| opendevreview | Julia Kreger proposed openstack/ironic master: Add a new 'category' field to the Portgroup object https://review.opendev.org/c/openstack/ironic/+/955713 | 01:20 |
|---|---|---|
| opendevreview | Julia Kreger proposed openstack/ironic master: Remove ovn vtep support https://review.opendev.org/c/openstack/ironic/+/966140 | 01:20 |
| TheJulia | ugggh | 01:20 |
| TheJulia | just a rebase then | 01:22 |
| opendevreview | Merged openstack/ironic master: Add a new 'category' field to the Portgroup object https://review.opendev.org/c/openstack/ironic/+/955713 | 04:20 |
| janders | good morning Ironic o/ | 07:25 |
| rpittau | good morning ironic! o/ | 07:29 |
| Ayo[m] | Best of mornings, Ironic o/ | 09:41 |
| iurygregory | good morning ironic | 11:04 |
| opendevreview | Merged openstack/ironic master: Improve logging when console unit doesn't start https://review.opendev.org/c/openstack/ironic/+/963324 | 11:17 |
| opendevreview | Merged openstack/ironic master: Replace Chrome/Selenium console with Firefox extension https://review.opendev.org/c/openstack/ironic/+/961434 | 11:25 |
| opendevreview | David Nwosu proposed openstack/ironic master: Move build_configdrive and get_configdrive_image to configdrive_utils https://review.opendev.org/c/openstack/ironic/+/965880 | 11:43 |
| opendevreview | David Nwosu proposed openstack/ironic master: Move configdrive tests to test_configdrive_utils https://review.opendev.org/c/openstack/ironic/+/966158 | 11:43 |
| TheJulia | Good morning | 14:49 |
| janders | hey TheJulia o/ | 15:19 |
| janders | Hey folks. We had a few changes/improvements related to post-firmware-upgrade BMC handling merge recently ( https://review.opendev.org/c/openstack/ironic/+/960230, https://review.opendev.org/c/openstack/ironic/+/960711, https://review.opendev.org/c/openstack/ironic/+/962560 ). With this, I am tempted to remove the unconditional 300s wait | 15:22 |
| janders | post-upgrade we have in place: https://opendev.org/openstack/ironic/src/commit/5c1ff3f892fc2575fcae8e368138b2cb0a331140/ironic/drivers/modules/redfish/firmware.py#L305-L320 ). For starters this can be done with simple change of defaults to 0. Would you be OK with such change? I'd appreciate your thoughts. | 15:22 |
| janders | Saving 300s there would also allow for some tuning of the logic added in https://review.opendev.org/c/openstack/ironic/+/960230 so that we can reliably tolerate the BMC going incommunicado half way through the checks and then recovering | 15:23 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent-builder master: Add simple-init by default https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/966192 | 15:27 |
| cardoe | So there was something recently about DHCP-ing on all the interfaces in IPA and then we've also talked about virtual-media support. I'm just curious how it's all suppose to work together? Cause looking at https://opendev.org/openstack/ironic-python-agent-builder/src/branch/master/dib/ironic-python-agent-ramdisk/element-deps which is the normal IPA we include "dhcp-all-interfaces". | 15:31 |
| cardoe | The result is that the static IP config that redfish-virtual-media injects into the ISO for example isn't used cause we instead DHCP on the interface. | 15:32 |
| dtantsur | alegacy: ^^ | 15:32 |
| TheJulia | Yeah, it is supposed to by default. For vmedia with injected config, glean actually injects/resets the all interfaces since (and I’m going from memory here, on boot it expands out the config to dhcp on all Intefaces by default. Furthermore, network manager likes to do that by default anyhow | 15:36 |
| TheJulia | “Ohhh an interface! Seeet, line carrier! Let’s dhcp it!” | 15:36 |
| TheJulia | err, that should be sweet | 15:36 |
| cardoe | well we need to use two separate IPAs to have the redfish-virtual-media work. We cannot use the ironic-python-agent-ramdisk DIB otherwise it won't use the static IP info. | 15:46 |
| cardoe | What's shared to me now is we've literally cloned ironic-python-agent-ramdisk and the only change is to delete "dhcp-all-interfaces" | 15:48 |
| TheJulia | Interesting | 15:48 |
| TheJulia | and depending on your OS it will still do it, kind of | 15:49 |
| TheJulia | The goal/intent is for them to operate together seamlessly, it sounds like your operational requirements led to removal of dhcp-all-interfaces ? | 15:49 |
| cardoe | We're using redfish-virtual-media and not using DHCP on the network. | 15:49 |
| cardoe | Ironic injects the IP info always into the ISO it generates but when you've got the "dhcp-all-interfaces", it won't use it. | 15:50 |
| TheJulia | That is a bug then because simple-init grabbing config early on should lockout dhcp-all-interfaces from making config changes | 15:50 |
| TheJulia | since it should only be applying such config to unconfigured interfaces | 15:50 |
| TheJulia | at least, that is intent/modeling | 15:51 |
| TheJulia | (also, network-manager default settings, if you've got network manager may be trying to dhcp all interfaces regardless and that is a separate knob which I believe got changed in the defaults (yay for cases with hosts with like 20+ network interfacces) | 15:52 |
| TheJulia | I guess what is needed is a bug clearly stating the context regarding the experience of dhcp-all-interfaces overriding simple-init applied configuration and from there figure out what changed and how to head off such a conflict | 16:01 |
| cardoe | So redfish-virtual-media never sets up dhcp even if you're using neutron or dnsmasq in your ironic config for other boot methods | 16:03 |
| opendevreview | Clif Houck proposed openstack/ironic master: Trait Based Networking Filter Expression Parsing and Base Models https://review.opendev.org/c/openstack/ironic/+/961498 | 16:09 |
| opendevreview | Clif Houck proposed openstack/ironic master: Configuration file for Trait Based Networking https://review.opendev.org/c/openstack/ironic/+/962598 | 16:09 |
| opendevreview | Clif Houck proposed openstack/ironic master: WIP: Generate network plan based on trait based networking config https://review.opendev.org/c/openstack/ironic/+/964895 | 16:09 |
| opendevreview | Clif Houck proposed openstack/ironic master: WIP: Trait Based Networking Simulator https://review.opendev.org/c/openstack/ironic/+/966202 | 16:09 |
| JayF | I will likely be filming a video today or tomorrow based on the current contents of the work items spec. I also think we should merge it on Friday. Please take time to review it ASAP | 16:12 |
| TheJulia | cardoe: that is correct, the ramdisk is supposed to discover the configdrive data and apply it. Independently the ramdisk might be trying to dhcp as well, but it should lockout that state overall. At least, should. | 16:14 |
| JayF | Also, it looks like we never sent a photo in. I took a team photo so I can send to that in, does anyone want to take a guess at attendance? | 16:14 |
| JayF | (for ptg) | 16:14 |
| opendevreview | cid proposed openstack/ironic master: Add IPMI address validation to node validation https://review.opendev.org/c/openstack/ironic/+/964451 | 16:18 |
| TheJulia | At peak, I counted like 5x6 tiles on screen | 16:19 |
| cid | Yeah. At peak, I saw 38 participants, during the cross-project sessions with Nova and Neutron | 16:19 |
| TheJulia | yeah, I think I did end up counting then and got 36-38 which I think was beyond the tiling | 16:22 |
| JayF | I'll pass that on | 16:36 |
| JayF | and BTW TheJulia cid are you both +2 to work items? Ideally I could send that link out as our summary... | 16:36 |
| cid | When I looked at the older patch, looked good except the few commends about assignees | 16:39 |
| JayF | please formalize the vote in gerrit if you are OK with the content :) | 16:39 |
| cid | ack'd | 16:40 |
| TheJulia | rpittau: looks like the debian bullseye build has been failing on ipa-b for ages | 16:42 |
| rpittau | TheJulia: I haven't had the time to check that yet, it's in my todo list for the ipa debian stuff | 16:49 |
| rpittau | TheJulia: oh wait, I ahve a note about that | 16:49 |
| TheJulia | *looks* like it is quietly failing under the hood (i.e. one of the hardest dib problems to follow-up on | 16:50 |
| opendevreview | Riccardo Pittau proposed openstack/ironic-python-agent-builder master: Fix bullseye job https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/966204 | 16:52 |
| rpittau | TheJulia: ^^^ | 16:52 |
| opendevreview | Jay Faulkner proposed openstack/ironic-specs master: 2026.1 Ironic work items https://review.opendev.org/c/openstack/ironic-specs/+/965865 | 16:52 |
| opendevreview | Jay Faulkner proposed openstack/ironic-specs master: Add work items template, including maint checklist https://review.opendev.org/c/openstack/ironic-specs/+/965249 | 16:53 |
| opendevreview | Jay Faulkner proposed openstack/ironic-specs master: 2026.1 Ironic work items https://review.opendev.org/c/openstack/ironic-specs/+/965865 | 16:54 |
| opendevreview | Riccardo Pittau proposed openstack/ironic-python-agent-builder master: Fix bullseye job https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/966204 | 16:54 |
| opendevreview | Clif Houck proposed openstack/ironic master: Fix docstring for SystemdConsoleContainer.start_container https://review.opendev.org/c/openstack/ironic/+/965980 | 18:18 |
| opendevreview | Clif Houck proposed openstack/ironic master: Fix docstring for SystemdConsoleContainer.start_container https://review.opendev.org/c/openstack/ironic/+/965980 | 18:54 |
| JayF | trivial docs fix needs +2A https://review.opendev.org/c/openstack/ironic/+/965980 | 19:45 |
| opendevreview | Jay Faulkner proposed openstack/ironic-python-agent-builder master: Fix bullseye job https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/966204 | 20:00 |
| iurygregory | JayF, +W | 20:38 |
| alegacy | just an fyi those that have been kindly reviewing my networking patches, i've got the top one refactored into 5 commits. Need to go back over it manually to make sure I didn't leave too many stray ends hangin' around and then I'll push up a new series. | 21:21 |
| TheJulia | alegacy: much appreciated, thanks! | 21:41 |
| cardoe | TheJulia: trying to write up a proposed storage piece for the bug report to make my notes less crummy. | 21:41 |
| cardoe | So tell me if I understand the volume connector stuff correctly... | 21:41 |
| cardoe | A node should have multiple connectors with different types for example. So If I'm doing multipath, I'd want a connector with type=port connector_id=<baremetal-port-uuid> for each port and then type=iqn with the host_nqn or the initiator iqn | 21:42 |
| TheJulia | type=port was never done, but I was thinking on our side we would just quite literally populate the records, at least based upon the RFE I wrote up | 21:44 |
| TheJulia | I was thinking type=iqn, and type=host_nqn, and type=nqn on various distinct records | 21:45 |
| cardoe | So we're still not doing the network side with Ironic having any real awareness. | 21:46 |
| TheJulia | would it be easier to talk through? | 21:47 |
| cardoe | openstack baremetal volume connector create --node $NODE_UUID --type iqn --connector-id 'nqn.2014-08.org.nvmexpress:uuid:$INSTANCE_UUID' | 21:47 |
| cardoe | Is what we're running right now in response to a node moving to "deploying" | 21:48 |
| TheJulia | so your doing after the fact, interesting | 21:48 |
| TheJulia | why the instance_uuid? | 21:48 |
| cardoe | But remember, we're not booting from that volume. | 21:48 |
| cardoe | Cause this is for the OpenStack on OpenStack case. | 21:48 |
| cardoe | We're not actually attaching a volume in this case. | 21:49 |
| TheJulia | ok | 21:50 |
| cardoe | But I don't wanna shoe horn my weird case | 21:50 |
| cardoe | So trying to do this for the regular case we discussed but still have some compat. | 21:50 |
| TheJulia | no, I think it might help to talk through it a little bit which was sort of why I suggested the idea | 21:51 |
| cardoe | yeah | 21:51 |
| cardoe | google? | 21:51 |
| TheJulia | yeah, I'll past a link in a moment, closing some tabs first | 21:51 |
| TheJulia | https://meet.google.com/vxd-fegz-ubc | 21:52 |
| cardoe | Closing tabs is for the weak. I keep them open until my box OOMs. | 21:52 |
| TheJulia | lol | 21:52 |
| cardoe | Lemme see who else is around on my side. | 21:52 |
| TheJulia | I already started to, which is why I closed some | 21:52 |
| opendevreview | Merged openstack/ironic master: Fix docstring for SystemdConsoleContainer.start_container https://review.opendev.org/c/openstack/ironic/+/965980 | 22:06 |
| JayF | alegacy: maybe abandon https://review.opendev.org/c/openstack/ironic/+/962493/6 and put a pointer on it to the new changes? | 22:43 |
| JayF | iurygregory: would love a specific review from you on https://review.opendev.org/c/openstack/ironic-prometheus-exporter/+/966082 | 23:11 |
| iurygregory | JayF, it's on my list of priorities o/ | 23:45 |
| opendevreview | Merged openstack/ironic master: Add support for multiple JSON-RPC configuration groups https://review.opendev.org/c/openstack/ironic/+/960593 | 23:56 |
| opendevreview | Merged openstack/ironic master: Add force_dhcp option for managed inspection https://review.opendev.org/c/openstack/ironic/+/960824 | 23:58 |
| opendevreview | Merged openstack/sushy master: Handle missing storage controller mode attribute https://review.opendev.org/c/openstack/sushy/+/966072 | 23:59 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!