| devananda | lifeless: ^ adds ephemeral partition data to nova.virt.driver. We need to add rebuild() method as well, though, to get the feature you need | 00:00 |
|---|---|---|
| jroll | devananda: lucas' change fixes that issue as well | 00:00 |
| jroll | jfyi | 00:00 |
| jroll | and yeah, it'll conflict | 00:00 |
| devananda | jroll: which patch? | 00:01 |
| jroll | the driverfactory thing | 00:01 |
| * jroll finds a link | 00:01 | |
| devananda | ahhh | 00:01 |
| devananda | https://review.openstack.org/#/c/86971/3/ironic/nova/virt/ironic/driver.py | 00:01 |
| rloo | russell_h: thx, looking now | 00:02 |
| devananda | that fixes the "sends too many patches" thing | 00:02 |
| jroll | yes | 00:02 |
| devananda | jroll: but it doesn't add ephemeral data | 00:02 |
| * devananda rebases | 00:02 | |
| *** derekh has quit IRC | 00:02 | |
| jroll | devananda: right, just letting you know your first patch is solved by that one :) | 00:02 |
| devananda | jroll: gotcha. thanks | 00:02 |
| jroll | :) | 00:02 |
| *** eguz has quit IRC | 00:08 | |
| *** shakamunyi has joined #openstack-ironic | 00:11 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: nova.virt.ironic passes ephemeral_gb to ironic https://review.openstack.org/87396 | 00:13 |
| Shrews | Devananda: sup? | 00:24 |
| *** nosnos has joined #openstack-ironic | 00:26 | |
| devananda | Shrews: wondering if you want to implement driver.rebuild() :) | 00:26 |
| Shrews | devananda: context? | 00:27 |
| *** blamar has joined #openstack-ironic | 00:28 | |
| devananda | Shrews: http://git.openstack.org/cgit/openstack/nova/tree/nova/virt/baremetal/driver.py#n308 | 00:28 |
| devananda | we don't have that in nova.virt.ironic | 00:28 |
| devananda | and so calling "nova rebuild --preserve-ephemeral" raises an exception | 00:28 |
| Shrews | ooh, that sounds complex and fun | 00:29 |
| *** matsuhashi has joined #openstack-ironic | 00:30 | |
| devananda | Shrews: https://bugs.launchpad.net/ironic/+bug/1307754 | 00:30 |
| Shrews | devananda: got it. thx | 00:30 |
| devananda | Shrews: you probably want to start by basing it on https://review.openstack.org/#/c/87396/ | 00:30 |
| *** nosnos has quit IRC | 00:31 | |
| devananda | nova isn't even passing the right fields to ironic today, so i got that part up | 00:31 |
| Shrews | ack | 00:31 |
| Shrews | devananda: btw, https://review.openstack.org/83105 is aching for a +2/+A | 00:31 |
| devananda | awesome | 00:32 |
| *** nosnos has joined #openstack-ironic | 00:32 | |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Add configdrive generation to Nova driver https://review.openstack.org/86192 | 00:33 |
| *** stevehuang has quit IRC | 00:34 | |
| *** blamar has quit IRC | 00:35 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Better handling of missing drivers https://review.openstack.org/83572 | 00:43 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Add a @sync_command() decorator https://review.openstack.org/87144 | 00:45 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Sync oslo logging https://review.openstack.org/87346 | 00:45 |
| devananda | right. I'm off for dinner... will be back later... | 00:47 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Add configdrive generation to Nova driver https://review.openstack.org/86192 | 00:48 |
| *** shakamunyi has quit IRC | 00:54 | |
| *** blamar has joined #openstack-ironic | 00:59 | |
| *** ekarlso has quit IRC | 01:11 | |
| *** ekarlso has joined #openstack-ironic | 01:13 | |
| *** ekarlso has quit IRC | 01:13 | |
| *** ekarlso has joined #openstack-ironic | 01:18 | |
| *** newell_ has quit IRC | 01:40 | |
| *** harlowja has quit IRC | 01:52 | |
| *** harlowja_ has joined #openstack-ironic | 01:52 | |
| *** pleia2 has quit IRC | 02:00 | |
| *** pleia2 has joined #openstack-ironic | 02:06 | |
| openstackgerrit | A change was merged to openstack/ironic: Encapsulate Ironic client retry logic https://review.openstack.org/83105 | 02:13 |
| *** rloo has quit IRC | 02:15 | |
| *** coolsvap|afk is now known as coolsvap | 02:30 | |
| *** Haomeng|2 has joined #openstack-ironic | 03:00 | |
| *** Haomeng has quit IRC | 03:00 | |
| *** Haomeng has joined #openstack-ironic | 03:06 | |
| *** Haomeng|2 has quit IRC | 03:07 | |
| lifeless | devananda: cool | 03:13 |
| *** harlowja_ is now known as harlowja_away | 03:21 | |
| *** matsuhashi has quit IRC | 03:27 | |
| *** eghobo has joined #openstack-ironic | 03:36 | |
| *** nosnos has quit IRC | 03:39 | |
| *** harlowja_away is now known as harlowja_ | 03:53 | |
| *** radsy has quit IRC | 04:06 | |
| *** Haomeng|2 has joined #openstack-ironic | 04:22 | |
| *** Haomeng has quit IRC | 04:23 | |
| *** matsuhashi has joined #openstack-ironic | 04:25 | |
| *** nosnos has joined #openstack-ironic | 04:31 | |
| *** lazy_prince has joined #openstack-ironic | 04:51 | |
| *** matsuhashi has quit IRC | 05:03 | |
| *** matsuhashi has joined #openstack-ironic | 05:04 | |
| *** eghobo has quit IRC | 05:05 | |
| *** eghobo has joined #openstack-ironic | 05:05 | |
| *** sabah has joined #openstack-ironic | 05:13 | |
| *** yuikotakada has joined #openstack-ironic | 05:17 | |
| devananda | https://review.openstack.org/#/c/83572/ is ready for reviewers | 05:19 |
| *** greghaynes has joined #openstack-ironic | 05:27 | |
| *** harlowja_ is now known as harlowja_away | 05:37 | |
| *** ewindisch has quit IRC | 05:45 | |
| *** ewindisch has joined #openstack-ironic | 05:46 | |
| *** Haomeng|2 has quit IRC | 05:48 | |
| *** yuikotakada has quit IRC | 05:49 | |
| *** Haomeng has joined #openstack-ironic | 05:49 | |
| *** yonglihe_ has quit IRC | 06:03 | |
| *** yongli has joined #openstack-ironic | 06:06 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/83956 | 06:07 |
| Mikhail_D_wk | Good morning folks! :) | 06:41 |
| *** ifarkas has joined #openstack-ironic | 06:44 | |
| *** GheRivero_ is now known as GheRivero | 06:50 | |
| GheRivero | morning all | 06:50 |
| devananda | g'morning! | 06:51 |
| Haomeng | morning all:) | 06:57 |
| *** matsuhashi has quit IRC | 07:01 | |
| *** matsuhashi has joined #openstack-ironic | 07:01 | |
| *** matsuhashi has quit IRC | 07:02 | |
| *** matsuhashi has joined #openstack-ironic | 07:03 | |
| *** romcheg has joined #openstack-ironic | 07:11 | |
| *** yuriyz has joined #openstack-ironic | 07:20 | |
| *** mrda is now known as mrda_away | 07:27 | |
| *** sabah has quit IRC | 07:47 | |
| openstackgerrit | Haomeng,Wang proposed a change to openstack/ironic: Implements send-data-to-ceilometer https://review.openstack.org/72538 | 07:51 |
| dtantsur|zzzZZZz | Morning Ironic | 07:54 |
| *** dtantsur|zzzZZZz is now known as dtantsur | 07:54 | |
| *** jistr has joined #openstack-ironic | 07:54 | |
| devananda | hmm, i should probably go to sleep before anyone notices that i'm still working ;) | 07:54 |
| dtantsur | devananda, good morning and good night :) | 07:55 |
| Haomeng | devananda: good night:) | 07:56 |
| Haomeng | dtantsur: morning:) | 07:56 |
| devananda | heh... just one more review.... ;) | 07:57 |
| Haomeng | devananda: :) | 07:57 |
| *** max_lobur has joined #openstack-ironic | 08:08 | |
| Mikhail_D_wk | dtantsur: morning :) | 08:10 |
| Mikhail_D_wk | devananda: good night :) | 08:10 |
| openstackgerrit | A change was merged to openstack/ironic: Overwriting node_is_available in IronicDriver https://review.openstack.org/84214 | 08:13 |
| *** matsuhashi has quit IRC | 08:13 | |
| *** matsuhashi has joined #openstack-ironic | 08:14 | |
| *** lucasagomes has joined #openstack-ironic | 08:18 | |
| *** matsuhashi has quit IRC | 08:18 | |
| *** max_lobur has quit IRC | 08:21 | |
| *** romcheg has quit IRC | 08:23 | |
| *** eguz has joined #openstack-ironic | 08:26 | |
| *** romcheg has joined #openstack-ironic | 08:28 | |
| *** matsuhashi has joined #openstack-ironic | 08:28 | |
| *** derekh has joined #openstack-ironic | 08:29 | |
| *** yuriyz has quit IRC | 08:29 | |
| *** eghobo has quit IRC | 08:30 | |
| *** yuriyz has joined #openstack-ironic | 08:30 | |
| *** dshulyak has quit IRC | 08:30 | |
| *** athomas has joined #openstack-ironic | 08:31 | |
| openstackgerrit | Vladimir Kozhukalov proposed a change to openstack/ironic-python-agent: Added metaclass statement https://review.openstack.org/86876 | 08:34 |
| *** sabah has joined #openstack-ironic | 08:36 | |
| *** mdickson has quit IRC | 08:37 | |
| *** mdickson has joined #openstack-ironic | 08:38 | |
| *** eguz has quit IRC | 08:42 | |
| devananda | lucasagomes: oh hai! i'm still up ... have a sec to look at my comment on https://review.openstack.org/#/c/86971/ ? | 08:43 |
| lucasagomes | devananda, morning | 08:43 |
| lucasagomes | devananda, sure lemme take a look | 08:43 |
| lucasagomes | devananda, ah nice, good idea about naming things (I'm really bad on it) | 08:44 |
| *** dshulyak has joined #openstack-ironic | 08:44 | |
| lucasagomes | devananda, 1 sec I'm going to rebase and making the necessary changes | 08:44 |
| devananda | lucasagomes: i've got it done here, actually | 08:45 |
| lucasagomes | devananda, btw, I saw ur comment on the seamicro thing... the idea is to have a standard way to let clients set the boot devices | 08:45 |
| devananda | lucasagomes: if you dont mind me posting on top of your patch | 08:45 |
| lucasagomes | seamicro case it uses hd0 for some reason (just ported from the other code) | 08:45 |
| lucasagomes | devananda, ack go ahead then :D | 08:45 |
| lucasagomes | devananda, thanks for it | 08:45 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Refactor the driver fields mapping https://review.openstack.org/86971 | 08:46 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: nova.virt.ironic passes ephemeral_gb to ironic https://review.openstack.org/87396 | 08:46 |
| devananda | lucasagomes: no worries. I needed it to test the patch for ephemeral anyway :) | 08:46 |
| lucasagomes | devananda, :) good stuff | 08:46 |
| devananda | also, i believe Shrews is adding the nova.virt.ironic.driver:rebuild() method | 08:47 |
| devananda | so we can start supporting "nova rebuild --preserve-ephemeral" | 08:47 |
| devananda | that's a blocker for tripleo's usage | 08:48 |
| lucasagomes | devananda, ah that would be great, yeah right now the only way to preserve is setting directly via our api... pxe_preserve_ephemeral=True | 08:49 |
| lucasagomes | I saw his client wrapper patch merged too, that's a good progress | 08:49 |
| devananda | right. and if you call "nova rebuild" it just raises an exception today | 08:50 |
| lucasagomes | driver is getting more robust :) | 08:50 |
| devananda | indeed! | 08:50 |
| lucasagomes | :( /me never tried to rebuild the instance | 08:50 |
| devananda | tripleo needs inplace upgrades ... | 08:50 |
| lucasagomes | I c | 08:53 |
| openstackgerrit | A change was merged to openstack/ironic: Sync oslo/config/generator https://review.openstack.org/87344 | 08:56 |
| *** romcheg has quit IRC | 08:56 | |
| *** romcheg has joined #openstack-ironic | 08:57 | |
| *** martyntaylor has joined #openstack-ironic | 08:59 | |
| *** romcheg has quit IRC | 09:01 | |
| *** romcheg has joined #openstack-ironic | 09:03 | |
| openstackgerrit | Sirushti Murugesan proposed a change to openstack/ironic: Adds support for deploying Disk Images https://review.openstack.org/73054 | 09:10 |
| devananda | lucasagomes: you see the doc proposal from sandhya? http://docs-draft.openstack.org/43/83343/7/check/gate-ironic-docs/64102a0/doc/build/html/deploy/userguide.html | 09:11 |
| devananda | i'm impressed -- it's come really far since I saw it last | 09:11 |
| lucasagomes | devananda, I haven't lemme take a look | 09:11 |
| lucasagomes | wow!!!!!! | 09:12 |
| devananda | right?!? | 09:12 |
| lucasagomes | that's pretty cool | 09:12 |
| lucasagomes | lemme read | 09:12 |
| devananda | there's a bit of formatting that could be improved, but damn, let's land this. I already +2'd ruby's install guide, too | 09:14 |
| lucasagomes | devananda, yeah I missed some bullet points | 09:15 |
| lucasagomes | but not a big deal | 09:15 |
| lucasagomes | I would +2 that as well | 09:15 |
| lucasagomes | there's some docs in the ironic client as well #85107 | 09:16 |
| devananda | oh ya | 09:17 |
| devananda | did we get the doc publish job added for the client yet? | 09:17 |
| devananda | i believe rloo posted it, but i am not sure what happened | 09:17 |
| devananda | (dont hve the link handy) | 09:17 |
| sirushti | Hi, I had some doubts on my patch for the disk image support blueprint, would this be the right time to bring it up? :-) | 09:18 |
| devananda | hi sirushti ! sure | 09:19 |
| openstackgerrit | A change was merged to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/83956 | 09:19 |
| devananda | (though it's 2am for me, so my responses may or may not make sense...) | 09:19 |
| *** ilives has joined #openstack-ironic | 09:20 | |
| sirushti | devananda, heh! So from the implementation perspective, I've currently written the patch so as to delay validation of root/swap/ephemeral until we figure out what type the image is! | 09:21 |
| sirushti | It doesn't seem too appropriate to do that however | 09:21 |
| lucasagomes | devananda, I don't think we have a job to publish the cli docs yet :( I can take a quick look into it | 09:21 |
| devananda | lucasagomes: https://review.openstack.org/#/c/86390/ | 09:22 |
| *** romcheg has quit IRC | 09:22 | |
| lucasagomes | ah patch is up :) lemme take a look and vote | 09:22 |
| lucasagomes | to bump it in the queue | 09:22 |
| devananda | :) | 09:22 |
| sirushti | devananda, it feels like we're supposed to do all the necessary validations in validate() itself? | 09:23 |
| sirushti | devananda, which brings me to the question, is it okay to have a separate _validate() function? | 09:24 |
| devananda | sirushti: hmmmm | 09:24 |
| devananda | sirushti: so there can be only one driver.validate() method, but that could be split into eg. _validate_netboot() + _validate_diskboot() | 09:25 |
| devananda | or something | 09:25 |
| *** max_lobur has joined #openstack-ironic | 09:25 | |
| devananda | sirushti: have you mapped out the different permutations of [pxe|local] + [partition|whole disk] + [chassis default power off | power on ] ? | 09:26 |
| sirushti | devananda, Yes, so i meant that validate() will call _validate() instead of calling _parse_driver_info! because I'd like to introduce a Glance API call for validating and calling _parse_driver_info evertime seems rather expensive? no? | 09:26 |
| devananda | I think we need to account for all eight permutations here | 09:26 |
| devananda | oooh yes, that's a good point | 09:27 |
| devananda | having validate() actually check with glance makes sense to me | 09:27 |
| devananda | just like IPMITool.validate() actually pings the BMC to check the credentials | 09:27 |
| openstackgerrit | Vladimir Kozhukalov proposed a change to openstack/ironic-python-agent: Added disk partitioner https://review.openstack.org/86163 | 09:27 |
| sirushti | devananda, yeah! and possibly port all validations from _parse_driver_info() to _validate() | 09:28 |
| devananda | sirushti: ++ | 09:29 |
| devananda | sirushti: oh, also, see https://review.openstack.org/#/c/87396/ and https://review.openstack.org/#/c/86971/ | 09:30 |
| devananda | sirushti: these conflict slightly with your patch | 09:30 |
| *** romcheg has joined #openstack-ironic | 09:30 | |
| sirushti | devananda, Great :-) So another concern I had was with "Nova rebuild". I see there's an option for --preserve-ephemeral? That's probably not going to work with disk images! | 09:30 |
| devananda | sirushti: right! so, two things. a) the nova.virt.ironic driver does not support rebuild() yet -- Shrews is working on it. b) the nova.virt.ironic driver will need to raise an exception if the user attempts to rebuild --preserve-ephemeral on a whole-disk image | 09:32 |
| sirushti | devananda, will look into those patches! :-) Thanks! | 09:32 |
| devananda | sirushti: for that matter, we should probably prevent any booting of a whole-disk image with an ephemeral partition, no? | 09:32 |
| sirushti | devananda, yeah, so will the driver know about it to raise an exception? | 09:33 |
| devananda | maybe? if not, it can propagate back from ironic, eg. when nova.virt.ironic calls "ironic node-validate" | 09:34 |
| sirushti | devananda, hmmm, okay! btw, why prevent any booting of a whole disk image with an ephemeral partition? | 09:37 |
| devananda | sirushti: hm, well, besides that we cant support --preserve-ephemeral with a whole disk image | 09:43 |
| lucasagomes | devananda, I will rebase #86971, it failed on the gate checks | 09:43 |
| devananda | sirushti: how do you think matching should be done between flavor and image, when doing whole disk images? what if the image has a partition layout that doesn't match the flavor? | 09:44 |
| devananda | lucasagomes: cheers! I'm fading pretty soon :) | 09:44 |
| lucasagomes | devananda, :) yeah it's quite late there | 09:44 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Refactor the driver fields mapping https://review.openstack.org/86971 | 09:44 |
| sirushti | devananda, yep! So I assumed that Linux will have cloud-init to expand the root partition and Windows will have cloudbase-init to the same! | 09:45 |
| devananda | make that, fading now... | 09:49 |
| devananda | sirushti: i hope i was some help, and didn't confuse matters. i can't think of a good reason for what i said about no-ephemeral-with-whole-disk. I'll think about it more tomorrow. | 09:51 |
| devananda | really good night this time! :) | 09:51 |
| sirushti | devananda, will investigate a bit more on that and send in another patch based on the above discussion :-) | 09:51 |
| sirushti | heh, done! | 09:51 |
| sirushti | devananda, gnight and Thanks a lott!! You were of great help! :-) | 09:52 |
| *** yuriyz has quit IRC | 09:58 | |
| *** sabah has quit IRC | 10:01 | |
| openstackgerrit | A change was merged to openstack/ironic: Add Ironic User Guide https://review.openstack.org/83343 | 10:02 |
| *** yuriyz has joined #openstack-ironic | 10:04 | |
| lucasagomes | g'night devananda | 10:10 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Add DiskPartitioner https://review.openstack.org/83396 | 10:20 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Use DiskPartitioner https://review.openstack.org/83399 | 10:20 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Get rid of the swap partition https://review.openstack.org/83726 | 10:21 |
| *** martyntaylor has quit IRC | 10:25 | |
| *** nosnos has quit IRC | 10:27 | |
| *** ilives has quit IRC | 10:33 | |
| *** matsuhashi has quit IRC | 10:41 | |
| *** ilives has joined #openstack-ironic | 10:42 | |
| *** martyntaylor has joined #openstack-ironic | 10:43 | |
| *** matsuhashi has joined #openstack-ironic | 10:44 | |
| *** matsuhashi has quit IRC | 10:54 | |
| *** coolsvap is now known as coolsvap|afk | 10:56 | |
| openstackgerrit | A change was merged to openstack/ironic: Fix conductor.manager test assertion order https://review.openstack.org/87035 | 11:02 |
| *** romcheg has quit IRC | 11:10 | |
| *** martyntaylor1 has joined #openstack-ironic | 11:18 | |
| *** martyntaylor has quit IRC | 11:18 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Add ManagementInterface https://review.openstack.org/86063 | 11:26 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: IPMITool to use the new ManagementInterface https://review.openstack.org/86092 | 11:26 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: SeaMicro to use the new ManagementInterface https://review.openstack.org/86328 | 11:26 |
| *** romcheg has joined #openstack-ironic | 11:28 | |
| *** lucasagomes is now known as lucas-hungry | 11:29 | |
| openstackgerrit | Andrey Kurilin proposed a change to openstack/python-ironicclient: Sync latest code and reuse exceptions from oslo https://review.openstack.org/71500 | 11:47 |
| openstackgerrit | Andrey Kurilin proposed a change to openstack/python-ironicclient: Reuse module `cliutils` from common code https://review.openstack.org/72418 | 11:47 |
| openstackgerrit | A change was merged to openstack/ironic: Remove a DB query for get_ports_by_node() https://review.openstack.org/84861 | 11:53 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Add create() and destroy() to Node https://review.openstack.org/84823 | 11:54 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Clean up calls to get_node() https://review.openstack.org/84573 | 11:54 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Make sync_power_states yield https://review.openstack.org/84862 | 11:54 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Refactor sync_power_states tests to not use DB https://review.openstack.org/87076 | 11:54 |
| openstackgerrit | Vladimir Kozhukalov proposed a change to openstack/ironic-python-agent: Deprecated extension name attribute https://review.openstack.org/87557 | 11:57 |
| dtantsur | Guys, what is the policy for exceptions naming with regards to using Error word? I see in ironic itself calling FooBarError is discouraged, while in ironic-python-agent all/most exception classes have Error in them | 11:59 |
| *** ifarkas has quit IRC | 12:04 | |
| *** ifarkas has joined #openstack-ironic | 12:04 | |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Add create() and destroy() to Node https://review.openstack.org/84823 | 12:10 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Clean up calls to get_node() https://review.openstack.org/84573 | 12:10 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Make sync_power_states yield https://review.openstack.org/84862 | 12:10 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic: Refactor sync_power_states tests to not use DB https://review.openstack.org/87076 | 12:10 |
| NobodyCam | good morning Ironic | 12:11 |
| dtantsur | NobodyCam, morning | 12:11 |
| NobodyCam | morning dtantsur | 12:11 |
| NobodyCam | :) | 12:12 |
| *** Alexei_987 has joined #openstack-ironic | 12:13 | |
| Mikhail_D_wk | NobodyCam: good morning :) | 12:13 |
| NobodyCam | morning Mikhail_D_wk :) | 12:14 |
| Mikhail_D_wk | NobodyCam: Do you looking for me? ("NobodyCam: Mikhail_D_wk: still around?") | 12:16 |
| NobodyCam | oh I think i had a question about a comment on the logging patch .. but you answered it :) | 12:19 |
| *** jdob has joined #openstack-ironic | 12:29 | |
| *** lucas-hungry is now known as lucasagomes | 12:33 | |
| *** martyntaylor1 has quit IRC | 12:34 | |
| *** linggao has joined #openstack-ironic | 12:34 | |
| *** martyntaylor has joined #openstack-ironic | 12:34 | |
| lucasagomes | morning/afternoon dtantsur NobodyCam Mikhail_D_wk | 12:35 |
| NobodyCam | mornin lucasagomes | 12:36 |
| *** ilives has quit IRC | 12:36 | |
| *** martyntaylor1 has joined #openstack-ironic | 12:39 | |
| *** martyntaylor has quit IRC | 12:40 | |
| *** jbjohnso_ has joined #openstack-ironic | 12:42 | |
| Mikhail_D_wk | lucasagomes: morning :) | 12:43 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: IPMINative to use the new ManagementInterface https://review.openstack.org/86588 | 12:46 |
| *** rloo has joined #openstack-ironic | 12:48 | |
| NobodyCam | humm do we not create our image holding directories... trying to figureout why https://review.openstack.org/#/c/87501/3/elements/ironic-conductor/os-refresh-config/post-configure.d/68-ironic-conductor whould be needed | 12:48 |
| openstackgerrit | A change was merged to stackforge/pyghmi: Auto degrade sessions to operator when admin is forbidden https://review.openstack.org/87370 | 12:55 |
| dtantsur | ifarkas, patch for devstack fixing IPv6 issue: https://review.openstack.org/#/c/87594 | 12:56 |
| ifarkas | dtantsur, looks goot to me ;-) | 12:56 |
| ifarkas | good even | 12:57 |
| openstackgerrit | Ruby Loo proposed a change to openstack/python-ironicclient: Add documentation for ironic CLI https://review.openstack.org/86127 | 13:00 |
| ifarkas | dtantsur, lucasagomes, looking at https://review.openstack.org/#/c/85387/ seems contradictory to https://bugs.launchpad.net/ironic/+bug/1231351 | 13:09 |
| ifarkas | not sure how to reconcile this | 13:09 |
| NobodyCam | well YES we do create the directories. https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L330-L331 | 13:10 |
| dtantsur | to be more precise bugs https://bugs.launchpad.net/ironic/+bug/1199665 and https://bugs.launchpad.net/ironic/+bug/1231351 contradict | 13:10 |
| ifarkas | exactly | 13:10 |
| dtantsur | lifeless, you here? You participated in both bugs ^^^ | 13:10 |
| lucasagomes | ifarkas, dtantsur er, yeah... | 13:10 |
| dtantsur | For me caching does make sense anyway, because undercloud deployment is something that can be really large, with nearly the same images. And changing image in Glance in the middle of deployment anyway seems stupid to me. But I'd like to hear from cores | 13:11 |
| lucasagomes | dtantsur, ifarkas hmm looking at both, I also think that keeping the image might not be an ironic thing... as the bug states, glance is the right place to do it | 13:12 |
| ifarkas | yeah, but how can you be sure that the image hasn't been changed for sure | 13:12 |
| dtantsur | lucasagomes, the point was not to kill the network on deployment datacenter with say 100 compute nodes, I think | 13:13 |
| *** matty_dubs|gone is now known as matty_dubs | 13:13 | |
| dtantsur | and also make it consume considerably less time | 13:13 |
| dtantsur | baremetal deployment seems special to me in this sense, as in overcloud you'll be rarely deploying such a large amount of the same images | 13:14 |
| ifarkas | dtantsur, but you pay the price of not deploying the right image probably | 13:14 |
| lucasagomes | yeah I'm a bit worried about the cache invalidation thing | 13:14 |
| ifarkas | or at least there's some chance to it | 13:14 |
| dtantsur | ifarkas, maybe, if we consider changing images in the middle of deployment the right thing to do | 13:15 |
| ifarkas | dtantsur, the a use case I was thinking of is that the image changed between two independent deployments | 13:15 |
| dtantsur | ifarkas, this will anyway leave you with half of you deployment with wrong images (you can't update image on baremetal afaik) | 13:17 |
| dtantsur | Anyway, can we think of a reasonable timeout for caching? | 13:17 |
| ifarkas | dtantsur, no, there might be nodes which are running an updated version of the image. it might not be wrong. btw, tripleo is planning to do baremetal update s | 13:18 |
| dtantsur | ifarkas, what do you mean by 'tripleo' here? Tuskar + Ironic? ;) | 13:19 |
| ifarkas | dtantsur, devtest which will use tuskar + ironic | 13:20 |
| dtantsur | What I don't like is us sacrificing the common case (large deployment, the same images) for uncommon (changing images in the middle, use case for 2 deployments can be handled by timeouts) | 13:20 |
| lucasagomes | does the images in glance have a checksum in their metadata? if so could we at least test the local image checksum with the glance metada checksum? | 13:21 |
| ifarkas | dtantsur, ok, so how about, adding this feature to ironic but disabling it by default? | 13:22 |
| ifarkas | and we can support both workflow | 13:22 |
| dtantsur | ifarkas, makes sense. And I also feel like adding some timeout, say 1 hour, so that we don;t cache old sh*t forever | 13:23 |
| ifarkas | lucasagomes, checking checksum also seems to be a good idea | 13:23 |
| dtantsur | NobodyCam, could you comment on this all ^^^ | 13:23 |
| lucasagomes | right, I will add a comment to the review... code wise it lgtm dtantsur | 13:24 |
| NobodyCam | sorry was in OoO channel | 13:24 |
| dtantsur | ifarkas, images do have checksums | 13:24 |
| ifarkas | yes | 13:24 |
| *** andreykurilin has quit IRC | 13:24 | |
| dtantsur | Idea: we can include checksum into local image path, so that a new one will be always downloaded (and an old one will be purged after some time) | 13:27 |
| dtantsur | ifarkas, lucasagomes ^^^ ? | 13:27 |
| lucasagomes | dtantsur, yeah this is one thing I was thinking | 13:28 |
| NobodyCam | they do seem to contradict one anohter | 13:28 |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: Install guide for Ironic https://review.openstack.org/81645 | 13:28 |
| lucasagomes | dtantsur, now... idk if glance does keep a checksum in the image metadata | 13:28 |
| dtantsur | lucasagomes, glance image-show shows it | 13:28 |
| ifarkas | dtantsur, adding checksum to the image path seems to be a good solution | 13:30 |
| lucasagomes | dtantsur, right yeah... I think the checksum would be a good way to go | 13:30 |
| NobodyCam | I would have to think about this. but I think checksum is a good short term solution | 13:31 |
| lucasagomes | anyway, I added lifeless and devananda to give their input on this as well | 13:31 |
| *** blamar has quit IRC | 13:31 | |
| lucasagomes | NobodyCam, right, why a short term solution? or you think that the idea of ironic caching images doesn't really belong to ironic | 13:32 |
| dtantsur | nobody else can cache things for you :) | 13:32 |
| NobodyCam | the latter.. I'm not convinced we should be cacheing images.. Thou I do see the beniffet to it | 13:32 |
| NobodyCam | but thats what glance is for | 13:33 |
| NobodyCam | if we get to a point where the agent is pulling images directly from glance we wont be able to cache them | 13:33 |
| lucasagomes | right | 13:34 |
| lucasagomes | I think we are going towards that direction anway | 13:34 |
| lucasagomes | anyway | 13:34 |
| *** blamar has joined #openstack-ironic | 13:34 | |
| NobodyCam | thus the "short term" comment | 13:34 |
| NobodyCam | :-p | 13:34 |
| lucasagomes | heh got ur point now | 13:34 |
| dtantsur | yeah, point about agent makes sense | 13:35 |
| *** tteggel has joined #openstack-ironic | 13:40 | |
| NobodyCam | anyone what to check out my review of https://review.openstack.org/#/c/87501/ make sure i'm not on crack | 13:42 |
| NobodyCam | or lack of coffee | 13:42 |
| lucasagomes | NobodyCam, sounds correct to me, ironic seems to be creating that dir... maybe worth checking devstack as well | 13:48 |
| NobodyCam | lucasagomes: ack thats what I thought | 13:49 |
| ifarkas | NobodyCam, lucasagomes, did you check the bug? https://bugs.launchpad.net/tripleo/+bug/1307907 | 13:49 |
| ifarkas | I also see the directories created beforehand but then the bug should be invalid | 13:50 |
| lucasagomes | ifarkas, hmm yeah... hmm lemme do a quick investigation | 13:51 |
| lucasagomes | just looked quickly in the ironic code it seesm to be calling fileutils.ensure_path to create such directory | 13:51 |
| lucasagomes | ah I see | 13:51 |
| lucasagomes | hmmm | 13:51 |
| NobodyCam | lucasagomes: are we not setting ownership when we create the dir's | 13:52 |
| lucasagomes | NobodyCam, yeah, and it might not have the right permission to create the dir as well | 13:53 |
| dtantsur | Mikhail_D_wk, hi again :) Thanks for great refactoring job, I just have a question: did you think how to merge these changes? I mean, every other merged change breaks them. | 13:53 |
| lucasagomes | see the bug that ifarkas posted https://bugs.launchpad.net/tripleo/+bug/1307907 | 13:53 |
| dtantsur | I wanted to bring this up on a meeting, but we didn't have time | 13:54 |
| NobodyCam | dtantsur: your seeing this too | 13:54 |
| ifarkas | lucasagomes, NobodyCam but the processes are running by the ironic user, so it should be fine, not? | 13:54 |
| NobodyCam | they should be | 13:55 |
| lucasagomes | depends whether the ironic user does have permission to create dirs under /var/lib/... | 13:56 |
| NobodyCam | lucasagomes: the review and tripleo run with /mnt/state/* | 13:56 |
| lucasagomes | right, is that the ephemeral partition? | 13:57 |
| NobodyCam | ya and we set that to be owned by the ironic user: https://github.com/openstack/tripleo-image-elements/blob/master/elements/ironic-conductor/install.d/68-ironic-tftp-support#L19 | 13:58 |
| Mikhail_D_wk | dtantsur: I think need control for the new patches that can conflict with this patches... and I think final rebase need do before squash-review... if it happens =) | 13:58 |
| NobodyCam | oh seems I need to do a quick walkies.. brb | 13:59 |
| lucasagomes | NobodyCam, hmm I see... heh idk gotta take a look at the elements to see how that could have happened | 13:59 |
| dtantsur | Mikhail_D_wk, well, the solution could be postponing all accepts for a few hours, while you and a couple of reviewers (I can help, but I don't have +2 power) quickly land all these large patches | 14:00 |
| lucasagomes | looks alright to me, seems that bug/patch is not needed | 14:00 |
| ifarkas | lucasagomes, btw, I deployed undercloud with the seed machine using ironic and didn't hit this issue | 14:02 |
| rloo | dtantsur, Mikhail_D_wk, I was wondering about that too. We should be able to land one at a time, no? starting with https://review.openstack.org/#/c/60025/? | 14:03 |
| rloo | Mikhail_D_wk, jenkins -1'd 60025. Not sure if I should review it now, or wait for a rebase? | 14:04 |
| dtantsur | rloo, rebase is one side of the problem. Another is to do it quickly (it = 2 * +2), otherwise some change is gonna land in between and you have to start from the beginning | 14:05 |
| dtantsur | we can ask to people with +2 power to just sit at some not-too-busy time every day and land one of these | 14:05 |
| *** jbjohnso_ has quit IRC | 14:05 | |
| rloo | dtantsur: I don't know that it is reasonable to expect ALL of them to be landed at the same time. I'd like to at least start with one ;) | 14:06 |
| rloo | dtantsur: not sure there is a 'not-too-busy' time. But I'd really like to get Mikhail's changes in. | 14:06 |
| dtantsur | rloo, one per day :) | 14:07 |
| romcheg | Morning guys | 14:07 |
| NobodyCam | lucasagomes: could you add comment to the review? | 14:07 |
| rloo | dtantsur: only if it keeps the doctor away ;) | 14:07 |
| NobodyCam | morning romcheg | 14:07 |
| rloo | morning romcheg! | 14:07 |
| romcheg | I'm going to specify the version of ironiclient here https://github.com/openstack/requirements/blob/master/global-requirements.txt#L86 | 14:08 |
| romcheg | I don't think it's going to break something | 14:08 |
| rloo | romcheg: minimum version of ironicclient? | 14:08 |
| NobodyCam | romcheg: are you seeing a issue? | 14:08 |
| romcheg | NobodyCam: basically I got an issue by using a mirror of pypi | 14:09 |
| romcheg | NobodyCam: Which was out of sync | 14:09 |
| romcheg | So I got an older version installed | 14:10 |
| NobodyCam | ahh :) | 14:11 |
| linggao | lucasagomes ping | 14:12 |
| romcheg | but probably that is not an issue | 14:12 |
| romcheg | *not an issue we should carry about | 14:12 |
| NobodyCam | :) | 14:13 |
| lucasagomes | linggao, pong 1 sec I'm in a call | 14:14 |
| rloo | romcheg: I think if you encountered that problem, it is probably worth fixing so it helps the next person. | 14:15 |
| NobodyCam | rloo: ++ | 14:15 |
| romcheg | Will make a patch then | 14:16 |
| *** pcargnel has joined #openstack-ironic | 14:17 | |
| pcargnel | Hi, could someone help me understand this bug: https://bugs.launchpad.net/ironic/+bug/1284781 | 14:17 |
| *** jbjohnso_ has joined #openstack-ironic | 14:19 | |
| NobodyCam | pcargnel: looks like wsme added readonly and mandatory options | 14:20 |
| NobodyCam | reading the bug we also have code that | 14:21 |
| NobodyCam | I read thee bug as clean up our (ironic's) code to use wsme's options | 14:22 |
| pcargnel | NobodyCam, so the idea is to remove from, in this case, JSonPatchtype, the validations for readonly and mandatory? | 14:22 |
| NobodyCam | reading Haomeng's comments that what I get.. but I have not look deeply at the code | 14:23 |
| pcargnel | Ok, I'll take a deeper look and propose a patch. It's better to discuss with code :) | 14:25 |
| openstackgerrit | Dirk Mueller proposed a change to openstack/python-ironicclient: Remove py3kcompat module https://review.openstack.org/87656 | 14:26 |
| *** foexle has joined #openstack-ironic | 14:29 | |
| NobodyCam | pcargnel: great! | 14:29 |
| *** martyntaylor has joined #openstack-ironic | 14:29 | |
| foexle | heyho guys :) a short question. Would be ironic in IceHouse production ready for mixed setups? It's hard to find informations about mixed clouds with virt compute aggregates and bare metal provisioning | 14:31 |
| NobodyCam | foexle: not sure what production ready is? | 14:32 |
| *** martyntaylor1 has quit IRC | 14:32 | |
| NobodyCam | multi tentant? | 14:32 |
| foexle | NobodyCam: yes and i mean more or less stable | 14:33 |
| NobodyCam | Ironic will not be in the ice house release. | 14:33 |
| foexle | oh really ? i thought its the new incubation project | 14:34 |
| NobodyCam | yes but Ironic did not graduate in this cycle | 14:34 |
| foexle | ahhh ok all right good to know | 14:34 |
| foexle | so i need to look for bm drivers for the moment -.- ... | 14:35 |
| NobodyCam | :) | 14:40 |
| NobodyCam | we will(should) be released in J cycle | 14:40 |
| NobodyCam | brb | 14:40 |
| foexle | NobodyCam: Juno i see, some good features comes to Juno release :). All right, last question: I'm looking since a while for a good howto/tutorial/explanation do you have a good target ? .... Thanks for your time | 14:43 |
| *** agordeev has quit IRC | 14:43 | |
| *** martyntaylor has quit IRC | 14:46 | |
| *** jgrimm has quit IRC | 14:49 | |
| *** sanek11 has joined #openstack-ironic | 14:50 | |
| *** sanek11 is now known as agordeev | 14:52 | |
| lucasagomes | linggao, pong back :) | 14:54 |
| linggao | lucasagomes, I am reviwing your patch 86588 (ipminative) and have a questions. | 14:54 |
| linggao | I saw you overwrote boot_devices.DISK from 'disk' to 'hd' to match pyghmi api in _BOOT_DEVICES_MAP. | 14:54 |
| linggao | So the question is: when set_boot_device is called, what is the value of device if it is boot_devices.DISK? 'disk' or 'hd'? | 14:54 |
| lucasagomes | linggao, oh... I think I haven't mapped... it should be hd https://github.com/stackforge/pyghmi/blob/master/pyghmi/ipmi/command.py#L215-L221 | 14:56 |
| lucasagomes | linggao, but I can see a prob with my patch alreayd | 14:57 |
| NobodyCam | foexle: tutorial for? | 14:57 |
| lucasagomes | I didn't convert it... | 14:57 |
| linggao | I was not sure adding _BOOT_DEVICES_MAP will automatcally convert it or not :) | 14:58 |
| *** jistr has quit IRC | 14:58 | |
| lucasagomes | linggao, heh not really urgh... thanks for pointing it out | 14:59 |
| lucasagomes | I'm going fix it quickly here | 14:59 |
| lucasagomes | linggao, I based on that docstring for the values... but there's one there that I think will cause problem on pyghmi | 15:00 |
| lucasagomes | optical is not being mapped | 15:00 |
| lucasagomes | jbjohnso_, https://github.com/stackforge/pyghmi/blob/master/pyghmi/ipmi/command.py#L215-L221 | 15:00 |
| lucasagomes | jbjohnso_, optical is not part of the boot_devices mapping | 15:01 |
| lucasagomes | https://github.com/stackforge/pyghmi/blob/master/pyghmi/ipmi/command.py#L25-L42 | 15:01 |
| *** rloo has quit IRC | 15:01 | |
| linggao | lucasagomes, I was talking to jbjohnso_ on the mapping. Could you also map BIOS to 'setup | 15:01 |
| *** rloo has joined #openstack-ironic | 15:02 | |
| lucasagomes | linggao, will do, I was unsure about that one | 15:02 |
| lucasagomes | by the description in the docstring | 15:02 |
| NobodyCam | rloo: 86127 just a rebase...oh I thought I had +2 that one | 15:02 |
| foexle | NobodyCam: so i've read the baremetal wiki and it seems to be deprecated at several places. I thought you know any maybe | 15:02 |
| *** jistr has joined #openstack-ironic | 15:02 | |
| jbjohnso_ | linggao, I can, but fyi, the reverse mapping can't be as forgivinG | 15:02 |
| *** rloo has quit IRC | 15:02 | |
| jbjohnso_ | lucasagomes, as, right, guess I'll add optical so it can be symmetric | 15:03 |
| *** rloo has joined #openstack-ironic | 15:03 | |
| lucasagomes | jbjohnso_, ack... I'm using cdrom anyway :) | 15:03 |
| jbjohnso_ | + 'optical': 0x14, | 15:03 |
| jbjohnso_ | + 'bios': 0x18, | 15:03 |
| rloo | NobodyCam: I don't think you looked at 86127. you're probably thinking of some other docn ;) | 15:03 |
| NobodyCam | foexle: have you read https://wiki.openstack.org/wiki/BaremetalSplitRationale | 15:04 |
| NobodyCam | its why we're (Ironic) here | 15:04 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Add optical and bios aliases for boot devices. https://review.openstack.org/87682 | 15:04 |
| jbjohnso_ | there, review that at your leisure | 15:04 |
| linggao | lucasagomes, there is no 'optical' in ironic.common.boot_devices | 15:04 |
| lucasagomes | linggao, yeah, there's CDROOM | 15:04 |
| lucasagomes | CDROM* | 15:04 |
| linggao | you have mapped CDROM already | 15:05 |
| jroll | foexle: just run ironic in production | 15:05 |
| jroll | foexle: what could go wrong? :) | 15:05 |
| jroll | morning ironic | 15:05 |
| lucasagomes | linggao, yeah I think it's better in ironic to only have 1 name for each boot device, and drivers internally can map it for whatever they need | 15:05 |
| NobodyCam | humm rloo: line 24.. of https://review.openstack.org/#/c/86127/2/doc/source/cli.rst .... what if I have no-auth set | 15:05 |
| NobodyCam | morning jroll :) | 15:06 |
| linggao | there is 'SAFE' in ironic.common.boot_devices, is it used in the code now? | 15:06 |
| jroll | heya :) | 15:06 |
| foexle | jroll: yeah .... i'll take a look on my test env. | 15:06 |
| foexle | NobodyCam: yeah i've read it | 15:06 |
| lucasagomes | linggao, it was being used by the ipmitool driver | 15:06 |
| NobodyCam | foexle: to jroll's comment.. ironic is being developed while Nova-bm is stale.. | 15:06 |
| lucasagomes | linggao, so I just added it to the boot_devices as well | 15:06 |
| rloo | NobodyCam: you'd only have no-auth set if you're testing I think. Even with no-auth set, you have to specify something here, for IRONIC_URL/OS_AUTH_TOKEN. | 15:07 |
| jroll | foexle: it's definitely not production ready, but it's getting there. having another person with a goal of running in production and finding bugs is super helpful :) | 15:07 |
| linggao | lucasagomes, SAFE is not supported yet in pygmi. jbjohnso_ can add that if it is needed. | 15:07 |
| jroll | rloo: one could run noauth with ironic if the network is secure | 15:08 |
| lucasagomes | linggao, jbjohnso_ that would be good :) | 15:08 |
| *** martyntaylor has joined #openstack-ironic | 15:08 | |
| lucasagomes | linggao, so right now I'm just not supporting safe in the native driver | 15:08 |
| lucasagomes | which is fine, as things are added we can just extend that mapping | 15:08 |
| rloo | NobodyCam: not sure I want to complicate this by saying IF bare-metal service is set up with config .. noauth, you should ... here. DId you want me to? | 15:08 |
| linggao | hey jbjohnso_, I found you a job :) | 15:08 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: IPMINative to use the new ManagementInterface https://review.openstack.org/86588 | 15:09 |
| jroll | rloo: oh, looked at the link. hmm. | 15:09 |
| *** yongli has quit IRC | 15:09 | |
| linggao | lucasagomes, fine with me. | 15:09 |
| NobodyCam | rloo: I can see that.. this is not developer docs | 15:09 |
| NobodyCam | was just muttering outlouad | 15:09 |
| rloo | NobodyCam. I was just wondering if this was developer or admin doc :-) | 15:09 |
| *** sseago has quit IRC | 15:09 | |
| jroll | NobodyCam: I hope that if someone is running with noauth, they know what they're doing well enough to not need that part of the docs :) | 15:09 |
| jroll | or know enough to know they don't need user/pass | 15:10 |
| NobodyCam | jroll: ++ yes | 15:10 |
| jroll | I think it's fine | 15:10 |
| linggao | lucasagomes and all, I am working on revising console API. Currently we have the following implementation: | 15:11 |
| linggao | PUT v1/nodes/<uuid>/states/console to change the console mode | 15:11 |
| linggao | GET v1/nodes/<uuid>/states/console to get the console information and it returns the url of the console when the console is enable. It throws exception when the console is disabled. | 15:11 |
| rloo | I think if we get people asking -- we might consider adding it in. | 15:11 |
| NobodyCam | rloo yes... +2'd | 15:11 |
| jroll | rloo: +1 | 15:11 |
| anteaya | o/ | 15:11 |
| * jroll heads off to the office... bbl | 15:11 | |
| NobodyCam | hi anteaya | 15:12 |
| rloo | thank you NobodyCam, jroll. And hopefully it'll be someone else that adds it in later :-) | 15:12 |
| linggao | lucasagomes, I feel that the GET console information should be v1/nodes/<uuid>/console as I mentioned in the team meeting. | 15:12 |
| linggao | It should return {console: {'console_enabled': 'true','url': 'blahblah', 'type': 'blah'}} if the console is enabled. | 15:12 |
| linggao | and return console: {'console_enabled': 'false','url': 'none', 'type': 'none'}} if the console is not enabled. | 15:12 |
| anteaya | NobodyCam: heya slow train wifi | 15:12 |
| jbjohnso_ | lucasagomes, ok, can add, have you ever seen safe actually work? | 15:12 |
| lucasagomes | linggao, right... yeah the /states/ URI doesn't makes much sense there... I agree with that | 15:12 |
| lucasagomes | jbjohnso_, haven't tested | 15:12 |
| NobodyCam | lol i'm on slow t-mobile mifi :) | 15:12 |
| anteaya | NobodyCam: I need to add a docs publishing job to one of our repos, I know you publish your docs, can you think of a patch I can copy? | 15:13 |
| anteaya | birds of a feather | 15:13 |
| rloo | anteaya: I just did that for python-ironicclient. sec. | 15:14 |
| anteaya | rloo: awesome thanks | 15:14 |
| NobodyCam | :)awesome ty rloo | 15:14 |
| linggao | lucasagomes, thanks. I'll make that happen and put out for review. | 15:14 |
| rloo | anteaya: is this what you want: https://review.openstack.org/#/c/86390/ | 15:14 |
| lucasagomes | linggao, the output seems easier to the parser for me the way you described | 15:14 |
| lucasagomes | linggao, I would +2 that yes | 15:14 |
| anteaya | rloo: thanks | 15:14 |
| rloo | anteaya: it is a bit different for the client docs vs the non-client docs. | 15:14 |
| anteaya | I'll look around, it gives me a place to start | 15:15 |
| anteaya | thanks | 15:15 |
| NobodyCam | ok going to get ready to hit yawn road.. please use gtalk/hangout to get me if I am need in the little while | 15:15 |
| anteaya | NobodyCam: happy trails | 15:15 |
| lucasagomes | linggao, thanks for cleaning that up, we needed the console + api done to get out of the incubation so I don't think it was very well engineered | 15:16 |
| rloo | anteaya: let me know if you have questions. but yeah, if you look in those files, you can see other projects. | 15:16 |
| anteaya | oh and I met a high school student at pycon | 15:16 |
| foexle | jroll: heheh yeah i know :) .... so let me try on my dev env, maybe it make me happy and i can use it. It's not for a high frequented boot and destroy setup i'm looking only fo a good solution to merge bare metal machines for swift data nodes and whatever to the existing tenants. so ill try then ironic instead of bm driver | 15:16 |
| NobodyCam | :) will be at the east end of texas | 15:16 |
| NobodyCam | hehehe | 15:16 |
| anteaya | getting him to join openstack contributions | 15:16 |
| foexle | NobodyCam and jroll thanks again for your time | 15:16 |
| anteaya | I will suggest he hang out in here | 15:16 |
| NobodyCam | ok be back later | 15:16 |
| NobodyCam | foexle: anythiime | 15:16 |
| linggao | lucasagomes, thanks for the encouragment. | 15:16 |
| NobodyCam | anytime even | 15:16 |
| NobodyCam | :-p | 15:16 |
| anteaya | rloo: thanks will ask | 15:17 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Add optical and bios aliases for boot devices. https://review.openstack.org/87682 | 15:17 |
| jbjohnso_ | there, added safe | 15:17 |
| anteaya | I'll introduce him once he gets back home and online | 15:17 |
| jbjohnso_ | don't think firmware/OS ever actually coordinate enough for that to work in practice, but that's an implementation detail | 15:17 |
| rloo | anteaya: one more think I should add. That's the patch, but I'm still waiting for it to be approved. I was told that it should work as long as your python setup.py build_sphinx works ;) | 15:17 |
| linggao | jbjohnso_ have your put it on the pip? | 15:17 |
| jbjohnso_ | linggao, I can tag if people like the review | 15:18 |
| openstackgerrit | Russell Haering proposed a change to openstack/ironic-python-agent: Decorated commands no longer take command_name https://review.openstack.org/87147 | 15:18 |
| *** hemna_ has joined #openstack-ironic | 15:19 | |
| jbjohnso_ | lucasagomes, linggao: feel free to review 87682 | 15:19 |
| jbjohnso_ | should be nice and simple | 15:19 |
| lucasagomes | jbjohnso_, will take a look | 15:19 |
| jbjohnso_ | hmm.. | 15:19 |
| anteaya | rloo /me nods | 15:19 |
| linggao | jbjohnso_, just did. (+1) | 15:20 |
| jbjohnso_ | I need to look more.. | 15:20 |
| lucasagomes | jbjohnso_, just one thing about the docstring in the set_bootdev methods | 15:22 |
| lucasagomes | jbjohnso_, would be good to update that as well | 15:22 |
| *** foexle has quit IRC | 15:22 | |
| jbjohnso_ | well, might want to take another look after lunch, I think I flupped the values | 15:22 |
| jbjohnso_ | I'll have to take a deeper look | 15:22 |
| lucasagomes | jbjohnso_, ack I'm on the reviewer list so it will pop up in my inbox... I will take a look once you put up a new patch set | 15:23 |
| jbjohnso_ | basically, think I might have already had a bug | 15:23 |
| openstackgerrit | A change was merged to openstack/python-ironicclient: Add documentation for ironic CLI https://review.openstack.org/86127 | 15:23 |
| linggao | Hi matty_dubs, thanks for +1 path 64100 (console support). I have checked in another set to correct the spelling errors where the exception name were used. Could you please review it when you have a chance? | 15:26 |
| *** Mikhail_D_ltp has joined #openstack-ironic | 15:28 | |
| linggao | :q | 15:35 |
| *** jgrimm has joined #openstack-ironic | 15:36 | |
| *** jbjohnso_ has quit IRC | 15:36 | |
| *** Mikhail_D_ltp has quit IRC | 15:45 | |
| *** yuriyz has quit IRC | 15:49 | |
| *** hemna_ has quit IRC | 15:50 | |
| *** hemna_ has joined #openstack-ironic | 15:51 | |
| devananda | good morning, all | 15:54 |
| Shrews | devananda: morning | 15:56 |
| jroll | morning devananda :) | 15:56 |
| dtantsur | devananda, morning | 15:57 |
| devananda | how's everyone today? | 15:57 |
| *** matty_dubs is now known as matty_dubs|lunch | 15:57 | |
| lucasagomes | devananda, morning | 15:59 |
| lucasagomes | morning Shrews jroll :) | 15:59 |
| jroll | heya lucas :) | 15:59 |
| lucasagomes | devananda, we still need GET in the vendor passthru? I had it in my todo list | 15:59 |
| *** coolsvap|afk is now known as coolsvap | 15:59 | |
| jroll | doing well, deva, how're you? | 15:59 |
| jroll | lucasagomes: to list the available methods? | 16:00 |
| lucasagomes | devananda, from a conversation we had last month I think | 16:00 |
| lucasagomes | jroll, no it was to support GET methods in the vendor_passthru | 16:00 |
| *** eghobo has joined #openstack-ironic | 16:00 | |
| jroll | also, thinking about design sessions... anything y'all would like to talk about related to the agent? | 16:00 |
| lucasagomes | right now we only support POST'ing | 16:00 |
| jroll | ahhh I see | 16:01 |
| lucasagomes | so vendors might want to add a method to return some state | 16:01 |
| lucasagomes | or something related to their hardware but right now it's not supported by our API | 16:01 |
| jroll | yep. makes sense | 16:01 |
| devananda | lucasagomes: hmm. vendor-passthru? probably not. driver-vendor-passthru? possibly? | 16:01 |
| lucasagomes | I can't remember the use-case of the discussion, I just had it in my list so I was taking a look at it | 16:01 |
| *** eghobo has quit IRC | 16:01 | |
| lucasagomes | devananda, right, this is another thing | 16:02 |
| devananda | right | 16:02 |
| lucasagomes | driver-vendor-passthru and vendor-passthru | 16:02 |
| lucasagomes | I think I mentioned it before, can't we merged both of it? | 16:02 |
| lucasagomes | make node to be a parameter of the driver-vendor-passthru? | 16:02 |
| devananda | ah | 16:02 |
| devananda | so i think a reason we didn't do that | 16:03 |
| devananda | ws the synchronous nature of it | 16:03 |
| lucasagomes | if node -> generates a rpc topic for a specific conductor | 16:03 |
| devananda | both of those have to use RPC call | 16:03 |
| lucasagomes | if not generate a random topic | 16:03 |
| devananda | and API has to wait for reply from conductor | 16:03 |
| lucasagomes | yeah... now it's solved | 16:04 |
| lucasagomes | we have a sync validation and all... | 16:04 |
| *** jistr has quit IRC | 16:09 | |
| *** jistr has joined #openstack-ironic | 16:10 | |
| *** eghobo has joined #openstack-ironic | 16:11 | |
| openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Complete wrapping ironic client calls https://review.openstack.org/86993 | 16:11 |
| Shrews | lucasagomes: i love it when i accidentally fix bugs that i didn't know existed :) | 16:13 |
| lucasagomes | Shrews, ah heh yeah that's good! | 16:13 |
| *** foexle has joined #openstack-ironic | 16:14 | |
| *** derekh has quit IRC | 16:15 | |
| *** eghobo has quit IRC | 16:16 | |
| devananda | lucasagomes: on the driver fields mapping patch | 16:19 |
| devananda | lucasagomes: somehow I didn't realize it was consolidating the patch instance_uuid until just now | 16:19 |
| *** eghobo has joined #openstack-ironic | 16:19 | |
| lucasagomes | devananda, right, yeah so it would send 1 request to update all | 16:20 |
| devananda | lucasagomes: that was a separate call because it acts as a lock -- it should raise an error if attempting to set the instance_uuid when one is already set. | 16:20 |
| devananda | do you know if that will still occurr? | 16:20 |
| lucasagomes | devananda, hmm... dunno, lemme give it a quick test | 16:21 |
| *** lazy_prince has quit IRC | 16:23 | |
| devananda | i'm looking at api/controllers/v1/node and conductor/manager:update_node and I dont see anything that is enforcing that any longer | 16:23 |
| lucasagomes | devananda, nothing happens if it's already set,http://paste.openstack.org/show/75792/ | 16:23 |
| *** foexle has quit IRC | 16:23 | |
| lucasagomes | yeah | 16:23 |
| devananda | lucasagomes: change the value | 16:24 |
| devananda | lucasagomes: like, what happens if nova-scheduler tries to spawn() two instances on the same node. | 16:24 |
| lucasagomes | devananda, ahhh yeah it fails | 16:25 |
| devananda | good! but why? lol | 16:25 |
| lucasagomes | http://paste.openstack.org/show/75793/ | 16:25 |
| lucasagomes | lemme see | 16:25 |
| lucasagomes | although the message is wrong | 16:25 |
| lucasagomes | it should say it's already associated with the other instance | 16:25 |
| lucasagomes | heh lemme dig into the code | 16:25 |
| devananda | ahhh | 16:26 |
| devananda | the check is buried in db/sqlalchemy/api | 16:26 |
| devananda | good | 16:26 |
| devananda | well, the exception message is patently wrong, lol | 16:26 |
| devananda | so that's a bug | 16:26 |
| lucasagomes | heh yeah | 16:26 |
| lucasagomes | will open a bug about it got the logs here | 16:27 |
| devananda | thanks! | 16:27 |
| *** dshulyak has quit IRC | 16:27 | |
| *** dwalleck has joined #openstack-ironic | 16:30 | |
| *** Alexei_987 has quit IRC | 16:31 | |
| *** max_lobur has quit IRC | 16:41 | |
| *** eguz has joined #openstack-ironic | 16:45 | |
| *** dwalleck has quit IRC | 16:45 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Fix message preventing overwrite the instance_uuid https://review.openstack.org/87731 | 16:48 |
| *** eghobo has quit IRC | 16:49 | |
| lucasagomes | devananda, quick q re https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L199-L209 | 16:51 |
| devananda | ya | 16:52 |
| devananda | ahh. i think we can remove that TODO | 16:52 |
| devananda | as we're clearly handling that fine already | 16:52 |
| lucasagomes | does it makes sense to prevent instance_uuid to be updated based on the power states? would be better to check whether the node is already provisioned (provision_state != nostate) and is in transition (target_power_state != nostate)? | 16:52 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Added metaclass statement https://review.openstack.org/86876 | 16:52 |
| lucasagomes | and not in transition* | 16:53 |
| devananda | lucasagomes: hm, right. dont allow changing instance_uuid if node is provision(ed|ing) | 16:54 |
| lucasagomes | yeah | 16:54 |
| lucasagomes | I think it would make more sense than looking at the power_state (probably that code is quite old, before we had 2 states for a single node) | 16:54 |
| devananda | I think we've covered some other cases already (and better) that this originally covered (eg, with maintenance mode) | 16:54 |
| devananda | right | 16:54 |
| devananda | thinking in text ... | 16:55 |
| jroll | looking at provision_state and target_power_state would fix this bug as well :) https://bugs.launchpad.net/ironic/+bug/1306202 | 16:56 |
| devananda | - allow instance uuid to be SET when (target_)provision_state are unset | 16:56 |
| *** newell_ has joined #openstack-ironic | 16:56 | |
| devananda | - dont allow instance_uuid to be updated (how will this interact with "nova rebuild" ???) | 16:56 |
| devananda | do maintenance and power state matter at all, then? | 16:57 |
| *** harlowja_away is now known as harlowja_ | 16:58 | |
| jroll | well, the power state check is somewhat valid, but should probably be moved to the driver | 16:58 |
| jroll | right now it is ensuring that the node is powered off when deployed | 16:58 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Support for v2 agent payloads https://review.openstack.org/86984 | 16:58 |
| jroll | which should be up to the driver | 16:58 |
| jroll | related: https://review.openstack.org/#/c/86744/ | 16:58 |
| *** coolsvap is now known as coolsvap|afk | 16:58 | |
| devananda | jroll: thnen that should be done in do_node_deploy, not in update_node | 16:59 |
| jroll | +1 | 16:59 |
| jroll | JoshNang: this conversation is relevant to your interests | 16:59 |
| JoshNang | +1 :) | 16:59 |
| devananda | brb | 16:59 |
| jroll | JoshNang: want to fix up your patch to do that? | 17:00 |
| JoshNang | though in my patch I added a validate_power_state method. I'm fine either way. | 17:00 |
| jroll | I like it in do_node_deploy | 17:00 |
| lucasagomes | yeah and we already check if the node is in maintenance mode in the rpc do_node-deploy | 17:00 |
| jroll | JoshNang: and maybe a second patch to check provision_state in update_node when changing instance_uuid as deva/lucas talked about | 17:00 |
| lucasagomes | +1 | 17:01 |
| JoshNang | jroll: I think I have that patch already written but lacking a few tests | 17:01 |
| jroll | JoshNang: oh? is it up? | 17:02 |
| JoshNang | not yet | 17:02 |
| jroll | ok | 17:02 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Better handling of missing drivers https://review.openstack.org/83572 | 17:02 |
| *** EmilienM has quit IRC | 17:03 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 17:05 | |
| openstackgerrit | A change was merged to openstack/ironic: Install guide for Ironic https://review.openstack.org/81645 | 17:07 |
| *** EmilienM has joined #openstack-ironic | 17:07 | |
| *** stevehuang has joined #openstack-ironic | 17:14 | |
| *** lucasagomes is now known as lucas-dinner | 17:19 | |
| *** zdiN0bot has joined #openstack-ironic | 17:22 | |
| *** lucasagomes has joined #openstack-ironic | 17:25 | |
| *** lucasagomes is now known as lucas-afk | 17:25 | |
| *** lucas-dinner has quit IRC | 17:28 | |
| openstackgerrit | Russell Haering proposed a change to openstack/ironic-python-agent: Decorated commands no longer take command_name https://review.openstack.org/87147 | 17:30 |
| *** zdiN0bot has quit IRC | 17:37 | |
| *** zdiN0bot has joined #openstack-ironic | 17:38 | |
| openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Complete wrapping ironic client calls https://review.openstack.org/86993 | 17:39 |
| *** lazy_prince has joined #openstack-ironic | 17:41 | |
| *** overlayer has joined #openstack-ironic | 17:41 | |
| *** rupsky has joined #openstack-ironic | 17:46 | |
| *** absubram_ has joined #openstack-ironic | 17:47 | |
| JoshNang | hey guys, I'm getting a lot of "OperationalError: (OperationalError) no such table: nodes" when trying to use get_test_node, then create_node. I'm kind of confused, it's not supposed to be hitting a DB in testing, right? | 17:48 |
| *** matty_dubs|lunch is now known as matty_dubs | 17:48 | |
| devananda | JoshNang: unit testing? | 17:49 |
| JoshNang | yup! | 17:49 |
| *** epim has joined #openstack-ironic | 17:49 | |
| devananda | JoshNang: so, some of our unit tests definitely touch the DB. There's even a special class for those -- https://github.com/openstack/ironic/blob/master/ironic/tests/db/base.py#L22 | 17:50 |
| *** sseago has joined #openstack-ironic | 17:50 | |
| JoshNang | ah hah | 17:51 |
| JoshNang | thanks devananda! | 17:51 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Decorated commands no longer take command_name https://review.openstack.org/87147 | 17:51 |
| devananda | np! | 17:52 |
| JoshNang | heh main problem was subclassing the openstack testcase, rather than Ironics. thanks again! | 17:54 |
| devananda | ahh, that'd do it :) | 17:54 |
| *** eguz has quit IRC | 18:03 | |
| *** eghobo has joined #openstack-ironic | 18:04 | |
| *** jistr has quit IRC | 18:07 | |
| rloo | devananda, lucasagomes, I'm ahem, reading some old emails, http://lists.openstack.org/pipermail/openstack-dev/2014-March/030392.html | 18:11 |
| rloo | devananda, lucasagomes, we just made some changes to return nodes that are powered off, not in maintenance, but that email suggests we might want to return maintenance too? | 18:12 |
| *** lnxnut has joined #openstack-ironic | 18:12 | |
| devananda | rloo: indeed... thanks for pointing that out | 18:12 |
| devananda | rloo: to do what that email discusses, we need the nova driver to not merely receive a list of nodes both in and not in maintenance | 18:13 |
| devananda | but also to differentiate those two sets and tag some extra_specs onto the nodes in maintenance | 18:13 |
| rloo | how do we keep track of things-to-do? | 18:14 |
| devananda | and validate that the nova scheduler won't match it on a normal request, but that, with the --force-hosts extension, it will | 18:14 |
| devananda | that'll be fun to test :) | 18:14 |
| devananda | rloo: bugs! | 18:14 |
| devananda | rloo: launchpad bugs are really a crappy way to manage todo's, IMHO, but that's what we've got today... | 18:15 |
| devananda | krotscheck is working on storyboard, which will make it sooooo much better | 18:15 |
| rloo | devananda. but it isn't a bug, it's a feature :-). Ok, I'll open a bug about it | 18:15 |
| Shrews | and here i've been using notes on napkins. pfft | 18:15 |
| russell_h | devananda: does anything exist today for reaping stale node locks, should a process die before releasing a lock? | 18:15 |
| rloo | Shrews; as you as you do them, I'm fine if you use napkins ;) | 18:15 |
| devananda | russell_h: no. we've discussed this for about six months now and haven't reached a solution with enough support yet | 18:16 |
| russell_h | devananda: heh, ok | 18:17 |
| devananda | russell_h: tldr; if we make an API for that, it is an invitation to abuse it. | 18:17 |
| russell_h | yep | 18:17 |
| devananda | russell_h: so we need it to be automatic but *smart* | 18:17 |
| russell_h | what does "smart" mean here? | 18:17 |
| devananda | avoid split brain | 18:17 |
| russell_h | I mean, never wrong | 18:17 |
| russell_h | right | 18:17 |
| devananda | don't clobber a machine that's mid-firmware-flash | 18:17 |
| devananda | because, if you do, you brick it | 18:17 |
| devananda | so "never wrong" is a good starting point ;) | 18:18 |
| russell_h | ha | 18:18 |
| *** jbjohnso has joined #openstack-ironic | 18:18 | |
| devananda | at the moment, the work around is simple enough: mysql ironic -e "update nodes set reserved_by = NULL where ..." | 18:18 |
| devananda | so an operator can do it if they know they need to | 18:18 |
| devananda | (is that documented anywhere, besides IRC and napkins? probably not.) | 18:19 |
| russell_h | right, but in real life knowing that you need to sounds difficult | 18:19 |
| russell_h | at the very least you'd need to know that no one is going to hit your API while you're mucking around | 18:20 |
| devananda | russell_h: if this becomes a problem for operators, I would expect a cron of some sort that queries, eg, for | 18:20 |
| devananda | SELECT .. WHERE reservation not in (select distinct hostname from conductors) | 18:21 |
| devananda | or | 18:21 |
| devananda | SELECT .. WHERE updated_at < $old AND reservation IS NOT NULL | 18:21 |
| russell_h | the conductor doesn't hold a reservation while its waiting on a callback from a node right? | 18:21 |
| devananda | both of those could find dropped locks. it could even be done via the REST API -- you just can't release them that way | 18:21 |
| devananda | russell_h: correct | 18:22 |
| devananda | it sets the provision_state to waitcallback | 18:22 |
| russell_h | yep | 18:22 |
| russell_h | ok, I'll think on it, I think I agree its not urgent | 18:22 |
| devananda | :) | 18:23 |
| devananda | rloo: https://storyboard.openstack.org/#!/ | 18:23 |
| devananda | rloo: still fairly early in development, but krotscheck is makiing awsome progress on it | 18:24 |
| rloo | ooo, for inter-related projects. | 18:24 |
| devananda | rloo: i think for now, filing a Wishlist bug is fine, but in time, i'd like to start using ^ | 18:25 |
| rloo | devananda: ok, I'll follow your lead ;) | 18:25 |
| lifeless | NobodyCam: so https://review.openstack.org/#/c/87501/ - | 18:26 |
| *** zdiN0bot has quit IRC | 18:29 | |
| *** epim has quit IRC | 18:29 | |
| *** mdickson has quit IRC | 18:30 | |
| *** zdiN0bot has joined #openstack-ironic | 18:31 | |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Add optical and bios aliases for boot devices https://review.openstack.org/87682 | 18:31 |
| *** epim has joined #openstack-ironic | 18:32 | |
| jbjohnso | lucas-afk, there is the change including docstring. | 18:33 |
| jbjohnso | I for one am so elated that jenkins hates commits for punctuating the commit title | 18:34 |
| *** jdob_ has joined #openstack-ironic | 18:35 | |
| openstackgerrit | Pablo Fernando Cargnelutti proposed a change to openstack/ironic: Clean up of duplicated mandatory fields logic in JsonPatchType https://review.openstack.org/87755 | 18:40 |
| *** dwalleck has joined #openstack-ironic | 18:42 | |
| *** mdickson has joined #openstack-ironic | 18:46 | |
| *** lnxnut has quit IRC | 18:47 | |
| *** mdickson2 has joined #openstack-ironic | 18:49 | |
| *** mdickson has quit IRC | 18:50 | |
| comstud | I was wondering what the official 'stuck lock' solution was also :) | 18:51 |
| *** tatyana has joined #openstack-ironic | 18:51 | |
| comstud | besides OOB DB query | 18:51 |
| comstud | hehe | 18:51 |
| *** zdiN0bot has quit IRC | 18:52 | |
| *** romcheg1 has joined #openstack-ironic | 18:52 | |
| *** kevinbenton has joined #openstack-ironic | 18:56 | |
| *** overlayer has quit IRC | 18:57 | |
| *** mdickson2 has quit IRC | 19:01 | |
| *** mdickson has joined #openstack-ironic | 19:01 | |
| *** romcheg1 has quit IRC | 19:01 | |
| openstackgerrit | linggao proposed a change to openstack/ironic: Modify the get console API https://review.openstack.org/87760 | 19:01 |
| *** pcargnel has quit IRC | 19:02 | |
| *** dwalleck has quit IRC | 19:03 | |
| *** zigo has quit IRC | 19:04 | |
| *** zigo has joined #openstack-ironic | 19:07 | |
| *** romcheg1 has joined #openstack-ironic | 19:08 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Some minor clean up of various doc pages https://review.openstack.org/87765 | 19:10 |
| devananda | rloo: ^ might interest you | 19:10 |
| *** dwalleck has joined #openstack-ironic | 19:10 | |
| rloo | devananda: great. did you change alpha to beta? | 19:11 |
| devananda | yep! | 19:11 |
| devananda | and added a link to release notes | 19:11 |
| *** jdob_ has quit IRC | 19:11 | |
| rloo | good. I'll take a look in a few minutess. doing some "local" docn now :-( | 19:12 |
| *** lazy_prince has quit IRC | 19:12 | |
| *** dwalleck has quit IRC | 19:17 | |
| *** martyntaylor has left #openstack-ironic | 19:17 | |
| *** mdickson2 has joined #openstack-ironic | 19:23 | |
| rloo | devananda: https://review.openstack.org/#/c/83572/10/ironic/common/driver_factory.py. | 19:23 |
| jbjohnso | sigh.. I can't scan a server in under half a second for failures | 19:23 |
| rloo | devananda: maybe I'm tired, but the driver that raises the DriverNotFound isn't the one that captures the exception. | 19:24 |
| jbjohnso | 500 ms for this particular server... | 19:24 |
| devananda | rloo: i'm not sure what you mean. the driver(s) raise it, it's caught by a hook in stevedore | 19:24 |
| rloo | devananda. so 'we capture that exception' == stevedore captures it, and who logs the helpful message? | 19:25 |
| devananda | rloo: the hook we pass to stevedore | 19:25 |
| devananda | ah | 19:25 |
| rloo | devananda. the hook == _catch_driver_not_found? | 19:25 |
| devananda | rloo: so i removed the "more helpful message" somewhere along the patch series | 19:25 |
| devananda | rloo: yes | 19:26 |
| rloo | that's where I'm confused. | 19:26 |
| devananda | rloo: let me clean up the comment. functionally, it's sound, but the comment is misleading | 19:26 |
| devananda | thanks! | 19:26 |
| rloo | ok, thx! | 19:26 |
| *** mdickson has quit IRC | 19:26 | |
| rloo | devananda. stevedore isn't logging any message, is it? | 19:26 |
| devananda | rloo: start the conductor without that patch, yoyu'll get a huge traceback -- from stevedore | 19:27 |
| devananda | grabbing a random log from jenkins | 19:27 |
| devananda | rloo: http://logs.openstack.org/87/85387/4/check/check-tempest-dsvm-virtual-ironic/3bb8931/logs/screen-ir-cond.txt.gz | 19:27 |
| rloo | devananda. that's the log w/o this change right? | 19:28 |
| devananda | rloo: compare to after that patch -- http://logs.openstack.org/72/83572/10/check/check-tempest-dsvm-virtual-ironic/86bf3a8/logs/screen-ir-cond.txt.gz | 19:28 |
| devananda | yes | 19:28 |
| rloo | geez, how are you so fast with the links? | 19:28 |
| devananda | had two patches open already | 19:28 |
| rloo | so the 2nd one, with the patch, doesn't log any msg at all? | 19:29 |
| devananda | rloo: it logs an INFO | 19:29 |
| devananda | showing which drivers were loaded | 19:29 |
| *** epim has quit IRC | 19:30 | |
| rloo | devananda. right. i thought the comment in your patch 'log a more helpful message' was referring to some msg about the driver-not-found-due-to-seamicro-missing-library | 19:30 |
| *** zdiN0bot has joined #openstack-ironic | 19:30 | |
| devananda | ahhh | 19:31 |
| devananda | lemme demonstrate | 19:31 |
| rloo | (just not clear to me what the comment is referring to, that's all.) | 19:31 |
| openstackgerrit | linggao proposed a change to openstack/python-ironicclient: node-get-console command use the new API https://review.openstack.org/87769 | 19:32 |
| devananda | rloo: so you've seen what the traceback looks like before the patch (and the service continues) | 19:33 |
| devananda | rloo: after the patch, if I explicitly set enabled_drivers=pxe_seamicro, i get this: http://paste.openstack.org/show/75813/ | 19:33 |
| rloo | devananda: yup. | 19:33 |
| devananda | so that's the helpful message | 19:33 |
| devananda | and when loading is successful, all that i get is | 19:34 |
| rloo | ahh, ok, thx. that explains it. | 19:34 |
| devananda | 2014-04-15 12:33:56.236 19437 INFO ironic.common.driver_factory [-] Loaded the following drivers: ['fake', 'pxe_ssh', 'pxe_ipmitool'] | 19:34 |
| devananda | hmm | 19:34 |
| devananda | ok, there's a missing unit test | 19:35 |
| *** romcheg1 has quit IRC | 19:35 | |
| devananda | service.start() should (and does) fail now, if a driver can't be loaded | 19:35 |
| devananda | but i dont see a test for that | 19:35 |
| rloo | personally, i'd get rid of the 'capture that exception and log a more helpful message'. that's really only useful if you knew that the messages before weren't that helpful. | 19:35 |
| devananda | yep | 19:35 |
| *** romcheg1 has joined #openstack-ironic | 19:36 | |
| rloo | does service.start() fail with a fatal exception? i think I commented about that (a test for it) | 19:36 |
| devananda | it does | 19:37 |
| rloo | can't even remember now what i think it is a fatal exception; must have read it somewhere... | 19:37 |
| rloo | oh yeah, the description for the new config 'enabled_drivers' | 19:38 |
| *** epim has joined #openstack-ironic | 19:41 | |
| openstackgerrit | Vladimir Kozhukalov proposed a change to openstack/ironic-python-agent: Added disk partitioner https://review.openstack.org/86163 | 19:42 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Better handling of missing drivers https://review.openstack.org/83572 | 19:55 |
| *** Mikhail_D_ltp has quit IRC | 20:00 | |
| rloo | devananda: the user guide has some formatting issues; are you going to leave those for the original author? http://docs.openstack.org/developer/ironic/deploy/userguide.html | 20:02 |
| devananda | rloo: i didn't want to tackle that in this patch | 20:02 |
| rloo | devananda: ok, no worries. | 20:02 |
| devananda | rloo: but i'll be happy to see anyone address them :) | 20:02 |
| devananda | also, just updated https://wiki.openstack.org/wiki/Ironic | 20:03 |
| rloo | devananda: I'll be happy too :-) | 20:04 |
| devananda | rloo: thanks :) | 20:04 |
| rloo | ha ha | 20:04 |
| rloo | devananda. it isn't actually clear to me what should go in wiki.openstack.org, and what should go in docs.openstack.org/developer. email/irc aren't mentioned in the wiki? | 20:06 |
| devananda | rloo: it's not clear to me either, but as a rule of thumb, i'm going with | 20:07 |
| devananda | if it should change when the code changes -- it goes in docs | 20:07 |
| devananda | if it changes in relation to other things (like release or election cycles) it goes in the wiki | 20:07 |
| devananda | or if it just changes when ever we want it to change -- again, wiki | 20:08 |
| rloo | ahh. so the contributing part should go in the wiki. | 20:08 |
| rloo | (not that i care to move that info.) | 20:08 |
| devananda | http://docs.openstack.org/developer/ironic/dev/contributing.html | 20:08 |
| devananda | yea, a lot of that is duplicated in the wiki now | 20:09 |
| rloo | and I copied a lot of that for python-ironicclient ;) | 20:09 |
| *** zdiN0bot has quit IRC | 20:10 | |
| rloo | have to leave something for others to clean up! | 20:10 |
| devananda | :) | 20:10 |
| devananda | actually i dont see things on dev/contributing that need to be changed / removed | 20:11 |
| devananda | that's all fairly static links | 20:11 |
| rloo | so with your rule of thumb, it belongs in wiki ;) | 20:11 |
| devananda | things like "link to current release notes" -- my sense is that doesn't belong in our developer doc tree -- if anything, it may eventually go in a versioned doc repo | 20:11 |
| devananda | oh | 20:11 |
| devananda | hah, i see how you interpreted that | 20:11 |
| rloo | oh that reminds me. are we going to have release notes for python-ironicclient? | 20:12 |
| devananda | if it is unlikely to change, or should change explicitly when the code changes -- it goes to the in-tree developer docs | 20:12 |
| devananda | rloo: we do. on pypi | 20:12 |
| rloo | devananda: https://pypi.python.org/pypi/python-ironicclient. Where? | 20:13 |
| devananda | er, yea. not there. sec | 20:13 |
| devananda | rloo: each release has release notes, eg http://git.openstack.org/cgit/openstack/python-ironicclient/tag/?id=0.1.3 | 20:14 |
| devananda | i'm not sure why those aren't visible on pypi | 20:14 |
| rloo | devananda. Good to know. Thx. | 20:15 |
| devananda | rloo: oooh, shiny! https://pypi.python.org/pypi/python-novaclient | 20:16 |
| devananda | rloo: so nova doesn't have a release notes up there either, but they do have pretty docs on pypi that we don't | 20:16 |
| rloo | devananda: yeah, I noticed that too. | 20:17 |
| rloo | devananda: the shiny stuff comes from the README. We don't have much there, but feel free to add. | 20:18 |
| devananda | ;) | 20:19 |
| rloo | It seems to me that unless we want to update the README all the time, maybe a link to the dev docs is sufficient. | 20:19 |
| rloo | I mean, once we have dev docs ;) | 20:19 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Adding a reference driver for the agent https://review.openstack.org/84795 | 20:26 |
| *** lnxnut has joined #openstack-ironic | 20:34 | |
| devananda | rloo: if the readme is mostly just the output of "ironic --help | 20:35 |
| *** lnxnut has quit IRC | 20:35 | |
| devananda | " then that's easy enough for me to script updating it every time i do a release | 20:35 |
| *** lnxnut has joined #openstack-ironic | 20:35 | |
| rloo | devananda: I think the readme is whatever we want to add ;). If I were a user, I'd want to know how to use it so yes, ironic --help would be useful. | 20:36 |
| rloo | and to know what versions/changes. that info should be avail somewhere... just need to find it. | 20:37 |
| rloo | devananda: fyi, the doc folks have a script that will generate xml docn for the 'ironic' command. | 20:38 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Adding swift temp url support https://review.openstack.org/81391 | 20:48 |
| *** jdob has quit IRC | 20:58 | |
| *** harlowja_ has quit IRC | 20:58 | |
| *** harlowja has joined #openstack-ironic | 21:03 | |
| *** tatyana_ has joined #openstack-ironic | 21:06 | |
| *** zigo has quit IRC | 21:06 | |
| *** epim has quit IRC | 21:07 | |
| *** tatyana has quit IRC | 21:07 | |
| *** tatyana_ is now known as tatyana | 21:07 | |
| *** zigo has joined #openstack-ironic | 21:08 | |
| *** zdiN0bot has joined #openstack-ironic | 21:10 | |
| *** linggao has quit IRC | 21:12 | |
| *** zdiN0bot has quit IRC | 21:15 | |
| *** dwalleck has joined #openstack-ironic | 21:16 | |
| *** jbjohnso has quit IRC | 21:16 | |
| *** epim has joined #openstack-ironic | 21:17 | |
| *** coolsvap|afk has quit IRC | 21:25 | |
| *** coolsvap|afk has joined #openstack-ironic | 21:26 | |
| *** matty_dubs is now known as matty_dubs|gone | 21:28 | |
| NobodyCam | hello From NeedVille Tx :) | 21:29 |
| * dwalleck checks to see if that's actually a town | 21:32 | |
| dwalleck | But since it ends in -ville, I'll buy it | 21:32 |
| dwalleck | and damn, it is. You're one giant open area away | 21:36 |
| adam_g | dwalleck, FYI the baremetal scenario: it ended up being a combo of firewalling and lack of sudo for $STACK_USER | 21:36 |
| *** absubram_ has quit IRC | 21:37 | |
| dwalleck | adam_g: That makes sense since it worked when we both deployed it | 21:37 |
| adam_g | dwalleck, yeah.. devstack-gate drops $STACK_USER's sudo privs as soon as its done with devstack | 21:38 |
| *** blamar_ has joined #openstack-ironic | 21:38 | |
| *** blamar has quit IRC | 21:38 | |
| *** blamar_ is now known as blamar | 21:38 | |
| *** zdiN0bot has joined #openstack-ironic | 21:39 | |
| dwalleck | Huh, so the stack user for devstack-gate isn't a passwordless sudo user like in most other Devstack deployments? | 21:39 |
| *** mrda_away is now known as mrda | 21:43 | |
| adam_g | dwalleck, it is pre and during the devstack run, but the jenkins job removes its sudo privs before running tempest | 21:49 |
| adam_g | dwalleck, which uncovered that we were missing https://review.openstack.org/#/c/87408/ | 21:50 |
| dwalleck | Ahhh, now I get it | 21:50 |
| dwalleck | Good catch! | 21:51 |
| NobodyCam | lifeless: I just saw your comment | 21:53 |
| NobodyCam | dwalleck: We are docked at a friends farm here in needville | 21:55 |
| *** epim has quit IRC | 21:56 | |
| *** zdiN0bot has quit IRC | 21:58 | |
| *** epim has joined #openstack-ironic | 22:00 | |
| *** rupsky has quit IRC | 22:01 | |
| *** zdiN0bot has joined #openstack-ironic | 22:02 | |
| *** epim has quit IRC | 22:02 | |
| *** harlowja has quit IRC | 22:03 | |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Drivers determine acceptable power states https://review.openstack.org/86744 | 22:04 |
| *** harlowja has joined #openstack-ironic | 22:05 | |
| devananda | soliciting feedback on https://wiki.openstack.org/wiki/Ironic/ReleaseNotes/Icehouse | 22:10 |
| dividehex | can someone help me wrap my head around how neutron ( and neutron dhcp-agent) is exposed to the bare metal nodes? | 22:10 |
| *** dwalleck_ has joined #openstack-ironic | 22:10 | |
| devananda | dividehex: sure. whatcha got so far? | 22:10 |
| dividehex | devananda: i've been using the old nova baremetal driver and i'm looking at move to ironic. on the old baremetal, i'm running a dnsmasq w/ static ip injection. I'm also running neurton (ovs + gre). How does ironic handle this? | 22:14 |
| *** dwalleck has quit IRC | 22:14 | |
| *** tatyana has quit IRC | 22:15 | |
| devananda | dividehex: ironic relies on neutron for dhcp, and for setting the dhcp boot option | 22:16 |
| devananda | dividehex: so, no dnsmasq. no static IP injection. no file injection of any sort. | 22:16 |
| *** romcheg1 has quit IRC | 22:16 | |
| devananda | http://docs.openstack.org/developer/ironic/api/ironic.common.neutron.html | 22:16 |
| devananda | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L493 | 22:17 |
| devananda | during deploy, the PXE driver updates Neutron and changes the port dhcp options | 22:18 |
| dividehex | devananda: thats awesome. although i'm not sure how to expose neutron dhcp-agent to the switch. Do i need to do some ovs brigde magic? or does the dhcp agent run directly on the ethX device attached to the baremetal node switch? | 22:20 |
| NobodyCam | devananda: could key feature #2 (Heterogeneous hardware support.) be confused with virt and baremetal in the same AZ? | 22:20 |
| devananda | dividehex: "expose neutron dhcpagent to the switch" -- this confuses me slightly | 22:23 |
| devananda | dividehex: is neutron-dhcpagent not on the same L2 broadcast domain as your bare metal nodes? | 22:23 |
| devananda | NobodyCam: thanks, i'll clarify that | 22:23 |
| NobodyCam | other then that I really like it :) | 22:24 |
| *** overlayer has joined #openstack-ironic | 22:25 | |
| dividehex | devananda: i might be confused about how neutron dhcpagent works. My only experience Ive seen is L2 being isolated to the virt compute node | 22:25 |
| *** athomas has quit IRC | 22:27 | |
| dividehex | devananda: is there any documentation regarding setting up neutron on a ironic conductor host or ironic environment? | 22:28 |
| devananda | dividehex: hm, not that i'm aware of. TripleO is doing that quite successfully, though | 22:29 |
| devananda | dividehex: in this case, the L2 is not isolated to a single host -- there is no virt compute node, after all. | 22:29 |
| *** zdiN0bot has quit IRC | 22:31 | |
| devananda | dividehex: all the testing I'm aware of to date relies on defining an AZ by putting all baremetal nodes into the same L2 domain, though you should be able to run multiple neutrondhcp agents. I don't know if anyone's tried spanning multiple L2 domains with Ironic (in a single AZ) yet. | 22:31 |
| devananda | NobodyCam: feel like reviewing https://review.openstack.org/#/c/86971 ? | 22:34 |
| dividehex | devananda: thanks for the help. I'll keep poking through docs and probably take a look at what tripleO is doing. I've been tasked with building out a baremetal provisioning system for out CI environment. So I appreciate all the help and advice :} | 22:34 |
| * NobodyCam looks | 22:34 | |
| devananda | dividehex: great! please don't hesitate to ask -- folks in here are generally very helpful | 22:35 |
| dividehex | thanks! | 22:35 |
| devananda | dividehex: you also might be interested in our upstream CI work. devstack can stand up Ironic. configure neutron, etc... | 22:35 |
| dividehex | ooohh that sounds interesting. do you have a link? | 22:36 |
| devananda | dividehex: hmm. well. it's in devstack, so... :) | 22:36 |
| dividehex | i haven't tried devstack yet. I take a look there | 22:37 |
| devananda | https://wiki.openstack.org/wiki/Ironic/Devstack | 22:37 |
| devananda | actually | 22:37 |
| devananda | that's a bit out of date | 22:38 |
| * devananda dislikes wiki's for that reason | 22:38 | |
| dwalleck_ | I've realize that anything I don't store in a repository is always old | 22:39 |
| devananda | dwalleck_: exactly | 22:39 |
| devananda | dividehex: https://etherpad.openstack.org/p/IronicAndDevstackAgain | 22:39 |
| dwalleck_ | And people don't like old data =P No good deed goes unpunished | 22:39 |
| devananda | dividehex: it's focused on our CI environment, but probably has enough data to get you started | 22:39 |
| dwalleck_ | adam_g actually has a better one....there's one in the Ironic docs | 22:39 |
| dividehex | devananda: excellent. Thank you | 22:40 |
| devananda | dividehex: if you've never used devstack before, here's a tip: *dont* run it directly on your workstation! | 22:40 |
| dwalleck_ | https://github.com/openstack/ironic/blob/master/doc/source/dev/dev-quickstart.rst is what I've been using | 22:40 |
| dividehex | haha.. i'll make sure to spin up a vm for it. | 22:40 |
| devananda | dwalleck_: ah! right. I was just looking at that earlier today, heh | 22:41 |
| * devananda annotates the wiki and 'pad to stop directing ppl there | 22:41 | |
| *** epim has joined #openstack-ironic | 22:48 | |
| *** iron1 has quit IRC | 23:00 | |
| *** jgrimm has quit IRC | 23:01 | |
| *** eguz has joined #openstack-ironic | 23:16 | |
| *** eguz has joined #openstack-ironic | 23:16 | |
| *** eghobo has quit IRC | 23:20 | |
| *** radsy has joined #openstack-ironic | 23:21 | |
| *** radsy has joined #openstack-ironic | 23:21 | |
| devananda | rloo: you said there's a tool to convert .conf into xml doc'n -- would that work for eg. http://git.openstack.org/cgit/openstack/ironic/tree/etc/ironic/ironic.conf.sample ? | 23:22 |
| rloo | devananda: the tool converts 'ironic' command to xml doc. sec, let me give you an example. | 23:23 |
| rloo | devananda: http://docs.openstack.org/cli-reference/content/cinderclient_commands.html | 23:24 |
| *** eguz has quit IRC | 23:24 | |
| devananda | ah, that's handy too | 23:25 |
| *** lucas-afk has quit IRC | 23:25 | |
| rloo | devananda: there is docn about config settings too, I didn't ask the doc folks how they generated those. I assume they grab from the .conf but don't know if they have a script or not. | 23:26 |
| devananda | rloo: i was thinking of eg http://docs.openstack.org/havana/config-reference/content/configuring-openstack-compute-basics.html#section_compute-config-overview | 23:26 |
| devananda | yea | 23:26 |
| rloo | devananda. yup. i can ask. matty_dubs started writing something like that, but I think there must be some automated way... | 23:27 |
| devananda | so eventually we'll need all that automata. I presume we'll be able to learn a lot more at the cross project doc'n sessions in atlanta | 23:27 |
| rloo | devananda. yup. when I 'spoke' with anne last week, she apologized for not having any docn for projects to do docn. She seemed really busy so I figured I'd ask more questions later. | 23:28 |
| devananda | heh | 23:28 |
| devananda | that's ironic :) | 23:28 |
| *** eghobo has joined #openstack-ironic | 23:28 | |
| rloo | devananda. I had this thought that I'd snoop around their stuff to figure things out, but ahh, I guess, that's low in my list of things I want to do soon. | 23:29 |
| openstackgerrit | A change was merged to openstack/ironic: Refactor the driver fields mapping https://review.openstack.org/86971 | 23:30 |
| rloo | devananda. ie how docn do we need "now", vs "later"? If we have enough now, I'd rather postpone the rest for later. | 23:30 |
| *** epim has quit IRC | 23:31 | |
| devananda | rloo: we have users, increasingly many of them | 23:31 |
| *** zdiN0bot has joined #openstack-ironic | 23:32 | |
| rloo | devananda: so it'd be good to find out from doc folks, how soon we can leverage their scripts etc, to get docn up. | 23:32 |
| *** openstackstatus has quit IRC | 23:32 | |
| devananda | i'm not sure how to quantify "how much" we need, though. clearly, we aren't integrated with the main openstack-doc projects right now... but, i believe, that is a goal of the incubation process | 23:32 |
| devananda | so that, when the TC votes to graduate ironic, we're already doc'd in all the right places and ways | 23:32 |
| *** openstackstatus has joined #openstack-ironic | 23:32 | |
| devananda | and, to users, feel like just another openstack project | 23:33 |
| devananda | that hasn't historically been the case ... but it's the goal the TC is pushing new projects towards | 23:33 |
| devananda | (and existing ones that aren't there yet) | 23:33 |
| *** epim has joined #openstack-ironic | 23:33 | |
| rloo | yes, it'd be nice to have docs before graduation. but we should since graduation is weeks away. | 23:34 |
| devananda | hm? | 23:34 |
| rloo | ie, we have time before graduation to get the docs done ;) | 23:34 |
| devananda | s/weeks/months/ -- yes | 23:35 |
| rloo | i'm being optimistic ;) | 23:35 |
| rloo | besides, the weeks go by so quickly. | 23:35 |
| devananda | lol | 23:35 |
| devananda | that they do | 23:35 |
| devananda | 20 weeks really isn't that long ;) | 23:35 |
| rloo | only 20?! better get movin' | 23:36 |
| *** zdiN0bot has quit IRC | 23:36 | |
| *** epim has quit IRC | 23:36 | |
| rloo | as far as the configuration docn goes, the stuff at http://docs.openstack.org/havana/config-reference/content/configuring-openstack-compute-basics.html#section_compute-config-overview looks mostly from the .conf.sample, so users can really just look at the .sample. | 23:37 |
| devananda | that's what i figured ... | 23:38 |
| rloo | what i think we need to do is make sure the docstrings for the config options are well-written. | 23:38 |
| *** eguz has joined #openstack-ironic | 23:39 | |
| rloo | the only additional useful info (from what i can tell) are the sample config settings. | 23:39 |
| *** nosnos has joined #openstack-ironic | 23:39 | |
| *** eghobo has quit IRC | 23:43 | |
| *** overlayer has quit IRC | 23:50 | |
| *** rloo has quit IRC | 23:54 | |
| *** mordred has quit IRC | 23:58 | |
| *** mordred has joined #openstack-ironic | 23:58 | |
| *** ChanServ changes topic to "Restarting gerrit really quick to fix replication issue" | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!