| arne_wiebalck | Good morning, Ironic! | 07:36 |
|---|---|---|
| arne_wiebalck | TheJulia: for the inspector issue? no, too many other issues I am struggling with at the moment ... | 07:37 |
| rpittau | good morning ironic! o/ | 07:43 |
| MahnoorAsghar | o/ | 07:43 |
| dtantsur | morning folks | 07:59 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Update /etc/keystone ownership on upgrade from Xena https://review.opendev.org/c/openstack/bifrost/+/828227 | 08:05 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: CI: properly report failures in the upgrade job https://review.opendev.org/c/openstack/bifrost/+/827687 | 08:06 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Only remove old services after they are stopped https://review.opendev.org/c/openstack/bifrost/+/828252 | 08:45 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Only remove old services after they are stopped https://review.opendev.org/c/openstack/bifrost/+/828252 | 08:47 |
| opendevreview | Merged openstack/bifrost stable/xena: Revert "Install libvirt-python from source instead of a wheel" https://review.opendev.org/c/openstack/bifrost/+/828178 | 08:47 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: Only remove old services after they are stopped https://review.opendev.org/c/openstack/bifrost/+/828252 | 09:17 |
| opendevreview | Dmitry Tantsur proposed openstack/bifrost master: CI: properly report failures in the upgrade job https://review.opendev.org/c/openstack/bifrost/+/827687 | 09:17 |
| opendevreview | Merged openstack/metalsmith stable/wallaby: Avoid non-Stream CentOS 8, it is EOL; make one job non-voting https://review.opendev.org/c/openstack/metalsmith/+/827234 | 09:23 |
| *** lifeless_ is now known as lifeless | 09:24 | |
| dtantsur | the outcome of a local testing: a (significantly larger) ubuntu image works, both debian and debian-minimal images do not... | 09:40 |
| parasiti1 | hi, i have one small question: i'm wondering what the 'my_ip' conf var is used for. My conductor is behind a FIP, so the conductor & horizon are sending me back the "wrong" URL for the ipmi console because it's built with the 'my_ip' conf var. Can i safely set the my_ip var to an "external ip" or is it important that it's kept to an internal network ip? thanks | 09:47 |
| dtantsur | parasiti1: I think it's also used for something in PXE | 09:48 |
| dtantsur | parasiti1: hmm, not, it's my_ipv6. anyway, there also an option [console]socat_address | 09:49 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: CI: use a custom centos partition image instead of cirros https://review.opendev.org/c/openstack/ironic/+/827882 | 10:02 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: [WIP] Fix dib ipa jobs https://review.opendev.org/c/openstack/bifrost/+/827293 | 10:02 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: [DNM] Testing the CI https://review.opendev.org/c/openstack/ironic/+/827500 | 10:02 |
| rpittau | seems like debian is causing some trouble for us :/ | 10:03 |
| parasiti1 | dtantsur: i'm using the shellinbox, which only sends back a url with my_ip. i'll try to switch to socat, but i have issues: it tells me it cant allocate ports.. | 10:12 |
| parasiti1 | dtantsur: with socat i hit this error: Error enabling the console on node 70daf724-4028-4e81-8abf-462651b2e857. Reason: Unable to allocate a free port on host localhost for IPMI terminal, not enough free ports.: ironic.common.exception.NoFreeIPMITerminalPorts: Unable to allocate a free port on host localhost for IPMI terminal, not enough free ports | 10:23 |
| parasiti1 | dtantsur: ok i think i know why, the conductor is a docker container with a pod IP, not the host Ip, and the socat impl checks that it can bind the port on the "my_ip" which is not set to the container ip, but to the host ip | 10:31 |
| parasiti1 | dtantsur: maybe i can use this socat_address and set it to the pod ip (or even try 0.0.0.0 :) | 10:33 |
| opendevreview | Mark Goddard proposed openstack/tenks master: docs: linux bridge, supported OS distros https://review.opendev.org/c/openstack/tenks/+/828278 | 10:57 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 11:16 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Change default boot mode to UEFI https://review.opendev.org/c/openstack/tenks/+/827553 | 11:16 |
| iurygregory | good morning Ironic o/ | 11:36 |
| janders | hey iurygregory o/ | 11:36 |
| TheJulia | good morning ironic | 13:04 |
| iurygregory | good morning TheJulia =) | 13:04 |
| rpittau | good morning TheJulia iurygregory :) | 13:05 |
| iurygregory | hey rpittau o/ | 13:05 |
| dmellado | ping dtantsur around? | 13:21 |
| * TheJulia tries to wake up | 13:24 | |
| iurygregory | anyone got this type of error with gerrit before: https://paste.opendev.org/show/812583/ ? | 13:36 |
| iurygregory | funny that complains about merge conflict (I've cloned the repo, created a new branch so it should match what master has LOL) | 13:38 |
| TheJulia | iurygregory: have you updated git-review? | 13:40 |
| iurygregory | TheJulia, nope, the F35 install is new (from January) all I did was install the git-review | 13:42 |
| opendevreview | Merged openstack/ironic-python-agent stable/xena: Make the standalone job voting again https://review.opendev.org/c/openstack/ironic-python-agent/+/827288 | 13:46 |
| TheJulia | dunno, maybe there is a compatability bug which hasn't been identified yet | 13:46 |
| TheJulia | ? | 13:46 |
| iurygregory | maybe .-. going to try a few more things to see if it helps | 13:47 |
| dtantsur | dmellado: around now | 13:50 |
| dtantsur | good morning TheJulia | 13:50 |
| rpittau | iurygregory: that's due to an old version of git-review | 13:51 |
| dtantsur | team, https://review.opendev.org/c/openstack/bifrost/+/828227/ https://review.opendev.org/c/openstack/bifrost/+/828252/ and https://review.opendev.org/c/openstack/bifrost/+/827687 are needed to fix the bifrost upgrade jobs. please review. | 13:51 |
| rpittau | iurygregory: you should install version 2.2.0 or newer | 13:51 |
| iurygregory | rpittau, oh | 13:51 |
| iurygregory | ok let me check the version I have | 13:51 |
| dtantsur | iurygregory: https://bugzilla.redhat.com/show_bug.cgi?id=2039224 | 13:51 |
| iurygregory | jesus 1.28.0 | 13:52 |
| dtantsur | there is a patch in bz that you can apply locally | 13:53 |
| rpittau | or just install git-review per user :) | 13:53 |
| iurygregory | yeah, git is 2.34.1 ... | 13:53 |
| iurygregory | tks! | 13:53 |
| rpittau | well I fixed the issue with the dhcp all interfaces, although now we have a new one! \o/ | 13:54 |
| dtantsur | sweeet | 13:54 |
| rpittau | EFI ftw -> https://paste.openstack.org/show/812585/ | 13:55 |
| arne_wiebalck | Bare Metal SIG meeting in 5 mins! | 13:55 |
| iurygregory | tks for the reminder arne_wiebalck o/ | 13:56 |
| dtantsur | rpittau: if this is a whole disk image, there is an element to add for UEFI support | 13:56 |
| dtantsur | rpittau: block-device-efi | 13:56 |
| * dtantsur will need to watch the meeting in recording.. | 13:57 | |
| rpittau | mmm I'll check that | 13:58 |
| rpittau | thanks | 13:58 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: [WIP] Fix dib ipa jobs https://review.opendev.org/c/openstack/bifrost/+/827293 | 14:09 |
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic master: [DNM] Testing CI https://review.opendev.org/c/openstack/ironic/+/828338 | 14:09 |
| iurygregory | \o/ pip ++ <3 | 14:10 |
| arne_wiebalck | Thanks TheJulia, this was a great session! | 14:51 |
| TheJulia | Thanks! | 14:52 |
| *** akaha|rover is now known as akahat|dinner | 14:55 | |
| NobodyCam | I concur arne_wiebalck | 14:58 |
| TheJulia | I was surprised how long that ran | 15:03 |
| dtantsur | iurygregory: wanna hear the explanation of the magical behavior yesterday? when my feature worked depending on the moon phase? | 15:09 |
| iurygregory | dtantsur, yes | 15:10 |
| iurygregory | :D | 15:10 |
| dtantsur | the key is https://review.opendev.org/c/openstack/bifrost/+/828252 | 15:10 |
| TheJulia | moon phasing features... sound awesome | 15:11 |
| dtantsur | I made a mistake in bifrost upgrade that resulted in the old ironic-api and ironic-conductor COEXISTING with the new ironic service | 15:11 |
| iurygregory | oh wow | 15:11 |
| dtantsur | so, depending on the random movements of the atoms, I ether got the old or the new conductor handling my deployment | 15:11 |
| dtantsur | I nearly lost the rest of my sanity on this issue :) | 15:11 |
| iurygregory | and how the starts are aligned =D | 15:12 |
| TheJulia | dtantsur: where did you get more sanity?!? | 15:12 |
| TheJulia | I need to order some | 15:12 |
| dtantsur | speaking of which, these patches are blocking bifrost release: https://review.opendev.org/c/openstack/bifrost/+/828227/ https://review.opendev.org/c/openstack/bifrost/+/828252/ and https://review.opendev.org/c/openstack/bifrost/+/827687 | 15:12 |
| dtantsur | TheJulia: welll... I found some on the desk, not sure how put it there | 15:12 |
| dtantsur | less fun stuff: I cannot build a working debian partition image. Ubuntu and CentOS seem to work, Debian does not :( | 15:13 |
| iurygregory | dtantsur, maybe zigo has some ideas? | 15:14 |
| zigo | o/ | 15:14 |
| zigo | About what? | 15:14 |
| dtantsur | maybe, but it's probably something DIB does rather than a Debian issue | 15:14 |
| dtantsur | zigo: so, in case you have ideas. I was building an image for Ironic with $ disk-image-create debian-minimal baremetal dhcp-all-interfaces cloud-init grub2 | 15:15 |
| dtantsur | the resulting partition (= only root) image can be deployed by ironic, but does not boot beyond grub (it drops in the shell) | 15:15 |
| dtantsur | do you have any ideas at least where to dig? | 15:15 |
| zigo | Did you try the grub-cloud variant? | 15:16 |
| zigo | grub-cloud-amd64 | 15:16 |
| * TheJulia wonders if the base image is whole disk vs partition | 15:16 | |
| dtantsur | zigo: mmm? what's the difference? | 15:16 |
| dtantsur | TheJulia: all base images are whole disk ones. | 15:16 |
| zigo | It has support for BOTH uefi AND bios. | 15:17 |
| zigo | Not sure if your issue is UEFI / BIOS related ... | 15:17 |
| dtantsur | zigo: okay, interesting, I'll give it a spin | 15:17 |
| TheJulia | dtantsur: so... according to stevebaker, base image extraction only actually works for partition source images | 15:17 |
| dtantsur | TheJulia: this cannot be true. I don't think we've ever had any partition source images except for cirros. | 15:17 |
| TheJulia | He metioned it to me yesterday and that he was going to talk to Ian before proceeding | 15:17 |
| dtantsur | I tested CentOS this morning. | 15:17 |
| TheJulia | Last centos image I pulled down was single partition in a whole disk image | 15:18 |
| TheJulia | maybe that is what he was talking about | 15:18 |
| zigo | dtantsur: Let me know how it goes, I may have more ideas. Also, make sure your setup really has GRUB_TERMINAL="console serial" so that Grub output to both. | 15:18 |
| dtantsur | TheJulia: this does not match my experience, lemme check the image DIB downloaded | 15:18 |
| TheJulia | dtantsur: ++ | 15:18 |
| dtantsur | TheJulia: okay, it's one partition, but it's still a whole disk image. Note the difference: https://paste.opendev.org/show/812592/ | 15:21 |
| dtantsur | (also "thank you" CentOS for not making a UEFI-compatible image in year 2022) | 15:22 |
| TheJulia | I think it was just disks with multiple partitions, tbh | 15:22 |
| dtantsur | so, this works; I suspect that Debian is built using debootstrap instead | 15:23 |
| TheJulia | is there a debian-minimal vs debian difference? | 15:23 |
| dtantsur | I tried both, the size difference was a few megabytes | 15:23 |
| dtantsur | zigo: so, grub-cloud-amd64, is it a package or a DIB element? I cannot find it in DIB | 15:24 |
| zigo | A package. | 15:24 |
| dtantsur | ah, hmm | 15:25 |
| dtantsur | zigo: this is what DIB installs: https://opendev.org/openstack/diskimage-builder/src/branch/master/diskimage_builder/elements/grub2/pkg-map#L3-L8 | 15:25 |
| dtantsur | also note that grub does actually start for me. it just drops in the shell. | 15:26 |
| rpittau | success!!! \o/ https://review.opendev.org/c/openstack/bifrost/+/827293 | 15:26 |
| dtantsur | w00t! | 15:26 |
| rpittau | sorry :D | 15:26 |
| rpittau | I had that in troubleshooting since mid december | 15:27 |
| zigo | dtantsur: That's a list that would work for Ubuntu, grub-cloud-amd64 only exists in Debian, so if you wish to switch to it, you'd need to differentiate. | 15:27 |
| zigo | https://packages.ubuntu.com/search?keywords=grub-cloud-amd64 <--- Gives no result ... | 15:27 |
| zigo | dtantsur: Does grub give you any clue of what's going on? | 15:28 |
| zigo | No error message or something? | 15:28 |
| dtantsur | nothing | 15:28 |
| dtantsur | please don't tell me we hardcode username "cirros" somewhere..... | 15:29 |
| dmellado | I hope it wasn't the case xD | 15:31 |
| dmellado | ok, so after all my fight and work with bifrost, I submitted a session for the summit | 15:32 |
| parasiti1 | hi again, i'm getting deeper into my comprehension of ironic+neutron integration. there seems to be one step i've completely missed. when ironic tries to plug in the tenant port, it seems that it has to deal with VIFs objects. (as described in ironic/ironic/drivers/modules/network/common.py:235) | 15:33 |
| parasiti1 | but i really dont now what VIFs objects are related to. | 15:33 |
| TheJulia | vifs are just neutron ports | 15:34 |
| parasiti1 | TheJulia: you mean i have to pre create a neutron port related to my ironic port ? | 15:34 |
| TheJulia | for standalone ironic+neutron, yes | 15:35 |
| TheJulia | if your nova integrated, nova has a bunch of cases where it will create VIFs automatically | 15:35 |
| parasiti1 | Ok. How my neutron port should be created so that it maps my ironic port ? by setting the same MAC ? | 15:37 |
| TheJulia | Ironic will reset the mac for you | 15:40 |
| TheJulia | realistically just a port on a network so an address is assigned and all | 15:40 |
| TheJulia | and provided to the node post-deployment | 15:40 |
| parasiti1 | ok thanks | 15:41 |
| parasiti1 | should i see the neutron ports i manually created when i run the "baremetal node vif list" command ? | 15:42 |
| parasiti1 | or does it list only "attached" vifs ? | 15:43 |
| TheJulia | parasiti1: only if they have been "attached" to the node | 15:44 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: CI: use a custom centos partition image instead of cirros https://review.opendev.org/c/openstack/ironic/+/827882 | 15:44 |
| TheJulia | parasiti1: if not needed, they should be removed generally. You can't have more vifs than ports or portgroups as well. | 15:45 |
| parasiti1 | TheJulia: i don't get it :) i tried to follow the "ironic/doc/source/admin/multitenancy.rst" doc. if i got what you said right, i should "openstack port create..." ports after "baremetal port create" but before deploying the node (with metalsmith for instance) ? | 15:49 |
| TheJulia | but ports I believe metalsmith will record the port attachment | 15:50 |
| parasiti1 | TheJulia: Oh, so i shall : 1. baremetal port create, 2. openstack port create, 3. baremetal node vif attach "neutron-port-id", 4. metalsmith deploy ? | 15:53 |
| TheJulia | parasiti1: I'm unsure because I don't know the internals or interaction with metalsmith | 15:54 |
| ameya49 | TheJulia : Morning, with reference to this document : https://docs.openstack.org/ironic/latest/install/configure-nova-flavors.html Question : Is cpu_arch used for scheduling? | 16:20 |
| TheJulia | ameya49: two places, nova instance scheduling by cpu type *and* it is used if you have to do pxe loaders by architecture | 16:27 |
| NobodyCam | Good Morning OpenStack folks | 16:31 |
| NobodyCam | Now that I am actually awake | 16:31 |
| dtantsur | and that's still only Tuesday - impressive job NobodyCam :D | 16:33 |
| * dtantsur hopes to become awake around tomorrow | 16:33 | |
| NobodyCam | hehehehe | 16:34 |
| NobodyCam | o/ Morning dtantsur | 16:34 |
| arne_wiebalck | TheJulia: ameya49: isn't the scheduling based on resource classes (and traits) ? | 16:42 |
| TheJulia | it is, but CPU aiui is still required | 16:42 |
| parasiti1 | TheJulia: effectively, , i didnt have to create any neutron port when using metalsmith. just add the 'metalsmith deploy --network NETWORK' option, and metalsmith creates the neutron ports and 'vif attach' them automatically. thanks for the info | 16:44 |
| TheJulia | parasiti1: ack, thanks | 16:44 |
| arne_wiebalck | TheJulia: setting the number of CPUs may be needed (but not for scheduling), yes, but cpu_arch? ameya49: where do you use cpu_arch? | 16:46 |
| dtantsur | I think cpu_arch is matched as part of extra specs | 16:47 |
| dtantsur | but my nova knowledge is pretty rusty nowadays | 16:47 |
| TheJulia | it is populated into extra specs | 16:47 |
| arne_wiebalck | we do not have this in our flavors ... and my understanding was all but resource classes (and traits) is irrelevant for scheduling | 16:48 |
| dtantsur | CPU/memory/disk are irrelevant, but cpu_arch (and other extra specs) may still be used | 16:48 |
| TheJulia | https://github.com/openstack/nova/blob/e537d90d6fc0977742f7126c3f8cfef6bf8b2a15/nova/virt/ironic/driver.py#L371 | 16:48 |
| dtantsur | somebody should check the nova code :) | 16:48 |
| dtantsur | yeah, exactly | 16:48 |
| TheJulia | :) | 16:48 |
| arne_wiebalck | and we're sure this is not a left-over which is ignored? | 16:50 |
| arne_wiebalck | it is just that anything but RCs is used would be total news to me | 16:51 |
| arne_wiebalck | which does not mean it is not the case ofc :-D | 16:52 |
| ameya49 | TheJulia: introspection basic test case is using it for verifying the introspection data with flavor | 17:01 |
| *** akahat|dinner is now known as akahat|rover | 17:05 | |
| *** sshnaidm is now known as sshnaidm|afk | 17:18 | |
| rpittau | good night! o/ | 17:32 |
| arne_wiebalck | bye everyone, see you tomorrow o/ | 17:47 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: [WIP] Create API documentation from docstrings https://review.opendev.org/c/openstack/ironic/+/827200 | 17:54 |
| opendevreview | Ruby Loo proposed openstack/ironic master: More fixes for anaconda deploy interface https://review.opendev.org/c/openstack/ironic/+/827924 | 17:55 |
| opendevreview | Ruby Loo proposed openstack/ironic master: Anaconda deploy handles configdrive correctly https://review.opendev.org/c/openstack/ironic/+/827933 | 17:58 |
| dtantsur | see you tomorrow folks o/ | 18:20 |
| TheJulia | o/ | 18:28 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/queens: Remove legacy experimental jobs https://review.opendev.org/c/openstack/ironic/+/827713 | 18:40 |
| stevebaker[m] | good morning | 19:09 |
| stevebaker[m] | ah I have a video to edit :) | 19:33 |
| opendevreview | Verification of a change to openstack/ironic stable/xena failed: ImageCache: respect Cache-Control: no-store https://review.opendev.org/c/openstack/ironic/+/826370 | 19:54 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/train: Remove redfish cache entry upon errors https://review.opendev.org/c/openstack/ironic/+/820937 | 20:28 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Auto-populate lessee for deployments https://review.opendev.org/c/openstack/ironic/+/818641 | 20:32 |
| TheJulia | good morning stevebaker[m] | 20:33 |
| stevebaker[m] | gah | 20:33 |
| opendevreview | mitya-eremeev-2 proposed openstack/ironic master: Delete resource provider after node deletion. https://review.opendev.org/c/openstack/ironic/+/827295 | 20:38 |
| opendevreview | Merged openstack/bifrost master: Update /etc/keystone ownership on upgrade from Xena https://review.opendev.org/c/openstack/bifrost/+/828227 | 20:47 |
| opendevreview | mitya-eremeev-2 proposed openstack/ironic master: Delete resource provider after node deletion. https://review.opendev.org/c/openstack/ironic/+/827295 | 20:48 |
| *** osmanlicilegi is now known as Guest1 | 20:56 | |
| opendevreview | Merged openstack/bifrost master: Only remove old services after they are stopped https://review.opendev.org/c/openstack/bifrost/+/828252 | 21:10 |
| opendevreview | Merged openstack/bifrost master: CI: properly report failures in the upgrade job https://review.opendev.org/c/openstack/bifrost/+/827687 | 21:29 |
| opendevreview | Verification of a change to openstack/ironic stable/queens failed: Remove legacy experimental jobs https://review.opendev.org/c/openstack/ironic/+/827713 | 21:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!