| iurygregory | I'm starting to think we need some guard when a clean/service fails only because it failed to power on the hardware... | 00:44 |
|---|---|---|
| iurygregory | Thoughts on have some retry mechanism in case this happens "Failed to tear down from service for node 24d15edd-c4fb-40f2-a772-942f314b23bf, reason: Failed to set node power state to power on." ? | 01:33 |
| iurygregory | s/have/having | 01:33 |
| TheJulia | we've sort of discussed it, could you give us a complete picture into the failure? | 03:09 |
| iurygregory | firmware update for bios and bmc, both succeed, but the machine failed to power on and went to service failed .-. | 03:25 |
| iurygregory | I've collected some logs, I can share tomorrow, going to get some rest now | 03:26 |
| opendevreview | Goutham Pacha Ravi proposed openstack/python-ironicclient master: Replace CLA with DCO https://review.opendev.org/c/openstack/python-ironicclient/+/956327 | 05:21 |
| opendevreview | Jacob Anders proposed openstack/sushy master: Improve handling of empty-string eTags. https://review.opendev.org/c/openstack/sushy/+/956204 | 06:31 |
| rpittau | good morning ironic! happy friday! o/ | 06:56 |
| queensly[m] | Good morning | 07:11 |
| abongale | gm ironic! | 07:41 |
| janders | hey rpittau and Ironic o/ | 07:56 |
| rpittau | hey janders :) | 07:56 |
| opendevreview | Merged openstack/ironic-python-agent bugfix/11.1: Update .gitreview for bugfix/11.1 https://review.opendev.org/c/openstack/ironic-python-agent/+/956235 | 09:00 |
| opendevreview | Verification of a change to openstack/ironic master failed: conf: Add '[api] response_validation' option https://review.opendev.org/c/openstack/ironic/+/945515 | 09:43 |
| opendevreview | Verification of a change to openstack/ironic master failed: api: Add schema for allocations API (versioning) https://review.opendev.org/c/openstack/ironic/+/945217 | 09:43 |
| iurygregory | good morning ironic | 11:35 |
| opendevreview | Merged openstack/sushy master: Improve handling of empty-string eTags. https://review.opendev.org/c/openstack/sushy/+/956204 | 12:11 |
| opendevreview | Verification of a change to openstack/ironic master failed: conf: Add '[api] response_validation' option https://review.opendev.org/c/openstack/ironic/+/945515 | 12:51 |
| cardoe | iurygregory: if you're messing with redfish... https://review.opendev.org/c/openstack/ironic/+/955537 | 14:07 |
| cardoe | that'll make redfish inspect put the right value in for cpu_arch in node properties. | 14:07 |
| iurygregory | cardoe, oh I will add that to my list! | 14:46 |
| opendevreview | Merged openstack/ironic master: conf: Add '[api] response_validation' option https://review.opendev.org/c/openstack/ironic/+/945515 | 14:48 |
| cardoe | iurygregory, dtantsur, TheJulia: redfish inspection question for you all... if the state is disabled of the port... do we expect it to still be created in the inventory or not? afaik disabled ports via the BIOS won't show up to the OS at all. | 15:57 |
| iurygregory | cardoe, that matches my understanding as well | 16:04 |
| cardoe | So do we expect the disabled port to show up in the inventory? | 16:04 |
| iurygregory | do we show the state in the inventory? if we can show it's disable it would be ok, but if we can't I would say to not have the port showing up in the inventory | 16:05 |
| cardoe | We can. | 16:05 |
| iurygregory | I would say the information can be useful (considering operator point of view..) | 16:06 |
| cardoe | In one place (sushy specifically) we filter out interfaces that are not Health.OK and in another spot in Ironic we filter out things that aren't State.ENABLED. Those are technically two different fields in the "status" in redfish. | 16:06 |
| cardoe | The code tests used a mock that didn't treat status as an object (like it actually is with those two values) and instead it just set it to a singular value of ENABLED/DISABLED | 16:07 |
| iurygregory | yeah redfish response for status have the state and health information on it | 16:09 |
| cardoe | So now I'm trying to update the code to create ports from inventory data (which is just a utils function cause the normal inspection path uses it) | 16:09 |
| cardoe | and we filter out in sushy interfaces that don't have Health.OK | 16:10 |
| iurygregory | if we filter out the ones that are not Health.OK, don't we exclude the ones that are not State.Enabled? | 16:10 |
| cardoe | So on my real hardware things that don't have Health.OK are cards without the SPFs in them or maybe just on board NICs without a header attached to plumb them out. | 16:12 |
| cardoe | enabled/disabled is something I can control BIOS wise. | 16:12 |
| cardoe | I think we should include stuff that's health.ok and both enabled/disabled in inventory cause it works in theory | 16:12 |
| iurygregory | ++ | 16:13 |
| iurygregory | makes sense to me | 16:13 |
| cardoe | but if you did it via IPA you wouldn't see the disabled stuff is what I was getting at. | 16:13 |
| cardoe | okay I'll go that route. | 16:13 |
| cardoe | This whole series is weird and sucks cause literally it'll be like 3 lines of code change and then hundreds of lines of behavior change to the tests. | 16:14 |
| cardoe | Which just feels wrong. | 16:14 |
| iurygregory | I know how you feel | 16:14 |
| opendevreview | Elod Illes proposed openstack/ironic unmaintained/xena: [CI][doc] Turn off sphinx's warning as error feature https://review.opendev.org/c/openstack/ironic/+/956385 | 16:25 |
| opendevreview | Goutham Pacha Ravi proposed openstack/bifrost master: Replace CLA instructions with DCO https://review.opendev.org/c/openstack/bifrost/+/956392 | 17:56 |
| opendevreview | Goutham Pacha Ravi proposed openstack/python-ironicclient master: Replace CLA with DCO https://review.opendev.org/c/openstack/python-ironicclient/+/956327 | 18:36 |
| opendevreview | Merged openstack/ironic unmaintained/xena: [CI][doc] Turn off sphinx's warning as error feature https://review.opendev.org/c/openstack/ironic/+/956385 | 20:21 |
| opendevreview | Elod Illes proposed openstack/ironic unmaintained/xena: OSSA-2025-001: Disallow unsafe image file:// paths https://review.opendev.org/c/openstack/ironic/+/949179 | 20:49 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: fix up redfish inspection mock ethernet interface data https://review.opendev.org/c/openstack/ironic/+/955536 | 20:54 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: fix redfish processor inspection https://review.opendev.org/c/openstack/ironic/+/955537 | 20:54 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: allow running inspection hooks on redfish interface https://review.opendev.org/c/openstack/ironic/+/933066 | 20:54 |
| iurygregory | cardoe, just wondering if we want a releasenote in https://review.opendev.org/c/openstack/ironic/+/955537 | 21:03 |
| iurygregory | but i'm ok with it =) | 21:04 |
| cardoe | I can add one. | 21:12 |
| cardoe | Wondering if you know if I can somehow trigger continue_inspection() from inspect_hardware()? Typically its triggered by IPA hitting the continue_inspection API endpoint | 21:13 |
| cardoe | But I'm trying to make the behavior between the two as similar as possible. But I'm seeing in the global conductor one before it calls the driver specific one there's quite a bit of stuff | 21:14 |
| cardoe | My running of inspection hooks just runs the function in inspect_hardware() while the other implementations put it in continue_inspection() | 21:14 |
| iurygregory | I would need to check things a bit before providing an answer, I'm about to wrap up for the weekend | 21:15 |
| -opendevstatus- NOTICE: All OpenDev IRC and Matrix bots will be offline briefly (including channel logging and meeting management) in order to move them to a newer server | 23:49 | |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!