| *** david-ly_ has joined #openstack-ironic | 00:06 | |
| *** ijw_ has joined #openstack-ironic | 00:07 | |
| *** david-lyle has quit IRC | 00:08 | |
| *** ijw__ has quit IRC | 00:10 | |
| *** ijw_ has quit IRC | 00:29 | |
| *** dprince has joined #openstack-ironic | 00:37 | |
| *** rwsu has quit IRC | 00:38 | |
| *** boris-42 has quit IRC | 00:42 | |
| *** r-daneel has quit IRC | 00:48 | |
| *** zhenguo has joined #openstack-ironic | 00:52 | |
| *** naohirot has joined #openstack-ironic | 00:53 | |
| *** amotoki has joined #openstack-ironic | 01:01 | |
| *** amotoki has quit IRC | 01:05 | |
| *** smoriya has joined #openstack-ironic | 01:06 | |
| *** thrash is now known as thrash|g0ne | 01:19 | |
| *** dprince has quit IRC | 01:22 | |
| *** david-ly_ is now known as david-lyle | 01:36 | |
| *** mitchjameson has quit IRC | 01:40 | |
| *** dontalton has quit IRC | 01:56 | |
| *** Nisha has joined #openstack-ironic | 01:57 | |
| *** yuanying has joined #openstack-ironic | 02:02 | |
| *** Nisha has quit IRC | 02:06 | |
| *** ramineni has joined #openstack-ironic | 02:40 | |
| *** mitchjameson has joined #openstack-ironic | 02:43 | |
| *** Sukhdev has joined #openstack-ironic | 03:01 | |
| *** amotoki has joined #openstack-ironic | 03:02 | |
| *** amotoki has quit IRC | 03:06 | |
| *** amit213 has quit IRC | 03:12 | |
| *** amit213 has joined #openstack-ironic | 03:13 | |
| *** yuanying has quit IRC | 03:16 | |
| *** mtanino has quit IRC | 03:31 | |
| *** korean101 has quit IRC | 03:42 | |
| *** yuanying has joined #openstack-ironic | 03:47 | |
| *** rameshg87 has joined #openstack-ironic | 03:56 | |
| *** yuanying has quit IRC | 03:59 | |
| *** coolsvap|away is now known as coolsvap | 04:04 | |
| *** yuanying has joined #openstack-ironic | 04:09 | |
| *** amotoki has joined #openstack-ironic | 04:10 | |
| *** hemna has joined #openstack-ironic | 04:19 | |
| *** achanda has joined #openstack-ironic | 04:23 | |
| *** mitchjameson has quit IRC | 04:35 | |
| *** lazy_prince has joined #openstack-ironic | 04:39 | |
| *** hemna has quit IRC | 04:40 | |
| *** hemna has joined #openstack-ironic | 04:42 | |
| *** chenglch has joined #openstack-ironic | 04:44 | |
| *** saripurigopi has joined #openstack-ironic | 04:51 | |
| saripurigopi | morning Ironic | 04:54 |
|---|---|---|
| lazy_prince | saripurigopi: morning.. | 04:54 |
| saripurigopi | lazy_prince: o/ | 04:54 |
| *** jrath4 has joined #openstack-ironic | 04:54 | |
| openstackgerrit | Shivanand Tendulker proposed stackforge/proliantutils: Adding RIS support for virtual media devices https://review.openstack.org/194570 | 04:58 |
| *** Sukhdev has quit IRC | 05:01 | |
| *** Marga_ has joined #openstack-ironic | 05:02 | |
| *** jrath4 has quit IRC | 05:03 | |
| *** hemna has quit IRC | 05:05 | |
| *** achanda has quit IRC | 05:06 | |
| *** Nisha has joined #openstack-ironic | 05:21 | |
| *** radek__ has joined #openstack-ironic | 06:01 | |
| *** yuikotakada has joined #openstack-ironic | 06:03 | |
| *** yuanying_ has joined #openstack-ironic | 06:15 | |
| *** yuanying_ has quit IRC | 06:17 | |
| *** yuanying_ has joined #openstack-ironic | 06:18 | |
| *** yuanying has quit IRC | 06:18 | |
| *** yuanying_ has quit IRC | 06:19 | |
| *** jefabo has joined #openstack-ironic | 06:23 | |
| *** jefabo has left #openstack-ironic | 06:26 | |
| *** ifarkas has joined #openstack-ironic | 06:27 | |
| *** yuanying has joined #openstack-ironic | 06:32 | |
| *** yuanying has quit IRC | 06:32 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add iRMC Virtual Media Deploy module for iRMC Driver https://review.openstack.org/151958 | 06:38 |
| *** romcheg has joined #openstack-ironic | 06:39 | |
| *** blixt_ is now known as hblixt | 06:41 | |
| *** romcheg has quit IRC | 06:43 | |
| *** achanda has joined #openstack-ironic | 06:44 | |
| *** ukalifon1 has joined #openstack-ironic | 06:44 | |
| *** jefabo_ has joined #openstack-ironic | 06:46 | |
| *** jefabo_ has quit IRC | 06:51 | |
| *** jefabo_ has joined #openstack-ironic | 06:52 | |
| *** jefabo_ is now known as jefabo | 06:53 | |
| openstackgerrit | Ramakrishnan G proposed stackforge/proliantutils: Check if virtual media is attached before ejecting https://review.openstack.org/195432 | 06:53 |
| rameshg87 | ramineni: ^^^ | 06:54 |
| *** yog_ has joined #openstack-ironic | 06:54 | |
| *** ifarkas has quit IRC | 06:57 | |
| *** achanda has quit IRC | 06:58 | |
| *** bradjones has quit IRC | 07:02 | |
| *** Nisha has quit IRC | 07:02 | |
| *** yuanying has joined #openstack-ironic | 07:02 | |
| *** bradjones has joined #openstack-ironic | 07:04 | |
| *** bradjones has quit IRC | 07:04 | |
| *** bradjones has joined #openstack-ironic | 07:04 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add localboot support for iscsi_irmc driver https://review.openstack.org/163738 | 07:05 |
| *** yuanying has quit IRC | 07:05 | |
| *** yuanying has joined #openstack-ironic | 07:05 | |
| *** athomas has joined #openstack-ironic | 07:06 | |
| *** achanda has joined #openstack-ironic | 07:08 | |
| *** gzhai1 has joined #openstack-ironic | 07:13 | |
| *** dasm is now known as dasm|afk | 07:13 | |
| *** dasm|afk is now known as dasm | 07:14 | |
| *** viktors|afk is now known as viktors | 07:15 | |
| *** romainh has joined #openstack-ironic | 07:21 | |
| *** gzhai1 has quit IRC | 07:25 | |
| *** lsmola has joined #openstack-ironic | 07:28 | |
| *** natorious is now known as zz_natorious | 07:39 | |
| *** dtantsur|afk is now known as dtantsur | 07:50 | |
| dtantsur | Morning Ironic! | 07:50 |
| yuikotakada | dtantsur, morning :) | 07:56 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: When boot option is not persisted, set boot on next power on https://review.openstack.org/177642 | 07:56 |
| *** gzhai1 has joined #openstack-ironic | 08:01 | |
| *** lucasagomes has joined #openstack-ironic | 08:03 | |
| dtantsur | yuikotakada, o/ | 08:03 |
| *** ndipanov has joined #openstack-ironic | 08:06 | |
| gzhai1 | dtantsur:hi | 08:07 |
| dtantsur | o/ | 08:07 |
| gzhai1 | dtantsur:Do you have time to revisit my bp? https://review.openstack.org/#/c/176109/ | 08:07 |
| dtantsur | probably :) I'm pretty busy recently, will try to find time (it's on my radar already) | 08:08 |
| gzhai1 | dtantsur: Thanks! | 08:09 |
| *** coolsvap is now known as coolsvap|away | 08:16 | |
| *** coolsvap|away is now known as coolsvap | 08:19 | |
| *** dguerri` is now known as dguerri | 08:21 | |
| *** e0ne has joined #openstack-ironic | 08:32 | |
| huats | huats | 08:32 |
| lucasagomes | morning folks | 08:34 |
| lucasagomes | if you have a time please review https://review.openstack.org/#/c/194132 | 08:34 |
| lucasagomes | needs another +2 :-) | 08:34 |
| huats | I am having some issues too be able to have both vm instances and ironic one on the same OpenStack, is it just me ? | 08:36 |
| huats | (and a misconfiguration) | 08:36 |
| *** e0ne has quit IRC | 08:37 | |
| saripurigopi | lucasagomes: o/ | 08:37 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Add iBoot driver documentation https://review.openstack.org/193546 | 08:37 |
| lucasagomes | huats, usually people use host aggregates to separate both clouds | 08:37 |
| lucasagomes | huats, or cells | 08:37 |
| lucasagomes | the problem is scheduling, for baremetal we need to tweak the scheduler to behave differently | 08:38 |
| huats | lucasagomes: yeah that was my understanding | 08:38 |
| huats | in that case do you have just one list of filters ? | 08:38 |
| lucasagomes | since we can't allocate partial resources like you could with a hypervisor. Baremetal you always get 100% of the resources | 08:38 |
| lucasagomes | huats, you also have to tweak memory allocate ration | 08:38 |
| lucasagomes | to 1.0 instead of 1.5 and use a custom host manager for ironic | 08:39 |
| *** stegranet has joined #openstack-ironic | 08:39 | |
| lucasagomes | so reporting available resources are more accurate | 08:39 |
| lucasagomes | huats, 1 sec | 08:39 |
| huats | lucasagomes: I have ssen that the memory allocation can be specified by host aggregate yes | 08:39 |
| lucasagomes | huats, http://docs.openstack.org/developer/ironic/deploy/install-guide.html#configure-compute-service-to-use-the-bare-metal-service | 08:39 |
| huats | but I am not sure for the schedulling part | 08:39 |
| *** romcheg has joined #openstack-ironic | 08:41 | |
| huats | lucasagomes: in the doc you are refering it is the same settings for the whole openstack | 08:42 |
| lucasagomes | huats, right, so the scheduling part I think it depends on ur use case | 08:45 |
| lucasagomes | by default we do exact matching of the resources (those exact filters are enabled by default when you set the IronicHostManager) | 08:45 |
| lucasagomes | because we don't want users to ask for X and get Y, if one ask for 4 CPUs and 2G of ram | 08:45 |
| *** jcoufal has joined #openstack-ironic | 08:46 | |
| lucasagomes | we want to give him that and not 8 CPUs and 8 GB of ram (because we can't allocate part of the resources) | 08:46 |
| huats | yep | 08:46 |
| lucasagomes | but if that's OK for you, sure yeah you can tweak the scheduler to behave that way too | 08:46 |
| huats | i have seen that | 08:46 |
| huats | that is why I am asking if I can have a different list of filters based on hostaggregates, so that I can have one hostaggregate for vm (and thus their own filters) and another for baremetal (with another set of filters with exact types in particular) | 08:48 |
| lucasagomes | I believe you can, because you'r going to have 2 scheduler services running right? One for each host aggregates I beleive | 08:48 |
| lucasagomes | I would ask in #openstack-nova for specifics | 08:48 |
| huats | ok | 08:49 |
| huats | thanks ! | 08:49 |
| *** derekh has joined #openstack-ironic | 08:51 | |
| *** stendulker has joined #openstack-ironic | 09:05 | |
| *** pelix has joined #openstack-ironic | 09:05 | |
| openstackgerrit | Shivanand Tendulker proposed stackforge/proliantutils: Adding RIS support for virtual media interfaces https://review.openstack.org/194570 | 09:06 |
| sambetts | Morning all o/ | 09:09 |
| stendulker | Good morning sambetts :) | 09:09 |
| *** e0ne has joined #openstack-ironic | 09:10 | |
| yuikotakada | sambetts, stendulker: o/ | 09:10 |
| huats | lucasagomes: apparently running 2 diffents scheduller (with different configuration I mean) is not really doable | 09:10 |
| stendulker | yuikotakada: o/ | 09:10 |
| huats | if someone is using hostaggregates to have ironic and instances I am really really interested :) | 09:11 |
| sambetts | o/ yuikotakada | 09:11 |
| huats | devananda apparently that is what you told me to do :) | 09:11 |
| lucasagomes | huats, oh :-/ | 09:14 |
| lucasagomes | perhaps cells would be the only way them? | 09:14 |
| huats | I imagine :( | 09:15 |
| huats | which is not really nice | 09:15 |
| *** siva has joined #openstack-ironic | 09:15 | |
| huats | I have to figure it out :( | 09:15 |
| *** siva is now known as Guest61470 | 09:15 | |
| *** e0ne is now known as e0ne_ | 09:16 | |
| *** Guest61470 has left #openstack-ironic | 09:17 | |
| openstackgerrit | Tan Lin proposed openstack/ironic-specs: Add clean task field to node https://review.openstack.org/195470 | 09:19 |
| *** e0ne_ has quit IRC | 09:22 | |
| dtantsur | sambetts, morning! | 09:25 |
| openstackgerrit | Ramakrishnan G proposed stackforge/proliantutils: Check if virtual media is attached before ejecting https://review.openstack.org/195432 | 09:31 |
| sambetts | Morning dtantsur | 09:31 |
| rameshg87 | lucasagomes: was checking https://review.openstack.org/#/c/194132/6 | 09:32 |
| rameshg87 | lucasagomes: hi there | 09:32 |
| lucasagomes | rameshg87, hi there | 09:32 |
| rameshg87 | lucasagomes: do you think it makes sense to set last_error there ? | 09:32 |
| *** e0ne has joined #openstack-ironic | 09:33 | |
| lucasagomes | rameshg87, we could, to say that the node was put into deploy fail state? | 09:33 |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic: Do not preserve partitions after ephemeral https://review.openstack.org/182688 | 09:34 |
| * rameshg87 checks | 09:36 | |
| rameshg87 | lucasagomes: oh, so node goes to deleted and then cleaning and back to available | 09:36 |
| rameshg87 | okay, got it. I thought we were going to fail deploy | 09:36 |
| rameshg87 | that seems not the case | 09:36 |
| lucasagomes | yeah it will be ready for available again | 09:36 |
| lucasagomes | did you test it locally? | 09:37 |
| stendulker | lucasagomes, dtantsur: Good morning ! | 09:45 |
| dtantsur | stendulker, morning | 09:46 |
| *** alexpilotti has joined #openstack-ironic | 09:46 | |
| stendulker | lucasagomes, dtantsur: Can you please review this spec when you get some time https://review.openstack.org/#/c/174295/ | 09:47 |
| lucasagomes | stendulker, morning | 09:47 |
| lucasagomes | ack | 09:47 |
| stendulker | lucasagomes, dtantsur: It's about adding grub2 bootloader for UEFI and secure boot support for pxe_ilo driver | 09:47 |
| stendulker | lucasagomes, dtantsur: thank you | 09:48 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Make task_manager logging more helpful https://review.openstack.org/194619 | 09:51 |
| dtantsur | lucasagomes, no more dirty hacks ^^^ | 09:51 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Add IPMI 1.5 support for the ipmitool power driver https://review.openstack.org/195157 | 09:55 |
| *** naohirot has quit IRC | 09:58 | |
| openstackgerrit | Sergey Vilgelm proposed openstack/ironic: Switch to oslo.service https://review.openstack.org/195008 | 10:11 |
| *** MattMan has quit IRC | 10:13 | |
| *** rameshg87 has left #openstack-ironic | 10:14 | |
| *** athomas has quit IRC | 10:15 | |
| *** husanu4 has joined #openstack-ironic | 10:19 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Clean nodes stuck in DEPLOYING state when ir-cond restarts https://review.openstack.org/194132 | 10:21 |
| *** rameshg87 has joined #openstack-ironic | 10:22 | |
| *** husanu4 has quit IRC | 10:25 | |
| *** husanu1 has joined #openstack-ironic | 10:26 | |
| *** husanu1 has quit IRC | 10:28 | |
| *** athomas has joined #openstack-ironic | 10:29 | |
| *** husanu1 has joined #openstack-ironic | 10:30 | |
| *** husanu1 has quit IRC | 10:31 | |
| *** e0ne is now known as e0ne_ | 10:31 | |
| *** MattMan has joined #openstack-ironic | 10:33 | |
| *** thrash|g0ne is now known as thrash | 10:33 | |
| *** e0ne_ has quit IRC | 10:36 | |
| zhenguo | lucasagomes dtantsur, can you help to review this https://review.openstack.org/#/c/177642/ if you have time. | 10:37 |
| dtantsur | if we have time... | 10:38 |
| dtantsur | :) | 10:38 |
| openstackgerrit | Merged openstack/ironic: Add iBoot driver documentation https://review.openstack.org/193546 | 10:38 |
| zhenguo | dtantsur: thanks :) | 10:39 |
| lucasagomes | zhenguo, will take a look | 10:39 |
| zhenguo | thanks lucasagomes | 10:39 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Image service should not be set in ImageCache constructor https://review.openstack.org/194752 | 10:40 |
| *** rameshg87 has left #openstack-ironic | 10:41 | |
| openstackgerrit | Sam Betts proposed openstack/ironic-inspector: Deprecate authenticate opt in favor of auth_strategy https://review.openstack.org/195082 | 10:44 |
| openstackgerrit | Merged openstack/ironic: Add unit test for ilo_deploy _configure_vmedia_boot() https://review.openstack.org/194047 | 10:46 |
| *** erelwell has joined #openstack-ironic | 10:46 | |
| *** erelwell is now known as bethelwell | 10:46 | |
| *** Marga_ has quit IRC | 10:50 | |
| *** ramineni has quit IRC | 10:53 | |
| *** e0ne has joined #openstack-ironic | 10:55 | |
| openstackgerrit | Merged openstack/ironic: Add ENROLL and related states to the state machine https://review.openstack.org/192269 | 11:09 |
| *** chenglch has quit IRC | 11:10 | |
| *** yuikotakada is now known as yuikot_go_home | 11:12 | |
| sambetts | dtantsur, yuikotakada: the reason I changed it do get_auth_strategy instead of just simply returning true and false is that it'll reduce the amount of reworking we need to do in the future, and we could even add new strategies before completely deprecating authenticate | 11:15 |
| *** e0ne is now known as e0ne_ | 11:15 | |
| dtantsur | sambetts, well, we're not planning on adding new strategies yet, but if we do, amount of reworking will still be high - it won't be enough to just return new string value :) | 11:15 |
| sambetts | dtantsur: We shouldn't have to modify the util function though, only the parts of the code actually dealing with the auth | 11:17 |
| *** david-lyle has quit IRC | 11:17 | |
| *** jrist has quit IRC | 11:18 | |
| dtantsur | well, I don't mind both variants actually | 11:18 |
| *** jrist has joined #openstack-ironic | 11:19 | |
| *** david-lyle has joined #openstack-ironic | 11:21 | |
| *** lucasagomes has quit IRC | 11:22 | |
| *** achanda has quit IRC | 11:25 | |
| *** e0ne_ has quit IRC | 11:26 | |
| *** e0ne has joined #openstack-ironic | 11:28 | |
| *** Marga_ has joined #openstack-ironic | 11:32 | |
| *** lucasagomes has joined #openstack-ironic | 11:34 | |
| *** bradjones has quit IRC | 11:38 | |
| *** EmilienM|off is now known as EmilienM | 11:38 | |
| *** saripurigopi has quit IRC | 11:39 | |
| *** bethelwell has quit IRC | 11:40 | |
| *** bradjones has joined #openstack-ironic | 11:41 | |
| *** bradjones has quit IRC | 11:41 | |
| *** bradjones has joined #openstack-ironic | 11:41 | |
| *** bethelwell has joined #openstack-ironic | 11:41 | |
| *** lucasagomes is now known as lucas-hungry | 11:43 | |
| *** bethelwe_ has joined #openstack-ironic | 11:45 | |
| *** Marga_ has quit IRC | 11:48 | |
| *** bethelwell has quit IRC | 11:48 | |
| *** bethelwe_ has quit IRC | 11:49 | |
| *** stegranet has quit IRC | 11:52 | |
| *** stegranet has joined #openstack-ironic | 11:53 | |
| *** amotoki has quit IRC | 11:57 | |
| *** yuanying has quit IRC | 11:57 | |
| jefabo | huats: thansk for asking :) | 11:58 |
| huats | jefabo: do you have any clue ? | 11:59 |
| jefabo | huats: not at all, unless using a single flavor / signle type of hardware | 12:03 |
| *** zhenguo has quit IRC | 12:04 | |
| *** stegranet has quit IRC | 12:08 | |
| *** stegranet has joined #openstack-ironic | 12:09 | |
| openstackgerrit | Shivanand Tendulker proposed stackforge/proliantutils: Adding RIS support for virtual media interfaces https://review.openstack.org/194570 | 12:17 |
| *** killer_prince has joined #openstack-ironic | 12:18 | |
| *** killer_prince has quit IRC | 12:21 | |
| *** lazy_prince has quit IRC | 12:22 | |
| *** stegranet has quit IRC | 12:24 | |
| *** stegranet1 has joined #openstack-ironic | 12:24 | |
| *** coolsvap is now known as coolsvap|away | 12:26 | |
| *** stegranet1 is now known as stegranet | 12:27 | |
| *** stendulker has quit IRC | 12:27 | |
| *** bethelwell has joined #openstack-ironic | 12:28 | |
| *** dprince has joined #openstack-ironic | 12:29 | |
| *** dtantsur is now known as dtantsur|brb | 12:32 | |
| *** coolsvap|away is now known as coolsvap | 12:35 | |
| *** david-ly_ has joined #openstack-ironic | 12:35 | |
| *** coolsvap is now known as coolsvap|away | 12:36 | |
| *** david-lyle has quit IRC | 12:39 | |
| *** trown|outttypeww is now known as trown | 12:41 | |
| TheJulia | Good morning! | 12:44 |
| *** lucas-hungry is now known as lucasagomes | 12:47 | |
| *** bradjones has quit IRC | 12:47 | |
| *** bradjones has joined #openstack-ironic | 12:50 | |
| *** bradjones has quit IRC | 12:50 | |
| *** bradjones has joined #openstack-ironic | 12:50 | |
| *** husanu5 has joined #openstack-ironic | 12:50 | |
| *** husanu5 has quit IRC | 12:52 | |
| *** rloo has joined #openstack-ironic | 12:55 | |
| *** absubram has quit IRC | 12:55 | |
| *** husanu1 has joined #openstack-ironic | 12:56 | |
| *** romcheg has quit IRC | 12:58 | |
| *** romcheg has joined #openstack-ironic | 12:58 | |
| lucasagomes | TheJulia, good ugt morning | 12:59 |
| *** smoriya has quit IRC | 12:59 | |
| *** coolsvap|away is now known as coolsvap | 13:00 | |
| *** husanu1 has quit IRC | 13:00 | |
| *** e0ne is now known as e0ne_ | 13:04 | |
| *** e0ne_ is now known as e0ne | 13:04 | |
| *** yuikotakada has joined #openstack-ironic | 13:06 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Add delete role for dynamic inventory testing https://review.openstack.org/193690 | 13:06 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Individual playbook for creating a test VM https://review.openstack.org/193691 | 13:06 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Update the dynamic test playbook https://review.openstack.org/193692 | 13:06 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Add dynamic inventory description to README https://review.openstack.org/193709 | 13:06 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Addition of dynamic enrollment and deployment playbooks https://review.openstack.org/193737 | 13:07 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Change Bifrost over to using dynamic inventory https://review.openstack.org/193885 | 13:07 |
| *** dtantsur|brb is now known as dtantsur | 13:07 | |
| dtantsur | much code wow ^^^ | 13:07 |
| TheJulia | yeah... :( Next time I chain a bunch of stuff together, I'm going to do Depends-On | 13:08 |
| TheJulia | and just make big list in the last few commits. :( | 13:08 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update cached images based on update time https://review.openstack.org/195017 | 13:12 |
| *** husanux7 has joined #openstack-ironic | 13:13 | |
| *** husanux7 has quit IRC | 13:13 | |
| *** husanux6 has joined #openstack-ironic | 13:14 | |
| *** degorenko has quit IRC | 13:16 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update cached images based on update time https://review.openstack.org/195017 | 13:17 |
| *** husanux6 has quit IRC | 13:18 | |
| *** husanux1 has joined #openstack-ironic | 13:20 | |
| *** e0ne is now known as e0ne_ | 13:20 | |
| *** romcheg has quit IRC | 13:20 | |
| *** husanux1 has quit IRC | 13:21 | |
| *** coolsvap is now known as coolsvap|away | 13:21 | |
| *** e0ne_ is now known as e0ne | 13:22 | |
| *** Marga_ has joined #openstack-ironic | 13:22 | |
| *** romcheg has joined #openstack-ironic | 13:22 | |
| *** martini has joined #openstack-ironic | 13:22 | |
| martini | Good morning Ironic | 13:23 |
| *** husanux5 has joined #openstack-ironic | 13:25 | |
| rloo | hi ironickers, dtantsur. g'morning martini, TheJulia | 13:25 |
| dtantsur | morning rloo, martini | 13:26 |
| *** husanux5 has quit IRC | 13:29 | |
| *** Marga_ has quit IRC | 13:31 | |
| *** Marga_ has joined #openstack-ironic | 13:31 | |
| martini | TheJulia- For a full offline install, scripts/env-setup needs to be able to clone ansible from an alternate url. Any suggestions on how to store/pass that? | 13:34 |
| *** degorenko has joined #openstack-ironic | 13:34 | |
| *** absubram has joined #openstack-ironic | 13:35 | |
| TheJulia | We store it as an variable with a default, if it is already defined, then we use the dfined variable | 13:39 |
| TheJulia | defined variable, in the environment | 13:39 |
| TheJulia | so you could set it by doing something like export ANSIBLE_GIT_URL="/tmp/ansible.git" && scripts/env-setup | 13:40 |
| martini | Cool, thanks | 13:41 |
| TheJulia | np | 13:42 |
| *** bethelwell has quit IRC | 13:44 | |
| *** jcoufal has quit IRC | 13:49 | |
| bcornec | Hello everybody | 13:49 |
| TheJulia | Good morning! | 13:49 |
| bcornec | Have a question around bifrost. Still trying to call it in a self-sufficient and isolated docker container. | 13:49 |
| TheJulia | okay :) | 13:50 |
| TheJulia | Whats up? | 13:50 |
| bcornec | At a moment during the install I get the following error: | 13:50 |
| bcornec | TASK: [ironic-install | Starting MySQL] *************************************** | 13:50 |
| bcornec | <localhost> REMOTE_MODULE service name=mysql state=started | 13:50 |
| bcornec | <localhost> EXEC ['/bin/sh', '-c', 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1435239030.01-233799260644744 && chmod a+rx $HOME/.ansible/tmp/ansible-tmp-1435239030.01-233799260644744 && echo $HOME/.ansible/tmp/ansible-tmp-1435239030.01-233799260644744'] | 13:50 |
| bcornec | <localhost> PUT /tmp/tmpjvZIm0 TO /home/bruno/.ansible/tmp/ansible-tmp-1435239030.01-233799260644744/service | 13:50 |
| bcornec | <localhost> EXEC /bin/sh -c 'sudo -k && sudo -H -S -p "[sudo via ansible, key=farvusjvqpxqrnxscqfwmwsjuskluvup] password: " -u root /bin/sh -c '"'"'echo BECOME-SUCCESS-farvusjvqpxqrnxscqfwmwsjuskluvup; LANG=en_US.UTF-8 LC_CTYPE=en_US.UTF-8 /usr/bin/python /home/bruno/.ansible/tmp/ansible-tmp-1435239030.01-233799260644744/service; rm -rf /home/bruno/.ansible/tmp/ansible-tmp-1435239030.01-233799260644744/ >/dev/null 2>&1'"'"'' | 13:50 |
| bcornec | failed: [localhost] => {"failed": true} | 13:50 |
| bcornec | msg: * Starting MySQL database server mysqld | 13:50 |
| bcornec | ...done. | 13:50 |
| bcornec | FATAL: all hosts have already failed -- aborting | 13:50 |
| bcornec | PLAY RECAP ******************************************************************** | 13:50 |
| bcornec | to retry, use: --limit @/home/bruno/install.yaml.retry | 13:50 |
| bcornec | localhost : ok=21 changed=19 unreachable=0 failed=1 | 13:50 |
| bcornec | INFO[0846] The command [/bin/sh -c su - $BIFROST_UID -c "./launch-bifrost.sh $http_proxy $https_proxy"] returned a non-zero code: 2 | 13:50 |
| *** dasm is now known as dasm|afk | 13:50 | |
| bcornec | However I already stopped my local MySQL so have no real clue why it's not startng correctly here | 13:51 |
| NobodyCam | good (UGT) Morning Ironicers | 13:51 |
| NobodyCam | bcornec: anything in the mysql log? | 13:53 |
| TheJulia | bcornec: what distribution/distribution version are you using. That seems reminiscent of an error I found ages ago on TripleO | 13:55 |
| *** stegranet has quit IRC | 13:58 | |
| *** athomas has quit IRC | 13:58 | |
| *** stegranet has joined #openstack-ironic | 13:59 | |
| lucasagomes | NobodyCam, morning | 14:07 |
| *** r-daneel has joined #openstack-ironic | 14:09 | |
| NobodyCam | morning lucasagomes :) | 14:10 |
| bcornec | My distro is Mageia, but the container is Ubuntu 15.04 | 14:11 |
| bcornec | Problem is that it's difficult to get logs as the container dies... I'll try o reproduce manually, but thought I could ask before if that was a known problem. Last time I did an update from your git repo it fixed the previous issue I had, so guessed it could be handled similarly... | 14:12 |
| *** e0ne is now known as e0ne_ | 14:13 | |
| NobodyCam | bcornec: I have not seen that in my testing. | 14:13 |
| *** romainh has quit IRC | 14:15 | |
| *** romainh has joined #openstack-ironic | 14:15 | |
| TheJulia | hmm | 14:17 |
| TheJulia | bcornec: have you tried 14.04? | 14:18 |
| bcornec | I also have other error msgs which seem less important: | 14:18 |
| bcornec | - name: "Install Ironic using pip" | 14:18 |
| bcornec | pip: name=/opt/stack/ironic state=latest extra_args='--proxy http://web-proxy.fra.hp.com:8080' | 14:18 |
| bcornec | error: Your local changes to the following files would be overwritten by checkout: | 14:18 |
| bcornec | Please, commit your changes or stash them before you can switch branches. | 14:18 |
| bcornec | Aborting | 14:18 |
| bcornec | Cleaning up... | 14:18 |
| bcornec | mkdir: cannot create directory '/opt/stack': Permission denied | 14:18 |
| bcornec | Cloning into 'ansible'... | 14:18 |
| bcornec | reading manifest file 'lib/ansible.egg-info/SOURCES.txt' | 14:18 |
| bcornec | reading manifest template 'MANIFEST.in' | 14:18 |
| bcornec | no previously-included directories found matching 'v2' | 14:18 |
| bcornec | no previously-included directories found matching 'docsite' | 14:18 |
| bcornec | no previously-included directories found matching 'ticket_stubs' | 14:18 |
| bcornec | no previously-included directories found matching 'packaging' | 14:18 |
| bcornec | no previously-included directories found matching 'test' | 14:18 |
| bcornec | no previously-included directories found matching 'hacking' | 14:18 |
| bcornec | no previously-included directories found matching 'lib/ansible/modules/core/.git' | 14:18 |
| bcornec | no previously-included directories found matching 'lib/ansible/modules/extras/.git' | 14:18 |
| bcornec | writing manifest file 'lib/ansible.egg-info/SOURCES.txt' | 14:18 |
| TheJulia | please use paste.openstack.org | 14:19 |
| lucasagomes | folks http://paste.openstack.org | 14:19 |
| NobodyCam | bcornec: please use service lisk paste.openstack.org | 14:19 |
| NobodyCam | lol | 14:19 |
| lucasagomes | hah | 14:19 |
| lucasagomes | NobodyCam, PixeBot should detect it | 14:19 |
| NobodyCam | hummm :) | 14:19 |
| dtantsur | NobodyCam, morning! | 14:20 |
| bcornec | Are you working with 14.04 as a preferred distro ? | 14:20 |
| NobodyCam | mornign dtantsur | 14:20 |
| NobodyCam | :) | 14:20 |
| bcornec | Relaunching with 14.04. Invalidates the full docker cache of course, so we'll have to wait a bit before having the result ;-) | 14:21 |
| devananda | morning, all | 14:21 |
| NobodyCam | good morning devananda | 14:21 |
| TheJulia | bcornec: yes, we developed and run ci on trusty, presently working on rhel/centos support | 14:21 |
| bcornec | morning dvvdv | 14:22 |
| TheJulia | bcornec: give me a few and i'll look at that error | 14:22 |
| bcornec | TheJulia: easy to change in the docker file, longer to have it run again completely. You have a few to look at the error if you have time to do so :-) | 14:23 |
| dtantsur | devananda, morning! I've update the task_manager debugging patch https://review.openstack.org/#/c/194619/ as we agreed | 14:23 |
| *** e0ne_ has quit IRC | 14:23 | |
| bcornec | TheJulia: My modifications are at https://github.com/bcornec/bifrost in case you want ot have a look at what I'm doing... | 14:24 |
| devananda | dtantsur: at a read through, I like it! | 14:26 |
| dtantsur | cool :) | 14:28 |
| yuikotakada | dtantsur, hi, what does 'NOT_REGISTERED' mean in https://review.openstack.org/#/c/174231/, do you know? | 14:29 |
| yuikotakada | sambetts, around here? | 14:30 |
| sambetts | yuikotakada: Yup :) | 14:31 |
| yuikotakada | sambetts, o/ I'm reading your patch and I'm a little bit confused >_< | 14:31 |
| sambetts | yuikotakada: oh? whats up? | 14:32 |
| yuikotakada | sambetts, Even if CONF.authenticate is 'noauth', utils.add_auth_middleware(app) will be executed in main.py? | 14:32 |
| devananda | dtantsur: heh, tiny nit - there's no punctuation | 14:33 |
| devananda | Attempting to get exclusive lock on node 34184b36-bae7-47a1-91a0-75dcdd7ec28e for power state sync __init__ /opt/source/openstack/ironic/ironic/conductor/task_man | 14:34 |
| sambetts | yuikotakada: No, if CONF.authenticate = False then get_auth_strategy should return 'noauth' and therefore add_auth_middleware(app) will be skipped | 14:35 |
| *** e0ne has joined #openstack-ironic | 14:36 | |
| TheJulia | bcornec: Interesting... seems like starting with a new base, I suspect you've been tracking with the same image from the earlier bifrost code, maybe nuking /opt/stack/ansible? But if your rebuilding with 14.04, I suspect it will largely disappear except for the mkdir I think | 14:36 |
| dtantsur | devananda, you mean dot in the end? because everything starting with __init__ is not my fault :) | 14:36 |
| NobodyCam | good (UGT) morning sambetts | 14:36 |
| TheJulia | bnemec: we've also put a number of fixes in, you may wish to refresh your master branch and rebase your docker branch when you have a little time :) | 14:37 |
| devananda | dtantsur: oslo debug logging always appends method name and file name | 14:37 |
| *** athomas has joined #openstack-ironic | 14:37 | |
| dtantsur | yuikotakada, probably job is not ready to run yet. I've rechecked to figure out | 14:37 |
| devananda | dtantsur: so yea, we need some separator there. since you've got (...) in the other two messages, wdyt about using paren like this: (for $purpose) | 14:37 |
| devananda | dtantsur: so that all three logs have "some words (...) {func} {file}" | 14:38 |
| huats | devananda: yesterday I was asking about how to use separate scheduler filters (some dedicated to vm and some dedicated to baremetal) and you told me to use host aggregate. But I clearly don't see how... I have been searching and I fear it is not possible :( | 14:38 |
| huats | do you have anythng precise in mind ? | 14:38 |
| dtantsur | devananda, why not, will do | 14:38 |
| devananda | dtantsur: also, this is really really a lot of logs now. I think these are very helpful debug logs, but it might be good to use a CONF flag to disable by default? | 14:38 |
| huats | actually I just want to be able to use at the same time ironic and kvm for vm on the same openatck | 14:38 |
| devananda | dtantsur: that ^ is a longer discussion about cleaning up our logging in general -- not related to your patch | 14:39 |
| dtantsur | devananda, it's actually let then before, when we had 5-10 attempts... I believe that's purpose of DEBUG - to show everything | 14:39 |
| dtantsur | devananda, re our logs, I think we should make INFO useful, not strip DEBUG | 14:40 |
| devananda | dtantsur: ++ | 14:40 |
| devananda | huats: http://docs.openstack.org/kilo/config-reference/content/section_compute-scheduler.html#host-aggregates | 14:42 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Make task_manager logging more helpful https://review.openstack.org/194619 | 14:42 |
| dtantsur | devananda, ^^^ with (for purpose) | 14:43 |
| *** mitchjameson has joined #openstack-ironic | 14:43 | |
| devananda | eep. fat fingered a +A on that | 14:43 |
| yuikotakada | sambetts, Ah! CONF.authenticate is boolean! :D I misunderstood it is 'keystone' or 'noauth' same as auth_strategy. Thanks a lot! | 14:43 |
| NobodyCam | TheJulia: got a sec for https://review.openstack.org/#/c/194893 | 14:44 |
| sambetts | yuikotakada: :D hehe yup | 14:44 |
| dtantsur | devananda, lol, Alt+3 instead of Alt+2? ;) | 14:44 |
| yuikotakada | dtantsur, Can I believe that it will work well someday? I'm worrying about I wrote job wrongly. | 14:45 |
| * dtantsur is checking zuul | 14:46 | |
| yuikotakada | sambetts, I will read your patch again carefully and write reply (maybe within 1h), so please wait for a while... | 14:46 |
| dtantsur | yuikotakada, the job is queued in zuul this time, so at least it's going to run :) | 14:47 |
| sambetts | yuikotakada: Thanks :D | 14:47 |
| bcornec | TheJulia: Going further with 14.04 !! Now I have an error with | 14:47 |
| bcornec | TASK: [ironic-install | RabbitMQ - Fixing /etc/hosts] ************************* | 14:47 |
| bcornec | failed: [localhost] => {"changed": true, "cmd": ["sed", "-i", "s/localhost/d3659c5e113e localhost/", "/etc/hosts"], "delta": "0:00:00.096540", "end": "2015-06-25 14:45:04.390499", "rc": 4, "start": "2015-06-25 14:45:04.293959", "warnings": ["Consider using template or lineinfile module rather than running sed"]} | 14:47 |
| bcornec | stderr: sed: cannot rename /etc/sedyP8Egx: Device or resource busy | 14:47 |
| bcornec | More logs at http://paste.openstack.org/show/320990/ | 14:48 |
| yuikotakada | dtantsur, could you please teach me how to check zool? | 14:48 |
| dtantsur | yuikotakada, go to http://status.openstack.org/zuul/ and enter your change number (see in gerrit URL, 174231 in this case) | 14:49 |
| dtantsur | yuikotakada, when you see your patch, click on it to expand | 14:49 |
| TheJulia | bcornec: wow... that.... just... is no fun :( | 14:49 |
| huats | devananda: hum. This is how aggregates are created. But the thing is that the filters are shared accross all you aggregates AFAIK so you cannot have some specific filter list to some hostaggregate | 14:49 |
| BadCub | devananda: I have added the date confirmation for mid-cycle to meeting | 14:49 |
| bcornec | TheJulia: why no fun ? If it was working out of the box it would not have been fun ;-) | 14:50 |
| TheJulia | bcornec: Due to rabbitmq, the node hostname has to be resolvable with localhost *facepalm*, thats what that change does | 14:50 |
| JayF | BadCub: is the date finalized? E | 14:50 |
| bcornec | Now I can learn something and you can fix something hopefully ! | 14:50 |
| BadCub | JayF: yep. We got confirmation from HP Facilities for Aug 12-14 | 14:50 |
| NobodyCam | morning JayF :) | 14:51 |
| *** mitchjameson has quit IRC | 14:51 | |
| JayF | morning | 14:51 |
| NobodyCam | :) | 14:51 |
| yuikotakada | dtantsur, I see. Thank you. | 14:52 |
| dtantsur | morning JayF, BadCub | 14:52 |
| JayF | morning | 14:52 |
| BadCub | morning dtantsur | 14:52 |
| *** david-ly_ is now known as david-lyle | 15:01 | |
| *** mtanino has joined #openstack-ironic | 15:05 | |
| *** martini has quit IRC | 15:12 | |
| *** mtanino_ has joined #openstack-ironic | 15:15 | |
| *** mtanino has quit IRC | 15:16 | |
| sambetts | Thanks for the review yuikotakada :D | 15:20 |
| *** ijw_ has joined #openstack-ironic | 15:21 | |
| yuikotakada | sambetts, you're welcome :) | 15:21 |
| rloo | Hi BadCub, wrt the midcycle. If it is definitely "on", maybe send email about it so people can start planning for it sooner. | 15:24 |
| *** hblixt has quit IRC | 15:27 | |
| *** mtanino has joined #openstack-ironic | 15:28 | |
| rloo | hi dtantsur, devananda: wrt is_hostname_safe(). https://review.openstack.org/#/c/193587/ | 15:30 |
| rloo | dtantsur, devananda: I guess 'we' should decide once and for all. is this something to bring up at the meeting? | 15:31 |
| *** mtanino_ has quit IRC | 15:31 | |
| rloo | dtantsur, devananda: or at least decide for the near-future :) | 15:31 |
| dtantsur | I'm just against breaking changes without serious reason. I don't care about names too much, though having underscore would be good IMO | 15:31 |
| JayF | Like I get what Chris is saying though. We're in the worst of all worlds right now | 15:32 |
| JayF | we don't restrict to the standard properly, and we don't allow whatever folks want | 15:32 |
| JayF | we've invented a new standard | 15:32 |
| rloo | dtantsur: i looked at the past changes to that. The underscore was added for 'domain' part, not hostname part. | 15:32 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: WIP: Test shade microversion lock ***DO NOT MERGE*** https://review.openstack.org/195637 | 15:32 |
| JayF | whether or not it's worth breaking changes to fix is still the question; but I do think the way it is may not be awesome | 15:33 |
| dtantsur | relaxing rules is not a breaking change, so I would go for that one :) | 15:33 |
| rloo | JayF: ++ | 15:33 |
| devananda | rloo: that's the plan (sending an email today) | 15:33 |
| rloo | thx devananda | 15:33 |
| dtantsur | cool! | 15:33 |
| * rloo wonders if we need to spend more time reading specs in detail but i don't know that it would help much and might cause more frustration. | 15:34 | |
| dtantsur | rloo, are you following [Nova] discussion about dropping specs? :) | 15:35 |
| dtantsur | seems like it's easy to overdo | 15:35 |
| rloo | dtantsur: no, didn't know they were having a discussion. but i'm behind in my emails too. | 15:35 |
| *** e0ne is now known as e0ne_ | 15:35 | |
| rloo | dtantsur: the whole 'ungrammatical' way of "writing' things/projects in OpenStack has me a bit depressed | 15:36 |
| dtantsur | me too | 15:36 |
| dtantsur | not only is it ungrammatical, it's also confusing a lot | 15:36 |
| rloo | dtantsur: http://lists.openstack.org/pipermail/openstack-docs/2015-June/007020.html | 15:37 |
| dtantsur | imagine user seeing "ironic inspector" in the middle of a sentence. how are they supposed to understand that this is a project name | 15:37 |
| *** mark-os has quit IRC | 15:37 | |
| rloo | dtantsur: some good comments there but i don't know how/who decides what there. | 15:37 |
| BadCub | rloo: I will do that today | 15:37 |
| *** e0ne_ is now known as e0ne | 15:37 | |
| rloo | thx BadCub | 15:37 |
| dtantsur | rloo, yeah, we're not the only confused | 15:38 |
| *** mtanino has quit IRC | 15:40 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update cached images based on update time https://review.openstack.org/195017 | 15:43 |
| BadCub | rloo: email sent :) | 15:45 |
| *** yog_ has quit IRC | 15:46 | |
| lucasagomes | rloo, heh luckily I got my pop corn to watch the comments on that patch changing "Bare metal" to "Bare Metal" | 15:50 |
| lucasagomes | I remember when we were discussing whether the first line of the commit message should end with a period or not | 15:50 |
| lucasagomes | at end we allowed both :-) | 15:50 |
| rloo | lucasagomes: when you originally said you were getting popcorn (was it only a few days ago), I thought, what do you mean? Now I know :-( | 15:51 |
| lucasagomes | rloo, heh yeah watching the show | 15:51 |
| NobodyCam | morning rloo :) | 15:52 |
| rloo | morning NobodyCam! | 15:52 |
| NobodyCam | :) | 15:52 |
| rloo | lucasagomes: I added a bit more comedy to it ;) | 15:52 |
| lucasagomes | hah will check | 15:53 |
| lucasagomes | I love the gif http://i.imgur.com/deAYzZe.gif | 15:53 |
| rloo | lucasagomes: ha ha | 15:53 |
| NobodyCam | :-p | 15:53 |
| *** yuikotakada has quit IRC | 15:54 | |
| lucasagomes | the cool thing is that people already stopped talking about whether it Bare Metal or Bare metal, they are talking about whether the nick name of the projects (Ironic) is a proper noun or not | 15:54 |
| lucasagomes | and so it should be capitalized | 15:54 |
| rloo | dtantsur: we need a spec for any change to a driver interface? or only certain kinds of changes? https://review.openstack.org/#/c/188466/ | 15:54 |
| jlvillal | Good morning Ironic. Four more work days until my (4 week) sabbatical :) | 15:54 |
| lucasagomes | jlvillal, w00t! | 15:55 |
| rloo | lucasagomes: what I don't understand (and doesn't seem to be documented) is where they came up with those 'conventions' in the first place. | 15:55 |
| rloo | jlvillal: stop making us jealous! :) | 15:55 |
| lucasagomes | rloo, well it's probably around for quite while, probably the docs team does make those conventions | 15:55 |
| jlvillal | lucasagomes, rloo: Thanks. But I have this feeling I will still come here to hang out at times :) | 15:56 |
| lucasagomes | jlvillal, you welcome, bring beer | 15:56 |
| jlvillal | lucasagomes: I will, if I can figure out what the beer emoticon is | 15:56 |
| *** e0ne is now known as e0ne_ | 15:56 | |
| rloo | jlvillal: you're always welcome ;) | 15:56 |
| jlvillal | Thanks rloo! | 15:57 |
| rloo | jlvillal: btw, before you take off, what's the status with testing? maybe you can update the etherpad. | 15:57 |
| NobodyCam | morning jlvillal | 15:57 |
| NobodyCam | jlvillal: 4 weeks of sabbatical wow | 15:58 |
| dtantsur | rloo, my intuition says yes. driver interface is a stable one in Ironic | 15:58 |
| dtantsur | at least we maintain it as such | 15:58 |
| dtantsur | so I'd like more visibility on it's changes, especially when we're in risk of breaking compatibility | 15:58 |
| rloo | dtantsur: so any change to an existing method there, or adding a new one, requires a spec? | 15:58 |
| JayF | jlvillal: I'm counting down too; something like 3 weeks until my 1m sabbatical | 15:58 |
| dtantsur | rloo, I believe so (others might disagree) | 15:58 |
| *** kozhukalov_ has joined #openstack-ironic | 15:59 | |
| rloo | dtantsur: I do agree that visibility is important. | 15:59 |
| jlvillal | rloo: Okay. I ended up spending far too much time trying (and failing) to get the Ubuntu Cloud image to work to run devstack. I want to spend more time trying to figure out the issue why the scheduler says it can't schedule. | 15:59 |
| jlvillal | JayF: Congrats! :) | 15:59 |
| rloo | dtantsur: I want to make sure we're all 'in agreement' so we are consistent. I sometimes feel sorry for our developers. | 15:59 |
| dtantsur | yeah, it's annoying, I agree | 16:00 |
| dtantsur | if we had more spec cores, we could move faster | 16:00 |
| dtantsur | rloo, e.g. Ruby Loo does wonderful reviews, do you think we could convince here to become a specs-core? ;) | 16:01 |
| rloo | dtantsur: oh, I was thinking more like 'all spec cores' need to review :D | 16:01 |
| dtantsur | :) | 16:02 |
| *** Sukhdev has joined #openstack-ironic | 16:03 | |
| rloo | dtantsur: the official name is 'Bare metal inspector' | 16:03 |
| TheJulia | JayF: Have fun! Will you be visiting NC? | 16:03 |
| JayF | TheJulia: nope; 1 week in Italy (Rome, then Venice); some time at home; and 5 days in Chicago | 16:04 |
| dtantsur | rloo, it sounds reasonable, but it's not official | 16:04 |
| TheJulia | JayF: Have fun! | 16:04 |
| JayF | TheJulia: there's any chance at all I'd go to NC to surprise family/friends, but I doubt I'll wanna travel that much as I'll also be away from hoem the week before and after my sabbatical | 16:04 |
| TheJulia | JayF: don't blame you there at all | 16:05 |
| *** e0ne_ has quit IRC | 16:06 | |
| rloo | dtantsur: it would be nice, but not sure it is that reasonable :) | 16:08 |
| JayF | morganfainberg: asking about the new policy.json framework in the v3 middleware - would policys configured using the new middleware still apply for folks using keystone v2? | 16:08 |
| JayF | s/ys/ies/ | 16:09 |
| morganfainberg | V3 uses policy.json, v2 is really "is an admin", "is not an admin" or "no access" | 16:09 |
| morganfainberg | Not much beyond that. | 16:09 |
| morganfainberg | And mostly hard-coded-ish | 16:10 |
| morganfainberg | In keystone that is. | 16:10 |
| morganfainberg | So we'd implement new things on v3 only | 16:10 |
| morganfainberg | V2 is froZen from a keystone perspective. | 16:10 |
| JayF | morganfainberg: so if I have a keystone v2 server, and using keystone v3 middleware to implment policies | 16:10 |
| JayF | will I be able to use those policies? | 16:10 |
| morganfainberg | So v2 and v3 keystone only Impact auth (minimally) and keystone crud | 16:11 |
| *** e0ne has joined #openstack-ironic | 16:11 | |
| devananda | morganfainberg: context here is JayF is interested in making security things around password access/storage/change-mgmt better, and I think we need to have policy in place before it makes any sense to start adding crypto features that we want to restrict access to | 16:11 |
| morganfainberg | As a non Keystone Service you shouldn't care which one is used. You would need to have middleware source the policy stuff from the v3 interface. | 16:11 |
| morganfainberg | But that should be done by our middleware so less of something you have to directly implement. | 16:12 |
| morganfainberg | devananda: sure. | 16:12 |
| *** greghaynes has joined #openstack-ironic | 16:12 | |
| morganfainberg | JayF: I recommend focusing on v3. In (every?) deployment that is adhering to keystone and defcore, v3 is available. | 16:13 |
| morganfainberg | Even if they use v2 primarily. | 16:13 |
| morganfainberg | For auth/crud | 16:13 |
| JayF | morganfainberg: well, I'm at Rackspace; so I can't change my "keystone" very easily | 16:13 |
| JayF | morganfainberg: mainly curious if work to implement policies in a "v3" way would lend me no direct value until that identity service was upgraded | 16:13 |
| morganfainberg | JayF: I'll just say that any new apis in keystone will be v3 only. | 16:14 |
| JayF | morganfainberg: and it sounds like the answer is yes? | 16:14 |
| morganfainberg | So. Again, you don't have a lot from ironic to deal with from a v2 or v3 difference. Unless you are asking keystone for crud things. (Or care about domains) | 16:15 |
| morganfainberg | You get the same roles, scope, etc data from the middleware. | 16:15 |
| devananda | morganfainberg: we dont have a lot -today-. we're discussing adding more | 16:15 |
| morganfainberg | V2 vs v3 in keystone should have no impact on ironic unless you cared about hierarchical multi tenancy or domains | 16:16 |
| devananda | morganfainberg: and I think JayF's question is, if we add the sorts of things we're talking about, using v3 APIs and the new keystonemiddleware, will his deployment be able to derive any benefit? | 16:16 |
| devananda | ahh | 16:16 |
| devananda | afaik, we do not | 16:16 |
| JayF | sweet | 16:16 |
| morganfainberg | devananda: unless you are doing crud work and interacting with keystone's crud interface (heat) | 16:16 |
| morganfainberg | It wouldn't matter much. | 16:17 |
| morganfainberg | I recommend assuming v3 is used if there is ever a question, but it shouldn't matter to you beyond how keystone middleware is configured ;) | 16:17 |
| devananda | think: create three "roles" with different "access controls" implemented for business reasons | 16:17 |
| morganfainberg | If ironic is trying to create roles in keystone vs operator leveraging roles. Then we can talk. | 16:18 |
| morganfainberg | And it's more interesting (use v3 :P) | 16:18 |
| devananda | someone's gotta CRUD those roles, but I imagine that ironic is just checking them from a middleware in our API layer | 16:18 |
| morganfainberg | Yah. You'll know if the user has those roles based on keystone middleware decoding the token. | 16:18 |
| devananda | or keystone is checking them via the keystonemiddleware, and maybe ironic doesn't know anything about them | 16:19 |
| devananda | right | 16:19 |
| devananda | so we add keystonemiddleware as a Pecan hook | 16:19 |
| morganfainberg | Yep. We will just pass on "roles are [x, y, z] | 16:19 |
| devananda | if there are no roles, it just passes through | 16:19 |
| morganfainberg | Should already be done when you use auth_token | 16:19 |
| morganfainberg | It gets added to the context and / env cars from __call__ | 16:20 |
| devananda | if there are roles (and an auth_toekn) then keystonemiddleware checks them -- this actually doesn't require any changes to ironic's logic, beyond just adding the keystonemiddleware hook | 16:20 |
| devananda | and the operator creating those roles and setting the policy | 16:20 |
| *** subscope has quit IRC | 16:20 | |
| morganfainberg | You already have the hook. Afaik | 16:20 |
| devananda | morganfainberg: how does said operator declare what Roles have what CRUD acces to which of Ironic's API resources? | 16:20 |
| morganfainberg | It is handed in the auth_protocol wrapper you have. | 16:20 |
| morganfainberg | That is done via policy.json (today) | 16:21 |
| morganfainberg | And the enforcers used that compare the rule to the auth context. | 16:21 |
| devananda | JayF: sounds like there may not be much work to do here | 16:21 |
| morganfainberg | So you pass the context to the enforcer which has loaded the rules from policy.json, and then you say "enforce on rule/api X" | 16:21 |
| JayF | devananda: just basically hooking up roles to endpoints | 16:22 |
| devananda | yup | 16:22 |
| JayF | devananda: sounds like 90% design 10% implementation | 16:22 |
| JayF | mariojv: around? | 16:22 |
| morganfainberg | Most projects use Oslo.policy already and you can just use them as an example. | 16:22 |
| lucasagomes | are we thinking about exposing the ironic api to the end user? | 16:22 |
| lucasagomes | (non operator) | 16:22 |
| morganfainberg | Nova has the more complex uses where they enforce on ownership of a resource from the dB as well | 16:22 |
| JayF | lucasagomes: well; "operator" is a broad term | 16:23 |
| devananda | lucasagomes: not really -- more like different types of operator | 16:23 |
| JayF | lucasagomes: at Rackspace I might have different "level" of operators I want to have different access | 16:23 |
| lucasagomes | devananda, JayF right, that sounds pretty neat | 16:23 |
| devananda | that ^ is a request i've seen from several large customers, too | 16:23 |
| JayF | lucasagomes: think also about the ability to have dashboard users which can't see credentials | 16:23 |
| JayF | well; I think we have that policy already | 16:23 |
| JayF | but more things like that+++ | 16:23 |
| devananda | JayF: it's a CONF option, not a policy, which is bad IMO | 16:24 |
| devananda | but yes, more things like that ++ | 16:24 |
| lucasagomes | ++ | 16:24 |
| lucasagomes | sounds good indeed | 16:24 |
| JayF | devananda: I think we have it in policy as well? | 16:24 |
| morganfainberg | devananda, JayF: so you just need to add in Oslo.plolicy and make sure the auth token middleware is properly adding to your request context, then just do enforce where you want to enforce. | 16:24 |
| JayF | devananda: you're going to make me wanna check to see if this is downstream-only now, lol | 16:24 |
| devananda | JayF: I believe some of those things are downstream, yea | 16:25 |
| morganfainberg | JayF: if it is part of your "public" offering it likely uses repose. Which isn't compatible with the upstream policy.json stuff | 16:25 |
| JayF | devananda: https://github.com/openstack/ironic/blob/master/etc/ironic/policy.json#L3 | 16:25 |
| morganfainberg | Repose acts before policy.json | 16:25 |
| JayF | devananda: I think the show_password role is what matters there, but imbw | 16:25 |
| JayF | morganfainberg: we don't front ironic with repose | 16:25 |
| morganfainberg | Ok cool | 16:26 |
| morganfainberg | That makes it easier. | 16:26 |
| *** stegranet has quit IRC | 16:26 | |
| JayF | I never put java in my production if given a choice :P | 16:26 |
| morganfainberg | Lol | 16:27 |
| openstackgerrit | Merged openstack/bifrost: Add the rest of the cookiecutter-generated files for testing https://review.openstack.org/194893 | 16:27 |
| openstackgerrit | Merged openstack/bifrost: Add a dynamic prepare for test role https://review.openstack.org/192001 | 16:27 |
| devananda | JayF: ah - you may be correct | 16:27 |
| openstackgerrit | Merged openstack/bifrost: Add README note about cleaning https://review.openstack.org/194783 | 16:28 |
| mariojv | JayF: yeah, stepped away for a minute | 16:30 |
| lucasagomes | dtantsur, ping any reason why not +a https://review.openstack.org/#/c/194132/ ? | 16:31 |
| *** achanda has joined #openstack-ironic | 16:31 | |
| JayF | mariojv: if you read scrollback, Deva + me + Morgan (PTL of Keystone) were talking about policy stuff. I think that might be a good quick win instead of the barbican integration you're working on now. | 16:31 |
| dtantsur | lucasagomes, was assuming devananda might want to have a look | 16:31 |
| JayF | mariojv: WDYT? (once you get more context) | 16:31 |
| mariojv | looking | 16:31 |
| lucasagomes | dtantsur, fair | 16:31 |
| lucasagomes | devananda, yeah please, if you have 1 min to take a look at https://review.openstack.org/#/c/194132/ | 16:32 |
| * devananda looks | 16:32 | |
| lucasagomes | it mitigates the problem of having to go to the database and unstuck the nodes in DEPLOYING | 16:32 |
| lucasagomes | devananda, on top of that, a timeout will be added | 16:32 |
| JayF | lucasagomes: we were talking some yesterday about possibly allowing a transition from almost any state into error for similar reasons | 16:33 |
| lucasagomes | JayF, yeah we def should do something | 16:34 |
| JayF | lucasagomes: in our case we have some things getting stuck cleaning (still not found a root cause) and right now it takes a DB update | 16:34 |
| *** romainh has left #openstack-ironic | 16:34 | |
| lucasagomes | JayF, actually our ERROR state can leave us stuck too | 16:34 |
| JayF | and everytime someone has to touch a DB to fix production software I die a little inside :P | 16:34 |
| devananda | JayF: I've hit that tuck-in-cleaning seeral times too | 16:34 |
| lucasagomes | JayF, because, from ERROR we only can move to DEPLOYING (rebuild) or DELETING (delete) | 16:34 |
| JayF | lucasagomes: o, really? even without locks? | 16:34 |
| lucasagomes | JayF, yup | 16:35 |
| JayF | lucasagomes: hrmph | 16:35 |
| lucasagomes | JayF, if something fails there you get in a loop where you can't get out of there | 16:35 |
| devananda | lucasagomes: huh? error -> deleting should be ok. why is that not? | 16:35 |
| lucasagomes | devananda, we had a case where ipmi commands where failing, tear_down was trying to power off the node | 16:35 |
| lucasagomes | erroring out and sending it back to error | 16:35 |
| *** romainh has joined #openstack-ironic | 16:36 | |
| lucasagomes | but problem was that (IIRC) it was in provision_state = ERROR and target_provision_state = DEPLOYED | 16:37 |
| lucasagomes | so any updates to the node was denied because it was "in transition" | 16:37 |
| lucasagomes | there's something there, I've a TODO to look into it a bit more and investigate | 16:37 |
| devananda | ahh :( | 16:37 |
| devananda | so yea, you may not have been able to UPDATE the ipmi creds | 16:37 |
| devananda | that's silly of us | 16:37 |
| lucasagomes | yeah exactly | 16:37 |
| *** romainh has quit IRC | 16:38 | |
| openstackgerrit | Merged openstack/bifrost: Add os_ironic_facts module https://review.openstack.org/191093 | 16:38 |
| lucasagomes | I will start looking into some of those stuck problems | 16:38 |
| openstackgerrit | Merged openstack/bifrost: Correct license header on zuul change parsing code https://review.openstack.org/195073 | 16:38 |
| lucasagomes | devananda, like having timeouts etc... that patch for DEPLOYING is just to mitigate the problem, since at that point we don't have any API that allows us to get out of it | 16:39 |
| openstackgerrit | Merged openstack/bifrost: Addition of shade support to bifrost_inventory.py https://review.openstack.org/194684 | 16:39 |
| lucasagomes | devananda, also, there's something that may require some thought... In nova it's allowed to do a destroy() when the instance is "spawning" | 16:40 |
| lucasagomes | basically you could abort the deployment, which is not allowed in Ironic | 16:40 |
| *** derekh has quit IRC | 16:40 | |
| *** romainh has joined #openstack-ironic | 16:40 | |
| lucasagomes | (unless you're at DEPLOYWAIT) | 16:40 |
| mariojv | JayF: using keystone here sounds like a good idea b/c we do need to have roles defined for accessing the secrets before we even worry about how they're securely stored | 16:41 |
| JayF | mariojv: well basically what we have to do is come up with more "roles" for Ironic and assign endpoints/actions to the roles | 16:41 |
| JayF | mariojv: so we could make a user that could, say, only have READ access to insecure information in the API | 16:41 |
| JayF | mariojv: vs an Admin who can CRUD any endpoitn | 16:41 |
| JayF | mariojv: the pieces are all in place (that's what the scrollback was figuring out) | 16:42 |
| mariojv | right, that sounds like a good idea | 16:42 |
| JayF | mariojv: honestly the biggest amount of work here is likely getting everyone to agree on roles for the policies :) | 16:42 |
| mariojv | lol | 16:42 |
| devananda | lucasagomes: on 194132, i think there's an edge case here | 16:42 |
| *** ndipanov has quit IRC | 16:42 | |
| devananda | lucasagomes: if I am running 2 conductors (A and B) and I kill -9 both of them, then only start up conductor A | 16:42 |
| devananda | lucasagomes: self.dbapi.clear_node_reservations_for_conductor(self.host) gets called for that host (A) | 16:43 |
| devananda | lucasagomes: so half the nodes fail | 16:43 |
| devananda | *half hte nodes that were in DEPLOYING move to DEPLOYFAIL | 16:43 |
| lucasagomes | right | 16:44 |
| devananda | lucasagomes: if i do not start conductor B, we still hvae no way to recover the other half of the nodes w/o a DB update | 16:44 |
| lucasagomes | yeah the other half will still locked | 16:44 |
| lucasagomes | that will be solved with a timeout | 16:44 |
| devananda | right | 16:44 |
| lucasagomes | so conductor A may check the timeout (when added) | 16:44 |
| lucasagomes | and will unbreak those nodes | 16:44 |
| devananda | what if I dont restart either A or B, but instead I start a new conductor C ? | 16:45 |
| devananda | none of the deploys are failed | 16:45 |
| lucasagomes | then nodes still stuck in DEPLOYING state | 16:45 |
| mariojv | JayF: do you think that addresses the issue of driver pw storage / retrieval though? i see how it's a good idea but don't understand how that would help us if we needed to do something like upgrade BMC credentials across part of a fleet | 16:45 |
| devananda | this is not solving the root of the problem -- it's making an assumption that the operators will restart conductors with the same hostname | 16:45 |
| lucasagomes | devananda, yeah the final solution is the timeout | 16:45 |
| JayF | mariojv: it doesn't address the same issue; but it addresses a similar issue that probably should be handled first | 16:45 |
| lucasagomes | devananda, yes, that's the assumption. It just mitigates because right now | 16:45 |
| JayF | mariojv: if we store the info not awesomely it's not quite as bad if we can at least isolate some api users from it | 16:46 |
| mariojv | i see | 16:46 |
| lucasagomes | devananda, if the operator starts a conductor with the same hostname what happens is that the nodes are in DEPLOYING state without lock | 16:46 |
| lucasagomes | devananda, and you just can't do anything there unless you change the db | 16:46 |
| devananda | mariojv: and in order to benefit from storing the passwords in a better way, you also need to control who has access to them (or else why did you secure them if anyone can access them?) | 16:46 |
| lucasagomes | devananda, at least with that patch, I can tell one to "restart the conductor and it will clean up the nodes it was managing" | 16:47 |
| lucasagomes | the timeout is on the way btw, that will solve the broader case | 16:47 |
| devananda | lucasagomes: right - so I agree this mitigates one issue today, but it's very limited. the real solution (timeouts) would actually solve both sets of problems | 16:47 |
| devananda | yea | 16:47 |
| lucasagomes | yeah it will | 16:47 |
| lucasagomes | this just fails fast and fails safe for that case | 16:48 |
| lucasagomes | while not perfect, it helps | 16:48 |
| devananda | lucasagomes: mind if I edit the commit message to word that a bit more clearly? | 16:48 |
| lucasagomes | devananda, go ahead | 16:48 |
| devananda | I agree it helps, just want it to be clear what it is doing | 16:48 |
| devananda | thx | 16:48 |
| mariojv | devananda: that makes sense, we ought to define some sort of rbac before jumping ahead and storing all passwords in one place, i think what you were discussing with policies + keystone will do that | 16:49 |
| *** ijw_ has quit IRC | 16:49 | |
| devananda | lucasagomes: one more thing -- this isn't actually FAILing the deploys from the conductor that restarted | 16:50 |
| devananda | lucasagomes: it's FAILing any node that was in DEPLOYING state and had no lock | 16:50 |
| devananda | lucasagomes: while the end result is probably the same, there's still a race if >1 conductor restarts at the same time | 16:50 |
| lucasagomes | yeah, but we only clean the locks for the nodes it was managing | 16:50 |
| devananda | mariojv: exactly | 16:51 |
| JayF | lucasagomes: I think a node being deployed by agent, it's not locked 100% of the time it's in DEPLOYING state | 16:51 |
| lucasagomes | JayF, yeah nodes are not locked on DEPLOYWAIT state | 16:51 |
| lucasagomes | devananda, perhaps I should handle InvalidTransitionState there? | 16:52 |
| *** pelix has quit IRC | 16:52 | |
| *** rwsu has joined #openstack-ironic | 16:52 | |
| lucasagomes | InvalidState* | 16:52 |
| openstackgerrit | Merged openstack/bifrost: Remove un-necessary checks from the dynamic test https://review.openstack.org/193687 | 16:52 |
| openstackgerrit | Merged openstack/bifrost: Add dynamic node unprovision role https://review.openstack.org/193688 | 16:52 |
| openstackgerrit | Merged openstack/bifrost: Correct ipv4 networking config issue in dynamic path https://review.openstack.org/193689 | 16:52 |
| openstackgerrit | Merged openstack/bifrost: Add example deploy all available playbook https://review.openstack.org/195100 | 16:53 |
| openstackgerrit | Merged openstack/bifrost: Follow-up fix for H306 in bifrost_inventory.py https://review.openstack.org/195324 | 16:53 |
| *** romcheg has quit IRC | 16:55 | |
| *** romainh has left #openstack-ironic | 16:56 | |
| *** athomas has quit IRC | 16:57 | |
| *** trown is now known as trown|lunch | 16:58 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Clean nodes stuck in DEPLOYING state when ir-cond restarts https://review.openstack.org/194132 | 16:59 |
| devananda | lucasagomes: let me know what you think of wording ^ | 16:59 |
| devananda | lucasagomes: if good, i'll +A it | 16:59 |
| *** igordcard_ has joined #openstack-ironic | 17:01 | |
| * lucasagomes looks | 17:01 | |
| devananda | ugh. one sec. formatting got mangled | 17:01 |
| lucasagomes | hhe yeah | 17:01 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Clean nodes stuck in DEPLOYING state when ir-cond restarts https://review.openstack.org/194132 | 17:03 |
| devananda | that'll teach me to use gerrit's web ui | 17:03 |
| lucasagomes | devananda, just one thing, we also call clean_up and stuff ( see the callback_method=utils.cleanup_after_timeout) | 17:03 |
| lucasagomes | devananda, that will make the node not stay in DEPLOYFAIL, it will actually clean it up | 17:03 |
| lucasagomes | and put it back to available | 17:03 |
| lucasagomes | so it will be read for scheduling again | 17:03 |
| lucasagomes | ready* | 17:03 |
| devananda | ah, right | 17:03 |
| lucasagomes | which is neat (IMO) | 17:03 |
| devananda | oooh | 17:04 |
| devananda | lucasagomes: so, a) what happens if it hasn't exceeded the timeout? b) what does that do to Nova? | 17:04 |
| lucasagomes | devananda, we currently don't have the timeout, but once we do since the node will leave the DEPLOYING state | 17:05 |
| lucasagomes | the loop that checks the timeout won't filter that node anymore | 17:05 |
| lucasagomes | devananda, the instance in nova will go to ERROR when spawning, so you can delete it | 17:05 |
| *** hemna_ has joined #openstack-ironic | 17:06 | |
| devananda | lucasagomes: so this is factually incorrect | 17:06 |
| devananda | lucasagomes: I can start a deploy, then restart conductor, and within a few seconds of starting the deploy it will log: Timeout reached while waiting for callback for node %s | 17:06 |
| *** jefabo has quit IRC | 17:06 | |
| devananda | and set node.last-error to that | 17:07 |
| devananda | see ironic.conductor.utils.cleanup_after_timeout | 17:07 |
| * lucasagomes looks | 17:07 | |
| devananda | we should not call that from conductor.startup | 17:07 |
| *** jefabo has joined #openstack-ironic | 17:07 | |
| dtantsur | g'night folks | 17:07 |
| *** dtantsur is now known as dtantsur|afk | 17:07 | |
| NobodyCam | night dtantsur|afk | 17:07 |
| lucasagomes | devananda, oh yeah hmm the log will be misleading | 17:08 |
| lucasagomes | devananda, yeah perhaps we should leave it in DEPLOYFAIL anyway | 17:08 |
| *** achanda has quit IRC | 17:08 | |
| *** e0ne has quit IRC | 17:08 | |
| *** ukalifon1 has quit IRC | 17:09 | |
| *** hemna_ has quit IRC | 17:09 | |
| *** Marga_ has quit IRC | 17:09 | |
| devananda | lucasagomes: i think that's better. comment posted | 17:09 |
| lucasagomes | devananda, ack | 17:10 |
| *** Marga_ has joined #openstack-ironic | 17:10 | |
| devananda | time to move coffee shops -- bbi15min | 17:10 |
| lucasagomes | devananda, thanks for looking into it, yeah good discussion. I will see if I can get the timeout soon too | 17:10 |
| devananda | lucasagomes: cheers | 17:10 |
| lucasagomes | devananda, enjoy | 17:11 |
| lucasagomes | I will call it a day as well | 17:11 |
| lucasagomes | have a good night everyone! | 17:11 |
| *** lucasagomes is now known as lucas-dinner | 17:12 | |
| bcornec | TheJulia: well, I think bifrost is doing the right stuff by using sed to modify /etc/hosts. THe problem I think is that /etc/hosts is generated by docker, so I think that's why we get the error stderr: sed: cannot rename /etc/sedyP8Egx: Device or resource busy. So I'd need to find a workaround to that.... | 17:12 |
| *** hblixt has joined #openstack-ironic | 17:14 | |
| *** blixt_ has joined #openstack-ironic | 17:14 | |
| *** ijw_ has joined #openstack-ironic | 17:15 | |
| TheJulia | bcornec: i guess people who have run rabbit in docker might know how to get it to run, and if we can detect it is in a container, or even just use a setting, it would be easy to support | 17:15 |
| *** hblixt has quit IRC | 17:18 | |
| *** romcheg has joined #openstack-ironic | 17:19 | |
| NobodyCam | night lucas-dinner | 17:19 |
| *** sambetts has quit IRC | 17:21 | |
| *** sambetts has joined #openstack-ironic | 17:24 | |
| *** achanda has joined #openstack-ironic | 17:25 | |
| *** bradjones has quit IRC | 17:26 | |
| *** bradjones has joined #openstack-ironic | 17:28 | |
| *** bradjones has quit IRC | 17:28 | |
| *** bradjones has joined #openstack-ironic | 17:28 | |
| *** blixt__ has joined #openstack-ironic | 17:33 | |
| *** ukalifon1 has joined #openstack-ironic | 17:34 | |
| *** blixt_ has quit IRC | 17:36 | |
| *** blixt__ has quit IRC | 17:37 | |
| *** sambetts has quit IRC | 17:38 | |
| *** sambetts has joined #openstack-ironic | 17:40 | |
| *** Marga_ has quit IRC | 17:42 | |
| *** dguerri is now known as dguerri` | 17:43 | |
| openstackgerrit | Chris Krelle proposed openstack/bifrost: Fix white space in readme file. https://review.openstack.org/195687 | 17:52 |
| NobodyCam | TheJulia: ^^^^^ | 17:52 |
| NobodyCam | super minor | 17:52 |
| TheJulia | Thank you! | 17:53 |
| *** boris-42 has joined #openstack-ironic | 17:54 | |
| rloo | NobodyCam: my first bifrost review :) | 17:54 |
| *** Marga_ has joined #openstack-ironic | 17:55 | |
| *** mitchjameson has joined #openstack-ironic | 17:58 | |
| *** hblixt has joined #openstack-ironic | 18:00 | |
| NobodyCam | rloo: w00t!!! | 18:01 |
| NobodyCam | biFrost needs more reviewers :) | 18:01 |
| *** husanu8 has joined #openstack-ironic | 18:02 | |
| rloo | TheJulia: you can +A that (we are good with only one +2 for typos in non-code) | 18:02 |
| *** dontalton has joined #openstack-ironic | 18:02 | |
| TheJulia | Merci | 18:02 |
| *** blixt_ has joined #openstack-ironic | 18:02 | |
| *** Sukhdev has quit IRC | 18:03 | |
| *** hblixt has quit IRC | 18:05 | |
| *** Sukhdev has joined #openstack-ironic | 18:07 | |
| *** husanu8 has quit IRC | 18:14 | |
| bcornec | TheJulia: I think I found why I have that issue: https://github.com/travis-ci/travis-build/pull/403 | 18:18 |
| *** lsmola has quit IRC | 18:19 | |
| rloo | TheJulia: I updated our wiki with that one +2 thing: https://wiki.openstack.org/wiki/Ironic/CoreTeam#Other_notes | 18:19 |
| rloo | others -- let me know if I misunderstood | 18:19 |
| NobodyCam | rloo: great update Thank you :) | 18:21 |
| rloo | NobodyCam: yw. I need to remember to keep our docs updated. | 18:21 |
| NobodyCam | rloo: we should remember to do that :) *BLUSH* | 18:22 |
| NobodyCam | s/we/we all/ | 18:22 |
| rloo | NobodyCam: no time for everything and my memory is so bad ;) | 18:23 |
| rloo | NobodyCam: yeah, all of us! | 18:23 |
| NobodyCam | +++ | 18:23 |
| jlvillal | BadCub: Do you have an approximate location for the sprint in August in Seattle? For trying to figure out lodging. | 18:23 |
| NobodyCam | 701 pike street | 18:23 |
| jlvillal | NobodyCam: Thanks :) | 18:24 |
| TheJulia | bcornec: looking | 18:24 |
| rloo | BadCub: were you going to look into getting some discount rate? (or maybe it was some other project) | 18:24 |
| BadCub | forgot to put that in the email :( | 18:24 |
| BadCub | rloo I can check into that, but most hotels want a minimum booking commit for group rates. | 18:25 |
| NobodyCam | I think we have like the most of if not the entire 24th floor | 18:25 |
| *** trown|lunch is now known as trown | 18:25 | |
| rloo | BadCub: oh. Don't you have an hp rate that folks can use ;) (I'm not going, just asking, so feel free to ignore me) | 18:26 |
| TheJulia | bcornec: I guess... in a sense, that does make sense. Just not a fan of using pipe to overwrite instead of in-place editing. :( | 18:26 |
| *** stegranet has joined #openstack-ironic | 18:27 | |
| NobodyCam | ahh :( | 18:27 |
| BadCub | rloo: HP dioes have employee rates at some hotel chains, but it is an employee rate that requires an emp. badge at check-in | 18:27 |
| BadCub | and usually it is actually higher than BAR | 18:27 |
| rloo | BadCub: oh, that is disappointing. Well, I suspect most people can expense the hotels so it doesn't really matter. | 18:28 |
| BadCub | rloo: indeed. Anyone coming to a summit should be expensing travel | 18:29 |
| *** jhova has joined #openstack-ironic | 18:29 | |
| rloo | BadCub: I believe there are OpenStack contributors that don't work for any company, but I don't think we have any such folks in ironicland. | 18:30 |
| devananda | * 21st floor | 18:31 |
| BadCub | rloo: I believe that to be the case as well. If there are any, I might suggest they call the Sheraton and see if there is a "HP Visitor" rate code. | 18:31 |
| *** blixt__ has joined #openstack-ironic | 18:31 | |
| *** krtaylor has quit IRC | 18:31 | |
| *** blixt_ has quit IRC | 18:35 | |
| *** hblixt has joined #openstack-ironic | 18:38 | |
| devananda | JayF: take a look at gov patch 191892 when you have a sec -- it's proposing the tag:service label to IPA, which I htink is misinformed | 18:38 |
| devananda | JayF: I will comment as such if you agree | 18:38 |
| *** blixt__ has quit IRC | 18:40 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update cached images based on update time https://review.openstack.org/195017 | 18:40 |
| JayF | devananda: I'm in agreement | 18:41 |
| devananda | ack | 18:41 |
| JayF | devananda: definately is not a long running REST API | 18:41 |
| devananda | comment posted | 18:43 |
| *** dontalton has quit IRC | 18:44 | |
| jlvillal | BadCub: Darn, 6000 attendees in town for the ASA 2015 Joint Statistical Meetings http://www.wscc.com/upcoming-events?field_event_start_date_value_1[value][month]=8&field_event_start_date_value_1[value][year]=2015 | 18:46 |
| BadCub | WOW | 18:46 |
| * jlvillal is not sure what the "ASA 2015 Joint Statistical Meetings" is ... | 18:46 | |
| * BadCub is not either | 18:46 | |
| jlvillal | BadCub: Statisticians it seems | 18:47 |
| BadCub | jlvillal: lol. it appears so | 18:47 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update cached images based on update time https://review.openstack.org/195017 | 18:52 |
| *** rbrooker has joined #openstack-ironic | 18:52 | |
| *** r-daneel has quit IRC | 18:54 | |
| *** r-daneel has joined #openstack-ironic | 18:55 | |
| *** dguerri` is now known as dguerri | 18:55 | |
| *** zz_natorious is now known as natorious | 18:58 | |
| bcornec | TheJulia: I have a workaround. I can add a line to the /etc/hosts file before running bifrost, thus avoiding your modification, thanks to your test, and then I sould go further... | 19:05 |
| *** kozhukalov_ has quit IRC | 19:07 | |
| *** gzhai1 has quit IRC | 19:08 | |
| *** gzhai1 has joined #openstack-ironic | 19:09 | |
| TheJulia | bcornec: good to hear :) | 19:09 |
| *** achanda has quit IRC | 19:11 | |
| *** dprince has quit IRC | 19:11 | |
| *** ukalifon1 has quit IRC | 19:15 | |
| *** romainh has joined #openstack-ironic | 19:18 | |
| harlowja | dtantsur|afk https://pypi.python.org/pypi/futurist | 19:24 |
| harlowja | yaaaa | 19:24 |
| bcornec | TheJulia: I'm now failing at DB user creation. Will have to look at that tomorrow. My repo is up to date if you want to try... | 19:28 |
| TheJulia | bcornec: I don't think I'll have time today, but I'll try to take a look in the morning. Have a good evening! | 19:28 |
| harlowja | dtantsur|afk so that should enable ironic starting to use it (along with https://review.openstack.org/#/c/195711/ ) | 19:29 |
| harlowja | btw, dtantsur|afk did u want to be core on that library (not sure i saw a respones to that a while ago) | 19:29 |
| NobodyCam | good (UGT) morning harlowja | 19:31 |
| harlowja | yo yo yo | 19:32 |
| NobodyCam | :) | 19:32 |
| harlowja | how goes things | 19:32 |
| NobodyCam | not to bad today :) | 19:33 |
| NobodyCam | harlowja: LOL: Code from the future, delivered to you in the **now** | 19:33 |
| NobodyCam | luv it | 19:33 |
| harlowja | :) | 19:33 |
| * harlowja tries to make descriptions that people will not completly say wtf to, so far most of the ones seem to slide under the rug without people noticing, lol | 19:34 | |
| NobodyCam | :) +++ | 19:34 |
| rloo | devananda: I hope you're ok with my comment, since this is your patch: https://review.openstack.org/#/c/147087/ | 19:34 |
| rloo | devananda: and they might 'plagiarize' your slides since I can't given him permission. | 19:36 |
| devananda | rloo: all my slides are CC-BY-SA. so as long as there is attribution, it's fine | 19:37 |
| devananda | without attribution, it is plagiarism, and I will be upset | 19:37 |
| rloo | devananda: what does 'attribution' mean, if they are adding that to our ironic docs? | 19:37 |
| rloo | devananda: maybe you should just comment in that patch :) | 19:37 |
| devananda | that the commit message and, if possible, the file indicate where it was copied from | 19:37 |
| devananda | i did :) | 19:37 |
| rloo | devananda: thx. always one step ahead of me :) | 19:38 |
| *** natorious is now known as zz_natorious | 19:44 | |
| *** ijw_ has quit IRC | 19:45 | |
| *** zz_natorious is now known as natorious | 20:01 | |
| *** mtanino has joined #openstack-ironic | 20:01 | |
| *** natorious is now known as zz_natorious | 20:02 | |
| *** Sukhdev has quit IRC | 20:07 | |
| *** krtaylor has joined #openstack-ironic | 20:09 | |
| devananda | https://www.eventbrite.com/e/openstack-ironic-sprint-august-2015-tickets-17533862254 | 20:14 |
| *** gabriel-bezerra has quit IRC | 20:17 | |
| BadCub | devananda: I added ^^ to the meeting agenda too | 20:19 |
| devananda | next ironic sprint should be hosted by !hp, !rackspace, but somewhere that is warm in the middle of winter. | 20:19 |
| devananda | who has offices in hawaii? :-D | 20:19 |
| *** gabriel-bezerra has joined #openstack-ironic | 20:20 | |
| NobodyCam | devananda: plam springs is warm this time of year | 20:20 |
| BadCub | ++ on Hawaii! | 20:20 |
| devananda | NobodyCam: there's no good airport there ... | 20:20 |
| NobodyCam | psp | 20:20 |
| BadCub | PSP is an awesome airport | 20:20 |
| NobodyCam | is most awesome airport | 20:20 |
| NobodyCam | https://en.wikipedia.org/wiki/Palm_Springs_International_Airport | 20:22 |
| NobodyCam | :-p | 20:22 |
| devananda | TheJulia: i tried your new dynamic playbook -- node went straight to 'cleaning' and hit the bug so it's stuck now | 20:22 |
| devananda | ohh | 20:24 |
| * NobodyCam eats a ham sandwich | 20:24 | |
| TheJulia | oh... :( | 20:26 |
| TheJulia | devananda: yeah.. cleaning is on by default. perhaps I should update destructions to tell the user they have to explicitly enable it | 20:27 |
| devananda | yea, looks like i'm hitting issues aorund cleaning more than anything else | 20:27 |
| NobodyCam | we call out cleaning here: https://github.com/openstack/bifrost/tree/master/playbooks/roles/ironic-install | 20:28 |
| devananda | once it gets stuck there, it's hard to un-cleaning-it | 20:28 |
| devananda | especially since manageable->available transition now goes through cleaning, too | 20:28 |
| devananda | and cleaning bugs it every time on my nuc | 20:28 |
| JoshNang | :/ | 20:29 |
| TheJulia | :( | 20:30 |
| NobodyCam | devananda: is it a bug in cleaning code or the NUC itself? | 20:31 |
| JoshNang | how does it get stuck? like is it not getting a list of clean_steps? not starting? | 20:31 |
| devananda | it's at least two things contributing to it | 20:32 |
| devananda | my "amttool" driver is pretty buggy -- but at least it doesn't block the GIL | 20:32 |
| devananda | so htings like "reboot" sometimes fail, because AMT is not responsive at times | 20:32 |
| devananda | but once it enters CLEANING state, it doesn't ever exit it (without me updating the DB) | 20:33 |
| NobodyCam | hummm | 20:34 |
| JoshNang | hrm. do you have any logs? I'll see if anything stands out. like Jay said, we hit a major bug with it, but it turned out to be something with our decom->cleaning transition | 20:34 |
| devananda | think i just found it | 20:35 |
| devananda | firewall | 20:35 |
| NobodyCam | ahh that could do it | 20:35 |
| JoshNang | ah | 20:35 |
| *** ijw_ has joined #openstack-ironic | 20:36 | |
| devananda | maaaybe | 20:37 |
| *** ijw_ has quit IRC | 20:41 | |
| lucas-dinner | devananda, for the next sprint there's a red hat office in Rio de Janeiro :-) | 20:50 |
| lucas-dinner | that would be great! | 20:50 |
| NobodyCam | lucas-dinner: +++ | 20:52 |
| NobodyCam | for Rio | 20:52 |
| devananda | TheJulia: I'm getting this: vagrant-ubuntu-trusty-64 dnsmasq-dhcp[1155]: no address range available for DHCP request via eth2 | 20:53 |
| devananda | TheJulia: even though I've modified groupvars/all to reference eth2 instead of virbr0 -- something isn't getting set up correctly | 20:53 |
| lucas-dinner | NobodyCam, (-: o/ | 20:53 |
| devananda | lucas-dinner: rio! yes! | 20:53 |
| lucas-dinner | I would love to, if we had a quorum I may be able to manage that | 20:54 |
| lucas-dinner | it will definitely be warm | 20:54 |
| devananda | travel will be the hardest thing -- if we dont get enough US/EU companie sending ppl, it won't really happen | 20:55 |
| lucas-dinner | yeah, we should put up to a vote just to see how it turns out | 20:55 |
| lucas-dinner | who knows | 20:55 |
| devananda | yah | 20:56 |
| * lucas-dinner brb | 20:57 | |
| *** ijw_ has joined #openstack-ironic | 21:02 | |
| * NobodyCam jumps on a call | 21:02 | |
| *** e0ne has joined #openstack-ironic | 21:05 | |
| *** zz_natorious is now known as natorious | 21:24 | |
| * trown wants to visit Rio | 21:28 | |
| *** dguerri is now known as dguerri` | 21:28 | |
| * BadCub votes for Rio | 21:38 | |
| *** ijw_ has quit IRC | 21:38 | |
| *** mtanino has quit IRC | 21:38 | |
| *** ijw_ has joined #openstack-ironic | 21:38 | |
| mrda | Morning Ironic | 21:38 |
| NobodyCam | morning mrda | 21:39 |
| hblixt | :) | 21:39 |
| mrda | o/ | 21:39 |
| *** e0ne is now known as e0ne_ | 21:39 | |
| NobodyCam | I see | 21:40 |
| hblixt | too easy | 21:40 |
| *** e0ne_ is now known as e0ne | 21:41 | |
| openstackgerrit | Merged openstack/bifrost: Fix white space in readme file. https://review.openstack.org/195687 | 21:43 |
| *** romcheg has quit IRC | 21:46 | |
| *** romcheg has joined #openstack-ironic | 21:46 | |
| *** romcheg has quit IRC | 21:47 | |
| *** krtaylor has quit IRC | 21:47 | |
| *** romainh has quit IRC | 21:47 | |
| *** romcheg has joined #openstack-ironic | 21:47 | |
| *** trown is now known as trown|outttypeww | 21:47 | |
| *** romcheg has quit IRC | 21:48 | |
| *** Pradip_ has joined #openstack-ironic | 21:55 | |
| *** Marga_ has quit IRC | 21:55 | |
| SpamapS | Hm | 21:56 |
| SpamapS | would it be an accurate statement that approximately 90% of all ironic-conductor's will want ipmitool? | 21:56 |
| SpamapS | the current packaging in Ubuntu does not even recommend ipmitool | 21:57 |
| Pradip_ | Hi, I installed ironic with devstack. But whenever I create a new instance, it doesn't go to the active state. also wen I delete the instance it hangs in that stage. Any ideas? | 21:58 |
| NobodyCam | SpamapS: depending on hardware... but ya i could go with *most* | 21:58 |
| SpamapS | NobodyCam: recommends says most | 21:58 |
| SpamapS | NobodyCam: and suggests is "a lot" | 21:59 |
| NobodyCam | lol | 21:59 |
| NobodyCam | Pradip_: hummm... have you checked the compute and conductor logs? | 21:59 |
| Pradip_ | NobodyCam: No | 22:00 |
| Pradip_ | I will do that. Also once I install the ironic, I can't create normal VM instance. It is saying that no host founf. Is it supposed to happen. i am fairly new in this thing. So I am really struggling | 22:01 |
| openstackgerrit | Stephanie Miller proposed openstack/bifrost: Disambiguate the roles of ci_testing, ci_testing_zuul etc https://review.openstack.org/195759 | 22:02 |
| *** jhova has quit IRC | 22:03 | |
| TheJulia | Pradip_: Sounds like nova-scheduler is another log you need to check | 22:09 |
| Pradip_ | TheJulia: Ok thanks | 22:11 |
| TheJulia | Pradip_: for deployment on baremetal though, you can directly attempt deployment via the ironic command line if you have enough details | 22:12 |
| TheJulia | i.e. instance_info defined | 22:12 |
| Pradip_ | i am doing that. But its not working. It is going upto open_callback stae | 22:15 |
| Pradip_ | and getting stuck there. Once I issue the delete command it get stuck in deleting mode. i tried waiting for 15-20 min in both cases | 22:15 |
| TheJulia | is that the state nova is reporting? | 22:16 |
| Pradip_ | nova is saying "spawning" and No state | 22:18 |
| *** hblixt has quit IRC | 22:18 | |
| Pradip_ | also when I try to add a normal VM, it is saying that no host found. Is it standard? | 22:19 |
| TheJulia | No, its not. That really sounds like the scheduler cannot find a host | 22:20 |
| Pradip_ | ohh. Then I don't know what to do? I will look into the logs for more details | 22:21 |
| TheJulia | Could you have applied the nova-compute config for using ironic to all of your hypervisor nodes? | 22:22 |
| TheJulia | the configurations should be different | 22:22 |
| TheJulia | i.e. ironic requires a separate nova-compute process to allow for scheduling on to baremetal nodes. | 22:23 |
| Pradip_ | ah. I guess that may be the reason. I have only one host at present, I am following the instruction in http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html for devstack to create 3 VMs to pose as bare metal nodes | 22:25 |
| TheJulia | Guys, why does http://docs.openstack.org/developer/ironic/deploy/install-guide.html#configure-compute-service-to-use-the-bare-metal-service not seem right to me ? Or have I lost my mind today? | 22:25 |
| TheJulia | OH | 22:25 |
| *** e0ne has quit IRC | 22:26 | |
| Pradip_ | I am trying to add another host. But before that I wanted to be familiar with the usage concepts. But I am stuck here for long now | 22:27 |
| TheJulia | okay | 22:28 |
| TheJulia | so have you defined baremetal nodes in ironic? | 22:28 |
| *** krtaylor has joined #openstack-ironic | 22:30 | |
| Pradip_ | I followed the instruction on the link to create 3 VM acting as bare metal node. I can see the loist too with ironic command. But when I am trying to provision them, the problem occcurs | 22:30 |
| Pradip_ | on that settings, I can;t add normal Vm too | 22:31 |
| *** dontalton has joined #openstack-ironic | 22:32 | |
| TheJulia | I understand that, you won't be able to launch VMs unless you configure multiple nova-compute processes. I guess the big question is what is the conductor log saying when your attempting to deploy on to a baremetal node | 22:35 |
| Pradip_ | INFO ironic.conductor.utils [-] Succesfully set node 28141b2a-988e-4f7a-bd3f-1934a4ced10a power state to power on. 2015-06-25 14:42:52.841 DEBUG ironic.conductor.task_manager [-] Attempting to reserve node 1 from (pid=25430) reserve_node /opt/stack/ironic/ironic/conductor/task_manager.py:179 | 22:36 |
| Pradip_ | and nova log: 2015-06-25 14:45:15.253 DEBUG nova.virt.ironic.driver [-] [instance: 26c17dd5-c4d7-490b-b4b1-d3f58f324602] Still waiting for ironic node 28141b2a-988e-4f7a-bd3f-1934a4ced10a to become ACTIVE: power_state="power on", target_power_state=None, provision_state="wait call-back", target_provision_state="deploy complete" from (pid=23843) _log_ironic_polling /opt/stack/nova/nova/virt/ironic/driver.py:150 | 22:37 |
| TheJulia | can you share with us the ironic node-show output for one of those nodes? | 22:37 |
| Pradip_ | instance_uuid | 26c17dd5-c4d7-490b-b4b1-d3f58f324602 | | target_power_state | None | | properties | {u'memory_mb': u'1024', u'cpu_arch': u'x86_64', u'local_gb': u'10', | | | u'cpus': u'1'} | | maintenance | 22:38 |
| TheJulia | please use paste.openstack.org :) | 22:38 |
| Pradip_ | sorry :) | 22:39 |
| TheJulia | No problem :) | 22:39 |
| Pradip_ | Paste #321203 | 22:40 |
| Pradip_ | can I share it here? | 22:41 |
| TheJulia | I just pulled it up | 22:41 |
| Pradip_ | ok | 22:42 |
| TheJulia | can you pull up the console and see whats happening to that VM? Basically it needs to boot and check-in after booting a deployment image | 22:42 |
| Pradip_ | ah yeah. it is saying no bootable device | 22:45 |
| TheJulia | another possibility is networking may not be right between the VM and your DHCP and TFTP services | 22:45 |
| Pradip_ | yeah I can see it | 22:45 |
| TheJulia | err, may not be | 22:45 |
| TheJulia | try rebooting the VM and see if it attempts to network boot at all | 22:45 |
| Pradip_ | it is also saying no more network device, connection timed out | 22:45 |
| *** dontalton has quit IRC | 22:45 | |
| *** dhellmann has quit IRC | 22:46 | |
| TheJulia | It would seem, that you likely are missing something with the network plumbing between your test VM and your dhcp/tftp services. Anything in your dhcp service log? | 22:46 |
| *** dhellmann has joined #openstack-ironic | 22:46 | |
| TheJulia | for the mac address of that test VM? | 22:46 |
| Pradip_ | same output. I guess It is not able to connect to the server for the pxelinux | 22:47 |
| TheJulia | That is what it seems | 22:47 |
| Pradip_ | yeah . I think so too | 22:48 |
| jlvillal | Pradip_: Just a question. How did you create your VM for devstack? I'm assuming you are running devstack in a VM | 22:48 |
| *** lucas-dinner has quit IRC | 22:48 | |
| Pradip_ | dhcp l;og is repeating same line over and over | 22:48 |
| *** ijw__ has joined #openstack-ironic | 22:49 | |
| Pradip_ | no i am usning devstack in real machine | 22:49 |
| jlvillal | Pradip_: How did you create your VM for devstack? I'm assuming you are running devstack in a VM. | 22:50 |
| jlvillal | Pradip_: I'm curious because I ran into similar sounding situation. | 22:50 |
| Pradip_ | jlvillal: I am running it on a real machine and followed the instruction in http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html | 22:52 |
| jlvillal | Pradip_: Oh. Okay. Thanks. | 22:52 |
| *** ijw_ has quit IRC | 22:52 | |
| jlvillal | Pradip_: Okay. Different from my case. Thanks. | 22:52 |
| Pradip_ | TheJulia: Thanks a lot for your help, I will check the networking issue | 22:53 |
| TheJulia | Pradip_: Your welcome, we're here to help :) | 22:59 |
| *** natorious is now known as zz_natorious | 22:59 | |
| *** stegranet has quit IRC | 23:02 | |
| *** ijw_ has joined #openstack-ironic | 23:10 | |
| *** ijw__ has quit IRC | 23:13 | |
| *** zhenguo has joined #openstack-ironic | 23:18 | |
| *** thrash is now known as thrash|g0ne | 23:21 | |
| *** Sukhdev has joined #openstack-ironic | 23:21 | |
| *** rbrooker has quit IRC | 23:22 | |
| SpamapS | $ ironic driver-list | 23:25 |
| SpamapS | 'NoneType' object has no attribute 'rfind' | 23:25 |
| SpamapS | anybody seen that? | 23:25 |
| NobodyCam | SpamapS: I have not seen that one | 23:26 |
| SpamapS | hrm | 23:27 |
| SpamapS | | publicURL | http:infra-cloud-baremetal0.localdomain:6385/ | | 23:28 |
| SpamapS | typeo | 23:28 |
| devananda | :) | 23:28 |
| devananda | SpamapS: how's it going, aside fromtypos? | 23:28 |
| SpamapS | not a useful error message tho ;) | 23:28 |
| SpamapS | devananda: so bifrost was an interesting challenge to get working with these crazy mellanox cards | 23:28 |
| SpamapS | devananda: they have a weird version of iPXE on them which gave TheJulia fits. ;) | 23:29 |
| devananda | SpamapS: haha | 23:29 |
| SpamapS | devananda: but once we worked around that, it's been mostly puppet wrangling. | 23:29 |
| *** absubram has quit IRC | 23:29 | |
| devananda | SpamapS: that's good! | 23:29 |
| SpamapS | | pxe_ipmitool | infra-cloud-baremetal0.localdomain | | 23:29 |
| SpamapS | there we go | 23:29 |
| *** ijw__ has joined #openstack-ironic | 23:30 | |
| devananda | TheJulia: dhcp issue aside, the deploy-all playbook seems to do the right thing | 23:31 |
| *** yuanying has joined #openstack-ironic | 23:32 | |
| *** ijw_ has quit IRC | 23:33 | |
| devananda | TheJulia: it's a bit odd that if it finds no nodes "available" it runs all the way through, but skips every step | 23:33 |
| *** UForgotten has joined #openstack-ironic | 23:34 | |
| TheJulia | devananda: awesome, sorry, didn't get a chance to even think about the dhcp issue | 23:42 |
| devananda | TheJulia: no worries. it's probably something in my local env | 23:42 |
| TheJulia | Yeah, that could be an issue. Wonder if I can make it fail easilly if there are no nodes available | 23:42 |
| devananda | JoshNang: I have a new theory about the clenaing failures -- I think this device requires passing the boot device every time | 23:43 |
| TheJulia | s/issue/source of confusion/ | 23:43 |
| JoshNang | devananda: ohhhh | 23:43 |
| devananda | JoshNang: yah. agent.py 228 manager_utils.node_set_boot_device(task, boot_devices.PXE, | 23:43 |
| devananda | JoshNang: nope - nvm. that's called from preprae_cleanning too | 23:44 |
| JoshNang | :/ | 23:45 |
| devananda | JoshNang: so there's a thing missing from our state machine -- "clean wait" | 23:45 |
| JoshNang | this is bifrost and agent-amttool? | 23:45 |
| TheJulia | pxe_amt | 23:45 |
| devananda | JoshNang: that time when a node isn't online yet, and the driver knows it should be cleaning, but it's waiting for it to finish POST | 23:45 |
| TheJulia | there is no agent_amt | 23:45 |
| JoshNang | devananda: yeah that would be a lot cleaner | 23:45 |
| JoshNang | TheJulia: gotcha, thanks! | 23:45 |
| devananda | it's my hacky agent-amttool | 23:46 |
| devananda | TheJulia: ^ yes there is. but not in trunk | 23:46 |
| TheJulia | heh, figures :) | 23:46 |
| devananda | TheJulia: https://review.openstack.org/#/c/193767/1 fyi | 23:47 |
| devananda | it exists because of https://bugs.launchpad.net/ironic/+bug/1454492 | 23:48 |
| openstack | Launchpad bug 1454492 in Ironic "openwsman library is not greenthread-aware, leads to driver lockup with both DRAC and AMT drivers" [Medium,Triaged] - Assigned to Imre Farkas (ifarkas) | 23:48 |
| TheJulia | devananda: ahhhh! | 23:51 |
| *** ijw_ has joined #openstack-ironic | 23:51 | |
| *** ijw__ has quit IRC | 23:53 | |
| *** Sukhdev has quit IRC | 23:56 | |
| *** smoriya has joined #openstack-ironic | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
