*** akhilaki has quit IRC | 00:12 | |
*** baoli has joined #openstack-ironic | 00:18 | |
*** baoli has quit IRC | 00:23 | |
*** hemna_ has quit IRC | 00:29 | |
*** rloo has quit IRC | 00:57 | |
*** gyee has quit IRC | 01:00 | |
*** liuzz has joined #openstack-ironic | 01:00 | |
*** zhangfei has joined #openstack-ironic | 01:09 | |
*** tiendc has joined #openstack-ironic | 01:12 | |
*** cjloader has joined #openstack-ironic | 01:36 | |
*** trungnv has quit IRC | 01:48 | |
*** trungnv has joined #openstack-ironic | 01:49 | |
*** etingof has quit IRC | 01:57 | |
*** cjloader has quit IRC | 02:07 | |
*** etingof has joined #openstack-ironic | 02:10 | |
*** fragatina has quit IRC | 02:19 | |
*** fragatina has joined #openstack-ironic | 02:19 | |
*** fragatina has quit IRC | 02:21 | |
*** cjloader has joined #openstack-ironic | 02:22 | |
*** cjloader has quit IRC | 03:09 | |
*** fragatina has joined #openstack-ironic | 03:30 | |
*** rh-jelabarre has quit IRC | 03:56 | |
*** hshiina|afk is now known as hshiina | 03:57 | |
openstackgerrit | zenghui.shi proposed openstack/python-ironicclient master: BIOS Settings support https://review.openstack.org/569318 | 04:01 |
---|---|---|
*** rcernin has quit IRC | 04:03 | |
*** rcernin has joined #openstack-ironic | 04:03 | |
*** gyan_ has joined #openstack-ironic | 04:12 | |
*** baoli has joined #openstack-ironic | 04:12 | |
*** baoli has quit IRC | 04:17 | |
*** hshiina2 has joined #openstack-ironic | 04:41 | |
*** hshiina has quit IRC | 04:41 | |
*** hshiina3 has joined #openstack-ironic | 04:43 | |
*** hshiina2 has quit IRC | 04:46 | |
*** early has quit IRC | 05:04 | |
*** marios has joined #openstack-ironic | 05:13 | |
*** early has joined #openstack-ironic | 05:15 | |
*** tiendc has quit IRC | 05:37 | |
*** tiendc has joined #openstack-ironic | 05:38 | |
*** mjura has joined #openstack-ironic | 05:53 | |
*** adreznec has quit IRC | 05:56 | |
*** tiendc has quit IRC | 05:59 | |
*** trungnv has quit IRC | 05:59 | |
*** tiendc has joined #openstack-ironic | 06:00 | |
*** trungnv has joined #openstack-ironic | 06:00 | |
*** masuberu has joined #openstack-ironic | 06:01 | |
*** adreznec has joined #openstack-ironic | 06:02 | |
*** masber has quit IRC | 06:04 | |
*** hshiina3 is now known as hshiina | 06:05 | |
*** masuberu has quit IRC | 06:06 | |
*** kzaitsev_pi has joined #openstack-ironic | 06:09 | |
openstackgerrit | Merged openstack/ironic master: Change exception msg of BIOS caching https://review.openstack.org/568193 | 06:25 |
*** zhangfei has quit IRC | 06:32 | |
*** zhangfei has joined #openstack-ironic | 06:34 | |
*** zhangfei has quit IRC | 06:39 | |
openstackgerrit | zenghui.shi proposed openstack/python-ironicclient master: BIOS Settings support https://review.openstack.org/569318 | 06:40 |
*** zhangfei has joined #openstack-ironic | 06:46 | |
*** rcernin has quit IRC | 07:01 | |
*** namnh has joined #openstack-ironic | 07:06 | |
*** e0ne has joined #openstack-ironic | 07:09 | |
*** tesseract has joined #openstack-ironic | 07:14 | |
*** namnh has quit IRC | 07:15 | |
*** threestrands has quit IRC | 07:17 | |
*** namnh has joined #openstack-ironic | 07:19 | |
*** jtomasek has joined #openstack-ironic | 07:19 | |
*** ajya has joined #openstack-ironic | 07:22 | |
*** toabctl has quit IRC | 07:26 | |
olivierb- | morning ironicers and thanks for the review TheJulia | 07:28 |
olivierb- | did not reconnect yesterday evening for once | 07:28 |
*** AlexeyAbashkin has joined #openstack-ironic | 07:46 | |
*** liuzz_ has joined #openstack-ironic | 07:46 | |
*** liuzz has quit IRC | 07:46 | |
openstackgerrit | Nguyen Van Trung proposed openstack/ironic master: Support raid configuration for BM via irmc driver https://review.openstack.org/512979 | 07:47 |
*** zhangfei has quit IRC | 07:51 | |
*** hshiina has quit IRC | 07:51 | |
*** namnh has quit IRC | 08:00 | |
*** namnh has joined #openstack-ironic | 08:00 | |
*** zhangfei has joined #openstack-ironic | 08:03 | |
*** milan has joined #openstack-ironic | 08:04 | |
*** Chandra has quit IRC | 08:14 | |
*** mgoddard has joined #openstack-ironic | 08:17 | |
etingof | good Friday morning everyone 0/ | 08:18 |
*** dougsz has joined #openstack-ironic | 08:18 | |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: apply fault https://review.openstack.org/556758 | 08:19 |
openstackgerrit | yolanda.robla proposed openstack/ironic-tempest-plugin master: Add manual clean step ironic standalone test https://review.openstack.org/568616 | 08:19 |
*** pmannidi has quit IRC | 08:24 | |
vdrok | good morning etingof ! | 08:28 |
*** arxcruz has quit IRC | 08:30 | |
*** lucas-afk is now known as lucasagomes | 08:32 | |
*** zhangfei has quit IRC | 08:35 | |
*** MattMan has quit IRC | 08:42 | |
*** MattMan has joined #openstack-ironic | 08:42 | |
*** derekh has joined #openstack-ironic | 08:44 | |
*** zhangfei has joined #openstack-ironic | 08:51 | |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: API implementation https://review.openstack.org/556015 | 08:56 |
openstackgerrit | Vladyslav Drok proposed openstack/python-ironicclient master: Allow to use none auth in functional tests https://review.openstack.org/508330 | 08:57 |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: Notification objects https://review.openstack.org/558152 | 09:15 |
*** sambetts|afk is now known as sambetts | 09:16 | |
openstackgerrit | Sam Betts proposed openstack/ironic master: Stop double json encoding the error message responses https://review.openstack.org/567837 | 09:16 |
dtantsur | morning ironic | 09:20 |
sambetts | o/ dtantsur, I've got a minute this morning so I'm going to update that port bug patch if your not already doing it | 09:21 |
dtantsur | sambetts: great! no, I'm not, just woke up :) | 09:21 |
*** arxcruz has joined #openstack-ironic | 09:23 | |
sambetts | dtantsur: you couldn't give me any insight on why this is happening on centos 7 could you http://192.133.156.18/37/567837/5/check/ironic-dsvm-tempest-ironic-cimc-current-centos-7-cimc/f2a7d20/logs/devstacklog.txt.gz#_2018-05-18_09_24_12_329 | 09:26 |
etingof | dtantsur, you woke up and it's Friday \o/ | 09:27 |
*** zhangfei has quit IRC | 09:27 | |
dtantsur | sambetts: not really.. it may be some temporary situation with EPEL. also note that yolanda_ experienced big problems with vbmc on centos just yesterday | 09:28 |
sambetts | dtantsur: I thought that it might have been temporary but it seems to have been stuck like this for about a week, feels weird there would be a dependency issue in the yum rpos | 09:29 |
sambetts | repos | 09:29 |
dtantsur | sambetts: ask on #infra? there are some Red Hatters interested in it IIRC | 09:30 |
*** serlex has joined #openstack-ironic | 09:33 | |
sambetts | dtantsur: hmmm, seems like it might be related to my CI images not getting updated correctly | 09:38 |
openstackgerrit | Ilya Etingof proposed openstack/ironic master: Adds boot mode support to ManagementInterface https://review.openstack.org/526773 | 09:40 |
dtantsur | pas-ha, vdrok, any objections to releasing ironic-staging-drivers stable/queens? | 09:41 |
*** zhangfei has joined #openstack-ironic | 09:41 | |
*** athomas has quit IRC | 09:43 | |
vdrok | dtantsur: seems like we had only a couple of fixes in ovirt since last tag, but no objections | 09:47 |
vdrok | also morning | 09:47 |
dtantsur | vdrok: morning! yep, I want the ovirt fixes released. | 09:48 |
openstackgerrit | Kaifeng Wang proposed openstack/python-ironicclient master: Power fault recovery: client support https://review.openstack.org/556774 | 09:49 |
pas-ha | dtantsur: nope | 09:53 |
dtantsur | good, releasing | 09:53 |
pas-ha | ship it :-) | 09:53 |
*** mjura has quit IRC | 09:54 | |
dtantsur | I'm also nagging our folks downstream with creating some kind of a CI job for the ovirt driver | 09:54 |
*** namnh has quit IRC | 09:56 | |
*** zhangfei has quit IRC | 09:57 | |
*** zhangfei has joined #openstack-ironic | 09:58 | |
openstackgerrit | Kaifeng Wang proposed openstack/ironic-tempest-plugin master: Power fault recovery: tempest tests https://review.openstack.org/558170 | 10:01 |
*** lbragstad has quit IRC | 10:15 | |
openstackgerrit | Nguyen Hung Phuong proposed openstack/ironic master: [WIP] Add CUSTOM_CPU_FPGA to Traits https://review.openstack.org/569377 | 10:18 |
*** lbragstad has joined #openstack-ironic | 10:19 | |
* milan is away: I'm busy | 10:35 | |
openstackgerrit | yolanda.robla proposed openstack/ironic-tempest-plugin master: Add manual clean step ironic standalone test https://review.openstack.org/568616 | 10:37 |
*** trungnv has quit IRC | 10:53 | |
openstackgerrit | Ilya Etingof proposed openstack/ironic master: Move boot-related code to boot_mode_utils.py https://review.openstack.org/569382 | 10:56 |
*** jcoufal has joined #openstack-ironic | 11:01 | |
*** zhangfei has quit IRC | 11:08 | |
*** zhangfei has joined #openstack-ironic | 11:10 | |
openstackgerrit | yolanda.robla proposed openstack/ironic master: Enable agent_ipmitool and pxe_ipmitool drivers https://review.openstack.org/569388 | 11:12 |
*** ajya has quit IRC | 11:14 | |
*** ajya has joined #openstack-ironic | 11:14 | |
*** tiendc has quit IRC | 11:21 | |
*** jcoufal has quit IRC | 11:35 | |
*** jcoufal has joined #openstack-ironic | 11:35 | |
jroll | morning | 11:35 |
dtantsur | hmm, folks this run time of 2 secs. is suspicious: http://logs.openstack.org/16/568616/11/check/ironic-dsvm-standalone/6486cef/job-output.txt.gz#_2018-05-18_11_14_18_095157 | 11:36 |
dtantsur | morning jroll | 11:36 |
jroll | dtantsur: we turned on nested virt for that job, look how much faster it is! | 11:36 |
dtantsur | though.. maybe it does not cound setUp? | 11:36 |
dtantsur | heh | 11:36 |
jroll | yeah, setUp was my first thought | 11:37 |
dtantsur | I wonder if we should move it out of setUp. this is confusing. what we tests is the whole boot process, not only the fact we can ping something.. | 11:38 |
dtantsur | like, how do I know we're not pinging locahost? :) | 11:40 |
jroll | 2018-05-18 11:14:18.090 10259 DEBUG tempest.lib.common.utils.test_utils [-] Call ping returns true in 2.016478 seconds call_until_true tempest/lib/common/utils/test_utils.py:109 | 11:40 |
jroll | so yes, it does not count setUp, it seems | 11:40 |
yolanda_ | dtantsur, so i'm trying to adapt the patches from Vasyl, for the cleanup | 11:40 |
yolanda_ | part of the original code was https://review.openstack.org/#/c/568616/11/ironic_tempest_plugin/tests/scenario/ironic_standalone/test_cleaning.py | 11:40 |
patchbot | patch 568616 - ironic-tempest-plugin - Add manual clean step ironic standalone test | 11:40 |
yolanda_ | see driver = 'agent_ipmitool' there, that's causing to skip that on master | 11:41 |
jroll | dtantsur: I guess I'm not opposed to moving it, but not sure it's worth the time | 11:41 |
dtantsur | jroll: yeah, same thought.. a low hanging fruit | 11:41 |
yolanda_ | so the question is , if those tests that carry agent_ipmitool on them, should have driver='ipmi' now | 11:41 |
dtantsur | but very low priority | 11:41 |
dtantsur | yolanda_: aha, I see. we need two copies of the same test: one using agent_ipmitool (for stable branches), one with ipmi (for master) | 11:42 |
dtantsur | this is similar to what we do with other tests: we maintain a copy per driver | 11:42 |
* dtantsur gets food, brb | 11:42 | |
yolanda_ | ah ok | 11:43 |
*** rh-jelabarre has joined #openstack-ironic | 11:43 | |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: API implementation https://review.openstack.org/556015 | 11:44 |
sambetts | dtantsur, TheJulia: are the ironic UTs for agent deploy broken? | 11:45 |
sambetts | or is it just my envionrment | 11:45 |
yolanda_ | dtantsur, i understand now, actually there are 3 copies of the test, one per each driver: pxe_ipmitool, agent_ipmitool, ipmi | 11:45 |
TheJulia | sambetts: they were good for me a few days ago, but i diddnt touch code yesterday really | 11:46 |
* TheJulia makes menatal note to fire tox -r off once out of bed. | 11:47 | |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: API implementation https://review.openstack.org/556015 | 11:49 |
dtantsur | sambetts: worked yesterday.. what's the failure? | 11:51 |
dtantsur | yolanda_: yep. it's probably fine to tests only one of the old drivers, they're not different wrt cleaning | 11:52 |
sambetts | dtantsur: I think I've worked it out, the agent deploy driver isn't mocking the calls to the network driver and has the flat network driver loaded resulting in neutronclient being sad | 11:53 |
sambetts | TheJulia: ^ | 11:53 |
sambetts | this is because of my port bug fixes btw | 11:53 |
dtantsur | I see, yeah | 11:53 |
sambetts | so upstream should be good | 11:53 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic stable/pike: Tear down console during unprovisioning https://review.openstack.org/568569 | 11:54 |
sambetts | but the error message coming from the traceback is horrible | 11:54 |
sambetts | AttributeError: '_SentinelObject' object has no attribute 'get_endpoint' | 11:54 |
sambetts | I was like woah.... | 11:54 |
dtantsur | morning TheJulia | 11:54 |
dtantsur | TheJulia: I hope https://review.openstack.org/569375 won't break bifrost (in that case the latter should probably be fixed) | 11:55 |
patchbot | patch 569375 - openstacksdk - baremetal: refuse to inspect associated machines | 11:55 |
* jroll adds a silly comment | 11:57 | |
*** slagle has quit IRC | 11:58 | |
*** slagle has joined #openstack-ironic | 12:00 | |
sambetts | dtantsur: shouldn't that be an API fix?? | 12:00 |
openstackgerrit | yolanda.robla proposed openstack/ironic-tempest-plugin master: Add manual clean step ironic standalone test https://review.openstack.org/568616 | 12:09 |
dtantsur | sambetts: what exactly? | 12:10 |
openstackgerrit | yolanda.robla proposed openstack/ironic-tempest-plugin master: Add manual clean step ironic standalone test https://review.openstack.org/568616 | 12:10 |
sambetts | dtantsur: preventing nodes with a instnace_uuid from going into inspect | 12:10 |
sambetts | dtantsur: can't it just be a check in the set node provisioning state | 12:11 |
*** jaganathan has quit IRC | 12:12 | |
sambetts | actually its probably worth adding that check in for other states too, making it so an available node can only move to deploying if it has a instance_uuid | 12:12 |
TheJulia | dtantsur: lgtm, I think a minor thing to release note and suggestion to indicate the field | 12:12 |
sambetts | making so that you have to remove the instace_uuid to move it to managable | 12:12 |
TheJulia | sambetts: can't, instance_uuid has to be set first as a different op and you may be the one setting the provision state | 12:13 |
TheJulia | well yo ucould with lots of checks that might not make sense there | 12:13 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic stable/ocata: Tear down console during unprovisioning https://review.openstack.org/568570 | 12:15 |
sambetts | TheJulia: not sure why setting instance_uuid first is an issue, I'm saying that we put the check in the top of node_set_provision_state, if node.instnace_uuid != None and new_state != "deploying": raise Exception("Nope") | 12:16 |
jroll | I'd be worried about standalone cases that use the instance_uuid field as ownership | 12:16 |
TheJulia | that would block undeploys, and bifrost users are not required to use instance_uuids | 12:17 |
TheJulia | all the instance uuid is nova's lock really | 12:17 |
TheJulia | or claim, that ties it back to nova | 12:17 |
sambetts | the way it would be implemented shouldn't prevent any state changes if instnace_uuid isn't set | 12:17 |
sambetts | only if its set | 12:18 |
sambetts | if its unset you can do anything | 12:18 |
jroll | I think some people tag instance_uuid for more than a single deploy, though | 12:18 |
jroll | and just leave it set | 12:18 |
sambetts | jroll: then they should use rebuild | 12:18 |
sambetts | not tear down an redeploy | 12:18 |
jroll | sambetts: well, no, as in "I own this machine" in standalone use cases | 12:19 |
sambetts | because we clear the instance_uuid anyway on tear down | 12:19 |
sambetts | along with instance_info | 12:19 |
jroll | ah do we? ok, I'm wrong then :) | 12:19 |
dtantsur | sambetts: this is a good point, but I don't think we take any actions like that based on instance_uuid.. | 12:20 |
dtantsur | except for blocking updating instance_uuid | 12:20 |
sambetts | why not, its a fundemental top level field in our node data strcuture | 12:21 |
TheJulia | we need to fix ownership concepts, I have a spec to start in that path, for what its worth | 12:21 |
TheJulia | which is just a silly field add | 12:21 |
*** EmilienM is now known as EvilienM | 12:21 | |
sambetts | my issue is that I don't think we should be implementing a check like dtantsur's patch client side, because it actually makes a lot of sense, and if its not in the API then anyone using the API directly doesn't get that check | 12:21 |
dtantsur | sambetts: well, changing API involves microversions, etc.. | 12:22 |
dtantsur | so people not the new API microversion won't get the fix without the client patch | 12:22 |
dtantsur | but, well, I don't thing that SDKs should do available->manageable transition automatically at all | 12:22 |
sambetts | I think thats fine, but we should error out if your trying to move a available node with an instance_uuid to manageble because the fact it has an instance uuid means its actually in the process of deploying | 12:24 |
*** trown|outtypewww is now known as trown | 12:24 | |
dtantsur | ++ I'd totally do it | 12:25 |
* dtantsur can file an RFE | 12:25 | |
dtantsur | or does it need a spec? | 12:25 |
sambetts | I'm bad at knowing if things do or not | 12:27 |
TheJulia | dtantsur: bug fix I think | 12:27 |
TheJulia | in other words, just do it | 12:28 |
jroll | we typically require one for anything with a microversion, this doesn't seem like it needs one though | 12:28 |
sambetts | a spec or a microversion ? | 12:28 |
jroll | both? :) | 12:28 |
jroll | we definitely require a microversion for anything with a microversion | 12:28 |
jroll | :P | 12:28 |
dtantsur | a microversion may be required, since it can be seen as a behavior change | 12:29 |
dtantsur | lol | 12:29 |
* jroll now reads sambetts' question correctly | 12:29 | |
sambetts | but it can also be seen as a bug in the old microversions right? and we can fix those without a microversion right? | 12:29 |
sambetts | micro-brain-splode | 12:30 |
dtantsur | hehe | 12:30 |
dtantsur | bug vs feature - an old debate | 12:31 |
TheJulia | microversion + bug is fine I think, spec I think is likely overkill time wise since the specs process is also designed to slow down things | 12:32 |
dtantsur | ++ this is my take on it | 12:33 |
jroll | yes, fine with me | 12:34 |
sambetts | :thumbs-up: | 12:34 |
TheJulia | heh | 12:35 |
* TheJulia goes back to packing | 12:36 | |
TheJulia | sambetts: i did tox -r, no errors like that, but I saw them in some tests I was working on where I wasn't properly capturing glance client stuffs and image verification stuffs with the pxe deploy driver that I put together | 12:37 |
* TheJulia now really goes back to packing | 12:37 | |
*** fragatina has quit IRC | 12:37 | |
sambetts | TheJulia: yeah, it seems like the AgentDeploy tests load more operation interfaces than just the "direct" deploy interface | 12:38 |
TheJulia | Yeah, they have some catches way down in the code :\ | 12:47 |
sambetts | I've switched it to use the noop network interface for now and it seems much happier | 12:48 |
sambetts | only loading the flat network interface in tests that are asking for it | 12:48 |
*** edleafe is now known as figleaf | 12:59 | |
dtantsur | +++++++ | 13:00 |
dtantsur | our tests are mess :( | 13:00 |
*** rloo has joined #openstack-ironic | 13:02 | |
*** jaypipes has joined #openstack-ironic | 13:04 | |
dtantsur | https://storyboard.openstack.org/#!/story/2002064 | 13:08 |
*** dtantsur is now known as dtantsur|brb | 13:11 | |
sambetts | dtantsur|brb: nice! | 13:13 |
*** racedo has joined #openstack-ironic | 13:14 | |
rloo | Good morning everyone! | 13:20 |
jroll | morning rloo \o | 13:22 |
* jroll is stepping away for an hour or two | 13:22 | |
*** rbudden has joined #openstack-ironic | 13:28 | |
*** rajinir has joined #openstack-ironic | 13:46 | |
*** Goneri has joined #openstack-ironic | 13:51 | |
*** ElCoyote_ has joined #openstack-ironic | 13:54 | |
openstackgerrit | Sam Betts proposed openstack/ironic master: Ensure we unbind flat network ports and clear BM mac addresses https://review.openstack.org/539981 | 14:05 |
*** baoli has joined #openstack-ironic | 14:11 | |
*** dtantsur|brb is now known as dtantsur | 14:14 | |
dtantsur | morning rloo | 14:14 |
rloo | dtantsur: :) | 14:14 |
*** toabctl has joined #openstack-ironic | 14:19 | |
*** trown is now known as trown|brb | 14:25 | |
rpioso | Good morning | 14:27 |
*** hoonetorg has quit IRC | 14:28 | |
*** trown|brb is now known as trown | 14:30 | |
openstackgerrit | Merged openstack/ironic-python-agent master: Fix for fatal error when GPT was used and only MBR was cleaned https://review.openstack.org/569086 | 14:30 |
openstackgerrit | Merged openstack/ironic-python-agent master: write byte objects when using os.write https://review.openstack.org/569159 | 14:30 |
*** hemna_ has joined #openstack-ironic | 14:33 | |
*** hoonetorg has joined #openstack-ironic | 14:45 | |
*** serlex has quit IRC | 14:46 | |
*** gyan_ has quit IRC | 14:47 | |
*** gyan_ has joined #openstack-ironic | 14:53 | |
*** cjloader has joined #openstack-ironic | 15:00 | |
* milan is back (gone 04:27:17) | 15:03 | |
*** gyan_ has quit IRC | 15:14 | |
*** jaypipes is now known as leakypipes | 15:17 | |
*** dansmith is now known as superdan | 15:18 | |
*** zhangfei has quit IRC | 15:19 | |
*** e0ne has quit IRC | 15:19 | |
*** jroll is now known as jrollinhatin | 15:23 | |
*** rloo is now known as waterloo | 15:27 | |
sambetts | dtantsur: which IRC do the DIB folks hang out in?? my CI is broken because this isn't released yet https://github.com/openstack/diskimage-builder/commit/abd63b01aac585d12eefcc5506065652a26243bb | 15:30 |
dtantsur | sambetts: #openstack-dib IIRC | 15:30 |
sambetts | dtantsur: thanks | 15:31 |
*** Goneri has quit IRC | 15:33 | |
*** baoli has quit IRC | 15:34 | |
*** baoli has joined #openstack-ironic | 15:35 | |
*** hemna_ has quit IRC | 15:37 | |
*** r-daneel has joined #openstack-ironic | 15:39 | |
*** baoli has quit IRC | 15:39 | |
*** hemna_ has joined #openstack-ironic | 15:41 | |
*** hemna__ has joined #openstack-ironic | 15:45 | |
*** baha has joined #openstack-ironic | 15:45 | |
NobodyCam | Good Morning Ironic'ers... | 15:46 |
NobodyCam | and .. | 15:46 |
*** tesseract has quit IRC | 15:46 | |
NobodyCam | TGIF. | 15:46 |
*** hemna_ has quit IRC | 15:47 | |
dtantsur | hey NobodyCam | 15:48 |
NobodyCam | hey hey dtantsur happy Friday | 15:48 |
NobodyCam | :) | 15:48 |
*** melwitt is now known as jgwentworth | 15:49 | |
NobodyCam | any fun plans for the weekend? | 15:49 |
*** baoli has joined #openstack-ironic | 15:50 | |
*** lbragstad is now known as elbragstad | 15:50 | |
dtantsur | not this time, just chilling out at home | 15:51 |
dtantsur | and you? | 15:51 |
NobodyCam | :) some times thats best | 15:51 |
NobodyCam | I have travel on sunday :( | 15:51 |
dtantsur | summit? | 15:52 |
NobodyCam | no not this time. :( just up to San Jose | 15:53 |
*** lucasagomes is now known as lucas-afk | 15:57 | |
*** baoli has quit IRC | 16:02 | |
*** baoli has joined #openstack-ironic | 16:02 | |
* etingof is going to enjoy the company of the medieval knights; though they might be busy saving a girl, as they figured this is meaning of life ;) | 16:02 | |
*** BernsO has joined #openstack-ironic | 16:03 | |
* dtantsur is curious | 16:03 | |
etingof | expect photos then ;) | 16:03 |
BernsO | Hi Guys, can someone help me with a problem I'm having with one of my ironic nodes? I have enabled automatic cleaning, but some of the nodes enter a clean_failed state after they are destroyed in nova. Is there any way to manually recover them and make them available again? If I just set the provisioning state to available they start cleaning again and I get stuck in a loop | 16:05 |
*** AlexeyAbashkin has quit IRC | 16:06 | |
*** baoli has quit IRC | 16:08 | |
*** baoli has joined #openstack-ironic | 16:08 | |
dtantsur | BernsO: if cleaning fails, you should understand why and fix it. It may be a bad idea to put nodes that failed cleaning back in the fleet. | 16:13 |
dtantsur | our API explicitly disallows the latter indeed | 16:13 |
*** marios has quit IRC | 16:14 | |
TheJulia | etingof: sounds fun! | 16:17 |
BernsO | dtantsur, I understand, but as far as I can see this is a bug | 16:20 |
BernsO | CleaningError', u'details': u'Error performing clean_step erase_devices: No HardwareManager found to handle method: Could not find method: erase_block_device | 16:20 |
dtantsur | BernsO: do you have a customized IPA image? | 16:21 |
BernsO | dtantsur, No default | 16:22 |
TheJulia | BernsO: so one you downloaded from tarballs.o.o? or built yourself? | 16:23 |
*** jaganathan has joined #openstack-ironic | 16:23 | |
*** jaganathan has quit IRC | 16:24 | |
*** jaganathan has joined #openstack-ironic | 16:24 | |
TheJulia | BernsO: What is confusing is us that we ship a GenericHardwareManager that is part of ironic-python-agent, and that that method is part of it.... | 16:25 |
BernsO | TheJulia, you mean the ironic-python-agent? I've made no customisations to the one we use | 16:25 |
BernsO | This problem only happens with a specific make of BMC/server, which is why I'm inclined to think its a bug | 16:26 |
TheJulia | BernsO: Oo | 16:26 |
dtantsur | BernsO: still, where is your IPA image coming from? | 16:26 |
*** e0ne has joined #openstack-ironic | 16:26 | |
* dtantsur suspects a vendor hardware manager that does not support this call | 16:26 | |
TheJulia | ditto | 16:27 |
* sambetts --> home | 16:29 | |
dtantsur | TheJulia: before I forgot: Monday is a public holiday here (again!), so I'll likely miss the meeting. | 16:29 |
dtantsur | c u sambetts | 16:29 |
TheJulia | dtantsur: meeting next week is cancelled | 16:29 |
TheJulia | dtantsur: have a wonderful weekend | 16:29 |
dtantsur | oh, right! even better :) | 16:29 |
sambetts | see those who are going to the summit in canada | 16:29 |
dtantsur | safe travels! | 16:30 |
sambetts | otherwise see y'all in a week | 16:30 |
*** sambetts is now known as sambetts|afk | 16:30 | |
sambetts|afk | dtantsur: I updated that port patch btw with tests and everything so it should be good to go | 16:30 |
dtantsur | great! | 16:30 |
*** hamzy has quit IRC | 16:31 | |
* sambetts|afk really gone now | 16:31 | |
*** e0ne has quit IRC | 16:32 | |
*** e0ne has joined #openstack-ironic | 16:33 | |
BernsO | TheJulia, dtantsur Im checking R.E. the image. Is there a way for me to verify manually whether this manager supports this call? | 16:35 |
*** etingof is now known as etingof|afk | 16:36 | |
TheJulia | BernsO: you could download the ramdisk and extract it | 16:36 |
dtantsur | BernsO: also during cleaning some information is populated in node.driver_internal_info | 16:36 |
dtantsur | hmm, ignore me, not the information you need | 16:36 |
TheJulia | there may be a container.tgz file, or you may find the ironic-python-agent source elsewhere in the image. tl;dr search what you extract for hardware.py | 16:36 |
TheJulia | it should be in there, as a interface class, and a full implementaiton | 16:37 |
BernsO | Ok I will have a look. Just to confirm though while Im getting to the root of this problem is there any way to stop 'cleaning failed' nodes from cleaning again when you set them to available state? or will I have to remove them from ironic and reenroll them? | 16:38 |
TheJulia | BernsO: All you can do is disable the cleaning functionality completely on the conductor, do it, and then change it back. It is not safe to operate without cleaning. | 16:39 |
*** dtantsur is now known as dtantsur|afk | 16:40 | |
dtantsur|afk | have a good weekend | 16:40 |
BernsO | Understood, thank you for your help TheJulia and dtantsur|afk :) | 16:41 |
*** mgoddard has quit IRC | 16:43 | |
*** trown is now known as trown|outtypewww | 16:45 | |
*** akhilaki has joined #openstack-ironic | 16:49 | |
*** dougsz has quit IRC | 16:54 | |
*** derekh has quit IRC | 17:00 | |
*** Goneri has joined #openstack-ironic | 17:01 | |
*** fragatina has joined #openstack-ironic | 17:01 | |
*** BernsO has quit IRC | 17:04 | |
*** racedo has quit IRC | 17:27 | |
*** bnemec is now known as beekneemech | 17:34 | |
prometheanfire | TheJulia: so... old version of diskimage builder (tested 2.10.1) works for me | 17:39 |
prometheanfire | so maybe new dib doesn't work | 17:39 |
TheJulia | prometheanfire: ugh, so sambetts|afk mentioned something along these lines earlier that he was hitting issues off of master dib | 17:42 |
prometheanfire | woo | 17:46 |
openstackgerrit | yolanda.robla proposed openstack/ironic-tempest-plugin master: Add manual clean step ironic standalone test https://review.openstack.org/568616 | 17:52 |
*** jcoufal has quit IRC | 17:56 | |
*** jcoufal has joined #openstack-ironic | 17:58 | |
*** fragatina has quit IRC | 18:05 | |
*** e0ne has quit IRC | 18:10 | |
prometheanfire | ironic is considering this a failure, conductor side | 18:16 |
prometheanfire | 2018-05-18 18:13:39.285 3106 INFO ironic_python_agent.extensions.standby [-] image (ab11b6da-e4d9-4a1d-8dc8-6130aac2e443) written to device /dev/sda root_uuid= | 18:16 |
prometheanfire | 2018-05-18 18:13:39.285 3106 INFO root [-] Command: prepare_image, result: {'result': 'prepare_image: image (ab11b6da-e4d9-4a1d-8dc8-6130aac2e443) written to device /dev/sda root_uuid='} | 18:16 |
prometheanfire | I do wonder if the uuid thing would fix it | 18:16 |
*** racedo has joined #openstack-ironic | 18:28 | |
*** racedo has quit IRC | 18:28 | |
TheJulia | prometheanfire: what deploy interface are you using? | 18:35 |
prometheanfire | ipmi/pxe | 18:36 |
prometheanfire | pike | 18:36 |
*** jcoufal has quit IRC | 18:39 | |
*** rnoriega has quit IRC | 18:46 | |
*** rnoriega has joined #openstack-ironic | 18:47 | |
*** milan has quit IRC | 18:48 | |
*** fragatina has joined #openstack-ironic | 18:54 | |
prometheanfire | TheJulia: in order to support PARTUUID testing I think I'll need the patch merged and a release made, while I can set ipa to install from source (and from my fork) I can't do the same for ironic_lib (where 1/2 of the partuuid change is) | 18:59 |
TheJulia | prometheanfire: partition image or whole disk image? | 19:01 |
prometheanfire | it should be a whole disk image, the image should be two partitions (boot and root) | 19:02 |
TheJulia | so if whole disk image, root_uuid doesn't matter, but I don't think we backported that | 19:02 |
* TheJulia looks | 19:02 | |
prometheanfire | I was using master IPA | 19:03 |
prometheanfire | https://github.com/prometheanfire/ironic-python-agent | 19:03 |
*** kzaitsev_pi has quit IRC | 19:10 | |
prometheanfire | I'll qemu-nbd the qcow and see what I find, because it's still not mounting /dev/sda2 to that tmp location | 19:11 |
TheJulia | I'm totally not seeing it :( | 19:11 |
prometheanfire | the mount error? | 19:11 |
TheJulia | prometheanfire: you could just inject your ironic-lib patch into the ramdisk manually | 19:12 |
TheJulia | prometheanfire: the root_uuid fixes that I thought were in place | 19:12 |
TheJulia | oh, what is your boot_option? | 19:12 |
prometheanfire | not sure | 19:12 |
prometheanfire | atm I'm having problems with the grub-install part | 19:12 |
prometheanfire | https://gist.github.com/prometheanfire/86552edc5a30fe277fdb6dd11594f864 | 19:13 |
TheJulia | yeah, you shouldn't be there if it is seen as a whole disk image :( | 19:13 |
TheJulia | that is what is so broken about this | 19:13 |
prometheanfire | 2018-05-18 19:07:29.252 3457 DEBUG root [-] TYPE did not match. Wanted: 'disk' but found: '' list_all_block_devices /usr/share/ironic-python-agent/venv/local/lib/python2.7/site-packages/ironic_python_agent/hardware.py:167 | 19:13 |
TheJulia | unless your actually doing a partition image | 19:13 |
prometheanfire | what do you mean by partition image | 19:14 |
prometheanfire | to me that means single partition | 19:14 |
TheJulia | partition image is just a filesystem | 19:14 |
prometheanfire | disk means mulit-partition | 19:14 |
TheJulia | with a bootloader at the front | 19:14 |
prometheanfire | ya | 19:14 |
prometheanfire | that's what it should be | 19:14 |
TheJulia | and the grub config should be such that the default filenames are there so nvram should load it up | 19:15 |
prometheanfire | I'll paste by dib command in a sec | 19:15 |
prometheanfire | on which image? | 19:15 |
TheJulia | whatever dib spits out | 19:15 |
prometheanfire | the ipa one or the final one? | 19:15 |
TheJulia | for the image to be deployed, not the ramdisk | 19:15 |
prometheanfire | ok | 19:16 |
* TheJulia closes the other laptop so she can pack it | 19:16 | |
prometheanfire | ok, it looks like it is a single partition | 19:17 |
*** kzaitsev_pi has joined #openstack-ironic | 19:17 | |
prometheanfire | Number Start End Size Type File system Flags | 19:17 |
prometheanfire | 1 1049kB 2264MB 2263MB primary ext4 boot | 19:17 |
TheJulia | prometheanfire: are you manually triggering deploy or doing it through nova? | 19:17 |
prometheanfire | and it's expecting a multi-part whole disk you say? | 19:18 |
prometheanfire | nova | 19:18 |
TheJulia | it shouldn't be doing grub_install afaik | 19:18 |
TheJulia | do you have swift and swift tempurls setup? | 19:18 |
prometheanfire | yep | 19:19 |
TheJulia | if with ironic, try the direct deploy interface | 19:19 |
*** fragatina has quit IRC | 19:19 | |
prometheanfire | if tempurls are set up with ironic? | 19:19 |
TheJulia | I bet it will just work, but then makes me wonder why the code is thinking that your image is not a whole disk iamge | 19:19 |
TheJulia | prometheanfire: correct | 19:19 |
prometheanfire | it's a qcow2 file with a single partition on it | 19:20 |
TheJulia | it has a partition table, which makes it a whole disk image | 19:20 |
prometheanfire | cool | 19:20 |
TheJulia | i bet the second partition is a config drive | 19:21 |
prometheanfire | oh, that'd make sense, but it should still be able to mount it | 19:21 |
prometheanfire | maybe fat utils aren't there | 19:21 |
TheJulia | or iso9660 support is missing | 19:22 |
TheJulia | and the uuid it was otherwise trying to mount was /dev/sda2 right? | 19:22 |
prometheanfire | oh, ya, could be an iso image | 19:22 |
prometheanfire | I use vfat at home, but this is OSA, so not the same | 19:22 |
prometheanfire | yes | 19:22 |
TheJulia | if I remember correctly, if the config drive comes out of nova, it is iso9660 by default | 19:23 |
*** ajya has quit IRC | 19:23 | |
prometheanfire | I can set config-drive=false on the nova boot as well | 19:24 |
TheJulia | so this might be making more sense now... and if so then... the bug really is that the dib images your handing it... for some reason are getting interpretted as partition images when they aren't | 19:24 |
TheJulia | worty trying | 19:24 |
* TheJulia resumes packing up | 19:24 | |
prometheanfire | conf trip? | 19:25 |
TheJulia | yeah, I fly out of phl to yvr at 0620 AM in th emorning | 19:25 |
prometheanfire | whee | 19:25 |
TheJulia | so leaving the future mother in law's house tonight so we don't have to get up at 3 AM | 19:25 |
prometheanfire | ah | 19:25 |
*** xhku_ has quit IRC | 19:43 | |
*** fbouliane has joined #openstack-ironic | 19:45 | |
*** KennethWilke has joined #openstack-ironic | 19:51 | |
*** e0ne has joined #openstack-ironic | 19:55 | |
*** KennethWilke has quit IRC | 20:02 | |
prometheanfire | TheJulia: what's direct deploy for pike? | 20:13 |
*** cjloader_ has joined #openstack-ironic | 20:14 | |
*** cjloader has quit IRC | 20:14 | |
TheJulia | Transfers down to the node via a swift tempurl. No iscsi or cross node interactiona. | 20:14 |
TheJulia | Direct is what we called the interface that were called agent_* drivers | 20:15 |
prometheanfire | ok, that's what it's doing | 20:15 |
prometheanfire | it fetched it via tempurl then qemu-image converted | 20:15 |
*** bdodd has quit IRC | 20:18 | |
*** bdodd has joined #openstack-ironic | 20:18 | |
*** cjloader_ has quit IRC | 20:18 | |
*** bdodd has quit IRC | 20:21 | |
*** bdodd has joined #openstack-ironic | 20:23 | |
*** bdodd has quit IRC | 20:29 | |
*** e0ne has quit IRC | 20:30 | |
*** bdodd has joined #openstack-ironic | 20:31 | |
*** Arkady has joined #openstack-ironic | 20:31 | |
*** Arkady has quit IRC | 20:35 | |
*** fragatina has joined #openstack-ironic | 20:37 | |
*** fragatina has quit IRC | 20:38 | |
*** fragatina has joined #openstack-ironic | 20:39 | |
*** cjloader has joined #openstack-ironic | 20:48 | |
*** cjloader has quit IRC | 20:53 | |
*** rbudden has quit IRC | 20:56 | |
*** baoli has quit IRC | 21:00 | |
*** baoli has joined #openstack-ironic | 21:01 | |
*** rbudden has joined #openstack-ironic | 21:02 | |
*** baoli has quit IRC | 21:05 | |
*** hemna__ has quit IRC | 21:07 | |
*** kzaitsev_pi has quit IRC | 21:12 | |
*** kzaitsev_pi has joined #openstack-ironic | 21:16 | |
*** baha has quit IRC | 21:31 | |
*** rh-jelabarre has quit IRC | 21:33 | |
*** figleaf is now known as edleafe | 21:40 | |
*** akhilaki has quit IRC | 21:50 | |
*** jtomasek has quit IRC | 22:06 | |
*** mitchjameson has joined #openstack-ironic | 22:07 | |
*** mitchjameson has quit IRC | 22:07 | |
*** jtomasek has joined #openstack-ironic | 22:12 | |
*** jtomasek has quit IRC | 22:17 | |
*** r-daneel has quit IRC | 22:22 | |
*** hemna__ has joined #openstack-ironic | 22:25 | |
*** rbudden has quit IRC | 22:30 | |
*** rajinir has quit IRC | 22:36 | |
*** d0ugal_ has joined #openstack-ironic | 23:12 | |
*** d0ugal has quit IRC | 23:13 | |
*** rnoriega has quit IRC | 23:17 | |
*** lhinds has quit IRC | 23:17 | |
*** zshi has quit IRC | 23:57 | |
*** jaganathan has quit IRC | 23:58 | |
*** cjloader has joined #openstack-ironic | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!