| *** naohirot has joined #openstack-ironic | 00:03 | |
| naohirot | good morning ironic | 00:05 |
|---|---|---|
| NobodyCam | morning naohirot :) | 00:06 |
| naohirot | NobodyCam: good evening :) | 00:07 |
| NobodyCam | :) | 00:07 |
| BadCub_ | Morning naohirot | 00:07 |
| NobodyCam | it's always morning for /me | 00:07 |
| naohirot | BadCub_: good morning/evening :) | 00:08 |
| naohirot | BadCub_: are you in US? | 00:08 |
| BadCub_ | naohirot, yes. Sunny Southern California :-) | 00:09 |
| naohirot | BadCub_: Okay, near LA | 00:11 |
| BadCub_ | naohirot Palm Springs actually. And yourself? | 00:11 |
| naohirot | BadCub_: I'm in Yokohama, Japan | 00:12 |
| BadCub_ | naohirot Very cool! :-) | 00:13 |
| naohirot | BadCub_: :) | 00:13 |
| BadCub_ | brb | 00:13 |
| *** jmanko has quit IRC | 00:17 | |
| Haomeng | naohirot: good morning:) | 00:20 |
| Haomeng | NobodyCam: good evening:) | 00:20 |
| Haomeng | BadCub_: morning:) | 00:20 |
| naohirot | Haomeng: good morning :) | 00:20 |
| Haomeng | naohirot: :) | 00:21 |
| NobodyCam | morning Haomeng | 00:21 |
| NobodyCam | :) | 00:21 |
| jroll | devananda: I pushed a tempest patch https://review.openstack.org/#/c/150635/ to pass tests on https://review.openstack.org/#/c/149754/ | 00:22 |
| jroll | would love eyes on ^^ | 00:22 |
| jroll | hiya naohirot :) | 00:23 |
| jroll | and morning Haomeng :) | 00:23 |
| naohirot | jroll: good evening :) | 00:23 |
| Haomeng | jroll: good morning:) | 00:24 |
| Haomeng | NobodyCam: :_ | 00:24 |
| Haomeng | NobodyCam: :) | 00:24 |
| NobodyCam | :) | 00:29 |
| *** zhenzanz has joined #openstack-ironic | 00:32 | |
| *** david-lyle is now known as david-lyle_afk | 00:33 | |
| *** krtaylor has joined #openstack-ironic | 00:35 | |
| *** ChuckC__ has quit IRC | 00:36 | |
| *** achanda has quit IRC | 00:37 | |
| *** ijw has joined #openstack-ironic | 00:37 | |
| NobodyCam | jroll: seems my if is failing to correctly detect a match. | 00:39 |
| *** achanda has joined #openstack-ironic | 00:39 | |
| *** ijw has quit IRC | 00:39 | |
| *** ijw has joined #openstack-ironic | 00:39 | |
| jroll | NobodyCam: hrm | 00:40 |
| NobodyCam | oh wait | 00:41 |
| NobodyCam | not I was looking at a old log | 00:41 |
| NobodyCam | http://logs.openstack.org/87/150187/8/check/check-tempest-dsvm-ironic-pxe_ssh/72dc764/logs/screen-ir-cond.txt.gz#_2015-01-28_00_26_03_282 | 00:41 |
| jroll | lol | 00:41 |
| * jroll now wonders why that would be failing | 00:42 | |
| *** saripurigopi has joined #openstack-ironic | 00:46 | |
| NobodyCam | this is very strange: http://paste.openstack.org/show/NrGlobfULmQAY92iprqp/ | 00:47 |
| NobodyCam | logs in to the node. verifys it active then logs out and deletes the connection | 00:47 |
| jroll | hrm, do we write the image that quickly? | 00:49 |
| *** ijw has quit IRC | 00:49 | |
| NobodyCam | I am looking thru what going on now. | 00:49 |
| NobodyCam | but that log looks like it does even try | 00:50 |
| NobodyCam | anything other then log out | 00:50 |
| *** ijw has joined #openstack-ironic | 00:50 | |
| jroll | oh right, we would see logs about writing the image and stuff | 00:51 |
| *** kbs has quit IRC | 01:04 | |
| *** kbs1 has joined #openstack-ironic | 01:04 | |
| NobodyCam | can you tag a patch with more then one buG? | 01:06 |
| jroll | yes | 01:07 |
| jroll | https://review.openstack.org/#/c/147523/ | 01:08 |
| jroll | this is surprisingly good | 01:08 |
| *** spandhe has quit IRC | 01:09 | |
| JayF | Is https://review.openstack.org/#/c/139517/ this a situation where I should land the spec then fix the description? | 01:10 |
| JayF | I know that was brought up in a meeting recently | 01:11 |
| JayF | and that seems like it might fit the bill | 01:11 |
| JayF | jroll: ^ have a thought? | 01:12 |
| jroll | was reading | 01:12 |
| jroll | this spec is about using these without IPMI right? | 01:13 |
| *** spandhe has joined #openstack-ironic | 01:13 | |
| JayF | yes | 01:13 |
| JayF | they can't use IPMI in the newer version because it's a PITA to enable | 01:13 |
| jroll | right | 01:14 |
| JayF | as the description enumerates ... for seemingly no reason | 01:14 |
| JayF | and both ruby and I thought this spec was going to do the enabling of the IPMI | 01:14 |
| JayF | until comments clarified | 01:14 |
| jroll | I think I'd land it and then tack on "without enabling or using IPMI" to the problem description | 01:14 |
| jroll | but link the follow up patch in your +A | 01:14 |
| JayF | that requires me to properly do the git-chained-patches dance | 01:15 |
| jroll | yeah | 01:15 |
| JayF | you know; I might wait | 01:15 |
| jroll | so? | 01:15 |
| jroll | lol | 01:15 |
| JayF | simply because it's not the author who made the comment | 01:15 |
| JayF | about how it's not using IPMI | 01:16 |
| jroll | eh, true | 01:16 |
| JayF | I just wanna land some things :( | 01:16 |
| jroll | maybe +2 but don't +A | 01:16 |
| jroll | and add your reason | 01:16 |
| JayF | jroll: wanna land a spec? https://review.openstack.org/#/c/136020/ | 01:16 |
| jroll | maybe | 01:17 |
| * jroll has some things queued up to review | 01:17 | |
| jroll | haaaaaaa | 01:18 |
| *** rloo has quit IRC | 01:18 | |
| jroll | that was my next tab | 01:18 |
| JayF | nice | 01:18 |
| jroll | oh wait | 01:18 |
| jroll | the actual code was my next one | 01:18 |
| jroll | irmc power, that is | 01:18 |
| jroll | whatever, I'll do it | 01:18 |
| JayF | then you should likely read and approve the spec first so you can +2 in good conscience :P | 01:18 |
| *** r-daneel has quit IRC | 01:19 | |
| jroll | they're different but yeah :P | 01:20 |
| *** arif-ali has quit IRC | 01:20 | |
| *** arif-ali has joined #openstack-ironic | 01:20 | |
| *** ijw has quit IRC | 01:22 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 01:24 |
| *** spandhe has quit IRC | 01:25 | |
| *** penick has quit IRC | 01:26 | |
| *** arif-ali has quit IRC | 01:27 | |
| *** arif-ali has joined #openstack-ironic | 01:27 | |
| jroll | NobodyCam: do you want another look at irmc management spec or should I just land it? | 01:32 |
| jroll | https://review.openstack.org/#/c/136020/ | 01:32 |
| *** Marga_ has joined #openstack-ironic | 01:33 | |
| *** rwsu is now known as rwsu-afk | 01:33 | |
| *** Marga_ has quit IRC | 01:37 | |
| *** Marga_ has joined #openstack-ironic | 01:38 | |
| *** arif-ali has quit IRC | 01:38 | |
| *** arif-ali has joined #openstack-ironic | 01:39 | |
| NobodyCam | jroll: lgtm. I see the spelling nits. which we should fix with a follow up | 01:40 |
| jroll | meh | 01:40 |
| NobodyCam | :-p | 01:40 |
| jroll | just a spec, I don't think it's a huge deal | 01:40 |
| jroll | go ahead and land it if you want | 01:40 |
| NobodyCam | will do | 01:41 |
| NobodyCam | naohirot: 136020 +a'd | 01:43 |
| openstackgerrit | Merged openstack/ironic-specs: iRMC Management Driver for Ironic https://review.openstack.org/136020 | 01:44 |
| NobodyCam | see ^^^^ | 01:45 |
| naohirot | NobodyCam: jroll: Oh great! | 01:45 |
| naohirot | NobodyCam: thanks :) | 01:45 |
| NobodyCam | okay all have a good night I'm out | 01:45 |
| jroll | night NobodyCam :) | 01:45 |
| naohirot | NobodyCam: good night ! | 01:45 |
| Haomeng | NobodyCam: good night:) | 01:49 |
| *** arif-ali has quit IRC | 01:50 | |
| *** arif-ali has joined #openstack-ironic | 01:51 | |
| *** zz_jgrimm is now known as jgrimm | 01:55 | |
| naohirot | jroll: good night :) | 01:58 |
| jroll | see ya naohirot :) | 01:58 |
| naohirot | jroll: :) | 01:58 |
| Haomeng | jroll: good night | 02:01 |
| openstackgerrit | Merged openstack/ironic: Agent driver fails with Swift Multiple Containers https://review.openstack.org/149451 | 02:04 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Improve test assertion for get_glance_image_properties https://review.openstack.org/150652 | 02:10 |
| *** chenglch has joined #openstack-ironic | 02:14 | |
| *** Marga_ has quit IRC | 02:14 | |
| *** Marga_ has joined #openstack-ironic | 02:15 | |
| *** BadCub_ has quit IRC | 02:27 | |
| *** saripurigopi has quit IRC | 02:27 | |
| *** jerryz has joined #openstack-ironic | 02:32 | |
| *** bradjones has quit IRC | 02:33 | |
| *** bradjones has joined #openstack-ironic | 02:36 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Fix drivers.rst doc format error https://review.openstack.org/150656 | 02:37 |
| openstackgerrit | Tan Lin proposed openstack/ironic-specs: Update amt-pxe-driver spec https://review.openstack.org/141269 | 02:42 |
| *** ramineni has joined #openstack-ironic | 02:46 | |
| *** eghobo has quit IRC | 02:47 | |
| *** harlowja is now known as harlowja_away | 02:57 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add iRMC Driver and its iRMC Power module https://review.openstack.org/144901 | 03:04 |
| *** ChuckC has joined #openstack-ironic | 03:27 | |
| *** Marga__ has joined #openstack-ironic | 03:34 | |
| *** Marga_ has quit IRC | 03:34 | |
| *** saripurigopi has joined #openstack-ironic | 03:43 | |
| *** slagle has quit IRC | 03:45 | |
| *** yuanying has joined #openstack-ironic | 03:48 | |
| *** stendulker has joined #openstack-ironic | 03:50 | |
| *** yuanying_ has quit IRC | 03:52 | |
| *** achanda has quit IRC | 03:55 | |
| *** yuanying has quit IRC | 03:55 | |
| *** yuanying has joined #openstack-ironic | 03:56 | |
| *** Nisha has joined #openstack-ironic | 04:03 | |
| *** bradjones has quit IRC | 04:10 | |
| *** bradjones has joined #openstack-ironic | 04:11 | |
| *** bradjones has quit IRC | 04:29 | |
| *** bradjones has joined #openstack-ironic | 04:30 | |
| *** saripurigopi has quit IRC | 04:40 | |
| *** achanda has joined #openstack-ironic | 04:55 | |
| *** achanda has quit IRC | 05:01 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic-specs: Ironic Management Interfaces to support UEFI Secure Boot https://review.openstack.org/135845 | 05:05 |
| *** pensu has joined #openstack-ironic | 05:07 | |
| *** pradipta_away is now known as pradipta | 05:08 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic-specs: Ironic Management Interfaces to support UEFI Secure Boot https://review.openstack.org/135845 | 05:08 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add states required for node-introspection https://review.openstack.org/147857 | 05:11 |
| *** Marga__ has quit IRC | 05:17 | |
| openstackgerrit | Zhenzan Zhou proposed openstack/ironic: Add config option to control password mask for node-show https://review.openstack.org/150688 | 05:17 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add states required for node-inspection https://review.openstack.org/147857 | 05:18 |
| *** pensu has quit IRC | 05:31 | |
| *** pensu has joined #openstack-ironic | 05:31 | |
| *** wendar_ has joined #openstack-ironic | 05:39 | |
| *** wendar has quit IRC | 05:40 | |
| openstackgerrit | Rakesh H S proposed openstack/python-ironicclient: ironicclient node-set-maintenance to accept true/false https://review.openstack.org/150692 | 06:00 |
| *** achanda has joined #openstack-ironic | 06:01 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add iRMC Driver and its iRMC Power module https://review.openstack.org/144901 | 06:13 |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add iRMC Management module for iRMC Driver https://review.openstack.org/146803 | 06:13 |
| * naohirot it seems the commit id of 144901 was changed accidentally when I squashed. | 06:22 | |
| *** pcrews has quit IRC | 06:24 | |
| openstackgerrit | sandhya proposed openstack/ironic-specs: Chassis Level Discovery Tool https://review.openstack.org/134866 | 06:27 |
| *** rakesh_hs has joined #openstack-ironic | 06:27 | |
| *** Nisha has quit IRC | 06:39 | |
| *** rameshg87 has joined #openstack-ironic | 06:55 | |
| *** Nisha has joined #openstack-ironic | 06:57 | |
| openstackgerrit | Merged openstack/ironic: Fix drivers.rst doc format error https://review.openstack.org/150656 | 07:00 |
| *** chenleji has joined #openstack-ironic | 07:01 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Enable agent_ilo for uefi-bios switching https://review.openstack.org/148792 | 07:01 |
| *** andreykurilin_ has joined #openstack-ironic | 07:05 | |
| *** achanda has quit IRC | 07:17 | |
| *** romcheg has joined #openstack-ironic | 07:17 | |
| *** romcheg has quit IRC | 07:17 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add InspectInterface for node-introspection https://review.openstack.org/149817 | 07:19 |
| *** achanda has joined #openstack-ironic | 07:20 | |
| *** andreykurilin_ has quit IRC | 07:22 | |
| *** jcoufal has joined #openstack-ironic | 07:23 | |
| openstackgerrit | Ramakrishnan G proposed openstack/ironic-specs: New driver interface for RAID configuration https://review.openstack.org/135899 | 07:24 |
| *** Guest69430 has joined #openstack-ironic | 07:28 | |
| yjiang5 | Nisha: hi | 07:31 |
| Nisha | yjiang5, hi | 07:32 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add InspectInterface for node-introspection https://review.openstack.org/149817 | 07:32 |
| yjiang5 | Nisha: for https://review.openstack.org/133534 , I'm not sure if I mis-understand. in your comments, what do you mean of value comes from ironic? | 07:33 |
| Nisha | yjiang5, for capabilities to be matched against node capabilities by the compute capabilities filter, flavor will have it defined as {'capabilities:key1' : 'value1'} | 07:35 |
| Nisha | while ironic have it defined as {'capabilities': 'key1:value1' } | 07:36 |
| *** killer_prince is now known as lazy_prince | 07:36 | |
| *** eghobo has joined #openstack-ironic | 07:37 | |
| *** coolsvap|afk is now known as coolsvap | 07:37 | |
| Nisha | yjiang5, the scenario discussed in the spec/code is when value1 = 'a b c d' i.e. space sepearted words/list of values applicable for that key. This is on ironic side | 07:38 |
| Nisha | on flavor side the requestor wants to selct the node like : { 'capabilities:key1' : 'c' } | 07:39 |
| Nisha | with this the above node shal get selected as the node supports 'c' as the property for key1 | 07:39 |
| *** eghobo has quit IRC | 07:40 | |
| Nisha | so as per code, extra_specs_op.match(str(cap), req) ---- in this 'cap' will have 'a b c d' while 'req' will hold 'c' | 07:41 |
| *** Haomeng|2 has joined #openstack-ironic | 07:41 | |
| *** Haomeng has quit IRC | 07:42 | |
| Nisha | extra_specs_op will do operations (as given in extra_specs_op.py) on the 'req' and not on 'cap' | 07:42 |
| Nisha | yjiang5, does this help? | 07:43 |
| yjiang5 | Nisha: if you set the extra spec as " <in> c", will that work? the <in> operation is: lambda x, y: y in x, so it should meet your requirement? | 07:44 |
| yjiang5 | Nisha: something like " nova flavor-key m1.small set "capabilities:cpu_info:features"="<in> sse4.1" | 07:45 |
| yjiang5 | Nisha: this example will find a host with sse4.1 in the value of "capabilities:cpu_info:features" that's returned by the hypervisor. | 07:46 |
| yjiang5 | Nisha: need go for sleep. Will talk to you tomorrow. | 07:53 |
| *** achanda has quit IRC | 07:59 | |
| *** yjiang5 is now known as yjiang5_away | 08:05 | |
| *** bradjones has quit IRC | 08:07 | |
| *** achanda has joined #openstack-ironic | 08:07 | |
| *** bradjones has joined #openstack-ironic | 08:08 | |
| openstackgerrit | Tan Lin proposed openstack/python-ironicclient: Add driver_internal_info to node-show output https://review.openstack.org/150724 | 08:16 |
| *** ifarkas has joined #openstack-ironic | 08:23 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-specs: iRMC Virtual Media Deploy Driver for Ironic https://review.openstack.org/134865 | 08:34 |
| Nisha | yjiang5_away, yes that works... | 08:44 |
| *** achanda has quit IRC | 08:50 | |
| *** smoriya has quit IRC | 08:51 | |
| *** ukalifon1 has joined #openstack-ironic | 08:56 | |
| *** ndipanov_gone is now known as ndipanov | 08:56 | |
| ukalifon1 | Hello #openstack-ironic. What can I do if I have a node stuck in "wait call-back" provisioning state? Thanks. | 08:57 |
| Haomeng|2 | ukalifon1: so far, we have no *force-delete* function via api/cmd call | 09:01 |
| Haomeng|2 | ukalifon1: so have to update db by manually | 09:01 |
| Haomeng|2 | ukalifon1: fyi, the sql - update nodes set power_state='power off', reservation=False,instance_uuid=null,instance_info=null,maintenance=False, provision_state=null,target_provision_state=null; | 09:01 |
| *** derekh has joined #openstack-ironic | 09:02 | |
| ukalifon1 | Haomeng|2: Thanks for that. Is there also something I can check to see why this is happening, if it will happen again? | 09:03 |
| *** athomas has joined #openstack-ironic | 09:05 | |
| *** erwan_taf has joined #openstack-ironic | 09:20 | |
| *** jistr has joined #openstack-ironic | 09:23 | |
| *** lucasagomes has joined #openstack-ironic | 09:30 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add InspectInterface for node-introspection https://review.openstack.org/149817 | 09:32 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Enable agent_ilo for uefi-bios switching https://review.openstack.org/148792 | 09:42 |
| *** MattMan has joined #openstack-ironic | 09:48 | |
| *** dtantsur|afk is now known as dtantsur | 09:56 | |
| dtantsur | Morning Ironic | 09:57 |
| lucasagomes | dtantsur, g'morning | 09:58 |
| *** pelix has joined #openstack-ironic | 09:58 | |
| dtantsur | lucasagomes, o/ | 09:58 |
| dtantsur | devananda, I'll double-check, but I believe https://review.openstack.org/#/c/147523 completes the spec | 09:59 |
| rameshg87 | hello lucasagomes, dtantsur | 10:02 |
| rameshg87 | ifarkas, hi | 10:02 |
| lucasagomes | rameshg87, hi there | 10:02 |
| ifarkas | rameshg87, hi | 10:02 |
| dtantsur | rameshg87, o/ | 10:02 |
| ifarkas | morning dtantsur | 10:02 |
| dtantsur | ifarkas, morning :) | 10:02 |
| dtantsur | so many of us | 10:03 |
| rameshg87 | ifarkas, i was checking the irc chat yesterday - seemed like you were planning to create the boot volume first and then the rest of raid volumes for raid configuration | 10:03 |
| rameshg87 | ifarkas, are you planning to propose it in https://review.openstack.org/135899 ? | 10:03 |
| ifarkas | rameshg87, yeah, I am planning to separate the creation of raid volumes in to 2: boot and the rest | 10:03 |
| rameshg87 | ifarkas, or are you planning to take care of this in the drac implementation ? | 10:04 |
| ifarkas | rameshg87, yep, I will update it | 10:04 |
| ifarkas | rameshg87, yes. unfortunately that's the only way to use the root device hint with drac | 10:04 |
| ifarkas | rameshg87, are you ok with this change? | 10:04 |
| rameshg87 | ifarkas, isn't it better to take care of it in the drac implementation itself ? | 10:04 |
| rameshg87 | ifarkas, do we need to change the general spec for this ? | 10:04 |
| *** pensu has quit IRC | 10:04 | |
| ifarkas | rameshg87, yeah, because it will be using the generic interface | 10:05 |
| ifarkas | rameshg87, if the generic interface doesn't support this 2 steps, then it can't be implemented in the drac driver | 10:05 |
| rameshg87 | ifarkas, but create_configuration() of drac implementation can take care of this right ? | 10:06 |
| rameshg87 | ifarkas, from the target_raid_config - it can figure out the boot volume, create it first, gather the required info, then create rest of the volumes | 10:07 |
| ifarkas | rameshg87, how create_config should know that know it should create the boot device or the rest? | 10:07 |
| ifarkas | rameshg87, gathering should be optional, we can't depend on discoverd | 10:07 |
| rameshg87 | ifarkas, then i might be missing something .. | 10:08 |
| rameshg87 | ifarkas, can you please tell what's your solution of root_device_hint for drac implementation ? | 10:08 |
| ifarkas | rameshg87, create the boot device with the drac driver. then optionally collect the root device hint with discoverd, and finally create the rest of the raid arrays | 10:09 |
| rameshg87 | ifarkas, and how/who will we optionally trigger discoverd inbetween ? | 10:10 |
| ifarkas | rameshg87, it depends how zapping/drac driver will be configured | 10:11 |
| *** Guest69430 has quit IRC | 10:12 | |
| rameshg87 | ifarkas, so will you have separate zapping task for rebooting and trigger discoverd introspection inbetween ? | 10:12 |
| ifarkas | rameshg87, yeah, but that's optional | 10:12 |
| rameshg87 | ifarkas, currently we have proposed an apply_configuration() method which will create all raid arrays | 10:14 |
| rameshg87 | ifarkas, do you propose to break it down into 2 zapping methods ? | 10:14 |
| ifarkas | rameshg87, yep | 10:14 |
| rameshg87 | ifarkas, and create_configuration() and delete_configuration() will remain the same ? | 10:15 |
| ifarkas | rameshg87, I think so | 10:16 |
| rameshg87 | ifarkas, okay, seems fine with me | 10:16 |
| *** romcheg has joined #openstack-ironic | 10:17 | |
| rameshg87 | ifarkas, can you please update the spec then .. | 10:17 |
| ifarkas | rameshg87, sure, it's on my list ;-) | 10:17 |
| rameshg87 | ifarkas, okay .. :) | 10:17 |
| rameshg87 | ifarkas, i would wish if the raid spec landed asap :) | 10:18 |
| ifarkas | rameshg87, yeah, same here | 10:19 |
| *** jerryz has quit IRC | 10:20 | |
| ifarkas | rameshg87, although it was mentioned during the last meeting that zapping is slipped to L. that might block our spec... | 10:20 |
| *** chenglch has quit IRC | 10:20 | |
| *** Nisha has quit IRC | 10:24 | |
| *** jerryz has joined #openstack-ironic | 10:24 | |
| *** athomas has quit IRC | 10:34 | |
| dtantsur | devananda, actually it does not, I need a couple more changes... | 10:37 |
| *** athomas has joined #openstack-ironic | 10:38 | |
| *** jiangfei|2 has quit IRC | 10:49 | |
| *** openstackgerrit has quit IRC | 10:50 | |
| *** openstackgerrit has joined #openstack-ironic | 10:50 | |
| openstackgerrit | Gopi Krishna S proposed openstack/ironic-specs: Add pxe_ucs driver to support Cisco UCS B/C/M-series servers https://review.openstack.org/139517 | 10:53 |
| openstackgerrit | Anusha Ramineni proposed openstack/ironic-specs: Implement Cleaning Operations for iLO drivers https://review.openstack.org/142401 | 11:01 |
| *** dlpartain has joined #openstack-ironic | 11:02 | |
| *** dlpartain has left #openstack-ironic | 11:04 | |
| *** ramineni has quit IRC | 11:05 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Add support for local boot https://review.openstack.org/146189 | 11:09 |
| *** dlpartain has joined #openstack-ironic | 11:10 | |
| *** rameshg87 has quit IRC | 11:15 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Add support for driver-specific periodic tasks https://review.openstack.org/147523 | 11:16 |
| dtantsur | devananda, now should be complete ^^^ | 11:16 |
| *** pensu has joined #openstack-ironic | 11:19 | |
| *** stendulker has quit IRC | 11:23 | |
| *** dlpartain has left #openstack-ironic | 11:25 | |
| *** xxj has quit IRC | 11:26 | |
| *** vdrok_ is now known as vdrok | 11:28 | |
| lucasagomes | devananda, jroll for the configdrive thing, I think it should be passed to Ironic when we trigger the node to be deployed | 11:29 |
| lucasagomes | like set-provision-state active, optionally give it a configdrive | 11:30 |
| *** yog has joined #openstack-ironic | 11:31 | |
| *** yog is now known as Guest76628 | 11:31 | |
| lucasagomes | devananda, jroll because if we add another endpoint in the API to pass the configdrive, there's the risk of us getting the configdrive, uploading it to swift and creating a tempurl | 11:31 |
| lucasagomes | but at the time the provision is started that temp url is already expired | 11:32 |
| lucasagomes | and if we don't wanna do that, if we wanna save it for uploading to swift later, it's better to not even upload | 11:32 |
| lucasagomes | and just store it within the Ironic database itself | 11:32 |
| lucasagomes | devananda, jroll thoughts? | 11:32 |
| *** jerryz has quit IRC | 11:39 | |
| *** xxj has joined #openstack-ironic | 11:42 | |
| *** pensu has quit IRC | 11:51 | |
| *** jerryz has joined #openstack-ironic | 11:51 | |
| *** pensu has joined #openstack-ironic | 11:52 | |
| *** slagle has joined #openstack-ironic | 12:06 | |
| *** Haomeng has joined #openstack-ironic | 12:06 | |
| *** Haomeng|2 has quit IRC | 12:07 | |
| *** timbyr_ has quit IRC | 12:17 | |
| *** timbyr_ has joined #openstack-ironic | 12:20 | |
| *** dtantsur is now known as dtantsur|brb | 12:31 | |
| *** subscope has quit IRC | 12:33 | |
| *** subscope has joined #openstack-ironic | 12:34 | |
| *** jerryz has quit IRC | 12:34 | |
| *** pensu has quit IRC | 12:43 | |
| *** pensu has joined #openstack-ironic | 12:44 | |
| *** jerryz has joined #openstack-ironic | 12:47 | |
| *** jerryz has quit IRC | 12:50 | |
| *** Marga_ has joined #openstack-ironic | 12:51 | |
| *** sambetts has quit IRC | 12:58 | |
| *** sambetts has joined #openstack-ironic | 12:59 | |
| *** lucasagomes is now known as lucas-hungry | 13:03 | |
| *** Marga_ has quit IRC | 13:04 | |
| *** Marga_ has joined #openstack-ironic | 13:05 | |
| *** dprince has joined #openstack-ironic | 13:06 | |
| *** Marga_ has quit IRC | 13:10 | |
| *** Marga_ has joined #openstack-ironic | 13:14 | |
| NobodyCam | Good Morning Ironicers, says the man maken coffee... | 13:14 |
| *** bradjones has quit IRC | 13:17 | |
| *** bradjones has joined #openstack-ironic | 13:18 | |
| *** bradjones has quit IRC | 13:23 | |
| *** Marga_ has quit IRC | 13:25 | |
| *** pensu has quit IRC | 13:27 | |
| *** bradjones has joined #openstack-ironic | 13:27 | |
| *** dtantsur|brb is now known as dtantsur | 13:28 | |
| dtantsur | NobodyCam, morning | 13:28 |
| NobodyCam | morning dtantsur :) | 13:32 |
| *** mjturek has joined #openstack-ironic | 13:44 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 13:47 |
| *** bradjones has quit IRC | 13:51 | |
| *** BadCub_ has joined #openstack-ironic | 13:53 | |
| BadCub_ | morning Ironic | 13:53 |
| NobodyCam | morning BadCub_ :) | 13:54 |
| *** bradjones has joined #openstack-ironic | 13:55 | |
| *** lucas-hungry is now known as lucasagomes | 13:57 | |
| devananda | morning, all | 13:59 |
| NobodyCam | morning devananda | 13:59 |
| NobodyCam | your up early | 13:59 |
| devananda | yah. all week. hope the hotels in EU are better | 14:00 |
| BadCub_ | Morning Devananda | 14:01 |
| lucasagomes | NobodyCam, BadCub devananda morning | 14:02 |
| NobodyCam | morning lucasagomes :) | 14:02 |
| lucasagomes | devananda, have a min to talk about how the configdrive should look in Ironic? | 14:02 |
| BadCub_ | Morning lucasagomes | 14:02 |
| * lucasagomes wants to tackle that quickly | 14:03 | |
| devananda | lucasagomes: yep | 14:03 |
| devananda | let's do it | 14:03 |
| lucasagomes | devananda, so, I'm not sure about exposing a new endpoint in the API. I was thinking more about making configdrive a optional parameter for the set provision state (when active) | 14:03 |
| lucasagomes | so as part of the request to set the provision state to active, we could pass the raw configdrive, and Ironic can then do whatever it needs to do (either upload it to swift or store in our db) with it | 14:04 |
| devananda | oh | 14:04 |
| devananda | an optional member in the request BODY | 14:05 |
| NobodyCam | lucasagomes: also not to distract you but are you planning on unWIPPING 142838? | 14:05 |
| devananda | that'd work | 14:05 |
| lucasagomes | devananda, my concern about adding a new endpoint, is that as we want Ironic to be standalone | 14:05 |
| lucasagomes | if we have an endpoint people will upload the configdrive, and then idk how long they will take untill trigger the deploy | 14:05 |
| lucasagomes | and that would potentially make the wsift temp url to expire | 14:05 |
| devananda | right | 14:05 |
| lucasagomes | brefore deployment starts, so we would have to add adittional logic there to check the experation time etc etc etc | 14:06 |
| lucasagomes | so being part of provision, sounds like a good way to do it | 14:06 |
| openstackgerrit | Tan Lin proposed openstack/ironic: Add Node.driver_internal_info https://review.openstack.org/150784 | 14:06 |
| lucasagomes | and that's how nova works too, it only generates and set the config drive when the instance is actually being spawn'ed | 14:06 |
| lucasagomes | NobodyCam, lemme check | 14:06 |
| lucasagomes | NobodyCam, oh so, Andre has some concerns about it :/ | 14:07 |
| lucasagomes | he says that didn't work for him | 14:07 |
| devananda | lucasagomes: i'm thinking through how that would look if I was driving a deployment from ansible (instead of nova) | 14:07 |
| devananda | lucasagomes: and it seems fine | 14:07 |
| lucasagomes | devananda, yeah, it's lightweight too (1 api call instead of 2) | 14:07 |
| devananda | sure | 14:07 |
| lucasagomes | if you agree with that, I'll modify the spec we have in Ironic and put up some patches for it | 14:08 |
| lucasagomes | then do the same with nova | 14:08 |
| devananda | ++ | 14:08 |
| lucasagomes | devananda, alright! and big thanks for bringing it to the mid-cycle | 14:08 |
| lucasagomes | really speeds things up | 14:08 |
| lucasagomes | NobodyCam, there's another RH guy that would check that bug fix, I will ping him and see if he actually got to test that | 14:09 |
| *** Marga_ has joined #openstack-ironic | 14:09 | |
| lucasagomes | NobodyCam, problem is, I can't reproduce that error consistently. It seems to appear from time to time | 14:09 |
| devananda | lucasagomes: thanks much! | 14:09 |
| lucasagomes | NobodyCam, we applied the patch for our PoC stuff internally, and AFAICT it fixed it... tho it didn't for Andre so... idk | 14:10 |
| NobodyCam | lucasagomes: ya, I believe i am hitting with my patch.. Let me look in to whats going on | 14:14 |
| NobodyCam | with it | 14:14 |
| *** rloo has joined #openstack-ironic | 14:15 | |
| lucasagomes | NobodyCam, right, cool yeah if u can test and let me know if it works would be great | 14:18 |
| NobodyCam | :) | 14:19 |
| devananda | looks like we also need a tempest patch to land to move forward with the NOSTATE rename | 14:19 |
| naohirot | devananda: NobodyCam: good morning | 14:22 |
| naohirot | lucasagomes: good evening | 14:22 |
| lucasagomes | naohirot, hey there, g'evening :) | 14:22 |
| NobodyCam | morning naohirot :) | 14:23 |
| NobodyCam | hehehe | 14:23 |
| dtantsur | devananda, naohirot, o/ | 14:23 |
| devananda | lintan__: typo in your patch ^ | 14:24 |
| naohirot | devananda: may I ask your advice? if you finished the conversation with lucasagomes | 14:24 |
| devananda | naohirot: sure | 14:24 |
| naohirot | dtantsur: good evening | 14:24 |
| naohirot | devananda: https://blueprints.launchpad.net/ironic/+spec/automate-uefi-bios-iso-creation is approved | 14:24 |
| naohirot | devananda: this is related to irmc deploy driver | 14:25 |
| naohirot | devananda: and I'd like to not to depend on the automatic uefi boot iso creation | 14:26 |
| naohirot | devananda: should I support automatic uefi boot iso creation in the initial irmc deploy driver? | 14:27 |
| devananda | naohirot: do you need UEFI support in the initial work? | 14:27 |
| naohirot | devananda: Yes | 14:27 |
| *** jerryz has joined #openstack-ironic | 14:28 | |
| naohirot | devananda: current iLO deploy driver assumes that deployer prepares iso in glance | 14:28 |
| naohirot | devananda: I'm thinking the same way | 14:28 |
| devananda | naohirot: ok, that seems fine, as long as it's documented | 14:29 |
| devananda | naohirot: for the deploy driver, i mean, ^ is fine | 14:29 |
| devananda | naohirot: what about boot phase? | 14:29 |
| naohirot | devananda: certainly I'm going to support automatic uefi iso creation after stabilized the implementation | 14:29 |
| devananda | naohirot: eg, after deploy driver completes the image deployment, will you use local boot or boot-from-virtual-media? | 14:29 |
| naohirot | devananda: I think both. | 14:30 |
| naohirot | devananda: if deployer chose agent_irmc, it's local | 14:31 |
| naohirot | devananda: if deployer chose iscsi_irmc, it's boot from virtual media | 14:32 |
| * devananda really wants those drivers to each support both modes | 14:32 | |
| * lucasagomes brb calls :/ | 14:32 | |
| naohirot | devananda: I documented the line 156 https://review.openstack.org/#/c/134865/19/specs/kilo/irmc-virtualmedia-deploy-driver.rst | 14:34 |
| devananda | naohirot: if booting from virtual media, the deployer will need to prepare two ISO images -- one for deploy phase, one for boot phase | 14:35 |
| devananda | naohirot: that's my only concern, and AFAIK, the reason why ilo team is implementing the automate-uefi-bios-iso-creation spec | 14:35 |
| naohirot | devananda: If booting from virtual media in UEFI mode, yes, deployer prepares two ISOs | 14:35 |
| rloo | hello ironickers | 14:37 |
| devananda | naohirot: sounds ok then. it's not the best user experience, which is why ilo team is addressing it, but also that shouldn't block your work | 14:37 |
| naohirot | devananda: I'll support automatic UEFI boot iso creation in the next development cycle | 14:37 |
| rloo | devananda: wrt tempest patch for NOSTATE stuff. Is there anything we can do to help there? | 14:37 |
| BadCub_ | Hi rloo! | 14:37 |
| rloo | hi BadCub_. (I always wonder how bad bad can be...) | 14:38 |
| naohirot | devananda: My concern is that current priority of the automatic UEFI boot ISO creation is low in https://blueprints.launchpad.net/ironic | 14:38 |
| devananda | rloo: nope. https://review.openstack.org/#/c/150635/is passing now. I'll poke folks today to land it | 14:38 |
| devananda | naohirot: sure | 14:39 |
| BadCub_ | rloo: LOL Very bad! | 14:39 |
| devananda | rloo: he's the worst! | 14:39 |
| BadCub_ | devananda: LOL, but I am sooooo "good" at it!!! | 14:39 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 14:40 |
| rloo | devananda: great! about the patch and BadCub_ ;) | 14:40 |
| BadCub_ | :-) | 14:40 |
| naohirot | devananda: thanks, BTW would you update the design status of iRMC Management to Approved? https://blueprints.launchpad.net/ironic | 14:40 |
| naohirot | devananda: this one https://blueprints.launchpad.net/ironic/+spec/irmc-management-driver | 14:41 |
| NobodyCam | morning rloo :) | 14:41 |
| rloo | morning NobodyCam! | 14:41 |
| naohirot | rloo: good morning | 14:42 |
| NobodyCam | :) | 14:42 |
| *** bradjones has quit IRC | 14:42 | |
| rloo | hi naohirot | 14:43 |
| *** bradjones has joined #openstack-ironic | 14:45 | |
| *** mjturek has quit IRC | 14:46 | |
| naohirot | NobodyCam: devananda may be busy, would you update the design status of iRMC Management to Approved? | 14:47 |
| *** r-daneel has joined #openstack-ironic | 14:48 | |
| *** r-daneel has quit IRC | 14:48 | |
| naohirot | NobodyCam: which is approved this morning, last evening your local time | 14:48 |
| *** r-daneel has joined #openstack-ironic | 14:48 | |
| NobodyCam | naohirot: this one: https://blueprints.launchpad.net/ironic/+spec/irmc-management-driver | 14:48 |
| naohirot | NobodyCam: Yes | 14:49 |
| *** jgrimm is now known as zz_jgrimm | 14:49 | |
| naohirot | NobodyCam: Definition is still "Review" | 14:49 |
| NobodyCam | naohirot: refresh | 14:50 |
| naohirot | NobodyCam: great thanks! | 14:51 |
| NobodyCam | :-p | 14:52 |
| naohirot | NobodyCam: rloo: devananda: dtantsur: lucasagomes: good night, see you tomorrow :) | 14:52 |
| dtantsur | g'night | 14:53 |
| NobodyCam | have a good rest naohirot | 14:53 |
| rloo | night naohirot | 14:53 |
| *** naohirot has quit IRC | 14:53 | |
| *** mjturek has joined #openstack-ironic | 14:54 | |
| *** zhenzanz has quit IRC | 14:54 | |
| *** EmilienM is now known as EmilienM|afk | 14:55 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add iRMC Driver and its iRMC Power module https://review.openstack.org/144901 | 15:05 |
| rloo | NobodyCam, devananda: you +2'd 144901. Should I fix the optional properties thing in a follow on patch? | 15:14 |
| devananda | jroll: for the NOSTATE->AVAILABLE change, we /could/ make this non-API-affecting | 15:16 |
| devananda | jroll: eg, that we have to have a tempest change here at all really means we're changing the API semantics. granted, it's in a small way, but still, we could avoid it | 15:16 |
| devananda | rloo: which thing? | 15:17 |
| devananda | rloo: I posted a fix for the doc merge issue already | 15:17 |
| devananda | ooh | 15:17 |
| devananda | just found your comment lower down | 15:17 |
| rloo | devananda: I commented. yeah, i was reviewing and didn't notice that you had +2 until i finished reviewing. | 15:18 |
| devananda | rloo: yea, I'm fine with that in a follow on | 15:18 |
| devananda | good catch, too | 15:18 |
| victor_lowther | Good morning, Ironic. | 15:18 |
| rloo | devananda: ok, I'll +2 and mention it. thx. | 15:18 |
| devananda | morning, victor_lowther | 15:18 |
| NobodyCam | morning victor_lowther | 15:18 |
| NobodyCam | Thank you rloo :) | 15:18 |
| victor_lowther | Time for another day of hating WSMAN. At least the XML parts of it. | 15:18 |
| NobodyCam | victor_lowther: :-p | 15:19 |
| devananda | I'm gonna start asking random people this question ^ :) | 15:21 |
| devananda | victor_lowther: for the NOSTATE->AVAILABLE change, we /could/ make this non-API-affecting. think that's important? | 15:21 |
| openstackgerrit | Rakesh H S proposed openstack/python-ironicclient: ironicclient node-set-maintenance to accept true/false https://review.openstack.org/150692 | 15:23 |
| *** erwan_taf has quit IRC | 15:23 | |
| devananda | we could also do a microversion bump | 15:23 |
| NobodyCam | devananda: which question? | 15:24 |
| devananda | NobodyCam: for the NOSTATE->AVAILABLE change, we /could/ make this non-API-affecting. think that's important? | 15:25 |
| victor_lowther | devananda: Meh. | 15:25 |
| victor_lowther | I like preserving backwards compatibility | 15:25 |
| devananda | so we've got backwards compat for the verbs | 15:26 |
| rloo | devananda, victor_lowther: if we want to preserve backwards compat, wouldn't non-api-affecting do that? | 15:26 |
| victor_lowther | but I also like APIs to not get too out of sync with our internals | 15:26 |
| devananda | like, a deploy is started by "PUT {active}" and such | 15:26 |
| devananda | none of the verbs are changing | 15:27 |
| devananda | what *is* changing in the current patch is the value of node.provision_state that indicates "hey this node is not in use" | 15:27 |
| *** jcoufal_ has joined #openstack-ironic | 15:27 | |
| devananda | it's going from "None" to "available" | 15:27 |
| victor_lowther | right | 15:28 |
| victor_lowther | How much relies on that exact value? | 15:28 |
| devananda | so the API semantics change a little bit, and sdague raised this when looking at the tempest patch to maintain compat between juno and kilo | 15:28 |
| devananda | apparently, both tempest and nova did | 15:28 |
| victor_lowther | figures. | 15:28 |
| devananda | as they key off that to know when a delete is complet | 15:28 |
| victor_lowther | Well, then, can we offer the current API and a version-bumped one at the same time? | 15:30 |
| *** jcoufal has quit IRC | 15:30 | |
| victor_lowther | because this is sounding like a "why not both" sort of situation to me. | 15:30 |
| devananda | *nod[B[B[B[B[BI suspect we can | 15:31 |
| devananda | gah ... i'm having wireless issues today | 15:32 |
| openstackgerrit | Merged openstack/ironic: Adds get_glance_image_properties https://review.openstack.org/146099 | 15:32 |
| NobodyCam | :( which hotel are you at | 15:32 |
| devananda | we can not-change it easily enough | 15:32 |
| devananda | NobodyCam: one too crappy to even remember its name. I'm on my cell | 15:32 |
| NobodyCam | lol :( | 15:32 |
| openstackgerrit | Merged openstack/ironic: Improve test assertion for get_glance_image_properties https://review.openstack.org/150652 | 15:32 |
| devananda | I'm looking into how easily I can do a microversion bump, eg. v1_1 | 15:32 |
| *** erwan_taf has joined #openstack-ironic | 15:34 | |
| devananda | victor_lowther: would we prevent any of the new state transitions in the version-bumped API? | 15:35 |
| rloo | devananda: why not make it non-api-changing for now so we can move forward and see what other wrinkles might be in store with the other changes. | 15:36 |
| victor_lowther | I was assuming that the version-bumped API was where the new state transitions would be visible | 15:36 |
| *** jerryz has quit IRC | 15:37 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: DO NOT MERGE - testing state mapping https://review.openstack.org/150821 | 15:38 |
| devananda | if I did it right, that should pass tempest | 15:38 |
| devananda | it hides the renaming of "None" -> "available" | 15:38 |
| devananda | victor_lowther: nodes *in* a new state still need to be visible in some way through the old API | 15:39 |
| devananda | victor_lowther: what about, in current API, a) don't change any current state or verb names, b) don't allow any new verbs, but c) still expose new states IFF a node is in one (eg, because someone else used the new API to put it there) | 15:41 |
| victor_lowther | That is why I said state transitions, not states. | 15:41 |
| rloo | devananda: we also changed the node's target_provision_state from NOSTATE to AVAILABLE when something failed. should we also map that too? | 15:41 |
| devananda | rloo: seems reasonable | 15:42 |
| rloo | devananda: I think your suggestion to victor_lowther makes sense and would cover that case | 15:44 |
| NobodyCam | brb | 15:44 |
| *** pcrews has joined #openstack-ironic | 15:46 | |
| rloo | devananda: I guess it wouldn't make sense to keep the existing provision_state & target_provision_state with whatever values it had before, and add some new foo_state & target_foo_state for the new world, and deprecate the provision_states later. | 15:46 |
| jroll | devananda: blah. | 15:47 |
| devananda | rloo: I think that would be much more complicated for users | 15:47 |
| *** bradjones has quit IRC | 15:47 | |
| devananda | and by users, I mean jroll | 15:48 |
| rloo | devananda: users could change their code to use the foo_states. vs changing their code to use a new v_1? | 15:48 |
| rloo | jroll: what say you? :-) | 15:48 |
| jroll | devananda: I understand why it would be nice to preserve the current semantics here, but it is a fairly minor thing and idk if I want such a small thing to hold us up for a cycle | 15:48 |
| devananda | jroll: what if it held us up for a day? | 15:48 |
| jroll | like, I'm going to need to make a single change in each external tool, not a huge deal | 15:48 |
| jroll | a day? | 15:49 |
| jroll | so you're saying keep it to get through tempest and then change it? | 15:49 |
| *** harlowja_away has quit IRC | 15:49 | |
| devananda | no | 15:49 |
| devananda | i'm saying add a v1_1 which does all the new things | 15:49 |
| *** ndipanov has quit IRC | 15:49 | |
| devananda | and some limiting functions in v1 that make it *look* the same | 15:49 |
| *** bradjones has joined #openstack-ironic | 15:49 | |
| devananda | all the code will actually be in v1, with a shim that says "if you asked for v1, run this backwards mapping" | 15:50 |
| devananda | and "if you asked for v1_1, then you get the things" | 15:50 |
| jroll | and the only backwards mapping would be AVAILABLE -> NOSTATE afaict with only one cup of coffee? | 15:50 |
| jroll | how much additional effort is this? | 15:51 |
| devananda | that and preventing the new verbs in the NodeStateController's PUT method | 15:51 |
| jroll | I guess it does provide an upgrade path without breaking all the things | 15:51 |
| devananda | right | 15:51 |
| jroll | idk if we need to block new verbs | 15:51 |
| rloo | would we show MANAGEABLE, or do MANAGEABLE -> NOSTATE | 15:51 |
| jroll | if you're using those new verbs, you clearly know about the new api and what you're doing | 15:52 |
| devananda | jroll: right. if you want to use new verbs, talk to v1_1 | 15:52 |
| jroll | we'd have to show MANAGEABLE, that has other implications, like "cannot start a deploy" | 15:52 |
| rloo | my guess is that if people have code checking provision_state, MANAGEABLE would break them too. | 15:52 |
| devananda | rloo: we can't *hide* the new states in v1 | 15:52 |
| devananda | if two clients are using different versions of the API, yea, one may affect the other | 15:52 |
| NobodyCam | ah ha! | 15:53 |
| rloo | devananda: but AVAILABLE is a new state and we're hiding that. | 15:53 |
| jroll | AVAILABLE is a new name | 15:53 |
| devananda | rloo: AVAILABLE is a renamed NOSTATE | 15:53 |
| jroll | not a new state | 15:53 |
| devananda | semantically identical to NOSTATE | 15:53 |
| jroll | devananda: I guess I think I would +2 either path, a version bump would make my life slightly easier as a deployer | 15:53 |
| jroll | but I could deal with just a straight rename | 15:54 |
| jroll | did I just kill the channel? :P | 15:57 |
| *** EmilienM|afk is now known as EmilienM | 15:58 | |
| NobodyCam | lol | 15:59 |
| NobodyCam | jroll: yes... and good morning | 15:59 |
| NobodyCam | :) | 15:59 |
| jroll | morning :D | 15:59 |
| rloo | I am thinking of jroll's comment about whether AVAILABLE <-> NOSTATE is the only thing | 15:59 |
| jroll | or maybe deva is just writing me some more code | 16:00 |
| *** ndipanov has joined #openstack-ironic | 16:00 | |
| jroll | rloo: I think that's the only renamed state, yes | 16:00 |
| * jroll chugs coffee before being sure of that | 16:00 | |
| rloo | jroll: right, and if that's the only thing we need to deal with wrt backwards compatibility | 16:00 |
| *** david-lyle_afk is now known as david-lyle | 16:01 | |
| vdrok | hi everyone! | 16:02 |
| rloo | jroll: so there are some other incompatibilities. Eg, I think DELETED -> ERROR in old, but DELETED -> DELETEFAIL in new | 16:02 |
| jroll | hmm | 16:03 |
| vdrok | devananda, one more question about https://review.openstack.org/#/c/130228/ , if the point is to PUT what we receive from GET, then internal attributes should be filtered? | 16:03 |
| rloo | jroll: maybe that's considered a rename too? | 16:04 |
| jroll | rloo: DELETING -> ERROR? | 16:04 |
| jroll | idk how it could end up DELETED -> ERROR, the only thing that happens once it's in DELETED is to move it to NOSTATE | 16:04 |
| rloo | jroll: maybe it is DELETING -> ERROR. I can't recall now. | 16:04 |
| jroll | probably | 16:04 |
| jroll | I don't think I've ever seen that though :) | 16:04 |
| NobodyCam | morning vdrok :) | 16:04 |
| vdrok | morning NobodyCam ! | 16:05 |
| jroll | hiya vdrok :) | 16:05 |
| rloo | jroll: yeah, just looked. we don't actually use DELETED except as a verb, ie not as a value in *provision_state. so yeah, DELETING | 16:05 |
| vdrok | morning jroll , rloo | 16:05 |
| rloo | hi vdrok | 16:05 |
| jroll | rloo: yeah, I guess we'll have to enumerate the differences somewhere and decide what to do with each. I can only assume devananda disappeared to hack on this and we can talk about it on the review :) | 16:06 |
| rloo | jroll: i think it might be worth doing the minimum to get further with the new states and see what/if there are other hiccups | 16:08 |
| jroll | rloo: the thing is, I don't think any functional tests look for error states | 16:08 |
| rloo | jroll: but does 'minimum' == maintaining backwards compatibility in the meantime? or breaking things 'temporarily' | 16:08 |
| jroll | soudns like the former | 16:09 |
| jroll | idk why we would put in extra effort for a version bump, if we're going to break something anyhow | 16:09 |
| rloo | i'd rather defer doing the version bump until we have to. I wish we could assume you're our only real user. so we'd need to satisfy you, nova, tempest, tripleO :-) | 16:10 |
| *** jistr has quit IRC | 16:11 | |
| *** ijw has joined #openstack-ironic | 16:12 | |
| *** jistr has joined #openstack-ironic | 16:12 | |
| jroll | rloo: well, having the version bump will keep my tooling from breaking. if we don't do a bump, I can fix my tooling quickly/easily, but it will be broken until I do | 16:13 |
| rloo | jroll: I was thinking no version bump and no api-change, so use devananda's mapping: https://review.openstack.org/#/c/150821/ | 16:14 |
| jroll | rloo: but then what's the path forward | 16:15 |
| jroll | ? | 16:15 |
| rloo | jroll: ie, even with a version bump, we'd need something like 150821 for v1 | 16:15 |
| rloo | jroll: that doesn't prevent new states from being added | 16:15 |
| jroll | rloo: but how do we expose e.g. AVAILABLE from there? | 16:15 |
| jroll | we'll eventually need another version that exposes that | 16:15 |
| jroll | we can't have the API return NOSTATE forever | 16:15 |
| rloo | right. I was thinking that we'd have a better idea after adding one or more new states. | 16:16 |
| rloo | jroll: at which time, we decide to go with v1_1 or just change to AVAILABLE. | 16:17 |
| rloo | jroll: I'd just like to know what else we might bump into. | 16:18 |
| jroll | hmm, yeah | 16:18 |
| NobodyCam | brb | 16:18 |
| rloo | whatever we do, we need to decide before kilo is released though | 16:18 |
| jroll | right | 16:19 |
| devananda | so | 16:21 |
| devananda | given that the semantics we're talking about changing here are very minor | 16:22 |
| devananda | I don't currently see any reason we couldn't keep both versions around for the same length of time | 16:22 |
| devananda | and if we needed to, give v1 a deprecation warning to encourage folks to use v1_1 | 16:23 |
| devananda | the oddity will be as we add new features (not new states) | 16:23 |
| devananda | they'll show up in both | 16:23 |
| jroll | I mean, we could do it for as long as v1 is alive | 16:23 |
| *** bradjones has quit IRC | 16:23 | |
| devananda | so v1_1, as I'm thinking about it currently, is just "v1 plus new states" | 16:23 |
| devananda | if we added a "make-some-coffee" interface, that'll show up in both | 16:24 |
| jroll | and that's fine, I think | 16:24 |
| jroll | adding an endpoint isn't backwards incompatible, generally | 16:24 |
| devananda | right | 16:24 |
| rloo | and the overhead of having/maintaining another version isn't much? | 16:25 |
| devananda | things like the new node.driver_internal_info attribute | 16:25 |
| devananda | that'll show up in both | 16:25 |
| *** erwan_taf has quit IRC | 16:25 | |
| devananda | rloo: another real version? yea, that's some overhead. I'm fairly sure this can be done as a shim layer | 16:25 |
| rloo | our client defaults to v1 I think. | 16:25 |
| devananda | less than a hundred LOC | 16:25 |
| jroll | right, we'll need to add client code etc | 16:26 |
| devananda | yup | 16:26 |
| jroll | change nova eventually | 16:26 |
| jroll | not a huge deal | 16:26 |
| devananda | well, the nova chagne alraedy landed | 16:26 |
| *** bradjones has joined #openstack-ironic | 16:26 | |
| jroll | I mean change nova to point at v1_1 and only use AVAILABLE | 16:26 |
| devananda | ah | 16:26 |
| jroll | but that's a ways off | 16:26 |
| devananda | yes | 16:26 |
| jroll | wonder if we should do a major version for the client when we default to 1_1 | 16:27 |
| rloo | would we need tempest tests for v1 and v1_1? | 16:27 |
| jroll | also, can we please call this v1.1 instead of v1_1 | 16:27 |
| jroll | rloo: we'll need tests for the differences | 16:27 |
| jroll | "need" | 16:27 |
| jroll | I'd love to get on the whole "move api tests to our tree" thing | 16:28 |
| rloo | we should probably use whatever versioning format that other openstack projects use. if there is consistency... | 16:28 |
| devananda | jroll: no | 16:29 |
| * jroll wonders what everett toews' irc nick is | 16:29 | |
| jroll | devananda: for why? | 16:30 |
| jroll | (just curious mostly) | 16:31 |
| *** ukalifon1 has quit IRC | 16:31 | |
| devananda | jroll: beacuse v1.1 is not a valid class name in python | 16:31 |
| jroll | -.- | 16:31 |
| jroll | devananda: do we use the string to evaluate the class? we can't just do if version == '1.1': api.v1_1() | 16:32 |
| jroll | ? | 16:32 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/150496 | 16:32 |
| *** Nisha has joined #openstack-ironic | 16:36 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Support for non-Glance image references https://review.openstack.org/136741 | 16:37 |
| *** penick has joined #openstack-ironic | 16:39 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add InspectInterface for node-introspection https://review.openstack.org/149817 | 16:40 |
| *** penick_ has joined #openstack-ironic | 16:42 | |
| *** penick has quit IRC | 16:44 | |
| *** penick_ is now known as penick | 16:44 | |
| *** zz_jgrimm is now known as jgrimm | 16:44 | |
| openstackgerrit | Merged openstack/ironic: Add iRMC Driver and its iRMC Power module https://review.openstack.org/144901 | 16:45 |
| *** coolsvap is now known as coolsvap|afk | 16:45 | |
| *** rakesh_hs has quit IRC | 16:45 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 16:46 |
| Nisha | NobodyCam, hi | 16:46 |
| Nisha | devananda, | 16:47 |
| Nisha | hi | 16:47 |
| Nisha | Good Morning | 16:47 |
| NobodyCam | morning Nisha :) | 16:48 |
| Nisha | NobodyCam, morning :) | 16:49 |
| NobodyCam | :) | 16:49 |
| Nisha | NobodyCam, i had a query or rather need advice from ironic on https://review.openstack.org/133534 | 16:49 |
| Nisha | s/ironic/ironic community | 16:49 |
| *** vdrok is now known as vdrok_afk | 16:50 | |
| *** Marga_ has quit IRC | 16:51 | |
| *** Marga_ has joined #openstack-ironic | 16:51 | |
| *** romcheg has quit IRC | 16:52 | |
| *** romcheg has joined #openstack-ironic | 16:52 | |
| Nisha | NobodyCam, in this spec we wanted to support the following: flavor side : {'capabilities:key1' : 'val2'} , ironic side : {'capabilities' : 'key1: val1 val2 val3' } | 16:53 |
| *** romcheg has quit IRC | 16:53 | |
| Nisha | NobodyCam, for above i had proposed changes in compute_capabilities_filter to Nova | 16:53 |
| Nisha | NobodyCam, today i got comment from one rreviewer, and as per that the above can be achieved without any modification in ironic and filter...just slight difference in how we give it at falvor side | 16:55 |
| Nisha | s/falvot/flavor | 16:55 |
| Nisha | s/falvor/flavor | 16:55 |
| NobodyCam | :) | 16:55 |
| *** Marga_ has quit IRC | 16:56 | |
| Nisha | NobodyCam, with modification it will look like : at flavor side : {'capabilities:key1' : '<in> val2' } , at ironic side : {'capabilities' : 'key1:val1 val2 val3'} | 16:56 |
| devananda | Nisha: what if there are multiple keys? | 16:57 |
| Nisha | the above is already supported by Nova and will/should not require any modification in filter. | 16:58 |
| Nisha | devananda, it will work | 16:58 |
| devananda | what will it look like in Ironic? | 16:58 |
| Nisha | i have tested above manulally at python shell using nova scheduler python files and it works...need to try with scheduler in stack.. | 16:59 |
| Nisha | devananda, it will look like following at ironic | 16:59 |
| *** absubram has joined #openstack-ironic | 16:59 | |
| *** Nisha has quit IRC | 17:02 | |
| *** Nisha has joined #openstack-ironic | 17:03 | |
| Nisha | devananda, sorry internet disconnected | 17:04 |
| Nisha | so at ironic side capabilities look like: | 17:04 |
| Nisha | {'capabilities': 'key1:val1 val2 val3,key2:v2,key3:v3'} This is how ironic supports today......in this example assumed that key1 has multiple possible values | 17:04 |
| *** rwsu-afk is now known as rwsu | 17:05 | |
| NobodyCam | oh brb | 17:07 |
| devananda | Nisha: got it, thanks | 17:08 |
| *** ifarkas has quit IRC | 17:08 | |
| Nisha | devananda, so do we really requirechanges in nova then? or we can live with the above kind of usage at flavor side | 17:09 |
| *** wendar_ is now known as wendar | 17:09 | |
| devananda | seems fine ... | 17:09 |
| Nisha | devananda, still before abandoning the spec and code i will just test it in nova in stack as it was just manual testing | 17:10 |
| *** jistr has quit IRC | 17:10 | |
| devananda | ++ | 17:10 |
| Nisha | devananda, thanks.... | 17:10 |
| lucasagomes | Nisha, when you get some time also please take a look at the #141012 patch | 17:11 |
| *** dtantsur is now known as dtantsur|afk | 17:11 | |
| dtantsur|afk | g'night everyone | 17:11 |
| devananda | jroll: looks like my 7 line patch is enough to pass tempest | 17:11 |
| jroll | devananda: yeah, I'm not surprised | 17:12 |
| devananda | nope | 17:12 |
| devananda | lucasagomes: API question for you | 17:12 |
| lucasagomes | devananda, hey there | 17:12 |
| * lucasagomes needs to catch up with the scrollback | 17:12 | |
| jroll | devananda: I think I like the v1_1 thing | 17:12 |
| devananda | lucasagomes: well, you should just read scrollback. v1_1 is discussed and so far everyone's liking it | 17:13 |
| devananda | I'm looking at how to implement cleanly, with the least maintenance burden | 17:13 |
| devananda | current ideas: | 17:13 |
| devananda | - add a parameter to several function calls that toggles the behavior, default to juno-states, and add controllers/v1_1 that pssses in the parameter to turn on kilo states | 17:14 |
| devananda | - try to do ^ using a mixin instead of a parameter | 17:14 |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Add 'irmc_' prefix to optional properties https://review.openstack.org/150849 | 17:15 |
| devananda | - wholesale move controllers/v1 to /v1_1, then put the mixin under /v1/ | 17:16 |
| devananda | (I dont see any real benefit to that) | 17:16 |
| devananda | - have Nodes and NodeStates objects auto-detect what controller they were called from and behave appropriately | 17:17 |
| devananda | (means no parameter or mixin, jsut a shim controller, but it looks like those classes don't have the request context) | 17:17 |
| devananda | (that might be possible in the NodeStatesController and NodeController classes though) | 17:18 |
| *** rloo is now known as rloo_afk | 17:19 | |
| lucasagomes | devananda, haven't read it fully. The problem is that for v1.1 the provision state that indicates that the nodes are available for deployment is AVAILABLE and for v1 is NOSTATE (None) right? | 17:19 |
| rloo_afk | back in a few hours, volunteering at my daughter's school... | 17:19 |
| * lucasagomes reads more | 17:19 | |
| devananda | lucasagomes: that's part of it. read more :) | 17:20 |
| devananda | lucasagomes: the mapping to preserve old API is trivially small | 17:20 |
| *** pensu has joined #openstack-ironic | 17:20 | |
| *** MattMan has left #openstack-ironic | 17:21 | |
| *** anderbubble has joined #openstack-ironic | 17:21 | |
| *** EmilienM is now known as EmilienM|afk | 17:22 | |
| *** Marga_ has joined #openstack-ironic | 17:22 | |
| *** anderbubble has quit IRC | 17:25 | |
| *** Marga_ has quit IRC | 17:26 | |
| lucasagomes | devananda, +1 on the v1_1 I would like that we kept the currrent v1 as-is so we don't break current users of the API | 17:28 |
| lucasagomes | for the long term, we may even want to add some microversioning to our API | 17:28 |
| jroll | +1 | 17:29 |
| NobodyCam | lucasagomes: yea | 17:29 |
| jroll | microversion all the things | 17:29 |
| lucasagomes | like if we do a GET / we now have a version which just points to "v1" but we could have some other field like micro_version="1.1" that clients could rely when interacting with our api | 17:29 |
| lucasagomes | actually better, we can also it as a header to the request | 17:29 |
| lucasagomes | X-Ironic-MicroVersion idk | 17:29 |
| lucasagomes | so we don't need to get / | 17:29 |
| lucasagomes | (can add both really) | 17:29 |
| jroll | lucasagomes: to be clear, we're still going to add the new states and whatnot to v1, but renamed states e.g. NOSTATE -> AVAILABLE will remain the old way for v1 | 17:30 |
| jroll | also, there's a standard header for api version | 17:30 |
| * jroll looks | 17:30 | |
| lucasagomes | jroll, yeah I got that, most of the other states are internal | 17:30 |
| devananda | lucasagomes: http://specs.openstack.org/openstack/nova-specs/specs/kilo/approved/api-microversions.html | 17:30 |
| lucasagomes | the NOSTATE -> AVAILABLE is what really changes things right? | 17:30 |
| * lucasagomes reads devananda links | 17:31 | |
| lucasagomes | ah... wow yeah gotta read that | 17:31 |
| jroll | yeah, maybe DELETING -> ERROR vs DELETING -> DELETEFAIL | 17:31 |
| lucasagomes | jroll, yeah we can translate whatever state from v1_1 to v1 using the shim | 17:31 |
| lucasagomes | yeah I'm good with that | 17:32 |
| jroll | yep | 17:32 |
| lucasagomes | devananda, the microversion is good, especially for us that have to interact with the deploy/discovery ramdisk | 17:33 |
| jroll | X-OpenStack-Compute-API-Version: 2.114 | 17:33 |
| jroll | I'm not entirely sure I understand the whole min/max version thing, those are min/max supported versions? | 17:34 |
| * lucasagomes have to read it, still finshing the configdrive thing | 17:34 | |
| jroll | in general I love this and I think we should do it | 17:35 |
| *** ndipanov has quit IRC | 17:36 | |
| jroll | make it way easier to change the api | 17:36 |
| lucasagomes | yeah this + using IPA as default ramdisk | 17:36 |
| * lucasagomes hates bash now | 17:36 | |
| jroll | lol | 17:37 |
| JayF | IPA replacing PXE ramdisk | 17:37 |
| JayF | might be a good hack choice for the sprint | 17:37 |
| jroll | ++ | 17:37 |
| jroll | lucasagomes: you coming to sf meetup? | 17:38 |
| lucasagomes | jroll, I won't make it unfortunately :( only the french one | 17:38 |
| JayF | BTW; if anyone is going to be in SF early for the mid-cycle; I'm doing an iptables primer 2/10 over lunch (noon-1pm). | 17:38 |
| jroll | and can that change if I spin up a bunch of openshift instances? | 17:38 |
| JayF | (day before the mid-cycle) | 17:38 |
| lucasagomes | JayF, yeah, it's not diffcult seems (I looked a bit into it) | 17:39 |
| lucasagomes | but requires some refactoring, like taking all the lookup() heartbeat() methods out of the agent driver | 17:39 |
| lucasagomes | putting in a common place etc | 17:39 |
| BadCub_ | JayF: I wont be arriving until later in the evening of the 10th :-( | 17:40 |
| lucasagomes | devananda, you're going to work on the v1_1? | 17:42 |
| *** romcheg has joined #openstack-ironic | 17:45 | |
| devananda | lucasagomes: I've started looking into it, yah. but I should really stop now | 17:45 |
| devananda | JayF: I arrive in SFO at like 8am | 17:46 |
| lucasagomes | Nisha, for the pass capabilities from flavor to Ironic, if u wanna test you can use https://review.openstack.org/#/c/146189/4/ironic/drivers/modules/deploy_utils.py | 17:46 |
| lucasagomes | Nisha, the local boot blueprint had a dependency on that spec, so I already wrote the code to consume it if u wanna take a look | 17:46 |
| devananda | lucasagomes: i'm tripping up right now on how to detect what root controller class was invoked | 17:47 |
| devananda | it seems like that would be an obvious thing, bu tit's not | 17:47 |
| *** eghobo has joined #openstack-ironic | 17:49 | |
| BadCub_ | We are driving up to SF. gotta put the new mods to the car to test *wg* | 17:49 |
| lucasagomes | devananda, hmm, right, because /nodes /ports are attributes to the controller class and not actully inherit from it | 17:50 |
| devananda | right | 17:50 |
| devananda | if I can detect, from within the NodesController and NodeStateController, what the *ROOT* controller class was | 17:50 |
| devananda | then it's easy | 17:51 |
| devananda | that should be on the pecan.request context somewhere, I'd think | 17:51 |
| lucasagomes | devananda, maybe we should have a shim class for the rest.RestController where the version can be stored | 17:51 |
| lucasagomes | and we keep that shim class in the base.py | 17:51 |
| devananda | lucasagomes: that wouldn't necessarily help | 17:51 |
| devananda | i mean, it wouldn't be accessible any differently | 17:51 |
| *** r-daneel has quit IRC | 17:52 | |
| devananda | oh! got it | 17:52 |
| lucasagomes | right, cause with that ^ we may be able to microversion the Controller itself | 17:52 |
| lucasagomes | say NodeController | 17:52 |
| lucasagomes | so for v1_1 the chassis, ports can even use the v1 controllers | 17:53 |
| lucasagomes | because nothing change on then, only the Nodes one (which holds the provision states) | 17:53 |
| * lucasagomes and I have to look in the code also, can't think much now, late and almost getting the configdrive thing working :D | 17:53 | |
| devananda | version hook | 17:53 |
| lucasagomes | oh | 17:54 |
| devananda | we need to make pecan's base app set the version | 17:54 |
| devananda | rather than try to parse it out of the request context later | 17:54 |
| jroll | devananda: +1 | 17:55 |
| lucasagomes | ah nice | 17:55 |
| jroll | devananda: eventually that should come out of a header and default to v1 | 17:55 |
| jroll | (would be a step towards microversion stuff) | 17:56 |
| devananda | yah | 17:56 |
| jroll | maybe we should just do that now | 17:56 |
| devananda | ... | 17:56 |
| devananda | probably | 17:56 |
| * jroll hopes and dreams | 17:56 | |
| jroll | it is our first version bump, don't see why not | 17:56 |
| lucasagomes | $ ls /mnt/ | 17:56 |
| lucasagomes | ec2 openstack !!!! | 17:56 |
| jroll | we can come back and clean up with decorators and stuff later | 17:56 |
| jroll | lucasagomes: u da man | 17:56 |
| *** achanda has joined #openstack-ironic | 17:57 | |
| lucasagomes | gotta test the ironic uploading to swift yet | 17:57 |
| lucasagomes | but cool, almost here | 17:57 |
| lucasagomes | there* | 17:57 |
| *** Marga_ has joined #openstack-ironic | 17:58 | |
| NobodyCam | fyi lucasagomes that error I am hitting is NOT from the initial deploy instead it appers to be comming from a rebuild | 17:59 |
| lucasagomes | NobodyCam, oh :/ | 17:59 |
| lucasagomes | NobodyCam, the rebuild is odd then... because we don't repartition the disk (AFAIR) | 17:59 |
| lucasagomes | so the partitions should be there no problem | 17:59 |
| NobodyCam | 2015-01-28 17:25:36.933 4928 INFO ironic.drivers.modules.pxe [-] Deployment to node 5b1f54c0-468d-4614-96fa-f7e6d5320c15 done | 18:00 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Add logic to store the config drive passed by Nova https://review.openstack.org/150877 | 18:00 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/python-ironicclient: Add --config-drive to node-set-provision-state https://review.openstack.org/150878 | 18:00 |
| NobodyCam | i'm looking in to what the issue | 18:00 |
| NobodyCam | is | 18:00 |
| lucasagomes | NobodyCam, so it finished deployment. Did it work? | 18:01 |
| devananda | sweet | 18:02 |
| devananda | got it | 18:02 |
| *** derekh has quit IRC | 18:03 | |
| *** trown is now known as trown|lunch | 18:04 | |
| devananda | http://paste.openstack.org/show/uh1a75NvS2c0oLPVcKe8/ | 18:05 |
| devananda | \o/ | 18:05 |
| devananda | jroll: lucasagomes: ^ | 18:05 |
| jroll | can I +A a paste? | 18:06 |
| devananda | totally | 18:06 |
| jroll | lgtm | 18:06 |
| jroll | that's awesome | 18:06 |
| lucasagomes | devananda, w00t! | 18:07 |
| lucasagomes | yeah, pretty good!! | 18:07 |
| jroll | devananda: mind making the version hook a separate patch when you do that? | 18:07 |
| jroll | and rebase the other stuff on top | 18:07 |
| * jroll abandons tempest patch | 18:07 | |
| lucasagomes | it's being a productive day for the project today, great ideas | 18:08 |
| lucasagomes | and with that, I will call it a day | 18:08 |
| lucasagomes | late and I'm hungry :) | 18:08 |
| *** lucasagomes is now known as lucas-dinner | 18:09 | |
| lucas-dinner | have a good night everyone | 18:09 |
| lucas-dinner | devananda, I will be in the channel and read the logs later :D but, yeah that looks awesome already! | 18:09 |
| jroll | night lucas-dinner :) | 18:09 |
| jroll | poor one out for v1 tonight | 18:09 |
| lucas-dinner | +1 | 18:10 |
| BadCub_ | night lucas-dinner | 18:10 |
| devananda | how is it that, this week, i've done the most work BEFORE coffee? | 18:12 |
| devananda | (yes, i haven't had coffee yet) | 18:12 |
| lucas-dinner | devananda, file a bug, there's something wrong with ur coffee | 18:13 |
| BadCub_ | devananda: NO COFFEE??? *mock look of horror* Oh the humanity! | 18:13 |
| * lucas-dinner stills here | 18:13 | |
| NobodyCam | night lucas-dinner | 18:13 |
| NobodyCam | *cough* no coffee! | 18:14 |
| NobodyCam | oh :( | 18:14 |
| *** openstackgerrit has quit IRC | 18:14 | |
| *** openstackgerrit has joined #openstack-ironic | 18:14 | |
| openstackgerrit | Gopi Krishna S proposed openstack/ironic-specs: Add pxe_ucs driver to support Cisco UCS B/C/M-series servers https://review.openstack.org/139517 | 18:14 |
| jroll | devananda: who knows, you could just be vomiting code everywhere and think you're doing work :P | 18:15 |
| *** jcoufal_ has quit IRC | 18:17 | |
| *** pradipta is now known as pradipta_away | 18:18 | |
| devananda | jroll: that's likely | 18:20 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: DO NOT MERGE - testing state mapping https://review.openstack.org/150821 | 18:20 |
| devananda | also, that's a very small patch ^ | 18:20 |
| jroll | nice | 18:21 |
| * jroll does not merge | 18:21 | |
| devananda | no, i need to reorder them real quick | 18:22 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 18:22 |
| *** eghobo has quit IRC | 18:23 | |
| *** lazy_prince is now known as killer_prince | 18:27 | |
| *** harlowja has joined #openstack-ironic | 18:28 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Rename NOSTATE to AVAILABLE https://review.openstack.org/149754 | 18:32 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add support for API microversions https://review.openstack.org/150821 | 18:32 |
| devananda | jroll: ok ... there :) | 18:32 |
| jroll | woot | 18:32 |
| devananda | I'm sure it should, you know, have unit tests, or soemthing | 18:32 |
| devananda | feel free to edit while I'm away ... bbiafh | 18:32 |
| jroll | I was gonna say | 18:32 |
| jroll | heh, alright | 18:32 |
| jroll | yeah I'll add tests | 18:33 |
| *** EmilienM|afk is now known as EmilienM | 18:38 | |
| *** achanda has quit IRC | 18:40 | |
| *** romcheg has quit IRC | 18:40 | |
| *** absubram has quit IRC | 18:41 | |
| *** achanda has joined #openstack-ironic | 18:42 | |
| *** romcheg has joined #openstack-ironic | 18:42 | |
| *** ijw has quit IRC | 18:43 | |
| * jroll wonders if this should only allow existing versions | 18:43 | |
| *** pelix has quit IRC | 18:43 | |
| *** spandhe has joined #openstack-ironic | 18:43 | |
| *** trown|lunch is now known as trown | 18:53 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Rename NOSTATE to AVAILABLE https://review.openstack.org/149754 | 18:54 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add support for API microversions https://review.openstack.org/150821 | 18:54 |
| jroll | added tests here | 18:54 |
| jroll | thinking on if we want to specify what versions are allowed | 18:54 |
| *** eghobo has joined #openstack-ironic | 18:55 | |
| devananda | yes | 19:00 |
| devananda | MIN and MAX | 19:00 |
| devananda | probably check those in the api/controllers/root.py or in api/controllers/v1/__init__.py | 19:01 |
| jroll | devananda: in this patch? | 19:01 |
| devananda | sure. set to 1 in teh first patch, increment in the second | 19:02 |
| jroll | ok | 19:02 |
| * jroll on a call now | 19:02 | |
| jroll | feel free to take that over | 19:02 |
| * devananda goes out in search of food and coffee | 19:03 | |
| openstackgerrit | Merged openstack/ironic-specs: Add pxe_ucs driver to support Cisco UCS B/C/M-series servers https://review.openstack.org/139517 | 19:14 |
| NobodyCam | nice! | 19:17 |
| *** pensu has quit IRC | 19:18 | |
| *** Nisha_away has joined #openstack-ironic | 19:22 | |
| *** Nisha has quit IRC | 19:22 | |
| *** ijw has joined #openstack-ironic | 19:27 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 19:29 |
| NobodyCam | brb | 19:33 |
| *** achanda has quit IRC | 19:36 | |
| *** krtaylor has quit IRC | 19:41 | |
| *** erwan_taf has joined #openstack-ironic | 19:42 | |
| *** romcheg has quit IRC | 19:46 | |
| *** romcheg has joined #openstack-ironic | 19:48 | |
| *** krtaylor has joined #openstack-ironic | 19:53 | |
| *** anderbubble has joined #openstack-ironic | 19:54 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Hardware Inspection https://review.openstack.org/149823 | 20:00 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add states required for node-inspection https://review.openstack.org/147857 | 20:00 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add InspectInterface for node-introspection https://review.openstack.org/149817 | 20:00 |
| *** achanda has joined #openstack-ironic | 20:02 | |
| *** absubram has joined #openstack-ironic | 20:07 | |
| *** r-daneel has joined #openstack-ironic | 20:11 | |
| *** absubram has quit IRC | 20:13 | |
| *** absubram has joined #openstack-ironic | 20:13 | |
| devananda | jroll: still good for me to take over those patches again? | 20:13 |
| * devananda takes no answer as "yes" | 20:14 | |
| jroll | devananda: go ahead | 20:14 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Improve testing of the Node's REST API https://review.openstack.org/150603 | 20:15 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Rename NOSTATE to AVAILABLE https://review.openstack.org/149754 | 20:15 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add support for API microversions https://review.openstack.org/150821 | 20:15 |
| *** r-daneel has quit IRC | 20:16 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 20:19 |
| Shrews | jroll: ping | 20:23 |
| jroll | heya Shrews | 20:24 |
| Shrews | should i be concerned that i cannot get neutron client to work against Rax? | 20:24 |
| Shrews | "publicURL endpoint for network service in IAD region not found" | 20:24 |
| * jroll cries | 20:24 | |
| jroll | does any client work against rax? | 20:25 |
| jroll | let me poke neutron peeps | 20:25 |
| Shrews | jroll: yeah, nova seems to work great | 20:25 |
| jroll | Shrews: they say rackspace-python-neutronclient | 20:26 |
| jroll | Shrews: you mean with the rax plugin it works great | 20:26 |
| *** romcheg has quit IRC | 20:26 | |
| Shrews | oh, so only with API key auth? hrm... /me tries | 20:26 |
| jroll | idk | 20:26 |
| Shrews | jroll: k. thx | 20:26 |
| jroll | np | 20:26 |
| *** lucas-dinner has quit IRC | 20:28 | |
| Nisha_away | i am getting error "pkg_resources.DistributionNotFound: ironic==2015.1.dev111" when i am using ironic-dbsync command.....how to resolve it | 20:28 |
| *** romcheg has joined #openstack-ironic | 20:32 | |
| *** romcheg1 has joined #openstack-ironic | 20:33 | |
| jroll | Shrews: it's basically the last three commits here: https://github.com/roaet/python-neutronclient/commits/master | 20:33 |
| jroll | (if you wanted a pointer to what to do in shade) | 20:34 |
| Shrews | jroll: ack. thx | 20:35 |
| *** romcheg has quit IRC | 20:36 | |
| anderbubble | Still working on our proof-of-concept ironic cluster. I’ve got a node enrolled, and validation is returning this error: “Cannot validate iSCSI deploy. The following parameters were not passed to ironic: ['root_gb', 'image_source’]” | 20:39 |
| anderbubble | Is this normal for this time? Looks like these might get populated once I tell it what I want to deploy. Is that right? | 20:39 |
| jroll | anderbubble: that's correct | 20:40 |
| jroll | the error is when you run "ironic node-validate $uuid" ? | 20:40 |
| anderbubble | Right. | 20:40 |
| jroll | yeah | 20:40 |
| devananda | jroll: ooh. so we need to add the microversion in the resposne header, too | 20:40 |
| jroll | we should like, fix that | 20:40 |
| jroll | devananda: yeah, and the Vary | 20:41 |
| devananda | gilliard just pointed out that caches need that | 20:41 |
| jroll | I didn't think about it when reviewing/updating | 20:41 |
| devananda | and register the min/max in the service catalog / keystone | 20:41 |
| anderbubble | jroll, I’ll add that to my notes as something to clarify in the installation guide. :) | 20:41 |
| jroll | anderbubble: cool, thanks! :) | 20:41 |
| devananda | jroll: also, pecan++. this is what Nova did to get similar functionailty https://review.openstack.org/#/c/136759/14/nova/api/openstack/wsgi.py,cm | 20:43 |
| jroll | devananda: woot | 20:43 |
| *** rloo_afk is now known as rloo | 20:48 | |
| devananda | ohhey, idea - what if our API-visible error codes included a) a unique ID that references our error condition, and b) a link to online documentation about that error | 20:50 |
| NobodyCam | devananda: that would be very kewl.. how much work to maintain? | 20:51 |
| *** dprince has quit IRC | 20:51 | |
| devananda | make it automatic | 20:51 |
| devananda | put all the error code docs on a single URL with inline links for each code | 20:52 |
| devananda | we add a new exception in tree, it would show up in docs right away | 20:52 |
| NobodyCam | oh very kewl +1 | 20:59 |
| *** ijw has quit IRC | 21:01 | |
| *** achanda has quit IRC | 21:02 | |
| *** achanda has joined #openstack-ironic | 21:12 | |
| *** maurosr has quit IRC | 21:14 | |
| NobodyCam | brb | 21:15 |
| *** maurosr has joined #openstack-ironic | 21:15 | |
| *** slagle has quit IRC | 21:20 | |
| kragniz | devananda: we've been talking about doing something like that in glance | 21:20 |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-specs: Implement Zapping States https://review.openstack.org/140826 | 21:21 |
| *** achanda has quit IRC | 21:26 | |
| *** slagle has joined #openstack-ironic | 21:31 | |
| *** absubram has quit IRC | 21:37 | |
| anderbubble | Getting really close to *actually* deploying a baremetal node: following the installation guide, but it doesn’t say how to actually deploy a node. Should I be selecting “sharednet1” or the original “private” network? I expected sharednet1, as that’s the network created in the guide; but that produced an error saying there was no subnet assigned. | 21:43 |
| *** yjiang5_away is now known as yjiang5 | 21:45 | |
| *** achanda has joined #openstack-ironic | 21:46 | |
| *** romcheg1 has quit IRC | 21:52 | |
| *** Nisha_away has quit IRC | 21:55 | |
| *** achanda has quit IRC | 21:57 | |
| *** achanda has joined #openstack-ironic | 21:58 | |
| *** achanda has quit IRC | 22:00 | |
| *** achanda has joined #openstack-ironic | 22:00 | |
| *** achanda has quit IRC | 22:01 | |
| *** erwan_taf has quit IRC | 22:02 | |
| *** achanda has joined #openstack-ironic | 22:03 | |
| *** eghobo has quit IRC | 22:04 | |
| *** jgrimm is now known as zz_jgrimm | 22:15 | |
| *** Marga_ has quit IRC | 22:17 | |
| anderbubble | This is obviously a corner case; but I tested my stack install by creating a vm instance first, then converting to ironic. Now I’ve got a vm-based instance that I can’t delete, even with force-delete | 22:46 |
| anderbubble | It’d probably be good to be able to clean up this kind of instance without (presumably) having to revert my nova config back to libvirt and friends. | 22:47 |
| jroll | mysql -c 'truncate table instances'? :P | 22:47 |
| jroll | (don't do that) | 22:47 |
| *** ijw has joined #openstack-ironic | 22:47 | |
| jroll | I really have no idea | 22:47 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Actively check iSCSI connection after login https://review.openstack.org/150187 | 22:48 |
| anderbubble | :) | 22:50 |
| anderbubble | Yeah, I could go troll through the db looking for the instance’s uuid; but I’m more likely to just ignore it for this install, since it’s proof-of-concept anyway. | 22:51 |
| anderbubble | Still, it’d be nice if force-delete could do it. | 22:51 |
| anderbubble | I guess this probably isn’t an ironic problem. :p | 22:51 |
| jroll | yeah :P | 22:52 |
| *** openstackgerrit has quit IRC | 23:06 | |
| *** openstackgerrit has joined #openstack-ironic | 23:06 | |
| openstackgerrit | Merged openstack/ironic-specs: Implement Zapping States https://review.openstack.org/140826 | 23:07 |
| victor_lowther | Another day, another BIOS bug | 23:08 |
| *** EmilienM is now known as EmilienM|afk | 23:08 | |
| JoshNang | \o/ zapping thanks NobodyCam | 23:09 |
| victor_lowther | Somehow, I doubt that ^[ -~]{0,MAX_ASSET_TAG_LEN}$ is really a valid regex | 23:09 |
| NobodyCam | victor_lowther: lol ... ahh the good old days weekly bios updates | 23:09 |
| NobodyCam | JoshNang: Thank you | 23:09 |
| victor_lowther | and yet, that is what the BIOS hands me to validate that a valid asset tag is being applied. | 23:10 |
| victor_lowther | So I get to fix a regex with a regex. | 23:11 |
| victor_lowther | JWZ would be proud. | 23:11 |
| * NobodyCam hates regex | 23:11 | |
| NobodyCam | I tend to have to learn it all over again every time | 23:12 |
| jroll | NobodyCam: regexr.com is gold | 23:12 |
| NobodyCam | ++ | 23:12 |
| NobodyCam | oh wait I haven't used that one | 23:14 |
| NobodyCam | thats nice | 23:14 |
| NobodyCam | TY jroll | 23:14 |
| jroll | :) | 23:14 |
| jroll | it's so good | 23:14 |
| NobodyCam | I think I got the No such file or directory thing | 23:18 |
| openstackgerrit | Victor Lowther proposed openstack/ironic: Add DRAC BIOS config vendor passthru API -- DO NOT MERGE https://review.openstack.org/150488 | 23:20 |
| victor_lowther | also, I would appreaciate a code review for where I am not using idiomatic python. | 23:23 |
| *** mrghort has joined #openstack-ironic | 23:25 | |
| mrghort | Hello everyone. Might I ask some n00b ironic questions, or is this more for dev discussion? | 23:30 |
| NobodyCam | mrghort: Happy to attempt to answer.. | 23:31 |
| *** anderbubble has quit IRC | 23:32 | |
| mrghort | Thanks! I have ironic installed, the commands work, I added my nodes, ports for my nodes, created images, set pxe_deploy_ramdisk & pxe_deploy_kernel in driver_info. IPMI commands work. When I reboot the node it tries to PXE but does not find an image to boot from. | 23:35 |
| mrghort | Looking in /tftpboot/pxelinux.cfg I don't see any node-specific files | 23:36 |
| mrghort | I sort-of assumed ironic should create them there for me? | 23:36 |
| NobodyCam | create / pull from glance | 23:37 |
| NobodyCam | can you paste the conductor log? | 23:37 |
| NobodyCam | we can pull the images out of glance. | 23:38 |
| NobodyCam | brb | 23:38 |
| mrghort | Jan 28 17:40:13 ironic journal: 2015-01-28 17:40:13.671 8947 INFO ironic.conductor.manager [-] Successfuly started conductor with hostname ironic.chameleon.tacc.utexas.edu. | 23:45 |
| mrghort | Jan 28 17:40:13 ironic ironic-conductor: 2015-01-28 17:40:13.671 8947 INFO ironic.conductor.manager [-] Successfuly started conductor with hostname ironic.chameleon.tacc.utexas.edu. | 23:45 |
| mrghort | Jan 28 17:40:13 ironic ironic-conductor: 2015-01-28 17:40:13.673 8947 INFO oslo.messaging._drivers.impl_rabbit [-] Connecting to AMQP server on ironic:5672 | 23:45 |
| mrghort | Jan 28 17:40:13 ironic ironic-conductor: 2015-01-28 17:40:13.728 8947 INFO oslo.messaging._drivers.impl_rabbit [-] Connected to AMQP server on ironic:5672 | 23:45 |
| mrghort | Jan 28 17:40:13 ironic journal: 2015-01-28 17:40:13.736 8947 INFO ironic.common.service [-] Created RPC server for service ironic.conductor_manager on host ironic.chameleon.tacc.utexas.edu. | 23:45 |
| mrghort | Jan 28 17:40:13 ironic ironic-conductor: 2015-01-28 17:40:13.736 8947 INFO ironic.common.service [-] Created RPC server for service ironic.conductor_manager on host ironic.chameleon.tacc.utexas.edu. | 23:45 |
| *** naohirot has joined #openstack-ironic | 23:45 | |
| NobodyCam | mrghort: please use paste.openstack.org | 23:46 |
| naohirot | good morning ironic, NobodyCam | 23:47 |
| mrghort | duh, sorry | 23:47 |
| NobodyCam | morning naohirot :) | 23:47 |
| mrghort | this is also in my log: http://paste.openstack.org/show/163407/ | 23:47 |
| naohirot | good morning devananda, rloo | 23:48 |
| NobodyCam | mrghort: how did you start the deploy? | 23:48 |
| mrghort | ironic node-set-power-state on | 23:49 |
| mrghort | do i need to deploy with nova? | 23:49 |
| NobodyCam | I would, | 23:50 |
| NobodyCam | have you used TripleO's devtest or devstack | 23:51 |
| mrghort | no, i already have my own openstack running so i tried to work with that | 23:53 |
| NobodyCam | that can work. have you build the deploy ram disk or agent image? | 23:54 |
| NobodyCam | you'll need to set the virt driver in nova to Ironic | 23:55 |
| naohirot | devananda: rloo: I really appreciate your generous treatment of driver.rst | 23:55 |
| mrghort | built the images per http://docs.openstack.org/developer/ironic/deploy/install-guide.html | 23:55 |
| NobodyCam | sorry have to stepaway agin | 23:56 |
| mrghort | compute_driver = nova.virt.ironic.IronicDriver | 23:57 |
| mrghort | i'll need to leave too. i'll get back on tomorrow. Thanks for the help! | 23:58 |
| *** mrghort has quit IRC | 23:58 | |
| naohirot | devananda: rloo: It seems I screwed up somehow when I submitted iRMC Management Driver which depends on iRMC Power driver. | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!