well, yeah | 15:36 |
russell_h | so when we make a backwards compatible change we just claim it doesn't require a version bump | 15:36 |
russell_h | feels good :) | 15:36 |
mordred | have I mentinoed that everythign about api versioning in openstack makes me want to kill bunnies? | 15:37 |
mordred | it CERTAINLY does not help end users deal with things | 15:37 |
jroll | I've heard throw wet cats, same idea I guess | 15:37 |
NobodyCam | morning mordred ... | 15:37 |
mordred | morning NobodyCam | 15:37 |
* NobodyCam pick up all the sharp things | 15:37 | |
NobodyCam | s/pick/picks/ | 15:38 |
mordred | jroll: if you can't kill bunnies by throwing wet cats, then really what good are you? | 15:38 |
jroll | :D | 15:39 |
*** igordcard has quit IRC | 15:39 | |
rloo | Shrews et al. I added a note about changing exception names, to https://wiki.openstack.org/wiki/Ironic/Developer_guidelines#Guidelines_for_Developers | 15:44 |
Shrews | rloo: ++. thx | 15:45 |
arif-ali | There is an issue with the RC release of ironic that has a missing dependancy of pbr; is that a known issue? | 15:47 |
devananda | arif-ali: no. bug # ? | 15:47 |
devananda | arif-ali: our requirements.txt file contains: pbr>=0.6,!=0.7,<1.0 | 15:48 |
arif-ali | I installed using python setup.py | 15:48 |
arif-ali | and that didn't install it, so had to manually install pbr by using pip | 15:48 |
devananda | arif-ali: if you're using python setup.py, you should first do "pip install -r requirements.txt" | 15:49 |
*** Haomeng|2 has joined #openstack-ironic | 15:49 | |
arif-ali | ah, thanks, I'm new to python and ironic, so testing things out | 15:49 |
devananda | :) | 15:50 |
*** Haomeng has quit IRC | 15:50 | |
kylestev | I like that the majority of ironic folks are genuinely helpful. I guess that's one of the ways you keep an open source project alive, though :P | 15:54 |
NobodyCam | :) | 15:54 |
kylestev | s/folks/folk/ | 15:55 |
jroll | maybe we're just desperate for users :P | 15:55 |
NobodyCam | :-p | 15:55 |
arif-ali | I've found the general OpenStack community up until now have been quite helpful, so my experience has been really good | 15:55 |
kylestev | hehe | 15:55 |
jroll | agree | 15:55 |
kylestev | arif-ali: good to hear :) | 15:56 |
arif-ali | Ironic is one project that does deffo interest me quite a bit due to HPC | 15:56 |
kylestev | I haven't interacted with many other projects, unfortunately. | 15:56 |
jroll | I love the HPC use case for ironic | 15:56 |
kylestev | hpc being high performance computing? | 15:56 |
arif-ali | So will be testing in anger over the next few weeks | 15:56 |
jroll | lol | 15:56 |
arif-ali | kylestev, correct | 15:56 |
jroll | kylestev: pq's cassandra cluster was similar, just different workload | 15:57 |
jroll | boot a metric crapton of servers and distribute a workload | 15:57 |
arif-ali | xCAT is already working it's way for an ironic driver, and that is something I use quite heavily in a lot of our projects | 15:57 |
kylestev | heh. I spent a week or so at my last internship (focusing on hpc) trying to convince people to use baremetal :o | 15:58 |
devananda | arif-ali: what does an "ironic driver for xcat" look like? | 15:58 |
*** jjohnson2_ has joined #openstack-ironic | 15:59 | |
arif-ali | devananda, the sourceforge website is not letting me go through source code, here's the github mirror https://github.com/xcat-org/xcat-core/tree/master/xCAT-OpenStack-baremetal | 16:00 |
arif-ali | I don't really know the status of it, but once I have the base ironic driver working, I will be testing it | 16:01 |
devananda | arif-ali: this is an implementation of nova.virt.driver: https://github.com/xcat-org/xcat-core/blob/master/xCAT-OpenStack-baremetal/lib/python/xcat/openstack/baremetal/driver.py | 16:02 |
*** jjohnson2 has quit IRC | 16:02 | |
devananda | very very minimal and probably non-functional implementation, fwiw | 16:02 |
*** jbjohnso__ has joined #openstack-ironic | 16:02 | |
arif-ali | ah, ok, the git messages seemed to suggest ironic | 16:02 |
devananda | I haven't yet found any ironic-related code in there ... | 16:03 |
jroll | it appears to be nova-bm | 16:03 |
arif-ali | ah, here it is https://github.com/xcat-org/xcat-core/tree/master/xCAT-OpenStack-ironic/ironic_baremetal | 16:03 |
jroll | so maybe the goal is to get the... oh | 16:03 |
devananda | arif-ali: ah, yep | 16:04 |
jroll | neat | 16:04 |
*** jjohnson2_ has quit IRC | 16:05 | |
devananda | https://github.com/xcat-org/xcat-core/tree/master/xCAT-OpenStack-ironic/ironic_baremetal/ironic/drivers/modules | 16:05 |
devananda | there we are. third-party out of tree driver. | 16:05 |
devananda | woot | 16:05 |
NobodyCam | nice | 16:05 |
devananda | I think that's grand. I also think they should just contribute that to Ironic. | 16:05 |
jroll | "just" | 16:06 |
kylestev | devananda: jroll: lol | 16:06 |
jroll | :P | 16:06 |
arif-ali | Is that the first driver out there, or are there any others? | 16:07 |
rloo | Shrews: look, an out of tree driver that uses exception.CatalogUnauthorized! | 16:08 |
jroll | arif-ali: it's the only driver that is still out of tree, that we know of | 16:08 |
jroll | arif-ali: the agent driver was run out of tree for a while, it's in tree now | 16:08 |
devananda | I suspect there are a few others | 16:09 |
jroll | I'm sure there are, we just don't know about them | 16:09 |
jroll | or I don't | 16:09 |
Shrews | rloo: ok | 16:09 |
*** ndipanov_gone has quit IRC | 16:12 | |
JayF | rloo: ahh yes, I'll take all those -2s and change them to -1s (pending the proposal being moved into kilo/) | 16:13 |
rloo | JayF: thx | 16:13 |
devananda | JayF: cheers | 16:15 |
*** Ng has quit IRC | 16:22 | |
*** aweeks has quit IRC | 16:22 | |
*** coolsvap is now known as coolsvap|afk | 16:23 | |
*** rushiagr_away is now known as rushiagr | 16:23 | |
JayF | Looks like I got them all, please ping me if I missed any :) | 16:23 |
*** igordcard has joined #openstack-ironic | 16:27 | |
*** ifarkas has quit IRC | 16:28 | |
*** jistr has quit IRC | 16:29 | |
*** marcoemorais has joined #openstack-ironic | 16:29 | |
*** dprince has joined #openstack-ironic | 16:34 | |
*** dtantsur is now known as dtantsur|afk | 16:35 | |
NobodyCam | ahh hands call done | 16:36 |
JayF | Does that mean you know if you'll be working for the H or the P yet? | 16:37 |
JayF | </possibly_bad_joke> | 16:37 |
NobodyCam | lol | 16:37 |
rloo | as long as NobodyCam still works in Ironic, I'm happy ;) | 16:38 |
NobodyCam | hehehe thats not changing at all.. | 16:38 |
rloo | :D | 16:39 |
NobodyCam | :) | 16:41 |
*** romcheg has quit IRC | 16:43 | |
*** zigo has quit IRC | 16:43 | |
*** zigo has joined #openstack-ironic | 16:46 | |
*** pensu has joined #openstack-ironic | 16:51 | |
*** derekh has quit IRC | 16:53 | |
NobodyCam | oh devananda ... question.. rloo found this: https://review.openstack.org/#/c/95025 is it still a thing didn't we handle this? | 16:54 |
* devananda looks | 16:54 | |
devananda | lol | 16:54 |
devananda | so we did all that, even if Nova never landed the spec | 16:55 |
devananda | yay process | 16:55 |
NobodyCam | heheheh :-p I thought we did that | 16:55 |
*** aweeks has joined #openstack-ironic | 16:57 | |
*** harlowja_away is now known as harlowja | 16:58 | |
*** Ng_ has joined #openstack-ironic | 16:59 | |
*** todd_dsm has joined #openstack-ironic | 17:01 | |
*** dlaube has joined #openstack-ironic | 17:04 | |
*** viktors is now known as viktors|afk | 17:05 | |
*** rushiagr is now known as rushiagr_away | 17:09 | |
openstackgerrit | Tom Hancock proposed a change to openstack/ironic: Switch default partition type to gpt https://review.openstack.org/126625 | 17:12 |
*** praneshp has joined #openstack-ironic | 17:14 | |
NobodyCam | hey jbjohnso__ you around? | 17:19 |
devananda | adam_g: do you have a patch up yet to make our pxe_ssh tests voting on other projects? | 17:22 |
devananda | adam_g: if not, I'm about to post one | 17:22 |
lucasagomes | aight I will call it a day here | 17:25 |
lucasagomes | have a good night everybody! | 17:25 |
devananda | NobodyCam: see https://review.openstack.org/#/c/126560/2 :) | 17:27 |
jbjohnso__ | NobodyCam, hello | 17:28 |
*** jbjohnso__ is now known as jjohnson2 | 17:28 | |
adam_g | devananda, haven't yet | 17:29 |
jjohnson2 | NobodyCam, what is up? | 17:30 |
devananda | adam_g: posted just now | 17:32 |
NobodyCam | hey hey jjohnson2 | 17:32 |
NobodyCam | I just had a off the wall question. | 17:32 |
NobodyCam | would you happen to know of a good doc in the ipmi hex codes? I was playing with standing up a port listener that would accept ipmi commands and then do *SOMETHING* based on the command it recived | 17:34 |
*** lucasagomes is now known as lucas-dinner | 17:35 | |
jjohnson2 | NobodyCam, like the well defined behaviors or arbitrary hex data? | 17:35 |
jjohnson2 | NobodyCam, I've had multiple requests to implement the converse side of the generic ipmi conversation | 17:36 |
NobodyCam | the defined behaviors. | 17:36 |
*** lucas-dinner has quit IRC | 17:36 | |
NobodyCam | I doing someting like: http://paste.openstack.org/show/0ntlCoCwTXfqFkbxvLpW/ | 17:36 |
*** lucasagomes has joined #openstack-ironic | 17:36 | |
jjohnson2 | yeah, so I have some ideas in terms of backend handlers under friendly names | 17:36 |
*** lucasagomes is now known as lucas-dinner | 17:36 | |
NobodyCam | night lucas-dinner | 17:37 |
lucas-dinner | NobodyCam, g'night :) | 17:37 |
jjohnson2 | oh, what is that all about? | 17:37 |
jjohnson2 | the repr of binary strings is a pain to discern... | 17:38 |
*** jcoufal has quit IRC | 17:42 | |
*** harlowja_ has joined #openstack-ironic | 17:42 | |
*** harlowja has quit IRC | 17:42 | |
*** andreykurilin_ has joined #openstack-ironic | 17:43 | |
*** rushiagr_away is now known as rushiagr | 17:46 | |
*** lucas-dinner has quit IRC | 17:47 | |
*** athomas has quit IRC | 17:49 | |
openstackgerrit | Ruby Loo proposed a change to openstack/ironic: KeyError from AgentVendorInterface._heartbeat() https://review.openstack.org/126636 | 17:49 |
rloo | did we decide that we were definitely not going to provide an API to change the maintenance mode of a node? https://review.openstack.org/#/c/125942/ | 18:00 |
JayF | We already do via the public API fwiw | 18:00 |
JayF | that's about RPC apiu | 18:00 |
JayF | (and no, you can't remove it from the public API, I use it too much :P) | 18:01 |
rloo | JayF: how do we provide it from the public API? That patch is removing it. I don't see any hooks in API code. | 18:01 |
*** lucasagomes has joined #openstack-ironic | 18:01 | |
JayF | rloo: I run | 18:01 |
jroll | rloo: ironic node-update uuid replace maintenance=True | 18:01 |
JayF | ironic node-update {UUID} replace maintenance=true # or false | 18:01 |
*** jcoufal has joined #openstack-ironic | 18:02 | |
JayF | jroll: it's slightly more upsetting when you finish my sentences on IRC /without/ sitting beside me :P | 18:02 |
rloo | JayF, jroll: thx. [Yes, I will vouch that you are two different people.] | 18:02 |
jroll | hehehe | 18:02 |
JayF | Hehehe | 18:02 |
JayF | Anyone looking for a job working on Ironic? Requirements: Name must start with a J (Note: We will /consider/ candidates with IRC nicknames that start with J even if their name doesn't) | 18:03 |
JayF | :P | 18:03 |
jroll | lol | 18:03 |
rloo | no, I think 3 J's is just right. (3 bears, 3 muskateers, 3 stooges,...) | 18:04 |
NobodyCam | lol | 18:06 |
jroll | lol | 18:09 |
*** andreykurilin_ has quit IRC | 18:13 | |
*** lucasagomes has quit IRC | 18:25 | |
NobodyCam | brb ... quick walkies time.. :) | 18:31 |
*** lucasagomes has joined #openstack-ironic | 18:39 | |
*** romcheg has joined #openstack-ironic | 18:41 | |
*** romcheg has quit IRC | 18:42 | |
openstackgerrit | Sirushti Murugesan proposed a change to openstack/ironic-specs: Whole Disk Image Support for PXE Deploy Driver https://review.openstack.org/97150 | 18:53 |
openstackgerrit | Sirushti Murugesan proposed a change to openstack/ironic: Adds support for deploying whole disk images https://review.openstack.org/73054 | 18:53 |
*** rushiagr is now known as rushiagr_away | 18:55 | |
NobodyCam | JayF: thank you for -1 and not just unblocking the specs :-) | 18:57 |
sirushti | Hi, could anyone take a look at these^? It'd be super helpful :-) | 18:57 |
NobodyCam | sirushti: will do in a bit | 18:58 |
sirushti | NobodyCam, thanks alot, I've also posted a WIP patch if you want to take a look at that too even though the spec isn't approved | 18:58 |
NobodyCam | :) yep I just saw that | 18:59 |
*** andreykurilin_ has joined #openstack-ironic | 18:59 | |
*** Ng_ is now known as Ng | 19:02 | |
NobodyCam | wb Ng :-p | 19:03 |
*** lucasagomes has quit IRC | 19:03 | |
Ng | :) | 19:03 |
NobodyCam | sirushti: just a question. is there a clean way to check that a whole disk image is infact not larger than the nodes phy. disk? | 19:07 |
*** todd_dsm has quit IRC | 19:08 | |
sirushti | NobodyCam, my assumption is that the flavor will have the nodes phy. disk size and we check in deploy_utils.check_image_size? | 19:10 |
sirushti | NobodyCam, is it not clean in that the disk_size ~ root_mb naming assumptions confusing? | 19:11 |
sirushti | NobodyCam, *are confusing? | 19:11 |
openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic-python-agent: Updated from global requirements https://review.openstack.org/114058 | 19:12 |
*** lucasagomes has joined #openstack-ironic | 19:15 | |
lifeless | NobodyCam: qemu-img an tell you, for qcow2 images. | 19:16 |
jroll | doesn't glance also store the size? | 19:17 |
jroll | oh I guess if we have to download it anyway, just check it ourselves | 19:17 |
* jroll doesn't like ironic downloading images :/ | 19:17 | |
lifeless | glance has the compressed size | 19:21 |
lifeless | last I checked | 19:21 |
jroll | mmm | 19:21 |
lifeless | should be able to annotate it in glance | 19:21 |
NobodyCam | ya I think glance is compressed size | 19:21 |
NobodyCam | thank you :) | 19:21 |
*** HenryG has quit IRC | 19:28 | |
JayF | NobodyCam: well, I figured being in the wrong dir was well worth a -1 :P | 19:28 |
NobodyCam | :) | 19:29 |
NobodyCam | anyone seen this one before? http://logs.openstack.org/50/97150/13/check/gate-ironic-specs-docs/957f927/console.html#_2014-10-07_18_57_52_957 | 19:34 |
*** lucasagomes has quit IRC | 19:36 | |
sirushti | NobodyCam, I did this http://paste.openstack.org/show/119471/ and it went away | 19:40 |
rloo | NobodyCam: that's 'normal'. | 19:40 |
rloo | NobodyCam: the sphinx error. oh wait. that's for ironic-specs. hmm... | 19:41 |
NobodyCam | rloo: ya | 19:41 |
sirushti | NobodyCam, also had to add a separate kilo folder | 19:41 |
NobodyCam | sirushti: the other option I see is to add :orphan: to th template | 19:41 |
NobodyCam | sirushti: yea | 19:42 |
rloo | NobodyCam: I thought someone had opened kilo for specs. There should be a kilo/ subdir. | 19:43 |
sirushti | NobodyCam, but then we'd have to have :orphan: in every rst file then | 19:43 |
NobodyCam | rloo: I think it not there because we havent landed anything for it. | 19:44 |
NobodyCam | sirushti: true... | 19:44 |
*** bradjones has quit IRC | 19:44 | |
JayF | Do we need to land specs/kilo/.gitkeep? hehe | 19:44 |
NobodyCam | :-p | 19:44 |
*** bradjones has joined #openstack-ironic | 19:46 | |
NobodyCam | brb | 19:48 |
*** romcheg has joined #openstack-ironic | 19:54 | |
*** achanda has joined #openstack-ironic | 19:54 | |
rloo | NobodyCam, sirushti: yes, the change to doc/source/index.rst is needed. | 19:57 |
NobodyCam | guy who is building a dog run is here... rloo you putting that patch up? :-p | 19:59 |
rloo | NobodyCam: I'm not sure a patch with that change will work. I suppose I could try. I got an error because there were no files under specs/kilo/*. But I don't know what kind of errors are 'Jenkins' errors vs sphinx errors. | 20:00 |
sirushti | rloo, yeah, sphinx complains if the kilo directory isn't present | 20:02 |
*** todd_dsm has joined #openstack-ironic | 20:02 | |
*** jjohnson2 has quit IRC | 20:03 | |
romcheg | Hey guys! | 20:03 |
rloo | NobodyCam: it didn't build any .html for me, so it probably won't work w/o a file in specs/kilo/* | 20:03 |
rloo | hi romcheg | 20:03 |
romcheg | looks like both Dmitry and Lucas are away now | 20:03 |
romcheg | Morning rloo! | 20:03 |
rloo | romcheg: what time zone are you in? :) | 20:04 |
romcheg | CEST | 20:04 |
NobodyCam | hey romcheg :) | 20:04 |
rloo | it is afternoon in california already | 20:04 |
romcheg | 22:04 here | 20:04 |
romcheg | rloo: It's just a habit to always say "morning" here :) | 20:05 |
NobodyCam | hehehehehe | 20:05 |
rloo | romcheg: ha ha | 20:05 |
NobodyCam | me too | 20:05 |
*** HenryG has joined #openstack-ironic | 20:06 | |
romcheg | bbiab | 20:08 |
*** harlowja_ is now known as harlowja_away | 20:09 | |
romcheg | and back | 20:22 |
NobodyCam | wb | 20:22 |
romcheg | could someone please take a look at https://review.openstack.org/#/c/126265/2/ironic/common/context.py | 20:23 |
romcheg | I think about that show_deleted in the context | 20:23 |
romcheg | as I understand that is from Nova for telling whether or not a request should return deleted items | 20:24 |
romcheg | but we do not use soft deletion so I think that attribute should be hidden | 20:24 |
romcheg | along with everything else that is not supported in Ironic | 20:24 |
rloo | romcheg: I haven't looked into this but I agree with lucas that it would be confusing to see it in the to_dict() but not in the init. does it hurt to pass it even if we don't use it? | 20:26 |
*** pensu has quit IRC | 20:27 | |
romcheg | rloo: I was actually thinking on whether I should remote it from to_dict() or add it back to __init__() | 20:27 |
rloo | romcheg: is that the *only* parameter that ironic doesn't use? | 20:28 |
romcheg | rloo: It's not about not using, it's about not supporting | 20:29 |
rloo | romcheg: is that the *only* parameter that ironic doesn't support? | 20:29 |
romcheg | rloo: It looks so | 20:30 |
rloo | romcheg: ok, i think the answer is easier. I looked at the commit to see what the patch was about. | 20:30 |
rloo | romcheg: if it is to update to use newer version of policy, and if we already had show_deleted before, then you should leave it alone. in this patch anyway. you can remove it in a diff patch. does that make sense? | 20:31 |
romcheg | rloo: It does. | 20:31 |
NobodyCam | rloo: ++ | 20:31 |
romcheg | rloo: One thing at a time :) | 20:31 |
rloo | romcheg: which doesn't answer whether we should or shouldn't have it. i just deferred the decision ;) | 20:31 |
romcheg | Thanks anyway :) | 20:32 |
*** pelix has quit IRC | 20:32 | |
jroll | romcheg: if we don't have a test that our public apis are un-authenticated, that might be nice to add | 20:40 |
jroll | not sure if we have one yet | 20:40 |
NobodyCam | oh good question jroll i'm not sure.. | 20:41 |
jroll | there's a certain driver that currently relies on that :/ | 20:41 |
* jroll can't wait to have client certs for that | 20:41 | |
jroll | or something | 20:42 |
* NobodyCam thinks of some kinda of acl but then stops... | 20:43 | |
romcheg | Hmm, I'm not sure I follow you guys | 20:43 |
jroll | romcheg: we have public APIs in ironic | 20:43 |
jroll | it would be nice to make sure those stay public with this change | 20:43 |
jroll | NobodyCam: I mean, we use network ACLs | 20:44 |
jroll | but yeah | 20:44 |
romcheg | jroll: There are unit tests for that… I don't remember anything in tempest | 20:44 |
jroll | romcheg: ok, that's what I wanted to know :) | 20:44 |
romcheg | jroll: https://github.com/openstack/ironic/blob/master/ironic/tests/api/test_acl.py | 20:45 |
romcheg | if that's what you mean :) | 20:45 |
jroll | romcheg: perfect, thanks | 20:45 |
*** romcheg has quit IRC | 20:48 | |
*** jjohnson2 has joined #openstack-ironic | 20:48 | |
*** bradjones has quit IRC | 20:52 | |
*** romcheg has joined #openstack-ironic | 20:54 | |
*** bradjones has joined #openstack-ironic | 20:55 | |
* romcheg is back after os x upgrade | 20:55 | |
*** jgrimm is now known as zz_jgrimm | 20:56 | |
romcheg | Now everything looks gorgeous, amazing, perfect, beautiful, unbelievably nice! | 20:57 |
romcheg | Especially Vim | 20:57 |
NobodyCam | lol ... maybe I'll do the upgrade..sometime | 20:57 |
romcheg | NobodyCam: It's still in beta | 20:58 |
NobodyCam | lol my air has wanted me to update for months now.. lol :-p | 20:58 |
romcheg | NobodyCam: I heard women love risky guys so I'm on Yosemite since beta 2 :) | 20:58 |
NobodyCam | nice !!! :-p | 20:59 |
*** marcoemorais has quit IRC | 20:59 | |
romcheg | So I'm going to revert any changes on the context in the policy patch and start the separate thread somehow | 21:00 |
*** marcoemorais has joined #openstack-ironic | 21:00 | |
*** linggao has quit IRC | 21:01 | |
*** jcoufal has quit IRC | 21:03 | |
arif-ali | `so trying to get ironic working in my RDO install, restarting all the services nova does not find the driver, all I keep getting is http://pastebin.com/D7RLjpAu | 21:04 |
*** harlowja_away is now known as harlowja_ | 21:04 | |
arif-ali | has anyone successfully installed, and have it working? thanks in advance | 21:04 |
NobodyCam | arif-ali: what is your compute driver in nova.conf set to | 21:05 |
arif-ali | NobodyCam, compute_driver=nova.virt.ironic.IronicDriver | 21:06 |
arif-ali | I followed the howto in http://docs.openstack.org/developer/ironic/deploy/install-guide.html | 21:07 |
NobodyCam | let me dbl check one thing | 21:10 |
*** Kui has joined #openstack-ironic | 21:12 | |
jroll | arif-ali: what version of nova are you on? e.g. is it new enough to have the ironic driver? | 21:13 |
openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic: Move database fixture to a separate test case https://review.openstack.org/125361 | 21:13 |
arif-ali | jroll, I am currently running the RDO Juno M3 release of the apps, let me nova-compute is at openstack-nova-compute-2014.2-0.4.b3 | 21:14 |
jroll | NobodyCam: did we make juno 3 or juno rc1 for nova? | 21:14 |
jroll | I can't remember what FFEs do | 21:14 |
jroll | arif-ali: try this: python -c 'import nova.virt.ironic.IronicDriver' | 21:16 |
arif-ali | hmm, no module found | 21:16 |
jroll | arif-ali: actually... is python-ironicclient installed? | 21:16 |
arif-ali | I installed that from git | 21:17 |
jroll | how, exactly? | 21:17 |
arif-ali | git clone from https://github.com/openstack/python-ironicclient.git, and then used "python setup.py install" | 21:18 |
arif-ali | maybe I am doing this wrong | 21:18 |
openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic: Move database fixture to a separate test case https://review.openstack.org/125361 | 21:21 |
*** dprince has quit IRC | 21:22 | |
*** jjohnson2 has quit IRC | 21:24 | |
NobodyCam | ahh ha! try : nova.virt.ironic.driver.IronicDriver | 21:25 |
NobodyCam | arif-ali: ^^^^^ | 21:25 |
jroll | oh ffs | 21:26 |
* jroll wonders what docs are wrong | 21:26 | |
NobodyCam | jroll: http://docs.openstack.org/developer/ironic/deploy/install-guide.html | 21:27 |
NobodyCam | compute_driver=nova.virt.ironic.IronicDriver | 21:27 |
jroll | fun. | 21:27 |
annegentle | bah | 21:28 |
jroll | mind fixing that? my git repo is in a state I don't want to touch right now :P | 21:28 |
* jroll suddenly wonders if annegentle pings on docs.openstack.org :P | 21:28 | |
NobodyCam | one sec | 21:28 |
annegentle | jroll: all the docs all the time | 21:29 |
jroll | :) | 21:29 |
arif-ali | NobodyCam, thanks trying now | 21:33 |
NobodyCam | arif-ali: let me know if that works... I have a patch ready to push up if it does | 21:33 |
arif-ali | previously it was failing at the driver, and only one of the services was bombing out, now none of the nova daemons are starting. There maybe further misconfiguration on my part. I'll keep fishing | 21:35 |
arif-ali | No module named driver | 21:36 |
arif-ali | as now I have compute_driver=nova.virt.ironic.driver.IronicDriver | 21:36 |
jroll | so, nova.virt.ironic.IronicDriver should actually work | 21:37 |
jroll | https://github.com/openstack/nova/blob/proposed/juno/nova/virt/ironic/__init__.py#L18 | 21:37 |
jroll | so should nova.virt.ironic.driver.IronicDriver, though | 21:37 |
arif-ali | python -c 'import nova.virt.ironic.driver.IronicDriver' --> ImportError: No module named driver.IronicDriver | 21:38 |
jroll | strange | 21:38 |
arif-ali | do I need to somehow reload the PYTHONPATH? | 21:38 |
NobodyCam | from nova.virt.ironic.driver import IronicDriver | 21:38 |
jroll | NobodyCam: https://github.com/openstack/nova/blob/proposed/juno/nova/virt/ironic/__init__.py#L18 | 21:39 |
jroll | oh | 21:39 |
jroll | I see what you mean | 21:39 |
*** andreykurilin_ has quit IRC | 21:47 | |
*** praneshp has quit IRC | 21:58 | |
*** praneshp has joined #openstack-ironic | 22:01 | |
*** igordcard has quit IRC | 22:08 | |
NobodyCam | bbiafm | 22:13 |
*** mikal has quit IRC | 22:21 | |
jroll | anybody ever seen this? | 22:22 |
jroll | api logs say: 2014-10-07 22:22:25.554 24401 WARNING wsme.api [-] Client-side error: No valid host was found. Reason: No conductor service registered which supports driver agent_ipmitool. | 22:22 |
jroll | ironic driver-list has: | agent_ipmitool | ironic-conductor01... | | 22:23 |
*** mikal has joined #openstack-ironic | 22:23 | |
jroll | this is pre-hash ring changes btw | 22:24 |
NobodyCam | jroll: hartbeat ok? | 22:33 |
jroll | yeah, afaict | 22:34 |
jroll | I don't think it would be constantly blipping | 22:34 |
jroll | this error is pretty constant | 22:34 |
arif-ali | did you guys manage to figure out the problem with my issue with ironic? | 22:35 |
devananda | jroll: api have a cached copy of the ring that's out of date? (I dont think API is caching it ...) | 22:36 |
jroll | devananda: I just deployed to and restarted the api and conductor servers | 22:36 |
jroll | deployed new code to* | 22:36 |
*** todd_dsm has quit IRC | 22:37 | |
devananda | huh | 22:38 |
jroll | yeah, dunno wtf is going on | 22:38 |
* jroll digs | 22:38 | |
*** marcoemorais has quit IRC | 22:40 | |
*** marcoemorais has joined #openstack-ironic | 22:40 | |
*** romcheg has quit IRC | 22:41 | |
*** marcoemorais has quit IRC | 22:41 | |
jroll | I wouldn't be so weirded out about this if driver-list matches | 22:43 |
*** marcoemorais has joined #openstack-ironic | 22:45 | |
*** marcoemorais has quit IRC | 22:45 | |
*** marcoemorais has joined #openstack-ironic | 22:46 | |
*** chuckC_ has quit IRC | 22:46 | |
*** marcoemorais has quit IRC | 22:46 | |
NobodyCam | arif-ali: i'm not sure what errors your encountering now.. can you paste (paste.openstack.org) them | 22:46 |
*** marcoemorais has joined #openstack-ironic | 22:46 | |
*** chuckC_ has joined #openstack-ironic | 22:47 | |
arif-ali | similar errors as before, where it wasn't able to find the compute_driver http://pastebin.com/PxL4eH7y | 22:50 |
*** chuckC_ has quit IRC | 22:51 | |
NobodyCam | arif-ali: and the driver is there? ls nova/virt/ironic/ | 22:53 |
arif-ali | NobodyCam, correct, it's there | 22:54 |
jroll | devananda: this almost feels like caching, just restarting the API fixes it... the conductor was started before the api service :/ | 22:54 |
jroll | really odd | 22:55 |
devananda | jroll: huh. odd. | 22:55 |
devananda | jroll: have some of the hash ring patches been applied? | 22:55 |
* JayF attests he's seen it happen and that jroll is not insane | 22:56 | |
jroll | devananda: no | 22:56 |
* jroll triple checks that | 22:56 | |
devananda | jroll: look for change-id Ib7ab55452499d1e1c362e4cd127f1e6e38106d6c | 22:56 |
arif-ali | NobodyCam, there's nothing in nova/virt/ironic/__init__.py | 22:56 |
jroll | devananda: wait, "Add HashRingManager to wrap hash ring singleton" | 22:57 |
arif-ali | But ./ironic/nova/virt/ironic/__init__.py does | 22:57 |
jroll | yep | 22:57 |
jroll | interesting... | 22:57 |
devananda | arif-ali: are you lookin in the nova tree or the ironic tree? | 22:57 |
*** marcoemorais has quit IRC | 22:57 | |
devananda | arif-ali: within nova, nova/virt/ironic/__init__.py should have two non-commented lines | 22:58 |
devananda | from nova.virt.ironic import driver | 22:58 |
devananda | IronicDriver = driver.IronicDriver | 22:58 |
NobodyCam | arif-ali: those sound like old versions | 22:58 |
* devananda prepares to jump on a conf call | 22:58 | |
devananda | jroll: make sure the reset is being called in the right places ... maybe it's not | 22:58 |
arif-ali | i'm in /usr/lib/python2.7/site-packages | 22:58 |
NobodyCam | arif-ali: https://github.com/openstack/nova/tree/master/nova/virt/ironic | 22:58 |
devananda | jroll: are you using the uwsgi framework or apache mod_wsgi? | 22:58 |
*** marcoemorais has joined #openstack-ironic | 22:58 | |
devananda | i dont *think* that'd matter .... | 22:59 |
jroll | devananda: I hate to say, just the wsgirefthing | 22:59 |
NobodyCam | have we not removed our driver | 22:59 |
NobodyCam | nm | 22:59 |
NobodyCam | lol | 22:59 |
devananda | jroll: hey, if that's good enough ... :) | 23:00 |
arif-ali | right, /usr/lib/python2.7/site-packages/nova/virt/ironic comes from the RDO RPM, maybe that's the problem | 23:00 |
devananda | jroll: I'd check that reset is being called. maybe it's not :( | 23:00 |
NobodyCam | arif-ali: ahh older package :) | 23:00 |
jroll | devananda: I was just going to say, I don't think it is | 23:00 |
devananda | jroll: that patch had the potential to introduce the behavior you're seeing, if reset wasn't called | 23:00 |
devananda | crap | 23:00 |
jroll | backport time | 23:01 |
jroll | whee. | 23:01 |
JayF | devananda: it's not good enough; it's just that we have other things that are more-bad at the moment :) | 23:01 |
JayF | devananda: re: wsgiref | 23:01 |
arif-ali | ok, back to drawing board, will tackle again tomorrow, thanks again for all your help !! | 23:01 |
*** ChuckC has quit IRC | 23:01 | |
NobodyCam | arif-ali: set compute mgr to ironic.nova.virt.ironic.driver.IronicDriver as a test | 23:02 |
jroll | devananda: so uh, I'm a bit confused as to where that should be called, when a conductor joins the ring? | 23:02 |
devananda | 78 # NOTE(deva): this is going to be buggy | 23:02 |
devananda | 79 self.ring_manager = hash_ring.HashRingManager() | 23:02 |
devananda | sh*t | 23:02 |
*** praneshp has quit IRC | 23:02 | |
jroll | I got a kick out of that comment | 23:02 |
jroll | fwiw | 23:02 |
devananda | it should be called in conductor/rpcapi.py get_topic_for() | 23:02 |
jroll | oh | 23:03 |
jroll | ohhhhhhh | 23:03 |
devananda | yah | 23:03 |
jroll | I misread before, I thought the hook instantiated the ConductorAPI on each request, guess not? | 23:03 |
devananda | L79 creates teh cache. it SHOULD be reset on each fetch at L91, to preserve current behavior | 23:03 |
devananda | it does | 23:03 |
devananda | the hook does that | 23:03 |
jroll | wait, it does | 23:03 |
jroll | oh, but it caches | 23:03 |
devananda | but now there's a global singleton for the hash ring | 23:03 |
devananda | ie, a cache | 23:04 |
jroll | oops | 23:04 |
jroll | you got this or you want me to? | 23:04 |
*** praneshp has joined #openstack-ironic | 23:05 | |
*** praneshp has quit IRC | 23:05 | |
*** marcoemorais has quit IRC | 23:05 | |
*** praneshp has joined #openstack-ironic | 23:06 | |
* devananda is now on a call | 23:07 | |
* jroll makes a patch | 23:09 | |
jroll | unsure how to best test this... will look | 23:10 |
JayF | jroll: apply patch to $downstream_fork, deploy and test? | 23:12 |
JayF | jroll: yes, I know that's an awful answer :P | 23:12 |
jroll | lol | 23:13 |
jroll | nah, I figured something out | 23:13 |
*** praneshp has quit IRC | 23:20 | |
*** ChuckC has joined #openstack-ironic | 23:23 | |
openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Fix ConductorAPI hash ring manager https://review.openstack.org/126751 | 23:25 |
jroll | devananda: ^ not sure why that isn't passing tests, I think I'm misunderstanding something | 23:25 |
JayF | jroll: ^ should you target that at stable/juno as well? | 23:25 |
JayF | jroll: or does it have to land in master first | 23:25 |
jroll | JayF: I have no idea, I just want to share a patch | 23:25 |
JayF | cool | 23:26 |
jroll | tests don't pass anyhow | 23:26 |
NobodyCam | jroll: also maybe add juno-rc-potential tag | 23:27 |
jroll | my main goal right now is for deva to look at this code | 23:27 |
NobodyCam | ack | 23:27 |
jroll | if I wanted it merged, I would have a proper commit message etc | 23:27 |
NobodyCam | :-p | 23:27 |
rloo | jroll: I thought the idea was to clear the hash-ring cache in get_topic_for()? | 23:30 |
jroll | oh, ha | 23:31 |
jroll | call reset() | 23:31 |
jroll | ? | 23:31 |
rloo | jroll: yeah. | 23:31 |
jroll | makes more sense | 23:31 |
rloo | jroll: cuz you'll just get a pointer to the same cached thing the way you have it. | 23:31 |
rloo | i guess i should -1 it... | 23:31 |
rloo | not that i understand the code that well. just basing on the discussion above ;) | 23:32 |
jroll | heh | 23:32 |
jroll | yeah, I misread the above | 23:32 |
jroll | you can -1 if you'd like, trying that now | 23:33 |
jroll | just need a commit message | 23:33 |
*** marcoemorais has joined #openstack-ironic | 23:33 | |
jroll | tests pass | 23:34 |
jroll | yay. | 23:34 |
*** marcoemorais has quit IRC | 23:34 | |
rloo | so basically, with this fix/change, we're creating the hash_ring with each API request. | 23:34 |
jroll | right... which we've always done | 23:34 |
rloo | oh, we did? I've already forgotten. ok then ;) | 23:35 |
jroll | I guess I should file a bug for this | 23:35 |
*** marcoemorais has joined #openstack-ironic | 23:36 | |
rloo | +1 for a bug. | 23:36 |
* devananda finishes call | 23:41 | |
jroll | devananda: is "juno-rc-potential" a thing I should put here? | 23:42 |
devananda | jroll: nope. it's not caling reset. all you're doing is getting the cached thing over and over again | 23:42 |
jroll | about to upload this | 23:42 |
jroll | devananda: right, fixed, just writing commit message | 23:43 |
devananda | jroll: file a bug. tag the bug with juno-rc-potential | 23:43 |
devananda | jroll: tag the commit with the bug | 23:43 |
devananda | jroll: target it to master | 23:43 |
*** praneshp has joined #openstack-ironic | 23:43 | |
devananda | jroll: I'll handle the backport later (separate thing). we can't do a backport until the fix is in master, to prevent divergence of the stable tree. | 23:43 |
NobodyCam | ohh tag the bug | 23:43 |
jroll | ok | 23:44 |
* devananda goes back to editing something else | 23:44 | |
openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Clear hash ring cache in get_topic_for* https://review.openstack.org/126751 | 23:44 |
jroll | devananda, rloo ^ | 23:44 |
jroll | tagging bug now | 23:44 |
jroll | hmm, how do I tag juno-rc-potential? don't see it as a milestone, propose for juno series? | 23:45 |
*** Haomeng has joined #openstack-ironic | 23:45 | |
*** Haomeng|2 has quit IRC | 23:46 | |
JayF | jroll: just add juno-rc-potential as a tag on the bug | 23:47 |
jroll | aha | 23:48 |
jroll | I really hate launchpad | 23:48 |
rloo | jroll: i just did it | 23:48 |
JayF | jroll: congrats on finding an RC bug, btw | 23:48 |
jroll | "congrats" | 23:48 |
*** r-daneel has quit IRC | 23:48 | |
JayF | jroll: would you have rather it have been something downstream we changed? heh | 23:48 |
rloo | +1. better now than after the release. | 23:49 |
jroll | devananda: I'm going to patch this into our staging environment, will let you know tonight / tomorrow | 23:49 |
devananda | jroll: awesome | 23:49 |
jroll | JayF: yes, there's no backport processes downstream :) | 23:49 |
JayF | yes there is | 23:49 |
JayF | it's called "hey jroll, can you backport this?" | 23:49 |
JayF | ;) | 23:49 |
jroll | right | 23:49 |
jroll | but no bugging ttx etc | 23:49 |
NobodyCam | lol | 23:50 |
NobodyCam | great catch jroll... /me calls it a day | 23:50 |
jroll | wow, it's 5 already | 23:50 |
* jroll deep breath | 23:50 | |
JayF | yeah man | 23:50 |
NobodyCam | lol and the day started at 5:30 am | 23:51 |
NobodyCam | lol | 23:51 |
jroll | 7 for me... | 23:52 |
* jroll throws NobodyCam a beer | 23:52 | |
NobodyCam | thats the normal time... :> TY jroll | 23:52 |
jroll | :D | 23:52 |
*** praneshp_ has joined #openstack-ironic | 23:53 | |
NobodyCam | speaking of... /me is out of suck... so now I'm running to the store too... g'night all :) | 23:53 |
NobodyCam | s/suck/such/ lol | 23:54 |
*** praneshp has quit IRC | 23:54 | |
*** praneshp_ is now known as praneshp | 23:54 | |
jroll | night NobodyCam :) | 23:54 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!