| devananda | mrda: i'm goign to open a BP for a generic timeout mechanism, as i suspect it's going to be somewhat complex | 00:02 |
|---|---|---|
| devananda | unfortunately we probably wont get to it in icehouse ... | 00:03 |
| mrda | so you just want the option removed in +bug/1270981 for now? With the generic "add config params for timeouts" to be handled seperately and independently? | 00:03 |
| devananda | i think taht's best | 00:10 |
| devananda | having an unused option in the config file is bad form | 00:10 |
| devananda | NobodyCam, lifeless - we're going to need to dogpile on getting timeouts and lock-breaking and such done during i-3 or tripleo will have problems moving to ironic | 00:12 |
| lifeless | devananda: ack; what do you need from me? | 00:13 |
| lifeless | mrda: so, I'm confused; the bug https://bugs.launchpad.net/ironic/+bug/1195073 - has it been fixed, or is it irrelevant to ironic ? | 00:14 |
| devananda | lifeless: i spotted 2 bugs cross referenced from nova, eg that one | 00:14 |
| devananda | which don't apply right now | 00:14 |
| devananda | that option is unused | 00:14 |
| devananda | basically ironic doesn't have timeouts today :( | 00:14 |
| mrda | lifeless: so I've created https://bugs.launchpad.net/ironic/+bug/1270981 to just remove the config option in Ironic | 00:14 |
| lifeless | devananda: will it apply in future ? | 00:14 |
| mrda | and so I was thinking that we can remove the link to Ironic in https://bugs.launchpad.net/ironic/+bug/1195073 | 00:15 |
| lifeless | mrda: based on what devananda is saying I suspect the answer is to make it invalid. | 00:15 |
| mrda | and devananda's creating a bp for adding in appropriate timeouts as future work (which I'm happy to help out with) | 00:15 |
| lifeless | though if there is no timeout, I'd argue the bug is valid, just premature :P | 00:16 |
| mrda | lifeless: so the Nova half is valid | 00:17 |
| devananda | lifeless: heh. you have a point -- the timeout should not default to "unset" | 00:17 |
| devananda | mrda: yes, nova half is valid (sorry if i wasn't clear on that) | 00:17 |
| devananda | lifeless: but it is premature to have a bug for something which isn't supported :) | 00:17 |
| mrda | please excuse my newbieness, but how do I mark the Ironic half as invalid or remove the link to Ironic? | 00:17 |
| devananda | mrda: have you joined the ironic-bugs team? | 00:18 |
| lifeless | mrda: you need to be in the bug triage team for ironic | 00:18 |
| mrda | most probably not :) | 00:18 |
| lifeless | devananda: is there a bug that pxe deploys don't timeout on ironic ? | 00:18 |
| devananda | lifeless: re timeouts, i will probably want an hour or so of your time to brainstorm on reasonably generic ways to handle timeouts across drivers' various needs | 00:18 |
| lifeless | devananda: I think it's important that there be one, tho I don't care whether it's this one or another | 00:19 |
| lifeless | devananda: ack, can do; later in the week perhaps ? | 00:19 |
| devananda | lifeless: there definitely needs to be something to track it. I'm filing a BP now and will start an etherpad to scratch notes in | 00:19 |
| devananda | as i think timeouts needs to span more than just PXE deploy | 00:19 |
| devananda | other drivers need timeouts, and different interfaces may need timeouts for passive things, too | 00:20 |
| lifeless | devananda: ok; my concern about not having a bug is that this is a defect right now, BP's are about future work [without getting into the craziness of splitting them out] | 00:20 |
| devananda | ack | 00:20 |
| mrda | devananda and lifeless : thanks for the LP advice | 00:20 |
| lifeless | so - if there isn't a bug, can we make one before we touch this existing one :) | 00:20 |
| devananda | lifeless: ack. i'll file a new bp that is clearer about the scope of this particula rproblem | 00:21 |
| lifeless | devananda: should I create a bug about it ? | 00:21 |
| devananda | er | 00:21 |
| devananda | i meant BUG | 00:21 |
| devananda | gah. bad fingers :) | 00:21 |
| lifeless | ok, cool - thank you :) | 00:21 |
| openstackgerrit | A change was merged to stackforge/pyghmi: Add 'get_health' to Command class https://review.openstack.org/67987 | 00:24 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1270986 | 00:25 |
| devananda | NobodyCam: you've been playing a lot with deployment. when you have a minute, can you look at / chime in on ^ ? | 00:26 |
| openstackgerrit | Michael Davies proposed a change to openstack/ironic: Remove unused config option - pxe_deploy_timeout https://review.openstack.org/67990 | 00:35 |
| devananda | lifeless: https://blueprints.launchpad.net/ironic/+spec/generic-timeouts | 00:37 |
| devananda | not a lot of detail yet | 00:37 |
| lifeless | ack | 00:39 |
| *** jbjohnso has quit IRC | 01:00 | |
| devananda | max_lobur_afk: commented on the exception traceback bug. i can still repro | 01:01 |
| mrda | devananda, should I be concerned that someone else is assigned to https://bugs.launchpad.net/ironic/+bug/1264596 ? Or should I just work on it? | 01:02 |
| devananda | hm | 01:03 |
| devananda | good point. didn't notice that as i was skimming, sorry | 01:03 |
| devananda | imre has been working on things. i'd check with him to avoid duplicating work | 01:03 |
| devananda | ifarkas on IRC | 01:03 |
| mrda | ok, thanks | 01:03 |
| *** greghaynes has quit IRC | 01:40 | |
| Haomeng | devananda: ping | 01:46 |
| *** coolsvap has quit IRC | 01:51 | |
| *** nosnos has joined #openstack-ironic | 01:59 | |
| *** jbjohnso has joined #openstack-ironic | 02:05 | |
| *** vkozhukalov has joined #openstack-ironic | 02:34 | |
| *** sjing has joined #openstack-ironic | 02:40 | |
| *** mrda_ is now known as mrda_away | 02:49 | |
| *** mrda_away is now known as mrda_ | 02:49 | |
| *** rloo has quit IRC | 03:27 | |
| *** greghaynes has joined #openstack-ironic | 03:40 | |
| *** datajerk has joined #openstack-ironic | 03:59 | |
| devananda | Haomeng: pong | 04:09 |
| *** jbjohnso has quit IRC | 04:10 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: API validates driver name for both POST and PATCH https://review.openstack.org/68018 | 04:24 |
| *** coolsvap has joined #openstack-ironic | 04:38 | |
| Haomeng | devananda: about this alembic patch - https://review.openstack.org/#/c/67415/ , I have concern | 04:54 |
| Haomeng | devananda: because there is no one project which remove all sqlalchemy migration scripts and use alembic | 04:54 |
| Haomeng | devananda: only one projet ceilometer is using alembic, but ceilometer use sqlalchemy at the same time, there should be some issue with alembic | 04:55 |
| *** datajerk has quit IRC | 05:03 | |
| *** Haomeng has quit IRC | 05:18 | |
| *** datajerk has joined #openstack-ironic | 05:34 | |
| *** Haomeng has joined #openstack-ironic | 05:41 | |
| devananda | Haomeng: sqlalchemy-migrate is no longer maintained upstream -- openstack-infra is temporarily maintaining it while all of openstack projects move to alembic | 05:42 |
| devananda | Haomeng: we can be one of the first to complete the move, since we have no prior release taht needs compat | 05:42 |
| Haomeng | devananda: ok, cool:) | 05:42 |
| Haomeng | devananda: yes, we have no prior release:) | 05:42 |
| *** datajerk has quit IRC | 05:42 | |
| Haomeng | devananda: got it, thank you:) | 05:42 |
| Haomeng | devananda: :) | 05:44 |
| devananda | Haomeng: np! | 05:44 |
| devananda | mrda: another thing you might be able to tackle: https://blueprints.launchpad.net/ironic/+spec/keep-powered-off-nodes-off | 05:44 |
| mrda | devananda: thanks, I'll take a look! | 05:44 |
| devananda | mrda: i should be sorta around for a while longer, even though it's late | 05:45 |
| devananda | i'll brain dump real quick. it'll make sense as you read the code :) | 05:45 |
| devananda | mrda: you'll want to edit ironic.conductor.manager _sync_power_states() | 05:46 |
| devananda | mrda: and probably add a config option to that file to influence the behavior of that method. it currently logs a warnign and updates the state in the DB | 05:47 |
| devananda | mrda: under certain circumstances, it should actually power the node off, instead of updating the db | 05:47 |
| mrda | ok, shall take a look | 05:47 |
| *** aignatov_ has joined #openstack-ironic | 05:52 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Sanitize node.last_error message strings https://review.openstack.org/64711 | 06:14 |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/68024 | 06:14 |
| *** kpavel has joined #openstack-ironic | 06:22 | |
| *** rongze has joined #openstack-ironic | 06:26 | |
| *** aignatov_ has quit IRC | 06:38 | |
| openstackgerrit | Sun Jing proposed a change to openstack/ironic: Support serial console access https://review.openstack.org/64100 | 06:40 |
| *** vkozhukalov has quit IRC | 06:47 | |
| * devananda will be back tomorrow | 06:52 | |
| openstackgerrit | Fengqian.gao proposed a change to stackforge/pyghmi: Add BMC bridge request extension https://review.openstack.org/60768 | 07:10 |
| *** aignatov_ has joined #openstack-ironic | 07:14 | |
| *** mrda has quit IRC | 07:27 | |
| *** romcheg has joined #openstack-ironic | 07:36 | |
| *** ifarkas has joined #openstack-ironic | 07:41 | |
| *** aignatov_ has quit IRC | 08:01 | |
| GheRivero | morning Ironic! | 08:04 |
| Haomeng | GheRivero: morning GheRivero: ) | 08:06 |
| *** jistr has joined #openstack-ironic | 08:14 | |
| *** mdurnosvistov has joined #openstack-ironic | 08:28 | |
| *** vkozhukalov has joined #openstack-ironic | 08:28 | |
| mdurnosvistov | good morning all! :) | 08:32 |
| *** romcheg has quit IRC | 08:34 | |
| Haomeng | mdurnosvistov: morning:) | 08:36 |
| *** sjing has quit IRC | 08:39 | |
| *** romcheg has joined #openstack-ironic | 08:49 | |
| *** aignatov_ has joined #openstack-ironic | 08:51 | |
| *** derekh has joined #openstack-ironic | 09:00 | |
| *** aignatov_ has left #openstack-ironic | 09:03 | |
| *** aignatov has joined #openstack-ironic | 09:09 | |
| *** athomas has joined #openstack-ironic | 09:37 | |
| *** kpavel has quit IRC | 10:00 | |
| *** yuriyz has joined #openstack-ironic | 10:00 | |
| *** martyntaylor has joined #openstack-ironic | 10:07 | |
| *** viktors has joined #openstack-ironic | 10:08 | |
| *** coolsvap has quit IRC | 10:09 | |
| *** coolsvap has joined #openstack-ironic | 10:12 | |
| *** coolsvap has quit IRC | 10:23 | |
| *** coolsvap has joined #openstack-ironic | 10:27 | |
| *** aignatov is now known as aignatov_ | 10:29 | |
| *** martyntaylor has quit IRC | 10:35 | |
| *** romcheg1 has joined #openstack-ironic | 10:49 | |
| *** rongze has quit IRC | 10:51 | |
| *** martyntaylor has joined #openstack-ironic | 10:51 | |
| *** romcheg has quit IRC | 10:54 | |
| *** Alexei_9871 has joined #openstack-ironic | 10:55 | |
| *** Alexei_9871 is now known as Alexei_987 | 10:58 | |
| *** aignatov_ is now known as aignatov | 11:01 | |
| *** max_lobur_afk is now known as max_lobur | 11:05 | |
| *** lucasagomes has joined #openstack-ironic | 11:22 | |
| *** rongze has joined #openstack-ironic | 11:22 | |
| *** romcheg has joined #openstack-ironic | 11:23 | |
| *** romcheg1 has quit IRC | 11:26 | |
| *** rongze has quit IRC | 11:27 | |
| viktors | lucasagomes: hi! | 11:27 |
| romcheg | Morning all! | 11:33 |
| *** rongze has joined #openstack-ironic | 11:35 | |
| max_lobur | morning Ironic | 11:35 |
| lucasagomes | morning guys | 11:38 |
| lucasagomes | viktors, pong | 11:38 |
| viktors | lucasagomes: I have a question about ironicclient. Tests on py33 are passing at master now. Does it makes sense to remove "non-voting" from py33 unittest on gerrit? | 11:41 |
| *** rongze has quit IRC | 11:41 | |
| *** martyntaylor has quit IRC | 11:42 | |
| romcheg | viktors: it does | 11:43 |
| romcheg | I can help you with making that patch | 11:43 |
| lucasagomes | viktors, I think it does yes :D we wanted it since from the beginning, we just left it non-vote because, afair, the keystone client didn't support py3 | 11:43 |
| viktors | romcheg: thanks, I'll ping you :) | 11:44 |
| romcheg | since you're 2 meters away from me it's easier to speak :) | 11:44 |
| viktors | lucasagomes, romcheg: ok, will do :) | 11:45 |
| lucasagomes | :D | 11:45 |
| *** martyntaylor has joined #openstack-ironic | 11:48 | |
| viktors | can somebody review patch to infra-config please? https://review.openstack.org/#/c/68092 | 12:05 |
| *** coolsvap has quit IRC | 12:22 | |
| *** rongze has joined #openstack-ironic | 12:35 | |
| *** rongze has quit IRC | 12:41 | |
| *** nosnos has quit IRC | 13:30 | |
| *** Alexei_987 has quit IRC | 13:30 | |
| *** jdob has joined #openstack-ironic | 13:31 | |
| *** nosnos has joined #openstack-ironic | 13:31 | |
| *** rongze has joined #openstack-ironic | 13:35 | |
| *** nosnos has quit IRC | 13:38 | |
| *** martyntaylor has quit IRC | 13:40 | |
| *** rongze has quit IRC | 13:40 | |
| *** rloo has joined #openstack-ironic | 13:47 | |
| *** rloo has quit IRC | 13:47 | |
| *** rloo has joined #openstack-ironic | 13:48 | |
| *** rloo has quit IRC | 13:48 | |
| *** rloo has joined #openstack-ironic | 13:48 | |
| *** martyntaylor has joined #openstack-ironic | 13:54 | |
| *** coolsvap has joined #openstack-ironic | 14:00 | |
| *** max_lobur is now known as max_lobur_afk | 14:04 | |
| *** jbjohnso has joined #openstack-ironic | 14:04 | |
| *** jdob_ has joined #openstack-ironic | 14:08 | |
| *** jrist has joined #openstack-ironic | 14:11 | |
| *** jrist has quit IRC | 14:12 | |
| *** jrist has joined #openstack-ironic | 14:15 | |
| *** jrist has quit IRC | 14:17 | |
| *** jrist has joined #openstack-ironic | 14:18 | |
| *** martyntaylor1 has joined #openstack-ironic | 14:26 | |
| *** max_lobur_afk is now known as max_lobur | 14:27 | |
| *** martyntaylor has quit IRC | 14:29 | |
| *** anniec has joined #openstack-ironic | 14:29 | |
| *** anniec_ has joined #openstack-ironic | 14:32 | |
| *** matty_dubs|gone is now known as matty_dubs | 14:33 | |
| *** jdob_ has quit IRC | 14:34 | |
| *** anniec has quit IRC | 14:34 | |
| *** anniec_ is now known as anniec | 14:34 | |
| *** rongze has joined #openstack-ironic | 14:35 | |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: ipmitool's _power_status raises IPMIFailure https://review.openstack.org/66894 | 14:37 |
| *** linggao has joined #openstack-ironic | 14:38 | |
| *** rongze has quit IRC | 14:39 | |
| *** coolsvap_away has joined #openstack-ironic | 14:45 | |
| *** aignatov is now known as aignatov_ | 14:45 | |
| *** coolsvap has quit IRC | 14:47 | |
| *** coolsvap_away is now known as coolsvap | 14:48 | |
| *** aignatov_ is now known as aignatov | 14:52 | |
| *** vkozhukalov has quit IRC | 15:04 | |
| *** rongze has joined #openstack-ironic | 15:12 | |
| openstackgerrit | Ghe Rivero proposed a change to openstack/ironic: Test to reproduce #1265518 https://review.openstack.org/66345 | 15:16 |
| *** martyntaylor1 has quit IRC | 15:22 | |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: SSHPower driver raises IronicExceptions https://review.openstack.org/66990 | 15:31 |
| *** lucasagomes is now known as lucas-hungry | 15:33 | |
| *** martyntaylor has joined #openstack-ironic | 15:38 | |
| *** martyntaylor has quit IRC | 15:38 | |
| *** martyntaylor has joined #openstack-ironic | 15:38 | |
| *** rloo has quit IRC | 15:39 | |
| *** rloo has joined #openstack-ironic | 15:39 | |
| openstackgerrit | A change was merged to openstack/ironic: Use same MANAGER_TOPIC variable https://review.openstack.org/66480 | 15:43 |
| *** rloo has quit IRC | 15:45 | |
| *** jbjohnso has quit IRC | 15:46 | |
| *** rloo has joined #openstack-ironic | 15:46 | |
| openstackgerrit | Ghe Rivero proposed a change to openstack/ironic: PXE instance_name is no longer mandatory https://review.openstack.org/67974 | 15:48 |
| *** rloo has quit IRC | 15:51 | |
| *** rloo has joined #openstack-ironic | 15:51 | |
| *** rloo has quit IRC | 15:52 | |
| *** rloo has joined #openstack-ironic | 15:54 | |
| *** rloo has joined #openstack-ironic | 15:54 | |
| NobodyCam | gah running late today.. | 15:55 |
| NobodyCam | morning Ironic | 15:56 |
| NobodyCam | brb | 15:56 |
| max_lobur | morning NobodyCam | 15:56 |
| NobodyCam | :) morning max_lobur | 15:56 |
| mdurnosvistov | morning NobodyCam | 15:57 |
| *** jbjohnso has joined #openstack-ironic | 16:00 | |
| max_lobur | lucas-hungry, could you please take a look when you have time https://bugs.launchpad.net/ironic/+bug/1244747 | 16:02 |
| max_lobur | and g'morning :) | 16:02 |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: SSHPower driver raises IronicExceptions https://review.openstack.org/66990 | 16:09 |
| *** rloo has quit IRC | 16:11 | |
| *** rloo has joined #openstack-ironic | 16:11 | |
| NobodyCam | morning mdurnosvistov :) | 16:13 |
| *** hemna has joined #openstack-ironic | 16:15 | |
| lucas-hungry | max_lobur, howdy! sure I gotta finish something I started first | 16:25 |
| lucas-hungry | but it's on my todo list for today :) | 16:25 |
| *** lucas-hungry is now known as lucasagomes | 16:25 | |
| max_lobur | cool, thanks:) | 16:25 |
| lucasagomes | NobodyCam, morning! | 16:25 |
| lucasagomes | max_lobur, morning/afternoon :D | 16:25 |
| max_lobur | afternoon :) | 16:25 |
| max_lobur | 18:26 in kharkiv :) | 16:26 |
| NobodyCam | lucasagomes: morning / afternoon | 16:26 |
| lucasagomes | :) afternoon here as well 14:26 | 16:26 |
| NobodyCam | anyone else having issues loading review.openstack.org? | 16:29 |
| NobodyCam | mn working now | 16:30 |
| *** aignatov is now known as aignatov_ | 16:30 | |
| lucasagomes | NobodyCam, it was working ok some hours ago | 16:34 |
| NobodyCam | ya just a strange burp here i guess | 16:35 |
| NobodyCam | lucasagomes: was the goal to replace ipmi tool with ipmi native? or to support both? | 16:38 |
| NobodyCam | was looking at https://review.openstack.org/#/c/66995/1/doc/source/developer/drivers.rst | 16:39 |
| lucasagomes | NobodyCam, I remember to with devananda a bit about it, we would like to have the native as the default one at least | 16:40 |
| lucasagomes | but now I don't know hw it's going to work for icehouse | 16:40 |
| lucasagomes | I mean, maybe if we get the native working fine we would think to drop the ipmitool | 16:40 |
| lucasagomes | but I think we are going to maintain both for icehouse at least | 16:40 |
| NobodyCam | yea I am recalling that now too | 16:41 |
| lucasagomes | for e.g the console access is being implemented on the ipmitool first | 16:41 |
| lucasagomes | NobodyCam, yea, so on that review I think it's important to expose both drivers on the developer documentation | 16:41 |
| lucasagomes | at least for this release | 16:41 |
| NobodyCam | lucasagomes: ++ | 16:41 |
| NobodyCam | gah down again | 16:43 |
| NobodyCam | wtf | 16:43 |
| lucasagomes | :( | 16:45 |
| devananda | morning, all | 16:45 |
| NobodyCam | good morninc devananda | 16:45 |
| lucasagomes | morning devananda | 16:46 |
| max_lobur | morning devananda | 16:46 |
| devananda | re: ipmitool vs ipminative, my goal is to support both right now, and deprecate ipmitool once we have testing for ipminative | 16:47 |
| max_lobur | devananda, I wanted to ask if you have someone from global requirements to help us merge the patch https://review.openstack.org/#/c/66349/3 (Add futures library to global requirements) | 16:48 |
| devananda | max_lobur: iirc, infra team basically manages that. my advice is to go raise awareness of it in #openstack-infra | 16:48 |
| devananda | and while you're at it, this one too - https://review.openstack.org/#/c/66077/ | 16:49 |
| devananda | :) | 16:49 |
| max_lobur | I already asked a few times, seems they don't pay much attention to such messages | 16:50 |
| max_lobur | ok, I'll spam :) | 16:50 |
| devananda | heh | 16:50 |
| devananda | max_lobur: they might be busy trying to unwedge teh gate (again) or something | 16:50 |
| max_lobur | maybe | 16:51 |
| max_lobur | also I have another question | 16:51 |
| max_lobur | do we plan to support xml in API? | 16:51 |
| max_lobur | somewhere in future | 16:51 |
| max_lobur | there are a few workpieces in tempest for xml made by romcheg, I was going to remove them | 16:52 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Remove the absolute paths from ironic-deploy-helper.filters https://review.openstack.org/66400 | 16:52 |
| NobodyCam | devananda: did you see lucasagomes' comment on https://review.openstack.org/#/c/66995/ | 16:52 |
| max_lobur | also romcheg described a bug, when he set content type to json and accept type to xml he was able to get xml response | 16:53 |
| max_lobur | I'm not sure if it's filled on launchpad | 16:53 |
| devananda | max_lobur: i dont have enough information yet to have a strong opinion on xml support | 16:53 |
| max_lobur | ok, then I will leave as is for now | 16:54 |
| devananda | max_lobur: bugs, we should fix. if xml support is broken, lets either fix it or remove it (temporarily) -- but that can wait until after i3 | 16:55 |
| devananda | source tripleo-undercloud-passwords | 16:55 |
| devananda | gah. wrong window | 16:55 |
| max_lobur | ok | 16:55 |
| max_lobur | also I remember lucasagomes pointed that because we're using things like jsonpatch it will be inconvenient (hard) to support xml requests | 16:56 |
| max_lobur | honestly I agree with this | 16:56 |
| *** rongze has quit IRC | 16:57 | |
| *** rongze has joined #openstack-ironic | 16:57 | |
| jbjohnso | devananda, btw, on console access | 16:58 |
| devananda | max_lobur: ah, good point | 16:58 |
| jbjohnso | I'm probably going to be releasing our console manager under apache2 in the relative near future, if anyone wants to actually have a demo.... | 16:59 |
| lucasagomes | max_lobur, yea, people might be able to input a json-patch in xml (that's pretty confusing) since we now have a wsme object that represents a json-patch | 16:59 |
| jbjohnso | apache license that is... | 16:59 |
| devananda | jbjohnso: sweet! | 16:59 |
| lucasagomes | but that might look pretty weird as well (and I;vent tested) | 16:59 |
| jbjohnso | devananda, also, should have the rest of sensors supported in a few hours (currently only does the analog ones) | 16:59 |
| jbjohnso | {'name': 'Riser 2A Temp', 'imprecision': 0.5, 'value': 32.0, 'states': [], 'health': 0, | 17:00 |
| jbjohnso | 'units': '\xc2\xb0C', 'type': 'Temperature'} | 17:00 |
| jbjohnso | is the sort of info in a 'SensorReading' object | 17:00 |
| jbjohnso | or Riser 2A Temp: 30.0 ± 0.5°C | 17:01 |
| jbjohnso | is a rendering of that that may make more sense | 17:01 |
| *** rongze has quit IRC | 17:02 | |
| jbjohnso | lunch time | 17:02 |
| *** wendar has joined #openstack-ironic | 17:02 | |
| devananda | jbjohnso: fantastic, thanks! | 17:03 |
| max_lobur | devananda, lucasagomes so remove, leave or postpone decision? :) | 17:04 |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Add ability to break TaskManager locks via REST API https://review.openstack.org/55549 | 17:04 |
| lucasagomes | I would vote to remove xml support | 17:04 |
| max_lobur | me too | 17:04 |
| devananda | max_lobur: what do we change to remove it // how hard will it be to add back? | 17:06 |
| *** vkozhukalov has joined #openstack-ironic | 17:06 | |
| max_lobur | devananda, in ironic I think we need just fix the bug that romcheg mentioned | 17:07 |
| max_lobur | to not to support xml completely | 17:07 |
| max_lobur | for tempest - just to remove one file | 17:07 |
| max_lobur | for client - need to investigate if we have something for xml | 17:08 |
| devananda | sounds good. eventually i think there will be folks who want it | 17:08 |
| devananda | so we shouldn't make it harder than it already is for it to be fixed :) | 17:08 |
| max_lobur | there's always folks wanting xml :) | 17:09 |
| *** lifeless has quit IRC | 17:09 | |
| *** mrda_ has quit IRC | 17:09 | |
| lucasagomes | I don't see much benefit in support xml, we would need tests for that, we would need API examples in XML as well, and it's not compatible with the way we partial update our resources | 17:09 |
| *** mrda has joined #openstack-ironic | 17:09 | |
| *** lifeless1 has joined #openstack-ironic | 17:09 | |
| *** hemna has quit IRC | 17:09 | |
| NobodyCam | rloo: question on 66990 ... looking at test_ssh.py you addressed several \ issues but I see several more. ie line # 209 | 17:09 |
| *** mdurnosvistov has quit IRC | 17:09 | |
| *** hemna has joined #openstack-ironic | 17:09 | |
| lucasagomes | devananda, max_lobur hw other openstack projects deal with this problem? | 17:10 |
| lucasagomes | do they support both versions? | 17:10 |
| lucasagomes | I think nova does (and maybe ceilometer) but idk the others | 17:10 |
| max_lobur | as I remember nova supported both | 17:10 |
| *** yuriyz has quit IRC | 17:11 | |
| rloo | NobodyCam. grrrr. Sorry. It would be nice if we had some guidelines about this. I mean: 1. SHOULD we be making changes to conform to changes-in-the-pipeline-that-we-hope-will-get-merged-but-what-if-doesn't | 17:11 |
| NobodyCam | I bring up only because you seem to have addressed several in the last round | 17:12 |
| rloo | NobodyCam. And 2. IF we should make changes, do we make changes to everything, OR just the code we are touching. The answer to this, from what I've seen in other reviews, is to only change your code. | 17:12 |
| NobodyCam | and no need to be sorry :) | 17:12 |
| rloo | NobodyCam. Ha ha. I'm not convinced we should even be making such changes until the global 'we're moving to' stuff has landed. Cuz what if it doesn't land. | 17:13 |
| rloo | NobodyCam. Anyway, did I answer your question? | 17:13 |
| NobodyCam | rloo: yes I agree about only changing code you touched. you did touch line 211 in my example | 17:13 |
| NobodyCam | but this is a good point | 17:14 |
| rloo | NobodyCam. hmm, line 211 isn't a '\' issue? | 17:14 |
| max_lobur | lucasagomes, devananda neutron - both xml and json | 17:14 |
| *** matty_dubs is now known as matty_dubs|lunch | 17:14 | |
| NobodyCam | nope thats on # 209 | 17:14 |
| NobodyCam | :-p | 17:15 |
| *** jistr has quit IRC | 17:15 | |
| lucasagomes | max_lobur, I see | 17:15 |
| NobodyCam | I am ok with lets wait until every thing lands | 17:15 |
| rloo | NobodyCam. right, since I didn't modify #209, I didn't think I should touch it. Presumably, that "other" review has fixed it. | 17:15 |
| lucasagomes | I think the tendency is to use both, also it might come for free on other projects that also is migrating/using to wsme/pecan | 17:15 |
| NobodyCam | rloo: that was the answer I was looking for :) | 17:15 |
| *** martyntaylor has quit IRC | 17:16 | |
| rloo | NobodyCam. good (I didn't actually check/look at the other review yet, on my list...) | 17:16 |
| max_lobur | lucasagomes, yes. So let's postpone decision? | 17:16 |
| max_lobur | I won't remove anything from tempest | 17:16 |
| NobodyCam | will see how slow mr J is today :) | 17:16 |
| lucasagomes | max_lobur, +1 | 17:17 |
| lucasagomes | maybe worth doing some more investigation | 17:17 |
| lucasagomes | for e.g trying update the resources with xml | 17:17 |
| lucasagomes | I think it might works (although will looks a bit confusing to say json-patch and support xml as well) | 17:17 |
| max_lobur | heh | 17:18 |
| max_lobur | true | 17:18 |
| lucasagomes | but as we have a wsme object that actually represents a json-patch object | 17:18 |
| NobodyCam | brb | 17:18 |
| lucasagomes | it might work fine :) | 17:18 |
| max_lobur | I remembered | 17:20 |
| max_lobur | I tested with xml content type | 17:20 |
| max_lobur | tempest.common.rest_client: DEBUG: Large body (9032) md5 summary: 4e819f9ebbae64dbb7a1144089cfcd5a | 17:20 |
| max_lobur | <max_lobur> tempest.common.rest_client: INFO: Request: POST http://192.168.122.224:6385/v1/chassis | 17:20 |
| max_lobur | <max_lobur> tempest.common.rest_client: DEBUG: Request Headers: {'Content-Type': 'application/xml', 'Accept': 'application/xml', 'X-Auth-Token': '<Token omitted>'} | 17:20 |
| max_lobur | <max_lobur> tempest.common.rest_client: DEBUG: Request Body: <?xml version="1.0" encoding="UTF-8"?> | 17:20 |
| max_lobur | <max_lobur> <chassis ><description >test-chassis--tempest-917975265</description></chassis> | 17:21 |
| max_lobur | <max_lobur> tempest.common.rest_client: INFO: Response Status: 500 | 17:21 |
| max_lobur | <max_lobur> tempest.common.rest_client: DEBUG: Response Headers: {'date': 'Wed, 15 Jan 2014 14:56:15 GMT', 'content-length': '59', 'content-type': 'text/plain', 'server': 'WSGIServer/0.1 Python/2.7.3'} | 17:21 |
| max_lobur | <max_lobur> tempest.common.rest_client: DEBUG: Response Body: A server error occurred. Please contact the administrator. | 17:21 |
| max_lobur | I think that's because of v1.media_types = [MediaType('application/json', | 17:21 |
| max_lobur | <max_lobur> 'application/vnd.openstack.ironic.v1+json')] | 17:21 |
| max_lobur | in our api | 17:21 |
| lucasagomes | hmm I see | 17:21 |
| lucasagomes | urgh yea will require some work | 17:22 |
| max_lobur | yes | 17:22 |
| max_lobur | let's just postpone | 17:22 |
| lucasagomes | maybe we can fix the gross problems with that, and leave it as "experimental" ? there's this possibility? | 17:22 |
| max_lobur | need to try | 17:23 |
| max_lobur | I'll spend some time on this tomorrow | 17:23 |
| lucasagomes | cheers | 17:23 |
| devananda | yea, i'd rather leave xml support as experimental | 17:24 |
| max_lobur | :) | 17:24 |
| devananda | than yank it completely | 17:24 |
| NobodyCam | lol Good catch on 66996 rloo :) | 17:24 |
| devananda | for now, anyway. if it's not fixable by i3, we should probably disable it (but, again, not yank it completely from the codebases) | 17:24 |
| devananda | since i think there's a high probability folks will want to work on it | 17:25 |
| lucasagomes | +1 | 17:25 |
| rloo | NobodyCam. Thx, not that I want a job as editor ;) | 17:25 |
| *** anniec has quit IRC | 17:26 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Add rootwrap filters for mkfs https://review.openstack.org/68182 | 17:27 |
| max_lobur | + | 17:27 |
| *** rongze has joined #openstack-ironic | 17:28 | |
| *** romcheg has quit IRC | 17:28 | |
| max_lobur | lol, happy hunting wished to me when I searched for someone from requirements core in infra :D | 17:29 |
| max_lobur | I'm going to mail to Doug Hellmann to ask if he can help | 17:30 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Run mkfs as root https://review.openstack.org/68182 | 17:30 |
| *** viktors has left #openstack-ironic | 17:30 | |
| *** rongze has quit IRC | 17:32 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Make the PXE driver understand ephemeral disks https://review.openstack.org/65750 | 17:33 |
| *** rongze has joined #openstack-ironic | 17:35 | |
| lucasagomes | devananda, btw, these are the patches needed for the deployment: #66389 #66390 #66400 | 17:39 |
| lucasagomes | (I forgot to send the # to you after the meeting yesterday) | 17:39 |
| devananda | lucasagomes: tyvm | 17:39 |
| * devananda looks now | 17:39 | |
| devananda | lucasagomes: i also posted a fix for the API last night | 17:40 |
| lucasagomes | devananda, the one about validate drivers? | 17:40 |
| *** rongze has quit IRC | 17:40 | |
| devananda | yea | 17:40 |
| lucasagomes | devananda, yea thanks for that, I liked ur approach of using the hash ring to validate that | 17:40 |
| lucasagomes | I already reviewed ur patch | 17:40 |
| lucasagomes | left some comments about the return code being returned | 17:40 |
| * devananda refreshes the correct page this time | 17:41 | |
| devananda | ahh yea | 17:41 |
| devananda | i thought about using DriverNotFound | 17:41 |
| devananda | lucasagomes: think that's sufficient? | 17:42 |
| lucasagomes | devananda, for the message yes | 17:42 |
| lucasagomes | but I think it should be HTTP 400 instead of 404 | 17:43 |
| lucasagomes | in case of the driver is not available | 17:43 |
| devananda | k | 17:43 |
| devananda | lucasagomes: hm, DriverNotFound is a NotFound which is 404 | 17:44 |
| devananda | lucasagomes: are you suggesting that be changed globally? | 17:44 |
| NobodyCam | brb post bbt walkies | 17:45 |
| devananda | after this patch lands, i dont think there are any other API requests which generate DriverNotFound -- though i haven't checked all the code paths for after a conductor fails and no more conductors provide that driver | 17:46 |
| *** jistr has joined #openstack-ironic | 17:47 | |
| lucasagomes | devananda, hmm I don't think it needs to be changed globally | 17:49 |
| lucasagomes | but catched on that api layer and e.code=400 set | 17:49 |
| lucasagomes | you can reraise that | 17:49 |
| lucasagomes | caught* | 17:50 |
| *** linggao has quit IRC | 17:51 | |
| lucasagomes | devananda, https://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L217-L221 | 17:51 |
| *** harlowja_away is now known as harlowja | 17:51 | |
| *** matty_dubs|lunch is now known as matty_dubs | 17:52 | |
| devananda | it's not a "driver not supported" -- it's really "no host found that supports this driver right now" | 17:53 |
| lucasagomes | hmm yea indeed | 17:53 |
| lucasagomes | not supported by any conductor | 17:53 |
| *** coolsvap is now known as coolsvap_away | 17:54 | |
| lucasagomes | devananda, yea the message is more fft, I just thought it would be a bit confusing to talk about "host not found" because of the driver | 17:55 |
| devananda | ack | 17:55 |
| lucasagomes | devananda, but I agree with u that, "driver not supported" is wrong | 17:55 |
| devananda | also 400 indicates a permanent failure, doesn't it? | 17:55 |
| devananda | i'm not sure 400 Bad Request is correct when the conductor may be only temporarily offline | 17:56 |
| lucasagomes | devananda, hmm I think it says that we should not resend the request without changing it | 17:56 |
| devananda | which is why I used 404 | 17:57 |
| lucasagomes | but I can't find other code for this case | 17:57 |
| lucasagomes | I see | 17:57 |
| devananda | which may be either temporary or permanent | 17:57 |
| devananda | i can clarify the message to "no conductor found at this time" or some such | 17:57 |
| devananda | ie, i can't satisfy the request right now. | 17:57 |
| *** derekh has quit IRC | 17:58 | |
| devananda | though i'm unclear about whether 404 is suitable when the URI is valid but the *content* is not | 17:58 |
| lucasagomes | right, the thing about 404 is because it more about the requested uri doesn't exist instead of this parameter is not available right now (I think( | 17:58 |
| devananda | yea... | 17:59 |
| devananda | ok | 17:59 |
| devananda | 400 it is | 17:59 |
| devananda | lucasagomes: or 500? | 18:00 |
| lucasagomes | devananda, lemme take a look | 18:00 |
| lucasagomes | it's always confusing this codes urgh heh | 18:00 |
| devananda | 400 bad request (client side error) | 18:01 |
| devananda | 500 internal server error | 18:01 |
| devananda | 400 don't retry without changing the request | 18:01 |
| NobodyCam | anyone see ironic check jobs running in zuul? | 18:01 |
| devananda | w3.org docs for 500 say: " the server is aware that it has erred or is incapable of performing the request ... the server should include an entity containing an explanation of the error situation, and whether it is a temporary or permanent condition." | 18:02 |
| devananda | lucasagomes: ^ looks appropriate to me | 18:02 |
| lucasagomes | indeed | 18:02 |
| lucasagomes | devananda, yea by the description it seems that it fits | 18:03 |
| lucasagomes | devananda, I was looking the 410 (Gone) as well | 18:03 |
| lucasagomes | in case the driver is temporarily unavailable | 18:03 |
| *** athomas has quit IRC | 18:04 | |
| max_lobur | devananda, FYI from Doug Hellmann regarding the global req: "I think we are unlikely to make any requirements changes this week because of the milestone cutoff. I'll take a look after Thursday, though." | 18:04 |
| lucasagomes | devananda, yea I think 500 seems fine | 18:05 |
| devananda | max_lobur: ack, thanks | 18:05 |
| devananda | max_lobur: though as a point of interest, smoe other projects are blocked on requirements changes much worse than we are | 18:05 |
| devananda | lucasagomes: ack | 18:06 |
| lucasagomes | devananda, although and if the client request a driver that actually doesn't exist at all, then it's a bad request or internal server error? | 18:06 |
| lucasagomes | urgh haha /me gotta stop thinking too much | 18:06 |
| lucasagomes | causes me more confusion | 18:06 |
| devananda | lucasagomes: just cause it doesn't exist now() -- but what if admin brings up a new conductor instance tomorrow that uses it? | 18:06 |
| max_lobur | :) | 18:06 |
| devananda | then viola! API will no longer error | 18:06 |
| devananda | with identical request | 18:06 |
| lucasagomes | devananda, yea I agree | 18:07 |
| lucasagomes | 500 then :) | 18:07 |
| devananda | hmmm | 18:07 |
| devananda | in this way of thinking, all our NotFound exc should be 500 | 18:07 |
| devananda | NodeNotFound, InstanceNotFound, etc | 18:07 |
| devananda | well, nvm | 18:07 |
| *** max_lobur is now known as max_lobur_afk | 18:07 | |
| lucasagomes | hah | 18:07 |
| lucasagomes | devananda, let's see | 18:08 |
| devananda | where it's part of the URI, it's a 404 | 18:08 |
| lucasagomes | users does have a way to | 18:08 |
| lucasagomes | know which driver is available at that moment? | 18:08 |
| devananda | eg, GET /v1/nodes/xxx --> NOT FOUND 404 makes sense | 18:08 |
| lucasagomes | they could use /drivers | 18:08 |
| devananda | ahhh yes | 18:08 |
| devananda | they could | 18:08 |
| lucasagomes | and if they input a driver which is not available | 18:08 |
| lucasagomes | that's a bad request | 18:08 |
| devananda | touche! :) | 18:09 |
| lucasagomes | :) | 18:09 |
| *** martyntaylor has joined #openstack-ironic | 18:09 | |
| devananda | lucasagomes: do i want to convert to a wsme.ClientSideError? | 18:09 |
| lucasagomes | devananda, yup it's 400 by default | 18:10 |
| lucasagomes | or if you just want to change the e.code | 18:10 |
| devananda | yea | 18:10 |
| lucasagomes | and reraise the same exception | 18:10 |
| lucasagomes | works as well | 18:10 |
| devananda | right, thanks | 18:10 |
| NobodyCam | we have patches submitted at ~6:30 this morning that mr J seems to be igonring | 18:17 |
| *** aignatov_ is now known as aignatov | 18:19 | |
| devananda | NobodyCam: gate has been having some serious issues | 18:19 |
| devananda | NobodyCam: see the busy-ness in -infra | 18:20 |
| devananda | it's MP day, so everyone is rushing to get things in, too | 18:20 |
| *** martyntaylor has left #openstack-ironic | 18:20 | |
| lucasagomes | jeez, just looked at zuul | 18:21 |
| lucasagomes | some patches might take "51 hr 42 min" | 18:22 |
| NobodyCam | I didn't even see our jobs in the check gate | 18:23 |
| lucasagomes | :/ | 18:23 |
| *** aignatov is now known as aignatov_ | 18:23 | |
| *** datajerk has joined #openstack-ironic | 18:24 | |
| *** wendar has quit IRC | 18:28 | |
| *** wendar has joined #openstack-ironic | 18:28 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: API validates driver name for both POST and PATCH https://review.openstack.org/68018 | 18:33 |
| devananda | lucasagomes: this is a bit risky: https://review.openstack.org/#/c/66400/2/etc/ironic/rootwrap.d/ironic-deploy-helper.filters | 18:42 |
| devananda | lucasagomes: granted, i understand why you want it, but is there precedent in other projects ? | 18:42 |
| devananda | NobodyCam: if you're not already slammed, can you take a peek at https://review.openstack.org/#/c/66389/ ? | 18:42 |
| * NobodyCam looking.. | 18:43 | |
| lucasagomes | devananda, nova already removed the absolute path of their commands as well | 18:43 |
| lucasagomes | https://github.com/openstack/nova/blob/master/etc/nova/rootwrap.d/baremetal-deploy-helper.filters | 18:43 |
| devananda | lucasagomes: hmm, yea. they did that for all their rootwrap filters | 18:44 |
| devananda | lucasagomes: ok! | 18:44 |
| lucasagomes | devananda, but yea I understand the risk | 18:44 |
| lucasagomes | you think we should go with a different approach? | 18:44 |
| lucasagomes | devananda, thanks for the reviews | 18:46 |
| devananda | nope, i'm fine with it. just logging my minor objection | 18:46 |
| lucasagomes | sure | 18:47 |
| devananda | lucasagomes: is 68182 critical for deploy, or for ephemeral? | 18:47 |
| lucasagomes | devananda, for ephemeral only | 18:48 |
| devananda | k. i'll -1 with my nits then :) | 18:48 |
| lucasagomes | :) heh ok | 18:48 |
| NobodyCam | devananda: looks good to me | 18:50 |
| NobodyCam | +a'd | 18:50 |
| devananda | nice, ty | 18:50 |
| lucasagomes | cheers guys, I've tested the rootwrap from oslo here | 18:51 |
| NobodyCam | did we ever come up with anything for the undocumented conf values? http://paste.openstack.org/show/T5bZNvawbA0r9D34Nk97/ | 18:51 |
| devananda | NobodyCam: also https://review.openstack.org/#/c/66400/, which should be quick | 18:51 |
| devananda | NobodyCam: humm, not sure I've seen this. why aren't they getting doc'd? | 18:52 |
| NobodyCam | was on the ageda some time back | 18:52 |
| NobodyCam | think we talked about them too | 18:52 |
| NobodyCam | ya I lookded at that one. should we land with out mr J's okay? | 18:53 |
| devananda | NobodyCam: yea. it's just a rebase | 18:55 |
| NobodyCam | looked ok with out mr J's check, so I +a'd | 18:58 |
| devananda | NobodyCam: re the keystone options missing -- i actually don't see ANY of those in our conf.sample | 18:59 |
| *** lifeless1 is now known as lifeless | 18:59 | |
| devananda | NobodyCam: looks like nova gets them from keystoneclient.middleware.auth_token | 18:59 |
| NobodyCam | that was the point I rasing | 19:00 |
| rloo | this is a dumb question, but aren't instance uuids unique? | 19:11 |
| NobodyCam | they should be | 19:11 |
| rloo | https://review.openstack.org/#/c/66858/. | 19:11 |
| rloo | it has already been merged, but I was just wondering. | 19:12 |
| NobodyCam | The page you requested was not found, or you do not have permission to view this page. | 19:12 |
| NobodyCam | doh | 19:12 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Run mkfs as root https://review.openstack.org/68182 | 19:12 |
| devananda | rloo: no | 19:12 |
| devananda | rloo: instance_uuid is unique within Nova. we aren't enforcing uniqueness on the node.instance_uuid column | 19:13 |
| devananda | rloo: though one might ask "why not?" | 19:13 |
| rloo | do we generate instance uuids? (we = ironic) | 19:14 |
| devananda | definitely not | 19:14 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Make the PXE driver understand ephemeral disks https://review.openstack.org/65750 | 19:14 |
| rloo | ok, so in practise, they will be unique, but ironic isn't assuming that they will be. | 19:14 |
| devananda | correct | 19:14 |
| devananda | and it's currently conceivable that a deploy might fail on node X and be retried on node Y for the same nova instance | 19:15 |
| rloo | got it. thx. | 19:15 |
| devananda | there may be some edge cases there that we have to handle ... | 19:15 |
| rloo | but the same nova instance will have the same uuid so it is still unique. | 19:15 |
| devananda | unique within nova. not within ironic | 19:16 |
| devananda | if the deploy on node X isnt fully cleaned up, i mean | 19:16 |
| rloo | hmm. ok. | 19:17 |
| devananda | let's say deploy instance AAAA on node X fails but is left in error state. retried on node Y. now both X and Y have same instance UUID, but different staets | 19:17 |
| devananda | this could be a terrible thing -- i haven't really thought it through until now | 19:17 |
| devananda | but looking at the code, i think it's possible | 19:17 |
| *** datajerk has quit IRC | 19:17 | |
| NobodyCam | brb.. bagel making time | 19:18 |
| devananda | rloo: if we imposed a UniqueConstraint on node.instance_uuid, we'd prevent that situation. | 19:19 |
| devananda | mrda: new LHF - https://bugs.launchpad.net/ironic/+bug/1271283 | 19:20 |
| rloo | devananda: had to look up what a UniqueConstraint was:-) Yes, I'd feel better if we did that. Why didn't you want to do it? Or are you OK with doing it now? | 19:27 |
| *** vkozhukalov has quit IRC | 19:27 | |
| devananda | rloo: it may just be one of those days when i want to pull everything apart, but ... lemme ask ... why do we even stash the instance_uuid on node? | 19:29 |
| devananda | rloo: why not just let the client stash it, if they want to, in the node.properties dict? | 19:29 |
| devananda | at what point does Ironic actually care about that information | 19:30 |
| rloo | devananda. Seems like the only reason for it is if ironic needs to communicate with nova/neutron. | 19:30 |
| devananda | nova is upstream from ironic. do we ever send commands /to/ nova? | 19:31 |
| rloo | rloo. but i will freely admit, that i'm not familiar with the ironic code that uses the instance name/uuid. | 19:31 |
| devananda | for neutron, we need the node <-> port id mapping. for cinder, we'll need the node <-> volume id mapping | 19:31 |
| rloo | devananda. seems more like if nova/someone wants to query ironic about instance uuid. eg, which node is instance uuid on? | 19:32 |
| devananda | ah | 19:32 |
| devananda | yep | 19:32 |
| devananda | so filtering | 19:32 |
| *** datajerk has joined #openstack-ironic | 19:32 | |
| devananda | so there's some special handling around "is_associated" | 19:33 |
| devananda | which is basically instance_uuid is || is not null | 19:33 |
| rloo | devananda. yeah, that rings a bell. i recall a discussion about the api. | 19:34 |
| devananda | mrda: another LHF - https://bugs.launchpad.net/ironic/+bug/1271291 | 19:40 |
| *** aignatov_ is now known as aignatov | 19:45 | |
| *** mdurnosvistov has joined #openstack-ironic | 19:46 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add comment about node.instance_uuid https://review.openstack.org/68220 | 19:47 |
| devananda | rloo: ^ :) | 19:47 |
| rloo | devananda. thx! What about adding a UniqueConstraint? | 19:48 |
| devananda | rloo: i'm not convinced that we should enforce that yet | 19:50 |
| rloo | devananda. haha, i'm not convinced we shouldn't, so will wait and see... | 19:50 |
| devananda | :) | 19:52 |
| lucasagomes | would it be allowed to deploy the same instance multiple times? | 19:54 |
| lucasagomes | if not I think we need the constraint | 19:54 |
| *** datajerk has quit IRC | 19:56 | |
| NobodyCam | brb | 19:57 |
| *** GheRivero has quit IRC | 19:57 | |
| *** GheRivero has joined #openstack-ironic | 19:58 | |
| devananda | lucasagomes: what about failed deploys? | 20:00 |
| devananda | lucasagomes: should we force client to unset instance_uuid before they can attempt to re-deploy on another node? | 20:01 |
| devananda | (seems logical to me...) | 20:01 |
| lucasagomes | devananda, yea seems logical... but hw it works on nova? | 20:01 |
| lucasagomes | to redeploy something | 20:01 |
| lucasagomes | user is going to re-issue the nova boot command right? | 20:02 |
| devananda | scheduler retries it | 20:02 |
| devananda | no | 20:02 |
| devananda | scheduler will retry N times (usually 3) | 20:02 |
| lucasagomes | right, and that's going to use the same instance uuid? | 20:02 |
| devananda | yes | 20:02 |
| devananda | if all fail, then user retries,which generates new uuid | 20:02 |
| *** datajerk has joined #openstack-ironic | 20:03 | |
| lucasagomes | right, so I agree that we should _not_ have a unique consstraint | 20:03 |
| *** romcheg has joined #openstack-ironic | 20:03 | |
| devananda | otoh, nova driver should clean up if a deploy fails | 20:03 |
| lucasagomes | otherwise the 2x and 3x is going to fail because it tried to update the node with the same instance uuid | 20:03 |
| lucasagomes | hmm | 20:03 |
| devananda | also true | 20:03 |
| devananda | so it's not clear to me what the edge cases are, which is why i haven't moved on this before :) | 20:04 |
| lucasagomes | would clean-up involve remove the instance uuid? or just remove the config files etc | 20:04 |
| devananda | dunno | 20:04 |
| devananda | it might be worth filing a bug to track this discussion | 20:04 |
| lucasagomes | right, yea | 20:04 |
| devananda | rloo: want to summarize ^ into a bug? | 20:04 |
| *** jbjohnso has quit IRC | 20:04 | |
| rloo | haha. i can try devananda. | 20:05 |
| rloo | it isn't quite a bug though. but sure. | 20:05 |
| lucasagomes | right so I think that we should not add the constraint right now, let's get things working | 20:05 |
| lucasagomes | and then we can observe the behaviour | 20:05 |
| rloo | lucasagomes, devananda. would it be sufficient to add a comment in the code (https://review.openstack.org/68220) to revisit whether it should be UniqueConstraint or not? | 20:06 |
| dkehn | devananda: in https://review.openstack.org/#/c/66071/7/ironic/common/neutronv2/__init__.py you reiew, would that be CONF.import_opt('url', 'ironic.neutron'), then x = CONF.url ? or would one drop the ironic and x = CONF.url ? | 20:06 |
| lucasagomes | devananda, side-note about the clean-up, if the driver clean-up the instance_uuid from the node if it fails, could the scheduler pick the same node to deploy again? (that would might cause it to fail again for the same reason) | 20:07 |
| devananda | dkehn: you wouldn't need to do that import if you merged the files :) | 20:07 |
| lucasagomes | rloo, hmm I would just open a bug | 20:07 |
| devananda | rloo: bug ++, w/ summary of the points raised here | 20:07 |
| dkehn | devananda: true, was trying to maintain the file from nova, can do that if you would prefer | 20:07 |
| rloo | ok lucasagomes, devananda. | 20:08 |
| devananda | lucasagomes: yep, exactly, unless somehow the node was marked failed in a different way ..... | 20:08 |
| lucasagomes | devananda, yea, which doesn't seems to be right now, we rely on the instance_uuid to know if the node is already associated or not | 20:08 |
| lucasagomes | so yea we can't add the constraint in the moment | 20:08 |
| dkehn | devananda: your call here | 20:09 |
| *** jistr has quit IRC | 20:11 | |
| devananda | dkehn: this is small enough, copying it from nova seems less important than making it maintainable locally | 20:14 |
| dkehn | devananda: ok, so I'll combine it into one and removed the common/neutronv2 and move a single neutron.py into the common directory, sound about right | 20:16 |
| *** epim has joined #openstack-ironic | 20:16 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Support preserve_ephemeral https://review.openstack.org/68236 | 20:19 |
| devananda | dkehn: common/network/neutron.py ++ | 20:23 |
| lucasagomes | have a g'night everybody | 20:24 |
| devananda | g'night, lucasagomes | 20:24 |
| lucasagomes | devananda, NobodyCam rloo night :) | 20:24 |
| NobodyCam | night lucasagomes :) | 20:24 |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/68243 | 20:24 |
| *** lucasagomes has quit IRC | 20:26 | |
| devananda | NobodyCam: is there an open bug for the missing keystone_authtoken conf section? | 20:27 |
| devananda | NobodyCam: ive fixed it :) | 20:27 |
| NobodyCam | not sure we ever filed a bug for that | 20:28 |
| devananda | k, filing one | 20:28 |
| NobodyCam | :) | 20:28 |
| *** datajerk has quit IRC | 20:29 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add [keystone_authtoken] to ironic.conf.sample https://review.openstack.org/68256 | 20:30 |
| NobodyCam | great patch :) | 20:31 |
| NobodyCam | Lucas is aready gone :-p | 20:33 |
| *** datajerk has joined #openstack-ironic | 20:34 | |
| NobodyCam | woo hoo ironic jobs are appering in zuul again :) | 20:35 |
| *** jdob has quit IRC | 20:35 | |
| *** jdob has joined #openstack-ironic | 20:36 | |
| NobodyCam | Queue lengths: 1194 events wow!!! | 20:36 |
| *** rongze has joined #openstack-ironic | 20:37 | |
| *** rongze has quit IRC | 20:42 | |
| devananda | max_lobur_afk, lucas - -http://lists.openstack.org/pipermail/openstack-tc/2014-January/000494.html | 20:45 |
| devananda | tl;dr -- let's just drop XML support and be done with it | 20:45 |
| *** jdob has quit IRC | 20:47 | |
| *** jdob has joined #openstack-ironic | 20:47 | |
| *** mrda_ has joined #openstack-ironic | 20:49 | |
| *** datajerk1 has joined #openstack-ironic | 20:52 | |
| *** datajerk has quit IRC | 20:54 | |
| devananda | romcheg: https://bugs.launchpad.net/ironic/+bug/1271317 | 20:55 |
| romcheg | devananda agree | 20:56 |
| *** mdurnosvistov has quit IRC | 20:57 | |
| mrda_ | morning all | 20:58 |
| mrda_ | devananda: thanks for the bug refs, will look at them today | 20:59 |
| NobodyCam | morning mrda_ | 20:59 |
| mrda_ | morning NobodyCam | 21:00 |
| NobodyCam | :) | 21:01 |
| devananda | mornin, mrda_ ! | 21:01 |
| mrda | hey devananda - are you ever away from keyboard? | 21:03 |
| devananda | mrda_: yep | 21:03 |
| mrda | seems to be rare :) | 21:03 |
| devananda | my tan is fading as we speak :) | 21:04 |
| NobodyCam | by the bouncing bubbie at the door I see its time for another quick walkies | 21:04 |
| mrda | :") | 21:04 |
| *** zul has quit IRC | 21:06 | |
| *** jbjohnso has joined #openstack-ironic | 21:16 | |
| *** mrda is now known as mrda__ | 21:16 | |
| *** mrda_ is now known as mrda | 21:16 | |
| *** romcheg1 has joined #openstack-ironic | 21:45 | |
| *** romcheg has quit IRC | 21:46 | |
| *** sameer has joined #openstack-ironic | 21:56 | |
| *** jdob has quit IRC | 21:58 | |
| * NobodyCam brb eating.... | 22:00 | |
| devananda | general question - should our API be printing a log message for every incoming HTTP request? | 22:00 |
| devananda | seems a little too verbose. but also, apache does this with the access log, so maybe it's just fine if it can be redirected appropriately (which, maybe it can, i'm not sure) | 22:01 |
| NobodyCam | at debug level sure | 22:01 |
| devananda | at non-debug level | 22:02 |
| rloo | what does nova do? | 22:02 |
| * devananda hasnt checked | 22:02 | |
| sameer | yes - especially during troubleshooting for users to know that request reached up to ironic. | 22:02 |
| jbjohnso | it's interesting that I notice things ipmitol missed... | 22:07 |
| jbjohnso | for example, I have a server with a sensor indicating 'non recoverable failure' and ipmitool describes state as 'ok' | 22:08 |
| NobodyCam | jbjohnso: nice | 22:09 |
| jbjohnso | {'states': ['Critical'], 'health': 2, 'type': 'Add-in Card', 'name': 'Mezz Exp 2 Fault'} | 22:09 |
| jbjohnso | 2 == const.Health.Critical... | 22:09 |
| jbjohnso | guess I was wrong about it non-recoverable, but still | 22:09 |
| jbjohnso | well, the short story is discrete sensors work | 22:13 |
| *** harlowja is now known as harlowja_away | 22:14 | |
| jbjohnso | and ipmictl shows that get_health does what I wanted... | 22:14 |
| jbjohnso | $ time IPMIPASSWORD=Passw0rd ./ipmictl.py 10.240.181.1 USERID health | 22:15 |
| jbjohnso | Logged into 10.240.181.1 | 22:15 |
| jbjohnso | {'badreadings': [{'name': 'Mezz Exp 2 Fault', 'imprecision': None, 'value': None, 'states': (), 'health': 2, 'units': '', 'type': 'Add-in Card'}], 'health': 2} | 22:15 |
| jbjohnso | so a function to command to ask just for the things that are wrong | 22:15 |
| *** aignatov is now known as aignatov_ | 22:23 | |
| NobodyCam | devananda: where are we logging those calls? | 22:24 |
| devananda | eg, 127.0.0.1 - - [21/Jan/2014 13:39:55] "GET /v1/nodes/eb55cbeb-59e6-45be-b7e3-c972ec564a4c HTTP/1.1" 200 798 | 22:25 |
| devananda | not sure where it's coming from | 22:26 |
| devananda | probably a pecan setting | 22:26 |
| NobodyCam | ya | 22:26 |
| NobodyCam | I dont see us logging such info | 22:26 |
| devananda | mikal: re: third party testing discussion and ironic, in case you havent seen it in the flood of other things -- http://lists.openstack.org/pipermail/openstack-dev/2014-January/025011.html | 22:26 |
| devananda | NobodyCam: right. it's not explicit in ironic's code | 22:27 |
| NobodyCam | devananda: tho in looking should this: https://github.com/openstack/ironic/blob/master/ironic/api/config.py#L32 be like https://github.com/openstack/ironic/blob/master/ironic/api/config.py#L40 ??? | 22:29 |
| devananda | you have an interesting point | 22:30 |
| *** harlowja_away is now known as harlowja | 22:35 | |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Add support for discrete sensors https://review.openstack.org/68279 | 22:39 |
| jbjohnso | devananda, ^^^ if you want to peruse. I have been self-approving a bit, but hey... | 22:40 |
| NobodyCam | jbjohnso: where would one find table 42-2? | 22:41 |
| jbjohnso | NobodyCam, the ipmi specification | 22:42 |
| NobodyCam | ahhh :) | 22:42 |
| devananda | jbjohnso: i peek now and then, but frankly don't have the depth of knowledge to assess functionality | 22:43 |
| devananda | jbjohnso: i would very much like to see an automated test suite for pyghmi, though | 22:43 |
| devananda | jbjohnso: we talked once about the (im)possibility of getting functional tests with a mock IPMI implementation ... what about unit tests? or tests on real hardware? | 22:44 |
| dkehn | devananda: in https://review.openstack.org/#/c/66071/7/ironic/common/neutronv2/neutron.py, the last comment covert this to a IronicException are we sure about that given this is going tocome from the neutronclient: https://github.com/openstack/python-neutronclient/blob/master/neutronclient/common/exceptions.py#L52 | 22:45 |
| devananda | jbjohnso: if some vendor(s) donated hardware to it, i don't see why we couldn't write a test harness that would, on each proposed commit to pyghmi, run a test suite on said hardware | 22:45 |
| devananda | jbjohnso: to ensure that, eg, the SDR queries actually returned something reasonable | 22:46 |
| devananda | jbjohnso: this would allow other developers to work on it who may not have your expertese or personal lab ;) | 22:46 |
| devananda | dkehn: yes, very sure about that | 22:47 |
| dkehn | devananda: ok | 22:47 |
| devananda | dkehn: the point is that higher level code in Ironic should only be concerned with catching IronicException or classes derived from it | 22:47 |
| devananda | dkehn: all drivers are supposed to catch lower-level, external exceptions, and convert them. They don't all do that today, but they should | 22:47 |
| devananda | dkehn: eg, the ConductorManager main thread shouldn't need to catch a NeutronException because taht's driver-specific. the proposed SeaMicro driver and our PXE driver should both raise the same exception for the same type of error, eg. FailedToAssociateNetworkPort ,or what ever | 22:49 |
| devananda | (just an example) | 22:49 |
| devananda | NobodyCam: if you're still looing at the logging stuff, http://pecan.readthedocs.org/en/latest/logging.html#logging-requests-with-paste-translogger might be interesting | 22:49 |
| devananda | NobodyCam: has example of redirecting to an apache-style access log | 22:50 |
| devananda | NobodyCam: it's also not a big issue now | 22:50 |
| NobodyCam | ack | 22:51 |
| jbjohnso | devananda, tests on real hardware would work | 22:58 |
| dkehn | devananda: in respect to the previous questiom , if I add an exception for this instance, e.g. FailedToUpdateDHCPOptOnPort would be ok? | 22:59 |
| devananda | dkehn: yep | 22:59 |
| dkehn | devananda: nothing that exists really seems to fit | 22:59 |
| dkehn | devananda: k | 22:59 |
| devananda | dkehn: taht's fine, then | 23:00 |
| devananda | jbjohnso: what do you think it'll take for that to be spun up, both hardware and human-wise? | 23:02 |
| jbjohnso | devananda, hmm... well the sensor data one would require the most verbose info to be able to compare well | 23:03 |
| jbjohnso | the other commands and session management stuff is easy enough and perfectly common (ipmitool had lan, lanplus, and things like -o intel to do session negotiation to different bmcs in different ways | 23:04 |
| jbjohnso | but I designed the stuff in pyghmi to be more adaptive | 23:04 |
| jbjohnso | thus far all the bmcs that required 'special' treatment that I can lay hands on don't require the user to understand that in pyghmi... | 23:05 |
| jbjohnso | but for example I compare the output of 'ipmictl.py' against expectations as a sort of test harness | 23:05 |
| jbjohnso | oh, fyi, the sensor value may still differ from ipmitool because of a design choice I made | 23:06 |
| jbjohnso | hypothetically, if a sensor says '8', in a certain scenario that could mean 'some value between 4 and 16' | 23:06 |
| jbjohnso | and ipmitool would say '8', I say '10, give or take 6' | 23:07 |
| *** harlowja has quit IRC | 23:07 | |
| *** matty_dubs is now known as matty_dubs|gone | 23:13 | |
| openstackgerrit | A change was merged to openstack/ironic: sync_power_states handles missing driver info https://review.openstack.org/65297 | 23:16 |
| NobodyCam | wow that was approved at 8:30 this morning | 23:17 |
| *** jbjohnso has quit IRC | 23:25 | |
| devananda | awesome taht _something_ finally got through | 23:29 |
| NobodyCam | ya | 23:30 |
| *** datajerk1 has quit IRC | 23:37 | |
| openstackgerrit | Michael Davies proposed a change to openstack/ironic: Minor documentation update https://review.openstack.org/66995 | 23:38 |
| openstackgerrit | Michael Davies proposed a change to openstack/ironic: Update method doc strings in pxe.py https://review.openstack.org/66996 | 23:38 |
| *** jbjohnso has joined #openstack-ironic | 23:38 | |
| devananda | mrda__: did you mean to do that ^ ? | 23:42 |
| mrda__ | I was just trying to save you some time | 23:42 |
| devananda | it doesn't look like you changed anything in either patch, though | 23:43 |
| devananda | just rebased, which, if there was no conflict, jenkins would do automatically | 23:43 |
| mrda__ | yes, I failed at git | 23:43 |
| devananda | ah :) | 23:43 |
| devananda | no worries | 23:43 |
| mrda__ | I was trying to address the review comment, but it didn't merge my change, for reasons | 23:43 |
| devananda | awesome. i mean, that you were trying to fix it up. not that it ate your change | 23:44 |
| Haomeng | morning/evening all:) | 23:44 |
| mrda__ | not trying to be a PITA, just trying to be helpful. I'll get better, promise. | 23:44 |
| devananda | in any case, i think i'm done for today -- been at this for far too many hours now :) | 23:44 |
| * devananda hopes the gate unwedges and some things land | 23:45 | |
| NobodyCam | I dont see any other ironic patches in the queue.. yet :( | 23:47 |
| *** blamar has quit IRC | 23:47 | |
| devananda | NobodyCam: let's not approve any more today | 23:48 |
| devananda | NobodyCam: there are already 3 in the +a queue, all of which are essential for deploy functionality | 23:48 |
| NobodyCam | yes! | 23:48 |
| NobodyCam | or should I say +a to that :-p | 23:49 |
| devananda | there are a whole bunch more I wish we could land, but they'll realistically need to wait until the gate queue is lower | 23:49 |
| devananda | which means, probably, a few days | 23:49 |
| NobodyCam | ya | 23:49 |
| * devananda realy goes AFK now | 23:50 | |
| NobodyCam | :) | 23:50 |
| *** romcheg1 has quit IRC | 23:51 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!