| devananda | heading out for a smoothie and a movie, ttyatmw | 00:14 |
|---|---|---|
| *** matsuhashi has joined #openstack-ironic | 00:15 | |
| *** datajerk has joined #openstack-ironic | 00:20 | |
| openstackgerrit | A change was merged to openstack/ironic: Accessing a subresource that parent does not exist https://review.openstack.org/54742 | 00:23 |
| *** datajerk has quit IRC | 00:25 | |
| rloo | devananda: ^^^ thanks! | 00:42 |
| rloo | NobodyCam: thank you too! | 00:42 |
| NobodyCam | :-p | 00:43 |
| NobodyCam | you did all the great work | 00:43 |
| *** hemna has quit IRC | 00:45 | |
| rloo | NobodyCam: to be honest, devananda is great (and a great communicator), and the feedback/comments I got from everyone really helped. | 00:47 |
| NobodyCam | :) | 00:47 |
| NobodyCam | seems the gate jobs are acting up today... -infra is work on thing as we type. so it should land soon (-ish) | 00:48 |
| NobodyCam | s/work/working/ | 00:49 |
| rloo | thx NobodyCam. I've waited this long. No worries :-) | 00:49 |
| openstackgerrit | A change was merged to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/56744 | 00:50 |
| openstackgerrit | A change was merged to openstack/ironic: Changes power_state and adds last_error field https://review.openstack.org/54466 | 00:51 |
| *** matsuhashi has quit IRC | 01:01 | |
| *** matsuhashi has joined #openstack-ironic | 01:02 | |
| *** matsuhashi has quit IRC | 01:02 | |
| *** matsuhashi has joined #openstack-ironic | 01:02 | |
| *** nosnos has joined #openstack-ironic | 01:21 | |
| *** sjing has joined #openstack-ironic | 01:25 | |
| *** sjing has quit IRC | 01:29 | |
| *** sjing has joined #openstack-ironic | 01:30 | |
| *** michchap has quit IRC | 01:38 | |
| *** michchap has joined #openstack-ironic | 01:38 | |
| *** epim has quit IRC | 01:50 | |
| *** xmltok has joined #openstack-ironic | 01:53 | |
| *** xmltok has quit IRC | 02:01 | |
| openstackgerrit | Haomeng,Wang proposed a change to openstack/ironic: Import missing gettext _ to fix Sphinx error https://review.openstack.org/56605 | 02:04 |
| openstackgerrit | Ruby Loo proposed a change to openstack/python-ironicclient: Shows 'last_error' property for a node https://review.openstack.org/57087 | 03:03 |
| *** sjing has quit IRC | 03:05 | |
| *** rloo has quit IRC | 03:05 | |
| *** sjing has joined #openstack-ironic | 03:06 | |
| *** epim has joined #openstack-ironic | 03:10 | |
| *** coolsvap has joined #openstack-ironic | 03:15 | |
| *** matsuhashi has quit IRC | 03:29 | |
| *** matsuhashi has joined #openstack-ironic | 03:30 | |
| *** matsuhashi has quit IRC | 03:35 | |
| sandeepr | ping NobodyCam | 04:19 |
| *** matsuhashi has joined #openstack-ironic | 04:46 | |
| *** nosnos has quit IRC | 05:23 | |
| *** nosnos has joined #openstack-ironic | 05:23 | |
| *** prekarat has joined #openstack-ironic | 05:26 | |
| *** prekarat has quit IRC | 05:35 | |
| *** prekarat has joined #openstack-ironic | 05:35 | |
| *** michchap has quit IRC | 06:02 | |
| *** michchap has joined #openstack-ironic | 06:04 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/57099 | 06:10 |
| *** matsuhashi has quit IRC | 06:41 | |
| *** matsuhashi has joined #openstack-ironic | 06:42 | |
| *** prekarat has quit IRC | 06:43 | |
| *** matsuhashi has quit IRC | 06:46 | |
| *** matsuhashi has joined #openstack-ironic | 06:48 | |
| *** epim has quit IRC | 06:56 | |
| *** epim has joined #openstack-ironic | 06:57 | |
| *** michchap has quit IRC | 07:03 | |
| *** michchap has joined #openstack-ironic | 07:04 | |
| *** nosnos_ has joined #openstack-ironic | 07:05 | |
| *** nosnos has quit IRC | 07:05 | |
| *** prekarat has joined #openstack-ironic | 07:07 | |
| *** matsuhashi has quit IRC | 07:09 | |
| *** matsuhashi has joined #openstack-ironic | 07:10 | |
| *** matsuhas_ has joined #openstack-ironic | 07:13 | |
| *** matsuhashi has quit IRC | 07:14 | |
| *** arata has joined #openstack-ironic | 07:17 | |
| *** romcheg has joined #openstack-ironic | 07:28 | |
| *** epim has quit IRC | 07:28 | |
| GheRivero | morning Ironic | 07:34 |
| *** sjing has quit IRC | 07:36 | |
| *** sjing has joined #openstack-ironic | 07:38 | |
| Haomeng | morning GheRivero | 07:47 |
| *** tatyana has quit IRC | 07:48 | |
| *** tatyana has joined #openstack-ironic | 07:49 | |
| Haomeng | GheRivero:ping | 07:55 |
| GheRivero | Haomeng: pong | 07:56 |
| Haomeng | Got your comments about my patch, will add into tox.ini as exception | 07:56 |
| Haomeng | thanks for your good idea:) | 07:56 |
| GheRivero | you are welcome :) | 07:56 |
| Haomeng | but not sure if this is working for our Sphinx error | 07:57 |
| Haomeng | let me have a try and update the result to you | 07:57 |
| GheRivero | btw, I have a comment in https://review.openstack.org/#/c/53991/7/ironicclient/v1/port_shell.py | 07:57 |
| GheRivero | are the fields print in order? | 07:58 |
| Haomeng | let me check | 07:58 |
| Haomeng | I am not sure the printing order, but this order is based on lifeless comments, he think this is easy to read and scan | 08:00 |
| GheRivero | ok... i-ll check it anyway | 08:01 |
| Haomeng | ok, thank you | 08:01 |
| Haomeng | I want to write an unittest for shell, to cover the shell output for these new fields involved by this patch | 08:02 |
| *** ndipanov_gone is now known as ndipanov | 08:02 | |
| GheRivero | nice :) | 08:05 |
| Haomeng | :) | 08:08 |
| Haomeng | GheRivero: I will setup new env to test if our toxini solution work for our doc-auto-gen with Sphinx by command "python setup.py build_sphinx ", how do you think if build_sphinx action will use our env controled by tox.ini ? | 08:13 |
| lifeless | Haomeng: oh, if it controls the UI order thats different | 08:15 |
| lifeless | Haomeng: I thought it was just an in-source list | 08:15 |
| Haomeng | ok, let me test | 08:17 |
| Haomeng | I think it is just the order in our source, not the output order | 08:18 |
| *** romcheg has quit IRC | 08:19 | |
| Haomeng | lifeless: I am trying to add shell command unittest, but not sure how to disable our keystone client connection in our unittest env, do you have idea? | 08:19 |
| *** coolsvap has quit IRC | 08:30 | |
| *** tatyana has quit IRC | 08:36 | |
| *** arata has left #openstack-ironic | 08:39 | |
| *** romcheg has joined #openstack-ironic | 08:40 | |
| *** tatyana has joined #openstack-ironic | 08:45 | |
| GheRivero | Haomeng: lifeless: the output is alphabatically ordered, the same that in other projects, so no need to touch anything | 08:47 |
| Haomeng | ok, thanks | 08:47 |
| Haomeng | I am running "python setup.py build_sphinx" to generate our api doc,got exception "error: 'source_dir' must be a directory name (got `/root/1251011/tox_solu/ironic/doc/source`", do you have idea? | 08:48 |
| GheRivero | no, sorry | 08:50 |
| Haomeng | ok, np:) | 08:50 |
| openstackgerrit | Haomeng,Wang proposed a change to openstack/ironic: Import missing gettext _ to fix Sphinx error https://review.openstack.org/56605 | 08:53 |
| *** jistr has joined #openstack-ironic | 08:55 | |
| openstackgerrit | Sascha Peilicke proposed a change to openstack/ironic: Support building wheels (PEP-427) https://review.openstack.org/57122 | 08:56 |
| *** tatyana has quit IRC | 08:59 | |
| *** ndipanov has quit IRC | 09:03 | |
| *** ndipanov has joined #openstack-ironic | 09:04 | |
| *** tatyana has joined #openstack-ironic | 09:05 | |
| *** lucasagomes has joined #openstack-ironic | 09:06 | |
| *** derekh has joined #openstack-ironic | 09:11 | |
| *** ndipanov_ has joined #openstack-ironic | 09:12 | |
| *** ndipanov has quit IRC | 09:13 | |
| openstackgerrit | Sascha Peilicke proposed a change to openstack/python-ironicclient: Support building wheels (PEP-427) https://review.openstack.org/57142 | 09:16 |
| *** ndipanov_ is now known as ndipanov | 09:19 | |
| *** ndipanov has joined #openstack-ironic | 09:19 | |
| *** tatyana has quit IRC | 09:21 | |
| *** michchap has quit IRC | 09:24 | |
| *** matsuhas_ has quit IRC | 09:25 | |
| *** matsuhashi has joined #openstack-ironic | 09:25 | |
| *** michchap has joined #openstack-ironic | 09:26 | |
| *** matsuhashi has quit IRC | 09:30 | |
| *** matsuhashi has joined #openstack-ironic | 09:30 | |
| *** nosnos_ has quit IRC | 09:33 | |
| *** nosnos has joined #openstack-ironic | 09:34 | |
| openstackgerrit | Sascha Peilicke proposed a change to stackforge/pyghmi: Support building wheels (PEP-427) https://review.openstack.org/57177 | 09:45 |
| *** coolsvap has joined #openstack-ironic | 10:03 | |
| *** sjing has quit IRC | 10:04 | |
| *** arata has joined #openstack-ironic | 10:07 | |
| *** prekarat1 has joined #openstack-ironic | 10:32 | |
| *** prekarat has quit IRC | 10:33 | |
| *** prekarat1 has quit IRC | 10:36 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: API ports update for WSME 0.5b6 compliance https://review.openstack.org/56682 | 10:37 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Rework patch validation on Ports https://review.openstack.org/56984 | 10:40 |
| *** prekarat has joined #openstack-ironic | 10:57 | |
| *** matsuhashi has quit IRC | 10:58 | |
| *** matsuhashi has joined #openstack-ironic | 10:59 | |
| *** prekarat has quit IRC | 11:03 | |
| *** matsuhashi has quit IRC | 11:04 | |
| *** coolsvap has quit IRC | 11:07 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/python-ironicclient: Replace node_id to node_uuid on Ports https://review.openstack.org/57194 | 11:16 |
| openstackgerrit | A change was merged to openstack/ironic: Sync common utils from Oslo https://review.openstack.org/56517 | 11:21 |
| openstackgerrit | A change was merged to openstack/ironic: sync common.local from oslo https://review.openstack.org/56518 | 11:21 |
| openstackgerrit | A change was merged to openstack/ironic: sync common.log from oslo https://review.openstack.org/56520 | 11:23 |
| openstackgerrit | A change was merged to openstack/ironic: sync common.notifier.* from oslo https://review.openstack.org/56521 | 11:25 |
| openstackgerrit | A change was merged to openstack/ironic: sync common.periodic_task from oslo https://review.openstack.org/56522 | 11:25 |
| openstackgerrit | A change was merged to openstack/ironic: sync common.service from oslo https://review.openstack.org/56523 | 11:25 |
| romcheg | Morning folks | 11:29 |
| shadower | hey | 11:30 |
| GheRivero | morning | 11:38 |
| Haomeng | morning | 11:43 |
| Haomeng | :) | 11:43 |
| lucasagomes | morning :) | 11:45 |
| *** ben_duyujie has joined #openstack-ironic | 11:48 | |
| *** arata has left #openstack-ironic | 11:57 | |
| max_lobur | Morning Ironic | 12:00 |
| lucasagomes | morning max_lobur | 12:02 |
| *** lucasagomes is now known as lucas-hungry | 12:02 | |
| Haomeng | morning, max_lobur | 12:17 |
| max_lobur | hi Haomeng | 12:18 |
| Haomeng | :) | 12:18 |
| Haomeng | lucasagomes: per your comments, that is good idea, I am trying to add our ironic shell test for my patch - https://review.openstack.org/#/c/53991/ | 12:40 |
| Haomeng | need some code to cover our ironic client shell unittest | 12:41 |
| *** prekarat has joined #openstack-ironic | 12:45 | |
| *** prekarat has quit IRC | 12:48 | |
| *** prekarat has joined #openstack-ironic | 12:48 | |
| Haomeng | tired, will go to sleep, nice day ironic team:) | 12:59 |
| *** gfa has left #openstack-ironic | 12:59 | |
| max_lobur | thanks, enjoy your rest :) | 13:00 |
| max_lobur | ping lucasagomes | 13:02 |
| Haomeng | :) | 13:03 |
| *** rloo has joined #openstack-ironic | 13:29 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Add last_error usage to deploy and teardown methods https://review.openstack.org/57227 | 13:31 |
| romcheg | Done the BP about RPC routing | 13:33 |
| romcheg | https://blueprints.launchpad.net/ironic/+spec/rpc-routing | 13:33 |
| romcheg | devananda: Please check it out when you're available ^ | 13:34 |
| GheRivero | \o/ | 13:34 |
| *** nosnos has quit IRC | 13:35 | |
| *** jdob has joined #openstack-ironic | 13:37 | |
| max_lobur | congrats! :) | 13:37 |
| *** lucas-hungry is now known as lucasagomes | 13:46 | |
| lucasagomes | max_lobur, Haomeng pong... sorry I was having lunch | 13:47 |
| lucasagomes | I gotta a call right now too, so I will be right back | 13:47 |
| max_lobur | k | 13:47 |
| *** linggao has joined #openstack-ironic | 13:52 | |
| *** arata has joined #openstack-ironic | 13:55 | |
| *** coolsvap has joined #openstack-ironic | 14:02 | |
| *** arata has left #openstack-ironic | 14:04 | |
| *** ben_duyujie has quit IRC | 14:14 | |
| *** jdob has quit IRC | 14:30 | |
| *** jdob has joined #openstack-ironic | 14:30 | |
| *** jdob has quit IRC | 14:44 | |
| *** jdob has joined #openstack-ironic | 14:44 | |
| *** lexx has joined #openstack-ironic | 14:49 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Add last_error usage to deploy and teardown methods https://review.openstack.org/57227 | 14:55 |
| *** yuriyz has joined #openstack-ironic | 14:58 | |
| *** datajerk has joined #openstack-ironic | 15:01 | |
| NobodyCam | good morning Ironic | 15:05 |
| max_lobur | morning NobodyCam :) | 15:05 |
| NobodyCam | :) | 15:05 |
| rloo | mornin' NobodyCam. | 15:05 |
| romcheg | Morning NobodyCam | 15:06 |
| NobodyCam | morning rloo romcheg | 15:06 |
| NobodyCam | heheh | 15:06 |
| * NobodyCam has no coffee | 15:06 | |
| NobodyCam | yet | 15:07 |
| NobodyCam | :-p | 15:07 |
| yuriyz | Morning All | 15:07 |
| NobodyCam | :) morning | 15:07 |
| linggao | morning all. | 15:07 |
| GheRivero | morning all | 15:08 |
| NobodyCam | morning | 15:10 |
| dkehn | NobodyCam: morning | 15:10 |
| NobodyCam | :) | 15:10 |
| NobodyCam | just started the coffee.. Was able to sleep in today | 15:10 |
| yuriyz | do you make a coffee ironically? :) | 15:12 |
| NobodyCam | not yet.. I have to go to see if my local home depot has the correct part. was going to do that this weekend | 15:14 |
| NobodyCam | walkies brb :) | 15:16 |
| * NobodyCam is back with coffee | 15:23 | |
| *** prekarat has quit IRC | 15:23 | |
| *** datajerk has quit IRC | 15:31 | |
| *** prekarat has joined #openstack-ironic | 15:33 | |
| *** prekarat has quit IRC | 15:36 | |
| *** prekarat has joined #openstack-ironic | 15:36 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Add last_error usage to deploy and teardown methods https://review.openstack.org/57227 | 15:39 |
| lucasagomes | yuriyz, sorry i marked ur patch as WIP by mistake, corrected already (idk why I can mark someone else patch as WIP) | 15:43 |
| yuriyz | ok :-) | 15:43 |
| *** SpamapS_ has joined #openstack-ironic | 15:48 | |
| max_lobur | lucasagomes, do you have a 5 minutes? | 15:50 |
| lucasagomes | max_lobur, sure :) | 15:50 |
| max_lobur | :) | 15:50 |
| max_lobur | https://bugs.launchpad.net/python-ironicclient/+bug/1251712 could you please take a look? | 15:50 |
| * lucasagomes clicks | 15:51 | |
| *** jamespage has quit IRC | 15:52 | |
| *** SpamapS has quit IRC | 15:52 | |
| *** mordred has quit IRC | 15:52 | |
| *** gohko has quit IRC | 15:52 | |
| *** datajerk has joined #openstack-ironic | 15:53 | |
| lucasagomes | max_lobur, I see.. it's a general problem. One thing about the DriverNotFound one is that 'driver' will be a mandatory field on the API when creating a node | 15:54 |
| lucasagomes | and will have a check there to see if that driver is supported | 15:54 |
| lucasagomes | creating/updating a node | 15:54 |
| lucasagomes | but yea, having an exception on the conductor and deserializing in the API is a general problem | 15:55 |
| lucasagomes | if it's async | 15:55 |
| max_lobur | yes, deserializing is a separate problem which I'm trying to fix now | 15:56 |
| max_lobur | but driver not found is just one case | 15:56 |
| max_lobur | I think there can be others | 15:56 |
| max_lobur | it will catch any exception with 404 code and print "Node not found" | 15:57 |
| max_lobur | so why not to print the original message (one that came from server) | 15:57 |
| lucasagomes | yes, I see urgh that's very misleading | 15:57 |
| lucasagomes | yea we can do that | 15:58 |
| max_lobur | just let ironicclient/shell.py to print it | 15:58 |
| lucasagomes | I think there's something in the client that grabs that message from the json | 15:58 |
| max_lobur | yea | 15:58 |
| max_lobur | it's stores message in exception | 15:59 |
| *** datajerk has quit IRC | 15:59 | |
| max_lobur | so ironicclient/shell.py L273 prints it properly | 15:59 |
| max_lobur | *L277 | 15:59 |
| max_lobur | print(str(e), file=sys.stderr) | 16:00 |
| *** mordred has joined #openstack-ironic | 16:00 | |
| lucasagomes | max_lobur, https://github.com/openstack/python-ironicclient/blob/master/ironicclient/common/http.py#L123-L131 | 16:00 |
| *** gohko has joined #openstack-ironic | 16:00 | |
| max_lobur | exactly | 16:01 |
| max_lobur | and then https://github.com/openstack/python-ironicclient/blob/master/ironicclient/common/http.py#L175 | 16:01 |
| lucasagomes | yes | 16:01 |
| lucasagomes | needs some work on those funcs | 16:01 |
| max_lobur | I meant if we remove all our try/excepts from shell routines (do_node_update etc.) it will work properly | 16:02 |
| max_lobur | similar is done in nova | 16:03 |
| *** datajerk has joined #openstack-ironic | 16:03 | |
| max_lobur | see https://github.com/openstack/python-ironicclient/blob/master/ironicclient/shell.py#L272-L278 | 16:03 |
| max_lobur | it will print all exceptions | 16:03 |
| lucasagomes | max_lobur, I see, well that's fine for me :) as there's this global try/except that will take care of it | 16:06 |
| max_lobur | yep | 16:06 |
| max_lobur | so, don't you mind if I fix that? :) | 16:07 |
| lucasagomes | max_lobur, sure not, i will thank you for that | 16:07 |
| lucasagomes | max_lobur, our client is heavily based on the ceilometer client, that's from where we started it | 16:08 |
| lucasagomes | and https://github.com/openstack/python-ceilometerclient/blob/master/ceilometerclient/v2/shell.py#L378-L387 | 16:08 |
| lucasagomes | so I think that's why we have this exception there | 16:08 |
| max_lobur | oh I see | 16:08 |
| max_lobur | that's a potential issue in ceilometer too | 16:08 |
| lucasagomes | yes | 16:08 |
| max_lobur | they can't always know what 404 can be thrown cc.alarms.update | 16:09 |
| max_lobur | so Alarm not found is not a valid messages for all cases | 16:09 |
| lucasagomes | we just organized a bit better our client (having a shell file per resource for example) but we do a lot of copy/cat on their client | 16:09 |
| lucasagomes | so yea if you find a problem in our side there's a good chance to happen on their side as well | 16:09 |
| lucasagomes | and vice versa | 16:09 |
| lucasagomes | :D | 16:09 |
| lucasagomes | max_lobur, +1 | 16:10 |
| max_lobur | thanks =) | 16:10 |
| lucasagomes | thank you :) | 16:10 |
| *** hemna has joined #openstack-ironic | 16:34 | |
| *** jistr has quit IRC | 16:34 | |
| NobodyCam | lucasagomes: https://github.com/ppouliot/ironic | 16:38 |
| NobodyCam | note the commit message | 16:39 |
| lucasagomes | o wow windows support! | 16:40 |
| NobodyCam | :) | 16:40 |
| lucasagomes | awesome! | 16:40 |
| NobodyCam | thats from primeministerp | 16:40 |
| lucasagomes | I will start watching this repo | 16:41 |
| NobodyCam | ya :) | 16:41 |
| lucasagomes | that's the beauty of opensource man | 16:42 |
| lucasagomes | people doing such things you know, adapting for whatever use case they have | 16:42 |
| NobodyCam | deva and I chatted with primeministerp at the summit | 16:43 |
| *** lexx has quit IRC | 16:46 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: API ports update for WSME 0.5b6 compliance https://review.openstack.org/56682 | 16:46 |
| lucasagomes | NobodyCam, oh cool | 16:47 |
| lucasagomes | how was it? | 16:47 |
| NobodyCam | actually he thought it was not going to be that big of a change for us | 16:48 |
| NobodyCam | basicly on the reboot just don't serve a kernel and rmdisk. but send a boot from local response | 16:49 |
| *** datajerk has quit IRC | 16:49 | |
| *** hemna has quit IRC | 16:49 | |
| romcheg | cool | 16:49 |
| lucasagomes | good stuff | 16:49 |
| *** hemna has joined #openstack-ironic | 16:52 | |
| *** tatyana has joined #openstack-ironic | 16:53 | |
| *** lexx has joined #openstack-ironic | 16:55 | |
| *** datajerk has joined #openstack-ironic | 17:01 | |
| *** xmltok has joined #openstack-ironic | 17:08 | |
| *** hemna has quit IRC | 17:09 | |
| openstackgerrit | Max Lobur proposed a change to openstack/python-ironicclient: Remove in-place try/except blocks from shell commands https://review.openstack.org/57285 | 17:10 |
| NobodyCam | bbt..brb | 17:12 |
| max_lobur | lucasagomes, please take a look when you have time https://review.openstack.org/#/c/57285/1 | 17:12 |
| lucasagomes | max_lobur, cool I will add myself as reviewer and give it a whirl soon | 17:13 |
| lucasagomes | max_lobur, thanks max! | 17:13 |
| max_lobur | note - once this lands you may see huge tracebacks from conductors right in error message, there is a bug for this and I'm working on it :) | 17:13 |
| max_lobur | yea, np =) | 17:13 |
| lucasagomes | yea, we need a middleware to remove the traceback | 17:14 |
| max_lobur | yea | 17:14 |
| lucasagomes | max_lobur, cool | 17:14 |
| lucasagomes | I think pecan has a hook called on_error | 17:14 |
| lucasagomes | idk if that would be the right place to strip off the traceback | 17:14 |
| xmltok | is it correct that you can build bare metal nodes in openstack now but the network can't do vlan tagging, and must be the default network? i think i am having a hard time finding documentation that relates to havana | 17:14 |
| max_lobur | if anyone interested, posted a solution https://bugs.launchpad.net/ironic/+bug/1244747 | 17:14 |
| lucasagomes | but might be something to look at | 17:14 |
| max_lobur | wanted to discuss with deva, but any comments are welcome | 17:15 |
| lucasagomes | max_lobur, will comment on that about the hook | 17:15 |
| max_lobur | I thought of rpc layer to remove this. If that won't be approved I'll research for pecan, thanks for this Idea =) | 17:15 |
| max_lobur | thanks! | 17:16 |
| lucasagomes | http://pecan.readthedocs.org/en/latest/pecan_hooks.html#pecan.hooks.PecanHook.on_error | 17:16 |
| max_lobur | TY, will add to bookmarks :) | 17:16 |
| *** hemna has joined #openstack-ironic | 17:16 | |
| lucasagomes | :) you welcome | 17:17 |
| *** datajerk has quit IRC | 17:17 | |
| max_lobur | xmltok, AFAIK yes, we're not able to do vlan tagging, pls correct me someone if that's not true ;) | 17:19 |
| max_lobur | xmltok, not sure if this documented | 17:19 |
| NobodyCam | max_lobur: I believe you are correct in your response | 17:20 |
| xmltok | thanks, that helps. i could get the tagging configuration done through salt, but then I guess i risk an ip allocation divergence with openstacks database | 17:21 |
| max_lobur | I think for you OS running on baremetal after provisioning you may apply any tags you want, but need to make sure that it's stays discoverable through initial ip so Ironic won't loose the control | 17:23 |
| xmltok | i'll look at the neutron api. if i can register an ip as being assigned to the node then i can just assign another ip and call it a day. i need to have my bare metal machines in the same vlan as VMs due to legacy system requirements. im sure its possible | 17:24 |
| max_lobur | yea, seems so | 17:26 |
| *** lexx_ has joined #openstack-ironic | 17:26 | |
| *** lexx has quit IRC | 17:29 | |
| *** SpamapS_ is now known as SpamapS | 17:31 | |
| *** prekarat has quit IRC | 17:32 | |
| NobodyCam | walkies .... brb | 17:39 |
| devananda | good morning, all | 17:42 |
| max_lobur | g'morning devananda | 17:46 |
| max_lobur | when you have time can you please take a look again https://review.openstack.org/#/c/56650/ | 17:46 |
| max_lobur | (Redefined default value of allowed_rpc_exception_modules) | 17:46 |
| max_lobur | devananda, I also posted results of my investigation of bug "API exposes python stack trace" here https://bugs.launchpad.net/ironic/+bug/1244747 . I'd like you to take a look on it too | 17:49 |
| *** derekh has quit IRC | 17:51 | |
| NobodyCam | and back | 17:51 |
| lucasagomes | devananda, morning | 17:53 |
| devananda | lucasagomes: morning | 17:53 |
| devananda | max_lobur: for 56650, is it possible to validate this fix in a unit test? | 17:57 |
| max_lobur | devananda, I researched for that - 1. for rpc level no (there should be a really massive mocks to interrupt rpc responce, and this will rely on serialize implementation) 2. for config, I wasn't successful when tried to test that setting is overriden. When I stop in debug I see it is, but in test - it's not | 18:00 |
| *** epim has joined #openstack-ironic | 18:01 | |
| devananda | max_lobur: 1. that's what I suspected. we'd need integration / devstack-style tests to see this | 18:03 |
| NobodyCam | sandeepr: pong :-p | 18:04 |
| devananda | max_lobur: would you mind opening a bug that rpc deserialization is not tested, low priority, with some description of how this could be validated once we get devstack/tempest tests going, then adding that bug # to the inline comment? | 18:04 |
| devananda | max_lobur: i'll happily approve after that. just want a breadcrumb trail so we come back and test it, so that we don't accidentally break it later | 18:05 |
| max_lobur | yea I'll do now | 18:06 |
| max_lobur | do you mean a bug for Ironic? | 18:06 |
| max_lobur | or for tempest or devstack maybe? | 18:06 |
| NobodyCam | devananda: do you happen to know James Slagle's irc nic | 18:08 |
| sandeepr | hey hey NobodyCam | 18:11 |
| NobodyCam | :) sandeepr I see your ping's but am not online to reply | 18:12 |
| sandeepr | np NobodyCam | 18:12 |
| romcheg | Morning devananda | 18:13 |
| max_lobur | for 2. config - yea probably it's possible to find why those didn't work for me, but will it be a useful test? there is a comment in the code pointing why it's needed so I assume no one remove those accidentally. And if they'll want to remove that they'll remove a config test as well | 18:13 |
| devananda | max_lobur: bug for ironic. | 18:16 |
| max_lobur | ok | 18:17 |
| *** xmltok has quit IRC | 18:18 | |
| devananda | max_lobur: re: 2, yea, I would assume no one will remove that comment & override. but having a test which validates the expected behavior so we _can_ remove that override with confidence | 18:18 |
| devananda | is what I'd like | 18:18 |
| devananda | granted, i agree it will need to be an integration (not unit) test | 18:18 |
| devananda | romcheg: good evening! | 18:19 |
| romcheg | devananda: How are you? | 18:19 |
| max_lobur | devananda, yea that's fair | 18:19 |
| devananda | romcheg: i'm still getting over the post-conference crud. got up much later than i expected | 18:19 |
| max_lobur | thanks, will create a bug | 18:19 |
| devananda | max_lobur: i swear, i'm not just arbitrarily stalling your patch :) | 18:19 |
| devananda | ideally, any time we fix a bug, there should be a new unit test to ensure the bug doesn't resurface | 18:20 |
| romcheg | devananda: We planned to talk about the RPC stuff. I filed a blueprint for it: https://blueprints.launchpad.net/ironic/+spec/rpc-routing | 18:20 |
| * devananda should be better about enforcing that | 18:20 | |
| devananda | romcheg: yes! I will look shortly ) | 18:20 |
| romcheg | I will have to go for about an hour. | 18:20 |
| max_lobur | haha :) it's totally OK | 18:20 |
| max_lobur | I agree =) | 18:20 |
| romcheg | After that I will be available | 18:21 |
| devananda | romcheg: ok, great. I'll make sure to read it in the next hour | 18:21 |
| romcheg | I also uploaded some changes to infra so you can check them out https://review.openstack.org/#/c/53917/ and https://review.openstack.org/#/c/53899/ | 18:22 |
| romcheg | I changed the behaviour so the tests for Ironic won't run for other projects. | 18:23 |
| romcheg | However, for Ironic they will run both in check and gate pipelines | 18:24 |
| romcheg | I need to check whether the recent changes have broken the tests or not | 18:24 |
| *** xmltok has joined #openstack-ironic | 18:27 | |
| devananda | romcheg: why does tempest-devstack-vm-ironic use neutron, but tempest-devstack-vm-ironic-postgres does nto? | 18:28 |
| romcheg | devananda: doh! | 18:31 |
| romcheg | forgot to add that to postgres job | 18:32 |
| romcheg | That was the last moment change :) | 18:32 |
| devananda | :) | 18:33 |
| devananda | romcheg: added comemnts to both patches | 18:34 |
| romcheg | thanks | 18:34 |
| romcheg | devananda: you didn't notice there's no IRONIC=1 either :) | 18:36 |
| devananda | ? | 18:36 |
| *** datajerk has joined #openstack-ironic | 18:36 | |
| devananda | oh! hah! | 18:36 |
| devananda | export DEVSTACK_GATE_IRONIC=1 | 18:36 |
| romcheg | I think I merged that wrong when rebasing | 18:37 |
| devananda | kinda matters :p | 18:37 |
| romcheg | Now have to rush | 18:40 |
| romcheg | See you in ~1 hour | 18:40 |
| romcheg | Bye max_lobur | 18:40 |
| NobodyCam | :) | 18:41 |
| max_lobur | bye romcheg | 18:41 |
| openstackgerrit | Max Lobur proposed a change to openstack/ironic: Redefined default value of allowed_rpc_exception_modules https://review.openstack.org/56650 | 18:41 |
| *** romcheg has quit IRC | 18:41 | |
| max_lobur | devananda, I've posted a bug but it seems I have no rights to set importance https://bugs.launchpad.net/ironic/+bug/1252824 | 18:42 |
| max_lobur | and I think it's can be set triaged already | 18:43 |
| *** blamar has quit IRC | 18:45 | |
| *** blamar has joined #openstack-ironic | 18:48 | |
| *** datajerk has quit IRC | 18:51 | |
| devananda | max_lobur: you probably need to be a member of ironic-bugs team on LP. it's open, anyone can join | 18:53 |
| *** jistr has joined #openstack-ironic | 18:54 | |
| max_lobur | devananda, ah I see | 18:54 |
| NobodyCam | deva question with reguard to your comment on 56605... is there a description of what # noqa does? | 18:54 |
| *** jistr has quit IRC | 18:54 | |
| lifeless | #noqa shuts flake8 up for things it would otherwise whinge about | 18:55 |
| devananda | NobodyCam: it tells pep8/flake8 to ignore any errors on that line | 18:55 |
| devananda | :) | 18:55 |
| NobodyCam | :) TY (both) hehehe | 18:55 |
| max_lobur | devananda, joined & updated | 18:56 |
| *** jistr has joined #openstack-ironic | 18:57 | |
| lifeless | devananda: https://review.openstack.org/#/c/56516/ needs another +2 | 18:57 |
| *** lucasagomes has quit IRC | 18:59 | |
| *** tatyana has quit IRC | 18:59 | |
| max_lobur | devananda, the last opened question to you is https://bugs.launchpad.net/ironic/+bug/1244747 (API exposes python stack trace). If you don't mind one of proposed approaches I'll do the patch tomorrow | 19:00 |
| NobodyCam | ahh I did not add mine. that was the one with the bad SqlA version | 19:00 |
| NobodyCam | lifeless: +2 /+a'd | 19:01 |
| devananda | max_lobur: thanks! | 19:01 |
| *** xmltok has quit IRC | 19:02 | |
| max_lobur | done for today, bye Everyone :) | 19:02 |
| NobodyCam | have a good night max_lobur | 19:02 |
| max_lobur | NICK max_lobur_afk | 19:03 |
| max_lobur | gah | 19:03 |
| NobodyCam | hehehe | 19:03 |
| max_lobur | does anybody know how to change nick in XChat? :) | 19:04 |
| NobodyCam | /raw nick <newnic> ?? | 19:05 |
| *** blamar has quit IRC | 19:06 | |
| *** max_lobur is now known as max_lobur_afk | 19:06 | |
| max_lobur_afk | hurraah | 19:06 |
| max_lobur_afk | TY NobodyCam | 19:06 |
| NobodyCam | :) | 19:06 |
| devananda | max_lobur_afk: i'll leave comments on the bug report - have a good evening :) | 19:07 |
| max_lobur_afk | TY :) | 19:07 |
| max_lobur_afk | bye again :) | 19:07 |
| *** blamar has joined #openstack-ironic | 19:12 | |
| *** epim has quit IRC | 19:16 | |
| *** epim has joined #openstack-ironic | 19:20 | |
| openstackgerrit | A change was merged to openstack/ironic: Update openstack/common/lockutils https://review.openstack.org/56719 | 19:27 |
| openstackgerrit | A change was merged to openstack/ironic: Redefined default value of allowed_rpc_exception_modules https://review.openstack.org/56650 | 19:27 |
| openstackgerrit | A change was merged to openstack/ironic: Import missing gettext _ to fix Sphinx error https://review.openstack.org/56605 | 19:27 |
| openstackgerrit | A change was merged to openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/54430 | 19:27 |
| *** datajerk has joined #openstack-ironic | 19:42 | |
| NobodyCam | brb | 19:43 |
| *** datajerk has quit IRC | 19:50 | |
| *** xmltok has joined #openstack-ironic | 20:01 | |
| *** datajerk has joined #openstack-ironic | 20:23 | |
| *** xmltok has quit IRC | 20:28 | |
| *** datajerk has quit IRC | 20:33 | |
| *** hemna has quit IRC | 20:39 | |
| *** hemna has joined #openstack-ironic | 20:43 | |
| *** datajerk has joined #openstack-ironic | 20:48 | |
| *** romcheg has joined #openstack-ironic | 20:53 | |
| *** datajerk has quit IRC | 20:53 | |
| romcheg | devananda: I'm here | 20:53 |
| romcheg | Sorry, it took me longer | 20:53 |
| devananda | romcheg: np | 21:02 |
| *** jistr has quit IRC | 21:08 | |
| * NobodyCam starts to rebuild his undercloudimage and looks for some food... | 21:11 | |
| *** ndipanov has quit IRC | 21:14 | |
| devananda | romcheg: have you looked at the conductor_service_record_keepalive bit? | 21:15 |
| devananda | romcheg: when a condcutor starts, it registers the list of drivers it can service | 21:16 |
| devananda | romcheg: taht table also serves as a heartbeat | 21:16 |
| devananda | romcheg: so you can get a list of available conductors & what driver each handles | 21:17 |
| devananda | romcheg: this solves your 2.1 idea already. I think 2.2 is unscalable and has race conditions taht make it unusable.. | 21:17 |
| devananda | romcheg: 2.3 is viable, too. | 21:18 |
| romcheg | I thought about using that table in 2.1 | 21:19 |
| devananda | romcheg: i would like to see the BP also describe where & how selection will happen | 21:19 |
| romcheg | IMO 2.3 is better because scheduling looks to be easier | 21:19 |
| devananda | eg, if there are 4 conductors, and 2 can service driver FOO, how does a single conductor get chosen? | 21:19 |
| romcheg | If we use 2.3 we might perform scheduling in the exhanger | 21:20 |
| devananda | in the rpcapi (client) layer? | 21:20 |
| NobodyCam | glance image-list | 21:20 |
| NobodyCam | gah | 21:20 |
| romcheg | In the case you described those two conductors will be subscribed to a certain topic | 21:21 |
| romcheg | I'm just thinking about how to delegate scheduling to rabbit | 21:25 |
| devananda | so | 21:26 |
| devananda | there's a wrinkle :) | 21:26 |
| devananda | romcheg: if scheduling a "spawn this instance" request, and >1 conductor can service it, should we consider load on the conductors? | 21:26 |
| devananda | romcheg: should we try to keep the # of instances managed by all conductors somewhat even? | 21:27 |
| devananda | romcheg: what happens when one conductor dies and we need to rebalance the load? | 21:27 |
| devananda | also, I think this raises the need to actually store the conductor->instance mapping | 21:27 |
| romcheg | We do that in neutron | 21:28 |
| romcheg | However this approach requires re-scheduling, if one of the conductors dies | 21:28 |
| romcheg | In neutron that's reasonable because agents host dhcp services | 21:29 |
| devananda | if a conductor dies mid-task, i think we can (for now) just abort, and let the rescheduling happen at a higher layer | 21:29 |
| devananda | what i mean is, right now, there is a strong association between a conductor and a deployed instance | 21:30 |
| devananda | when conductor X deployes instance A on node N, it maintains the TFTP directory to be able to boot instance A, and Neutron (will be) configured to route DHCP BOOT requests from node N to conductor X. | 21:31 |
| devananda | if X dies, part of our HA task is to update neutron to route DHCP BOOT requests from node N to conductor Y | 21:32 |
| devananda | and to tell Y to rebuild the TFTP directory for instance A | 21:32 |
| devananda | (or to validate it, if we use shared file system) | 21:32 |
| romcheg | I see | 21:33 |
| devananda | right now, we store the I<->N relationship, but not X<->N or X<->I | 21:33 |
| devananda | s/I/A/ | 21:33 |
| romcheg | Then I think we need to bind conductors to instances | 21:33 |
| devananda | yep | 21:34 |
| NobodyCam | will nova need to know that binding ? | 21:34 |
| devananda | NobodyCam: no | 21:34 |
| NobodyCam | :) | 21:34 |
| romcheg | I don't think so | 21:34 |
| romcheg | The scheduler can aggregate the data and use it for scheduling the requests | 21:36 |
| devananda | Nova knows about nodes for scheduling purposes (can I fulfil this requset for an ARMFH with 16GB RAM?). It then knows what instance was placed on what node.. I don't think it needs to know about /how/ that instance was deployed | 21:36 |
| NobodyCam | was thinking about nova pause N & nova start N | 21:37 |
| NobodyCam | should include nova stop n | 21:37 |
| devananda | NobodyCam: nova shouldn't expose nodes in its API, IMO | 21:38 |
| devananda | what you might mean is, nova pause A | 21:38 |
| NobodyCam | yes A | 21:38 |
| devananda | so that's fine -- it'll be mapped by the nova-ironic driver from the nova instance uuid to the ironic node uuid | 21:39 |
| NobodyCam | for which ironic use conductor balh | 21:40 |
| NobodyCam | *ironic KNOWS TO use... | 21:40 |
| devananda | yea. so again, ironic needs to know something about the conductor <-> instance relationship internally | 21:41 |
| devananda | it's conceivable that certain operations on an instance will only work from the conductor which deployed that instance | 21:41 |
| devananda | romcheg: is all that helpful? have more questiosn about the BP direction? | 21:43 |
| romcheg | I think it's quite clear now | 21:43 |
| romcheg | I also think that we have to think about the way of getting rid of the conductor <-> instance relationship for the future | 21:44 |
| romcheg | Because it adds certain amount of complexity to (re)scheduling | 21:45 |
| *** jdob has quit IRC | 21:48 | |
| anteaya | devananda: ttx is calling for an ironic rep in -meeting | 21:49 |
| lifeless | devananda: https://review.openstack.org/#/c/56516/ needsa rebase | 21:51 |
| devananda | anteaya: thansk for the explicit ping | 21:54 |
| anteaya | np | 21:54 |
| anteaya | figured you wanted to be there for thtat | 21:54 |
| devananda | romcheg: shared filesystem that lucas is working on is one way to solve that for PXE, but i'm not sure we can impose taht requirement on all drivers | 21:54 |
| devananda | anteaya: yea. was followign along with the meeting intermittently :) | 21:55 |
| anteaya | cool | 21:55 |
| devananda | taht reminds me, i need to learn all teh release things | 21:55 |
| *** xmltok has joined #openstack-ironic | 21:58 | |
| romcheg | devananda: thank you | 21:58 |
| *** xmltok has quit IRC | 21:58 | |
| romcheg | devananda: it's quite clear for me now. | 21:59 |
| romcheg | If I have any questions I will definitelly ask :) | 21:59 |
| devananda | romcheg: great! is Jan 23 a reasonable target to have that done by? | 22:01 |
| devananda | romcheg: https://wiki.openstack.org/wiki/Icehouse_Release_Schedule :) | 22:02 |
| romcheg | devananda: I think so | 22:02 |
| devananda | NobodyCam, think you're 2 or 7 weeks from the nova driver? | 22:03 |
| devananda | romcheg: great :) | 22:03 |
| *** xmltok has joined #openstack-ironic | 22:03 | |
| NobodyCam | the dhcp issue a blocker right now | 22:03 |
| romcheg | Oh no! Ukraine-France 1-3 :( | 22:04 |
| NobodyCam | romcheg: :( | 22:04 |
| NobodyCam | devananda: prob not two weeks | 22:04 |
| romcheg | Folks, it's quite late here | 22:06 |
| *** linggao has quit IRC | 22:06 | |
| romcheg | So I have to go to bad | 22:06 |
| romcheg | bed | 22:06 |
| romcheg | So have a nice day/evening | 22:07 |
| NobodyCam | Have a good sleep romcheg | 22:07 |
| devananda | g'night romcheg ! thanks for staying up to chat about the RPC BP :) | 22:08 |
| devananda | NobodyCam: how can I help unblock you on that? | 22:08 |
| NobodyCam | got a patch that adds dhcp setup to pxe? | 22:09 |
| NobodyCam | :-p | 22:09 |
| *** lexx_ has quit IRC | 22:12 | |
| NobodyCam | brb | 22:16 |
| NobodyCam | back | 22:22 |
| NobodyCam | gah now looks like another walkies...brb.. again :-p | 22:24 |
| *** romcheg has left #openstack-ironic | 22:36 | |
| devananda | lifeless: was there a reason you felt that a shared filesystem could not be used across ironic-conductors? | 22:36 |
| NobodyCam | and back again | 22:41 |
| lifeless | devananda: performance | 22:42 |
| devananda | lifeless: that's a drawback, not a blocker :) | 22:45 |
| devananda | lifeless: IIRC, you had some concern about networking | 22:46 |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: Adds doc strings to API FunctionalTest class https://review.openstack.org/57344 | 22:49 |
| lifeless | devananda: well, in that we can't split things up at all | 22:52 |
| lifeless | or it becomes a big central point | 22:52 |
| lifeless | or we do one ip per things we're deploying too | 22:52 |
| lifeless | devananda: whats the attraction of it? | 22:53 |
| lifeless | devananda: like, why is it at all interesting? | 22:53 |
| devananda | lifeless: HA for the ir-cond service | 22:53 |
| *** tatyana has joined #openstack-ironic | 22:53 | |
| devananda | lifeless: less work rebuilding the TFTP dir. all we need to do is update neutron | 22:54 |
| lifeless | devananda: which shifts all the load to another server rather than rebalancing | 23:00 |
| lifeless | devananda: I think you shouldn't do that approach at all: let deployers do that one; aim for something much better as integrated code | 23:00 |
| lifeless | devananda: deployers can already do what you describe with floatingip + corosync/pacemaker | 23:01 |
| lifeless | e.g. treat it as a scheduling thing and always provision 2 nodes with the tftp boot files, then when one fails update neutron | 23:02 |
| devananda | lifeless: it doesn't shift all load to another; that can be rebalanced | 23:04 |
| lifeless | a shared file system is a tonne of complexity: locking, concurrent updates, if it fails the node is down | 23:04 |
| lifeless | devananda: if you have more than two nodes sharing the one shared fs | 23:05 |
| devananda | lifeless: correct | 23:05 |
| lifeless | devananda: but that then means that when the shared fs fails more than 2 nodes fail | 23:05 |
| devananda | lifeless: it assumes the FS is HA. which, granted, isnot always true, but the point is to get the easiest solution in now and iterate for more complex ones | 23:06 |
| lifeless | devananda: so the easiest solution is to do nothing : deployers can use existing layer-on-top solutions for HA | 23:06 |
| lifeless | I think we can do a lot better than that, but not by adding more failure modes :) | 23:07 |
| devananda | i like "do nothing" :) | 23:08 |
| *** xmltok has quit IRC | 23:23 | |
| *** jdob has joined #openstack-ironic | 23:24 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add sphinxcontrib-pecanwsme https://review.openstack.org/57346 | 23:25 |
| devananda | Haomeng: thank you for the fix for sphinx // gettextutils._ | 23:26 |
| devananda | Haomeng: that was a blocking, but separate, bug. I have talked with dhellman and figured out how to start auto-generating our API docs. a WIP is ^ | 23:26 |
| devananda | updating commit message | 23:28 |
| devananda | and then walking home, bbiab | 23:29 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add hooks to auto-generate REST API docs https://review.openstack.org/57346 | 23:30 |
| Haomeng | devananda: yes, we just fix our api docs, but for restapi docs, we have to run sphinx with the plugin which dhellman told us | 23:33 |
| Haomeng | no problem, anytime | 23:33 |
| *** jdob has quit IRC | 23:54 | |
| *** epim has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!