| opendevreview | Merged openstack/sushy-tools master: Translate VM name to UUID to avoid redirect issue https://review.opendev.org/c/openstack/sushy-tools/+/924297 | 00:05 |
|---|---|---|
| opendevreview | Doug Goldstein proposed openstack/sushy master: drop usage of pkg_resource for newer python compat https://review.opendev.org/c/openstack/sushy/+/925304 | 01:07 |
| opendevreview | Doug Goldstein proposed openstack/sushy master: drop usage of pkg_resource for newer python compat https://review.opendev.org/c/openstack/sushy/+/925304 | 01:09 |
| opendevreview | Doug Goldstein proposed openstack/sushy master: drop usage of pkg_resource for newer python compat https://review.opendev.org/c/openstack/sushy/+/925304 | 01:14 |
| cardoe | Shhh nobody saw that | 01:14 |
| opendevreview | Doug Goldstein proposed openstack/sushy master: avoid MAC addresses that are the empty string https://review.opendev.org/c/openstack/sushy/+/925130 | 01:21 |
| opendevreview | Kamil Gustab proposed openstack/sushy master: Adds Ethernet Interfaces for manager. https://review.opendev.org/c/openstack/sushy/+/909451 | 06:33 |
| jssfr | good morning Ironic | 06:42 |
| rpittau | good morning ironic! o/ | 07:04 |
| rpittau | clarkb, JayF, cardoe, TheJulia: I have a different opinion on that, we're still testing Python 3.8 now and up to Python 3.11, currently Python 3.12 is not required (it's not-voting), while in 2024.2 we'll stop testing Python 3.8, the minimum req will be Python 3.9 and we'll still test Python 3.12 as non-voting | 07:28 |
| rpittau | keeping compatibility with Python 3.8 while not testing it does not sound like a good idea to me | 07:28 |
| rpittau | TL;DR it's not worth to keep compatibility with Python 3.8 in 2024.2, but we should just make an effort to be compatible with Python 3.12, so adding importlib-resources to global-requirements is not needed | 07:28 |
| jssfr | the ironic-inspector migration looks easy enough---seems like most stuff is now supported by ironic "core". | 07:44 |
| dtantsur | TheJulia, JayF, I used to tell you that iLO 5 does not support Redfish RAID. Based on what we are discovering now, this information might be false. If this is confirmed, our prospects of deprecating the iLO drivers are somewhat better. | 08:50 |
| opendevreview | Merged openstack/sushy master: avoid MAC addresses that are the empty string https://review.opendev.org/c/openstack/sushy/+/925130 | 08:57 |
| opendevreview | Robert Hoffmann proposed openstack/ironic-inspector master: Fix: ignore resolve error while looking up node https://review.opendev.org/c/openstack/ironic-inspector/+/925326 | 09:15 |
| opendevreview | Adam Rozman proposed openstack/ironic-specs master: propose root partition and configdrive encryption https://review.opendev.org/c/openstack/ironic-specs/+/924993 | 09:30 |
| Nisha_Agarwal | dtantsur, hi | 10:35 |
| dtantsur | o/ | 10:35 |
| Nisha_Agarwal | I saw Mahnoor's mail | 10:36 |
| dtantsur | cc masghar | 10:36 |
| Nisha_Agarwal | Have you tested the API for RAID? | 10:36 |
| dtantsur | Nisha_Agarwal: no, we're only about to. Our iLO 6 machines don't seem to support it (could be too few disks), but we've discovered a Gen 10+ machine with Redfish API present. | 10:37 |
| Nisha_Agarwal | dtantsur, ilo6 should have worked with redfish driver | 10:38 |
| Nisha_Agarwal | we tested it | 10:38 |
| Nisha_Agarwal | i mean RAID configuration on ilo6 should have worked | 10:40 |
| dtantsur | Nisha_Agarwal: we see missing SupportedRAIDTypes in the Storage resource. Can it be because of only one Drive or is anything else the cause? | 10:40 |
| Nisha_Agarwal | with redfish driver | 10:40 |
| iurygregory | good morning ironic | 10:47 |
| Nisha_Agarwal | dtantsur, ilo6 should have worked....you can see the documentation at https://hewlettpackard.github.io/ilo-rest-api-docs/ilo6/#creating-volumes | 10:57 |
| Nisha_Agarwal | if it is not as per standards, i guess you can mention that on the mail | 10:57 |
| dtantsur | Nisha_Agarwal: we're more interested in the iLO 5 situation at this point. Was standard RAID really implemented at some point? Does it only concern Gen 10+ or the normal Gen 10 too? | 10:59 |
| opendevreview | Robert Hoffmann proposed openstack/ironic-inspector master: Fix: ignore resolve error while looking up node https://review.opendev.org/c/openstack/ironic-inspector/+/925326 | 10:59 |
| Nisha_Agarwal | dtantsur, after looking at the documentation for both ilo5 and ilo6 it looks like it should work and even "SupportedRAIDTypes" should also be there....but it is better if the answer comes from ilo team copied onto the mail....myself and Mallik have moved out of openstack ironic due to internal movements.... | 11:04 |
| Nisha_Agarwal | in general standard Storage API was always present in ilo5 but RAID configuration was not supported | 11:05 |
| dtantsur | Nisha_Agarwal: ah, got it, thank you! | 11:08 |
| Nisha_Agarwal | dtantsur, :) welcome...i have asked owners of openstack ironic ...they will respond on the mail... | 11:27 |
| Nisha_Agarwal | owners of openstack ironic in HPE ^^^ | 11:28 |
| masghar | Thanks Nisha for the help! | 11:30 |
| opendevreview | Robert Hoffmann proposed openstack/ironic-inspector master: Fix: ignore resolve error while looking up node https://review.opendev.org/c/openstack/ironic-inspector/+/925326 | 12:52 |
| opendevreview | Verification of a change to openstack/ironic master failed: [devstack] Ensure downgrade_dnsmasq is idempotent https://review.opendev.org/c/openstack/ironic/+/924978 | 13:49 |
| cardoe | Well rpittau I'd be happy with that answer. Cause I can't tweak the global-requirements.txt. The instructions they provide don't work. It fails on Python 3.8 for me. | 13:50 |
| cardoe | I created a fresh Debian container as well. Installed pyenv and installed ALL the versions in there. | 13:50 |
| cardoe | ERROR: Could not find a version that satisfies the requirement pyasn1-lextudio | 13:57 |
| rpittau | cardoe: that should be in requirements anymore | 14:24 |
| rpittau | should not* | 14:29 |
| TheJulia | not in requirements *but* they said from debian | 14:36 |
| cardoe | All I ran was "tox -e generate" per the docs. | 14:47 |
| cardoe | https://review.opendev.org/c/openstack/requirements/+/925306 is my change originally | 14:48 |
| cardoe | https://docs.openstack.org/project-team-guide/dependency-management.html#update-processes that's the instructions I followed | 14:49 |
| cardoe | pyenv install 3.8.19 && pyenv install 3.9.19 && pyenv install 3.10.14 && pyenv install 3.11.9 && pyenv global 3.11.9 3.10.14 3.9.19 3.8.19 was my local setup. | 14:51 |
| TheJulia | dtantsur: hopefully we have the email addresses for the folks who are still working on ironic @ hpe for future needs as well | 14:55 |
| dtantsur | seems so | 14:55 |
| rpittau | cardoe: there's a leftover in global-requirements, pyasn1-lextudio should not be there | 14:56 |
| cardoe | okay so I guess I'll wait for the sushy fix until upstream fixes itself? | 14:59 |
| rpittau | cardoe: which sushy fix ? | 15:04 |
| cardoe | https://review.opendev.org/c/openstack/sushy/+/925304 | 15:05 |
| rpittau | cardoe: I don't see that as a fix, in the next dev cycle Python 3.8 won't be supported anymore, the min tested version will be Python 3.9 | 15:07 |
| rpittau | as I wrote before, keeping compatibility with an untested Python version is not a good idea | 15:07 |
| cardoe | clarkb: ^ | 15:12 |
| cardoe | dtantsur: Node 2f594efa-bf65-4aa7-bc85-57d834977a98 failed step {'interface': 'raid', 'step': 'delete_configuration', 'abortable': False, 'priority': 0}: HTTP GET https://10.19.192.4/redfish/v1/Systems/1/Storage returned code 400. iLO.0.10.ExtendedInfo: See @Message.ExtendedInfo for more information. Extended information: [{'MessageArgs': ['5, (84,00,00)'], 'MessageId': 'iLO.2.23.ResourceNotReadyRetry'}] | 15:12 |
| cardoe | based on your convo above I tried an HPE DL380 Gen10 box with redfish and told it to delete it's RAID | 15:12 |
| cardoe | works fine with ilo5 | 15:12 |
| dtantsur | work with ilo6 you mean? Gen 10 is ilo5 | 15:20 |
| cardoe | No. I meant ilo5. You were asking if it worked with iLO5 above. So I was testing it for ya. | 15:49 |
| dtantsur | cardoe: but you showed a failure above? | 15:50 |
| cardoe | Yes it failed. | 15:51 |
| opendevreview | Merged openstack/ironic master: Remove ironic-inspector from bugfix branches https://review.opendev.org/c/openstack/ironic/+/925200 | 15:51 |
| clarkb | rpittau: that statement makes it sound like it is an issue now though. And as mentioned here and subsequently on the requirements chnage libraries should go last | 15:54 |
| clarkb | or at least after the applications do. There will probably be things that go later like PBR | 15:54 |
| clarkb | so I guess it comes down to the timing of when you're able to coordinate all of that. In the broader openstack sphere there are things still using py38 they probably don't depend on sushy though so the scope is smaller here | 15:55 |
| clarkb | (I just want to make sure people are thinking about the downstream fallout of these changes and making them in a somewhat coordinated manner to avoid the massive fires that resulted the last time this was attempted) | 15:56 |
| rpittau | clarkb: I totally understand what you're saying, but the moment we remove the tests from upstream we can't guarantee any software will work anymore with that Python version, also considering that Py3.8 is EoL in October | 16:09 |
| rpittau | cardoe: btw this https://review.opendev.org/c/openstack/requirements/+/925405 will help with your testing | 16:11 |
| rpittau | good night! o/ | 16:12 |
| clarkb | yes, I am not arguing aginast removing things. I'm merely saying the removal needs to be somewhat coordinated and that doesn't appear to be happening | 16:12 |
| JayF | rpittau: clarkb: One interesting difference: sushy, as a library, is only used by Ironic. So I'm in the weird spot of agreeing with both of you. I'd like to keep python 3.8 supported for as long as possible, but I also think the situation is materially different than oslo given we basically just have to coordinate with ourself :) | 16:12 |
| clarkb | until thatdoes happen we need to err on the side of caution | 16:12 |
| clarkb | JayF: ah ok if the libraryisn't widely adopted then it does simplify things quite a bit | 16:12 |
| JayF | and this is the part where someone goes "oh, you didn't know that OpenStack CoconutWater uses sushy?" | 16:12 |
| JayF | Communication with redfish is all sushy does, so it's mostly Ironic, might wanna check to ensure starlingx or similar isn't using it | 16:13 |
| JayF | but I think they basically did their own thing and disregarded our previous work in the BM provisioning field | 16:13 |
| clarkb | ack | 16:13 |
| TheJulia | TYVM, OpenStack FizzyWater does! | 16:17 |
| TheJulia | :) | 16:17 |
| TheJulia | (joking, of course) | 16:17 |
| opendevreview | OpenStack Release Bot proposed openstack/ironic bugfix/26.0: Update .gitreview for bugfix/26.0 https://review.opendev.org/c/openstack/ironic/+/925410 | 16:30 |
| opendevreview | Doug Goldstein proposed openstack/ironic-python-agent master: update dynamic-login to mention the sshkey option https://review.opendev.org/c/openstack/ironic-python-agent/+/925414 | 17:22 |
| opendevreview | cid proposed openstack/ironic-python-agent master: Add support for burnin-gpu https://review.opendev.org/c/openstack/ironic-python-agent/+/925415 | 17:39 |
| opendevreview | Verification of a change to openstack/ironic master failed: [devstack] Ensure downgrade_dnsmasq is idempotent https://review.opendev.org/c/openstack/ironic/+/924978 | 17:55 |
| cardoe | rpittau: So can I ask how to change that global-requirements.txt and have it pass validate? It bombs telling me I don't have an upper-constraints but "tox -e generate" doesn't generate that like the docs say. | 19:01 |
| clarkb | cardoe: you added the dep to global-requirements.txt then ran tox -e generate and it didn't update the upper constraints file? | 19:04 |
| cardoe | nope | 19:04 |
| clarkb | the tox target is redirecting stdout to upper-constraints.txt in the local directory maybe the dir tox was invoked from was different? | 19:06 |
| cardoe | I'm in the top level of the git repo | 19:06 |
| cardoe | https://review.opendev.org/c/openstack/requirements/+/925306 is my change | 19:06 |
| cardoe | I'm re-running it with deleting the file first. | 19:07 |
| clarkb | you could edit the tox target to remove the > redirect and have it print to stdout to see what it does maybe | 19:07 |
| cid | Good night ironic o/ | 19:13 |
| JayF | \o | 19:13 |
| opendevreview | Merged openstack/ironic master: [devstack] Ensure downgrade_dnsmasq is idempotent https://review.opendev.org/c/openstack/ironic/+/924978 | 19:52 |
| TheJulia | \o/ | 21:07 |
| JayF | I'll note if anyone has a free moment. Runbooks needs a set of eyes that isn't mine, and CID posted a video demo of it too. I know we're all busy but if anyone wants a change of pace that'd be very helpful. Really the same for my ironic-guest-metadata patch, but I care more about unblocking CID in the short term. | 22:48 |
| iurygregory | JayF, I will give priority to that (sorry been a bit busy figuring out downstream things and my presentation for OIS Asia) | 23:07 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!