| opendevreview | Doug Goldstein proposed openstack/ironic master: doc: highlight firmware interface and deprecate management for updates https://review.opendev.org/c/openstack/ironic/+/986233 | 01:47 |
|---|---|---|
| opendevreview | norman shen proposed openstack/python-ironicclient master: Add portgroup uuid to vif attach OSC client https://review.opendev.org/c/openstack/python-ironicclient/+/990409 | 01:58 |
| opendevreview | Doug Goldstein proposed openstack/python-ironicclient master: Bump openstacksdk lower bound https://review.opendev.org/c/openstack/python-ironicclient/+/990412 | 02:34 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: Crazy idea: RCOE Enabled BMaaS https://review.opendev.org/c/openstack/ironic-specs/+/990413 | 02:52 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: Crazy idea: RCOE Enabled BMaaS https://review.opendev.org/c/openstack/ironic-specs/+/990413 | 03:05 |
| opendevreview | OpenStack Proposal Bot proposed openstack/ironic master: Imported Translations from Zanata https://review.opendev.org/c/openstack/ironic/+/990418 | 03:49 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: Crazy idea: RCOE Enabled BMaaS https://review.opendev.org/c/openstack/ironic-specs/+/990413 | 04:02 |
| cardoe | TheJulia: ^ is how my setup works. | 04:44 |
| cardoe | That’s the green vs purple network in trait based networking. | 04:45 |
| cardoe | That’s why I was suggesting fabric aware scheduling of ports. | 04:45 |
| opendevreview | Himanshu Roy proposed openstack/ironic master: Disallow certain steps for users in different deployments https://review.opendev.org/c/openstack/ironic/+/989017 | 05:29 |
| opendevreview | Himanshu Roy proposed openstack/ironic master: Add conductor-level disallow step filtering https://review.opendev.org/c/openstack/ironic/+/990422 | 05:29 |
| opendevreview | Himanshu Roy proposed openstack/ironic master: Add operator-configurable step disallow lists https://review.opendev.org/c/openstack/ironic/+/989017 | 05:49 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Fix send_sensor_data periodic with oslo.service spawn workers https://review.opendev.org/c/openstack/ironic/+/990251 | 07:11 |
| opendevreview | Saad Zia proposed openstack/ironic master: Fix sensitive properties returned unredacted in POST and PATCH HTTP responses https://review.opendev.org/c/openstack/ironic/+/990430 | 08:25 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Document the Metal3 ironic-standalone-operator job https://review.opendev.org/c/openstack/ironic/+/989014 | 10:00 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Add a Metal3 baremetal-operator functional job https://review.opendev.org/c/openstack/ironic/+/988766 | 10:05 |
| dtantsur | JayF: updated per your feedback ^^ | 10:05 |
| opendevreview | Saad Zia proposed openstack/ironic master: Fix TARGET_SCHEMA bounds to prevent storage exhaustion (F-85) https://review.opendev.org/c/openstack/ironic/+/990440 | 10:10 |
| abongale | g | 11:13 |
| abongale | please ignore ^^ | 11:13 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Add a Metal3 baremetal-operator functional job https://review.opendev.org/c/openstack/ironic/+/988766 | 11:48 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Fix send_sensor_data periodic with oslo.service spawn workers https://review.opendev.org/c/openstack/ironic/+/990251 | 11:54 |
| opendevreview | Saad Zia proposed openstack/ironic master: Cap LLDP TLV and interface limits to mitigate DoS risk https://review.opendev.org/c/openstack/ironic/+/990452 | 12:08 |
| iurygregory | good morning ironic | 12:09 |
| opendevreview | Konstantinos Karampogias proposed openstack/ironic-specs master: Firmware endpoint hardware identity https://review.opendev.org/c/openstack/ironic-specs/+/985800 | 12:25 |
| opendevreview | Esther Domfeh proposed openstack/ironic master: feat: add state and timestamps to node history https://review.opendev.org/c/openstack/ironic/+/989994 | 12:50 |
| opendevreview | Saad Zia proposed openstack/ironic master: Cap host-supplied failure reason to 1024 https://review.opendev.org/c/openstack/ironic/+/990458 | 12:51 |
| opendevreview | Merged openstack/tenks master: CI: add aarch64 jobs https://review.opendev.org/c/openstack/tenks/+/989768 | 13:09 |
| opendevreview | Michal Nasiadka proposed openstack/tenks master: WIP: Support setting OS_CLOUD https://review.opendev.org/c/openstack/tenks/+/990462 | 13:18 |
| opendevreview | cid proposed openstack/ironic master: Allow disabling port pre-creation during inspection https://review.opendev.org/c/openstack/ironic/+/990340 | 13:34 |
| opendevreview | Saad Zia proposed openstack/ironic master: Fix accumulation of any_left flag across storage controllers https://review.opendev.org/c/openstack/ironic/+/990471 | 13:42 |
| JayF | TheJulia: others: I know in the past we've indicated performance issues with Serial consoles. Do we have any elaboration on that? | 14:15 |
| JayF | I don't mean from an Ironic standpoint; I mean from a hardware perf standpoint | 14:15 |
| cardoe | Yeah it slows down the box to have to thunk out every byte of dmesg out the UART | 14:33 |
| TheJulia | yeah, which means an interrupt trigger | 14:33 |
| TheJulia | which means every char is an interrupt triggering. Basically, avoid serial consoles or even defining it in config at all | 14:34 |
| JayF | console=ttyS0 enough to activate it? | 14:34 |
| JayF | or do you have to have something attached | 14:34 |
| TheJulia | yup. you can hit it even when not doing pure serial console but when your not using a graphical framebuffer. The framebuffer works without a interrupt trigger since its just appending out to a section of RAM the graphics card is reading from. | 14:35 |
| JayF | that's what I thought, I've passed it on | 14:36 |
| TheJulia | some other kernel command line arguments which modify that console behavior and all can also set one up for such a condition. I think we've edited our defaults to try and guide folks away from ending up in such a state | 14:36 |
| opendevreview | Harald Jensås proposed openstack/networking-generic-switch master: Add netconf_models package with OpenConfig lib https://review.opendev.org/c/openstack/networking-generic-switch/+/990059 | 15:15 |
| opendevreview | Harald Jensås proposed openstack/networking-generic-switch master: Add NetconfSwitch base class for NETCONF transport https://review.opendev.org/c/openstack/networking-generic-switch/+/990060 | 15:15 |
| opendevreview | Harald Jensås proposed openstack/networking-generic-switch master: Add NetconfOpenConfigSwitch driver https://review.opendev.org/c/openstack/networking-generic-switch/+/990061 | 15:15 |
| opendevreview | Harald Jensås proposed openstack/networking-generic-switch master: Add NETCONF OpenConfig driver documentation https://review.opendev.org/c/openstack/networking-generic-switch/+/990062 | 15:15 |
| opendevreview | Harald Jensås proposed openstack/networking-generic-switch master: Pass physnet VLAN segments to device drivers https://review.opendev.org/c/openstack/networking-generic-switch/+/990510 | 15:15 |
| opendevreview | Esther Domfeh proposed openstack/ironic master: feat: add state and timestamps to node history https://review.opendev.org/c/openstack/ironic/+/989994 | 15:28 |
| cardoe | Are we okay depreciating management interface for firmware updates? | 16:01 |
| TheJulia | I'm unsure if we reached consensus when it was discussed last | 16:02 |
| JayF | Hit the mailing list | 16:06 |
| JayF | that's consensus that needs to be gotten in an operator-visible place | 16:07 |
| TheJulia | oh, okay | 16:09 |
| TheJulia | cool | 16:09 |
| cardoe | Email sent. | 16:22 |
| JayF | BTW I am very +1 | 16:23 |
| JayF | just mailing list gives us a thing to point at later if someone complains :) | 16:23 |
| cardoe | https://review.opendev.org/c/openstack/ironic/+/986233 is the patch | 16:24 |
| dougszu | hey cid / JayF, here's the RFE for docker / podman container provider support https://bugs.launchpad.net/ironic/+bug/2154494 | 16:59 |
| dougszu | please reach out if we can help / clarify anything | 16:59 |
| JayF | cid: ^ can you put this on your list? No specific urgency | 16:59 |
| dougszu | many thanks, we're on hand to test / debug | 17:00 |
| JayF | dougszu: oooh, this is shaped differently than I expected | 17:00 |
| JayF | you want something that uses docker/podman directly, no systemd? | 17:00 |
| JayF | I don't know if we'll need a spec for that, but we'll probably have to get a little more detail in the RFE | 17:00 |
| dougszu | yeah, so it is much more of a headache to start systemd units from inside the ironic container | 17:00 |
| JayF | yeah, I think that's what the k-a folks were complaining about originally | 17:01 |
| JayF | so just something that invokes docker/podman directly | 17:01 |
| dougszu | indeed - kind of similar to the k8s provider | 17:01 |
| dougszu | it seems much simpler to do it this way, than try and bend Kolla. I'm happy to add more to the RFE tomorrow | 17:02 |
| dtantsur | cardoe: at the very least, we cannot leave the management interface for firmware updates as it is now. It's absolutely untested and probably unused by anyone. | 17:04 |
| JayF | dougszu: more detail is always better! | 17:07 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: Crazy Idea Part Deux! Discussing tighter storage fabric integration https://review.opendev.org/c/openstack/ironic-specs/+/990532 | 17:19 |
| opendevreview | Esther Domfeh proposed openstack/ironic master: feat: add state and timestamps to node history https://review.opendev.org/c/openstack/ironic/+/989994 | 17:22 |
| TheJulia | cardoe: so https://review.opendev.org/c/openstack/ironic-specs/+/990532 is more discussion oriented trying to find the basic paths which would work the best, so you may want to drop some thoughts on there | 17:26 |
| shermanm | had a NGS feature come up internally, is there any interest in support for the case where the BMC piggy-backs off an in-band interface, on a tagged vlan? | 17:29 |
| JayF | Those setups pretty much completely invalidate a lot of the network security Ironic bakes in -- I'm not opposed to supporting it for NGS, but that's not a configuration I'd suggest anyone use. | 17:31 |
| shermanm | so, I do agree, but it's come up a lot for our support of HPC clusters, they don't like extra cables :( | 17:31 |
| JayF | of all the reasons I've heard to use that architecture ... that one is the most unique :| | 17:32 |
| JayF | "I don't want to buy extra switches" I've heard, but not cables :D | 17:32 |
| shermanm | to be fair, in our case it's not greenfield deployments, it tends to be repurposing existing clusters with ironic+openstack orchestration | 17:33 |
| shermanm | so it's the difference between "change the head node" and "physically recable things" | 17:33 |
| shermanm | I would be interested in putting together some reliability/security best practices for the "so you've decided to do this" case | 17:35 |
| TheJulia | There is a whole other headache there which I know dtantsur has encountered a lot, some bmcs become inaccessible when they are piggy backing on an interface even when the host is powered down in general. :\ | 18:00 |
| TheJulia | But, yeah, different vendors anda ll | 18:00 |
| TheJulia | shermanm: I think that would be good and a great document to have in general because I've had several "so you've decided to do $x when our docs say $y is best" | 18:01 |
| TheJulia | in the last few weeks | 18:01 |
| opendevreview | cid proposed openstack/ironic master: Allow disabling port pre-creation during inspection https://review.opendev.org/c/openstack/ironic/+/990340 | 18:19 |
| cid | JayF, dougszu. I will take a look at that. | 18:28 |
| opendevreview | Steve Baker proposed openstack/ironic master: DNM re-enable graphical console testing https://review.opendev.org/c/openstack/ironic/+/987442 | 21:03 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!