| rpittau | good morning ironic! o/ | 07:11 |
|---|---|---|
| sylvr | good morning! | 07:25 |
| sylvr | in this line https://review.opendev.org/c/openstack/ironic-inspector/+/860731/18/ironic_inspector/db/api.py#460 shouldn't orm.exc be replaced with orm_errors ? I don't know if this caused my issue, but seems inconsistant ? | 07:35 |
| sylvr | or is it on purpose because it's in a try/except ? | 07:37 |
| rpittau | sylvr: if you look at the import you'll see that orm.exc = orm_errors, so it's interchangeable | 08:08 |
| sylvr | rpittau: so no consistency issue then ! :) | 08:12 |
| rpittau | sylvr: a huge consistency issue! but technically it's the same, so in general it's not an issue :) | 08:14 |
| sylvr | well, I'm trying to fix an issue I encountering, maybe I'll change it if I submit a patch :) | 08:15 |
| sylvr | if I modify the code on site and then use systemctl restart ironic-inspector, the modified code should be run right ? | 08:38 |
| sylvr | welp... seems like the error vanished | 08:52 |
| masghar | JayF: yes, inspection rules being the biggest one I know of | 09:36 |
| masghar | JayF: From Dmitry's notes downstream, I also see these 4-5 items: 'Add an Ironic CI job with in-band inspection without inspector', 'Migrate "introspection interface" CLI commands', 'New Node field to distinguish auto discovered nodes', 'RPC API for the new PXE filter service' and 'Bonus: migrate JSON RPC back into Ironic (from ironic-lib)' | 09:52 |
| rpittau | JayF: re https://review.opendev.org/c/openstack/ironic-lib/+/908119 we can make the test result based on python version | 12:29 |
| TheJulia | good morning | 13:13 |
| rpittau | JayF: we convert the properties only in get_endpoint, which is successfully tested, so I guess it's ok to distinguish between python version <3.9 and >=3.9 | 13:13 |
| * TheJulia tries to wake up | 13:13 | |
| sylvr | TheJulia : good morning ! | 13:15 |
| TheJulia | masghar: I guess there is no plan for actual data miration? (don't assume this is requirement, not trying to scope creep, but we likely need to set the correct expectation for folks) | 13:21 |
| masghar | TheJulia: Not that I am aware of, no | 13:24 |
| masghar | But it does seem useful for those who want to eventually take the leap from inspector | 13:25 |
| masghar | New deployments arent a problem, but we will have lots of old deployments | 13:25 |
| TheJulia | I'm against trying to take that on, at least over some recent pain I got to feel. https://review.opendev.org/c/openstack/ironic-inspector/+/924172 | 13:28 |
| TheJulia | the data being optional for almost everyone... and all that :) | 13:33 |
| masghar | Oh there is a data migration tool in inspector? Thats nice! | 13:44 |
| masghar | the data is optional for everyong | 13:44 |
| masghar | sorry, I mean the data is optional for everyone? Wouldnt users be in trouble if their inventory information was in inspector DB and they suddenly stopped inspector? | 13:45 |
| opendevreview | cid proposed openstack/ironic master: Allow disabling specific boot modes during deployment/enrollment https://review.opendev.org/c/openstack/ironic/+/922243 | 13:50 |
| TheJulia | masghar: inspector is not an inventory tool. | 13:53 |
| TheJulia | it is just a data collection tool | 13:53 |
| masghar | TheJulia: oh well, in that case migration is not as important as I thought | 13:55 |
| masghar | data migration I mean | 13:55 |
| TheJulia | Yeah, it can always be recreated. We view ironic as the source of truth of if there is hardware inventory and it's state | 13:56 |
| masghar | Very nice, makes the inspector removal even better | 13:56 |
| opendevreview | cid proposed openstack/python-ironicclient master: Add CLI support for runbooks https://review.opendev.org/c/openstack/python-ironicclient/+/924101 | 14:30 |
| JayF | masghar: I think cid may have some cycles to help with that inspector retirement work over the next few weeks | 14:58 |
| JayF | cid: ^ If you're interested, that inspector retirement work is important and it'd be a big win if we could get it completely migrated this cycle | 14:59 |
| JayF | basically means one less service (almost) every Ironic deployment has to put out there | 14:59 |
| cid | Sure thing. | 15:03 |
| cid | I will be taking a look at that from next week. Currently trying to address review feedbacks and getting runbook to work on devstack. | 15:03 |
| * cid cli | 15:03 | |
| JayF | Yes absolutely, no rush just helping line it up as we talked about yesterday | 15:04 |
| masghar | cid, JayF: That would be very much appreciated, thank you! I have been caught up in downstream work | 15:05 |
| cid | No problems at all. | 15:06 |
| cid | JayF: ++ | 15:06 |
| rpittau | good night! o/ | 16:00 |
| masghar | o\ | 16:20 |
| surbhi | frickler: You got sometime to login and look into things? I am in EST time zone. Let me know yours too so, I can ping you during your works hours not after that. | 17:21 |
| opendevreview | cid proposed openstack/ironic master: Allow disabling specific boot modes during deployment/enrollment https://review.opendev.org/c/openstack/ironic/+/922243 | 17:34 |
| opendevreview | cid proposed openstack/ironic master: Allow disabling specific boot modes during deployment/enrollment https://review.opendev.org/c/openstack/ironic/+/922243 | 17:40 |
| anshul | hi! i'm trying to fix this error which comes on running some of our commands in our code. the error is "A default microversion for service baremetal of 1.88 was requested, but the cloud only supports a minimum of 1.1 and a maximum of 1.72. The default microversion was set because a microversion formatted version string, '1.88', was passed for the api_version of the service. If it was not intended to set a default microversion please r | 18:14 |
| anshul | emove anything other than an integer major version from the version setting for the service.". | 18:14 |
| anshul | our client calls the openstacksdk directly; and we're wondering how to default a version the way python-ironicclient does | 18:14 |
| TheJulia | it is supposed to try and negotiate the maximum available if memory serves | 18:16 |
| anshul | where/when is it supposed to do that? | 18:17 |
| TheJulia | I don't remember how it is done with the sdk | 18:20 |
| TheJulia | but some piece thinks version 1.88 is available, and then something else things 1.72 is the max | 18:21 |
| TheJulia | that seems wrong | 18:21 |
| TheJulia | hmmm | 18:21 |
| anshul | is there any documentation for this? | 18:21 |
| TheJulia | 1.88 is the maximum version of the sdk | 18:21 |
| TheJulia | it, in part, is covered by the sdk docs themselves | 18:22 |
| TheJulia | see, the sdk was never really designed for users to really do call level overriding on versions if memory serves | 18:22 |
| TheJulia | if memory serves, data gets loaded in through the cloud config | 18:23 |
| TheJulia | so it looks like as an api client user, you can assert the desired effect by influencing hte session client | 18:25 |
| TheJulia | https://github.com/openstack/openstacksdk/blob/9145dcec6480a06622afc986731619e6d0a52691/openstack/config/cloud_region.py#L819-L835 | 18:25 |
| cid | o/ | 19:23 |
| TheJulia | g'night | 20:50 |
| JayF | Random idea: post to node history to add history items from external sources | 21:54 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!