| opendevreview | Verification of a change to openstack/ironic master failed: Allocation candidates prefer matching name https://review.opendev.org/c/openstack/ironic/+/845856 | 00:03 |
|---|---|---|
| opendevreview | Merged openstack/ironic master: Make anaconda non-image deploys sane https://review.opendev.org/c/openstack/ironic/+/834709 | 01:28 |
| opendevreview | Merged openstack/ironic master: Trivial: log which state the node is in https://review.opendev.org/c/openstack/ironic/+/848874 | 01:37 |
| opendevreview | Verification of a change to openstack/ironic-python-agent master failed: Drop support for instance netboot https://review.opendev.org/c/openstack/ironic-python-agent/+/848969 | 01:37 |
| opendevreview | Aija Jauntēva proposed openstack/ironic stable/yoga: Fix markup typo in Redfish driver docs https://review.opendev.org/c/openstack/ironic/+/849810 | 08:32 |
| opendevreview | Aija Jauntēva proposed openstack/ironic bugfix/20.2: Fix markup typo in Redfish driver docs https://review.opendev.org/c/openstack/ironic/+/849811 | 08:32 |
| dtantsur | morning folks | 09:01 |
| janders | good morning dtantsur o/ | 09:03 |
| iurygregory | good morning Ironic | 11:29 |
| timeu | Hi Ironic, did anybody notice a recent change in the trunk port implementation in the latest OSP 16.2.3 release ? For some reason the subport of the trunk has 'binding:vnic_type': 'normal' instead of 'binding:vnic_type': 'baremetal' which causes our SDN's ML2 driver not to properly bind the port. This worked in a previous OSP (16.2.2) release. | 11:32 |
| dtantsur | I guess only OSP folks (TheJulia, hjensas, stevebaker[m]) know the diff | 11:35 |
| timeu | I am trying to bisect the changes and see if the issue came via a neutron minor upgrade or an upgrade of our SDN's plugin. | 11:43 |
| timeu | the parent port is properly bound but it also has barmetal as vnic_type while the subport has normal. I will dig through the code. | 11:44 |
| timeu | In other news the latest sushy 4.2.0 breaks redfish on the Lenovo hardware. I added a comment with the error message here: https://storyboard.openstack.org/#!/story/2009680 | 11:45 |
| opendevreview | kamlesh chauvhan proposed openstack/ironic-tempest-plugin master: Add iDRAC RAID cleaning steps tests https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/841601 | 11:47 |
| timeu | is there an easy way to find out which RPM version (i.e. openstack-neutron) corresponds to which upstream version ? | 11:57 |
| iurygregory | I think you have the information about the commit in the rpm name | 12:42 |
| iurygregory | so you could check the branch | 12:42 |
| TheJulia | Openstack releases with major and minor | 12:46 |
| timeu | iurygregory: thanks. This is the version that don't work neutron-15.3.5.dev132-py3.6 and this is the version it worked last time: neutron-15.2.1.dev12-py3.6. Not sure if this corresponds to the neutron git tags | 12:51 |
| timeu | It's also really hard for me right now to figure out what the expected behavior is. Before binding the port the first time to a baremetal node the vnic_type is normal. Once ironic does its thing the vnic_type of the parent bound is set to baremetal. Not sure who should update the vnic_type of the subport ? Is it the trunk service ? ironic or the mechanism driver ? | 12:53 |
| iurygregory | 15.3.5 is probably the latest release from train (train-em tag...) | 12:53 |
| iurygregory | https://opendev.org/openstack/releases/src/branch/master/deliverables/train/neutron.yaml | 12:53 |
| TheJulia | Ironic does the update | 12:54 |
| timeu | ah thansk for the link I guess with this I can figure out the commits and see what changed | 12:54 |
| timeu | ok so it might not be a neutron thing but maybe a change in ironic | 12:54 |
| timeu | I will try to compare the 2 versions I have and see what I can figure out. So the last time it worked (at least when I tested it) it was still with OSP 16.1. I tried it now again with OSP 16.2.3 and it didn't. I have to leave for today. Will debug it further tomorrow | 12:57 |
| TheJulia | timeu: I can look through changes in the packaging, but most likely taking tomorrow off | 12:59 |
| timeu | Thanks. that would be great. it's not urgent. We are just testing the feature in our dev environment. One last question is networking-baremetal responsible for updating the port information or ironic itself ? | 13:00 |
| kaloyan | Hey guys, I wanted to ask something else. We have this strange setup where we're migrating from an inhouse baremetal provisioning solution to Ironic and OpenStack in general (Nova, Neutron, Cinder). I have a machine with 6 ports on it, all described in Ironic. I've marked a port as PXE capable and configured the switches so that it and the rest of Neutrons' stack are in the | 13:06 |
| kaloyan | same vlan. The provider network is of type "flat" as well as the networking_interface field for that machine. When the machine boots, it loads up cloud-init and talks with Nova to fetch metadata. The network metadata received from Nova contains only one link as well as one network which is expected, since this is the only information Nova knows about. My question is the | 13:06 |
| kaloyan | following: How can I make Nova send information for the remaining ports? Shall I add ports in Neutron that aren't attached to a network? | 13:06 |
| kaloyan | Hmmm, looking at the Ironic driver for Nova, it looks like it it generates a config drive that gets passed to ironic when the spawn method of the driver is called | 13:11 |
| kaloyan | However, when booting from volume (using iSCSI) this is no longer the case, as cloud-init would talk to the metadata service which would return network metadata that only Nova knows | 13:14 |
| TheJulia | timeu: networking-baremetal does not update port info in ironic, it updates physical network mapping daata in neutron and tells port binding for flat networks that it is "successful" | 13:20 |
| TheJulia | timeu: ironic does the update actions. fwiw, I don't think anything really changed there, but I'm aware of an issue with OVN in the mix that I'm wondering if maybe your encountering. Do you happen to be trying to use OVN? | 13:21 |
| TheJulia | kaloyan: I guess, for our context, the question for us to ask is what would you use the other interfaces for? if my memory is serving me correctly, it matches the mac address in cloud init and configures the interface... | 13:23 |
| kaloyan | We intend to use them for testing our product (distributed storage software) so it requires networking | 13:24 |
| TheJulia | so you can issue multiple vif attachments.... | 13:27 |
| TheJulia | but I don't think the metadata gets in nova, since it is for the machine to boot, the other interfaces work via dhcp | 13:27 |
| * TheJulia checks the docs | 13:28 | |
| TheJulia | kaloyan: so you can create an instance with multiple nics on launch | 13:29 |
| TheJulia | and you'd need to have it's networks in neutron as well of course for that to work | 13:29 |
| TheJulia | I would assume that the additional ports would be in the cloud-init metadata as well as have been setup configuration wise | 13:30 |
| kaloyan | TheJulia: hmm... I understand. Now, given the fact that Ironic "injects" the rest of the ports in the network metadata when using a config drive (https://opendev.org/openstack/nova/src/commit/de65131f92ba5ba812e33e6ff63be0991687413a/nova/virt/ironic/driver.py#L994) isn't it a good idea to do the same when requesting metadata from Nova? | 13:41 |
| kaloyan | I'm really trying to avoid declaring phantom Neutron networks and ports | 13:42 |
| TheJulia | well, they aren't really phantom but it is a one or the other sort of thing | 13:46 |
| TheJulia | we have no means to inject/modify the metadata | 13:46 |
| kaloyan | I beleive this is a question that the team behind Nova should answer, isn't it? | 13:47 |
| TheJulia | likely | 13:47 |
| TheJulia | however, that pattern requires vifs to be registered which are neutron ports | 13:47 |
| kaloyan | yeah, I missed the part of the code that actually | 13:50 |
| ashinclouds[m] | I was reading it and going “wow, that is kind of not clear at a glance” | 13:51 |
| ashinclouds[m] | Err wrong client | 13:51 |
| TheJulia | \o: | 13:51 |
| TheJulia | Gah | 13:51 |
| TheJulia | I need more coffee | 13:51 |
| dtantsur | huh, CS9 broken in Bifrost? | 14:41 |
| iurygregory | new results or something from yesterday? | 14:58 |
| iurygregory | yesterday there was the problem with tarbals website | 14:59 |
| dtantsur | No DHCP during iPXE | 15:00 |
| dtantsur | UTF-16 stream does not start with BOM <== has been fixed long ago, no? | 15:01 |
| dtantsur | mmm, yeah, I've got some old IPA in my metal3-dev-env | 15:04 |
| dtantsur | oh, https://images.rdoproject.org/centos8/master/rdo_trunk/current-tripleo/ hasn't been updated since February. oh. | 15:05 |
| dtantsur | anyone looking into the new jsonschema? | 15:33 |
| dtantsur | - Failed: 134 | 15:37 |
| dtantsur | OMFG | 15:37 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Fix compatibility with jsonschema>=4.0.0 https://review.opendev.org/c/openstack/ironic/+/849882 | 16:08 |
| TheJulia | wow | 16:37 |
| TheJulia | that is all the wow | 16:37 |
| * TheJulia goes and gets water to prepare for next call of the day | 16:42 | |
| dtantsur | o/ | 17:00 |
| TheJulia | goodnight dtantsur | 17:01 |
| opendevreview | Merged openstack/ironic stable/yoga: Update known issue for iDRAC Swift firmware update https://review.opendev.org/c/openstack/ironic/+/849634 | 17:23 |
| opendevreview | Merged openstack/ironic bugfix/20.2: Update known issue for iDRAC Swift firmware update https://review.opendev.org/c/openstack/ironic/+/849633 | 17:23 |
| erbarr | does OCP 4.10 use wallaby or xena? | 18:36 |
| TheJulia | dunno | 19:20 |
| TheJulia | iurygregory: do you know? | 19:20 |
| iurygregory | erbarr, none of them =) bugfix branches =) | 19:37 |
| erbarr | ok, which one and how do they relate to release branches then? XD | 19:38 |
| iurygregory | erbarr, for ironic is bugfix/19.0 | 19:39 |
| iurygregory | bufgix branches we cut them between stable releases | 19:39 |
| iurygregory | erbarr, dos this answer your question ? | 19:43 |
| erbarr | yea, thanks | 19:44 |
| iurygregory | yw =) | 19:46 |
| opendevreview | Merged openstack/ironic-python-agent master: Drop support for instance netboot https://review.opendev.org/c/openstack/ironic-python-agent/+/848969 | 22:00 |
| JayF | https://review.opendev.org/c/openstack/ironic/+/847136 trivial review of a doc update if someone wants to take a look | 22:26 |
| opendevreview | Verification of a change to openstack/ironic master failed: Allocation candidates prefer matching name https://review.opendev.org/c/openstack/ironic/+/845856 | 23:02 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!