| openstackgerrit | Julia Kreger proposed openstack/bifrost master: DNM Enable remaining ubuntu jobs https://review.opendev.org/701851 | 00:11 |
|---|---|---|
| TheJulia | https://review.opendev.org/#/c/703522/1 <-- wow, both xenial and bionic passed | 00:12 |
| patchbot | patch 703522 - bifrost - Try shorthand format for console - 1 patch set | 00:12 |
| *** igordc has quit IRC | 00:25 | |
| *** igordc has joined #openstack-ironic | 00:25 | |
| *** igordc has quit IRC | 00:49 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Try shifting virtualbmc start https://review.opendev.org/703528 | 00:50 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: DNM Enable remaining ubuntu jobs https://review.opendev.org/701851 | 00:50 |
| TheJulia | etingof: any idea why with python3, virtualbmc blows up trying to start a second virtualbmc instance? | 00:51 |
| *** ricolin has joined #openstack-ironic | 00:55 | |
| openstackgerrit | Bob Fournier proposed openstack/ironic master: Don't require root partition when installing a whole disk image https://review.opendev.org/702293 | 01:01 |
| TheJulia | etingof: if you look at the logs here https://d64b7553b36edb75caae-206dd41ab15ee6b59e703e19aac8d4d8.ssl.cf1.rackcdn.com/701851/4/check/bifrost-integration-dhcp-ubuntu-xenial/506964c/ you'll see the weirdness. Note, it looks like if I only try to start one vbmc instance, things _may_ be fine... but that seems... weird | 01:14 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Keystone: Add make /etc/keystone/fernet-keys https://review.opendev.org/703529 | 01:17 |
| *** jistr has quit IRC | 01:17 | |
| *** jistr has joined #openstack-ironic | 01:19 | |
| TheJulia | dtantsur|afk: https://d486eaf604ecde1cb103-b9f037744a20529c427d420cc2ffc597.ssl.cf2.rackcdn.com/701851/4/check/bifrost-integration-dibipa-debian-ubuntu-xenial/a3053ad/job-output.txt <-- look at the bottom when disk-image-create fires off :( | 01:20 |
| TheJulia | etingof: yeah, shifting it didn't seem to fix it :\ | 01:24 |
| *** dsneddon has quit IRC | 01:48 | |
| *** Lucas_Gray has quit IRC | 01:49 | |
| *** hoonetorg has quit IRC | 02:17 | |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: Fix and return the CentOS 7 job https://review.opendev.org/703371 | 02:21 |
| *** tzumainn has quit IRC | 02:22 | |
| *** dsneddon has joined #openstack-ironic | 02:24 | |
| *** goldyfruit has quit IRC | 02:30 | |
| *** hoonetorg has joined #openstack-ironic | 02:31 | |
| *** ociuhandu has joined #openstack-ironic | 02:47 | |
| *** dsneddon has quit IRC | 02:49 | |
| *** ociuhandu has quit IRC | 02:51 | |
| *** dsneddon has joined #openstack-ironic | 03:16 | |
| *** dsneddon has quit IRC | 03:22 | |
| *** hwoarang has quit IRC | 03:39 | |
| *** hwoarang has joined #openstack-ironic | 03:40 | |
| *** dsneddon has joined #openstack-ironic | 03:50 | |
| *** dsneddon has quit IRC | 03:59 | |
| *** mbeierl has quit IRC | 04:03 | |
| *** mkrai_ has joined #openstack-ironic | 04:18 | |
| *** cdearborn has quit IRC | 04:21 | |
| *** dsneddon has joined #openstack-ironic | 04:26 | |
| *** dsneddon has quit IRC | 04:31 | |
| *** goldyfruit has joined #openstack-ironic | 04:34 | |
| *** dsneddon has joined #openstack-ironic | 05:03 | |
| *** dsneddon has quit IRC | 05:08 | |
| *** jdandrea has quit IRC | 05:19 | |
| *** dsneddon has joined #openstack-ironic | 05:41 | |
| *** dsneddon has quit IRC | 05:46 | |
| *** yolanda has quit IRC | 06:04 | |
| *** dsneddon has joined #openstack-ironic | 06:20 | |
| *** dsneddon has quit IRC | 06:25 | |
| *** dsneddon has joined #openstack-ironic | 06:57 | |
| *** TxGirlGeek has joined #openstack-ironic | 06:59 | |
| *** dsneddon has quit IRC | 07:02 | |
| *** TxGirlGeek has quit IRC | 07:07 | |
| *** yolanda has joined #openstack-ironic | 07:25 | |
| *** jtomasek has joined #openstack-ironic | 07:25 | |
| *** ociuhandu has joined #openstack-ironic | 07:30 | |
| *** dsneddon has joined #openstack-ironic | 07:33 | |
| *** yolanda has quit IRC | 07:33 | |
| arne_wiebalck | Good morning, ironic! | 07:34 |
| *** yolanda has joined #openstack-ironic | 07:34 | |
| *** ociuhandu has quit IRC | 07:35 | |
| *** dsneddon has quit IRC | 07:38 | |
| *** yolanda has quit IRC | 07:48 | |
| *** bnemec has joined #openstack-ironic | 08:06 | |
| *** iurygregory has joined #openstack-ironic | 08:11 | |
| iurygregory | good morning o/ | 08:11 |
| *** dsneddon has joined #openstack-ironic | 08:13 | |
| *** dsneddon has quit IRC | 08:19 | |
| *** tesseract has joined #openstack-ironic | 08:20 | |
| *** yolanda has joined #openstack-ironic | 08:21 | |
| sziviani | Good morning :) | 08:22 |
| *** priteau has joined #openstack-ironic | 08:29 | |
| gudrutis21 | O/ | 08:30 |
| openstackgerrit | Thomas Bechtold proposed openstack/ironic master: Drop rootwrap.d/ironic-lib.filters file https://review.opendev.org/684270 | 08:44 |
| *** dsneddon has joined #openstack-ironic | 08:48 | |
| *** rpittau|afk is now known as rpittau | 08:48 | |
| rpittau | good morning ironic! o/ | 08:48 |
| *** dsneddon has quit IRC | 08:53 | |
| iurygregory | morning rpittau o/ | 08:56 |
| rpittau | hey iurygregory :) | 08:57 |
| *** dtantsur|afk is now known as dtantsur | 08:57 | |
| dtantsur | morning ironic | 08:57 |
| iurygregory | morning dtantsur | 08:57 |
| dtantsur | TheJulia: something with venv handling inside DIB? | 08:58 |
| dtantsur | I wonder if it requires python-virtualenv, not python-venv.. | 08:58 |
| *** iurygregory has quit IRC | 08:59 | |
| rpittau | I wonder if we still need xenial considering that it s upports only python3.5 | 09:02 |
| rpittau | dtantsur: I believe you're right, the "problem" is how DIB manages virtualenvs | 09:08 |
| *** goldyfruit has quit IRC | 09:11 | |
| *** iurygregory has joined #openstack-ironic | 09:13 | |
| *** lucasagomes has joined #openstack-ironic | 09:14 | |
| * iurygregory is back after rebooting, since my fedora decided to freeze | 09:14 | |
| *** dougsz has joined #openstack-ironic | 09:17 | |
| *** khansa has joined #openstack-ironic | 09:17 | |
| *** k_mouza has joined #openstack-ironic | 09:22 | |
| *** alexmcleod has joined #openstack-ironic | 09:24 | |
| *** dsneddon has joined #openstack-ironic | 09:25 | |
| *** yolanda has quit IRC | 09:27 | |
| *** derekh has joined #openstack-ironic | 09:28 | |
| *** mkrai_ has quit IRC | 09:28 | |
| *** dsneddon has quit IRC | 09:30 | |
| *** yolanda has joined #openstack-ironic | 09:33 | |
| *** mbeierl has joined #openstack-ironic | 09:42 | |
| openstackgerrit | Zygimantas Matonis proposed openstack/ironic-inspector master: Added scope to introspection rules. https://review.opendev.org/702215 | 09:48 |
| *** yolanda has quit IRC | 09:49 | |
| iurygregory | dtantsur, rpittau, I'm backporting the uefi patch to train, I'm wondering if we also want to backport patches that were add before like https://review.opendev.org/#/c/679333/ https://review.opendev.org/#/c/686382/ | 09:50 |
| patchbot | patch 679333 - ironic-python-agent - Don't try to put a bootloader in place when bootable (MERGED) - 15 patch sets | 09:50 |
| patchbot | patch 686382 - ironic-python-agent - Set rd.md.uuid kernel parameter when deploying on ... (MERGED) - 8 patch sets | 09:50 |
| *** dsneddon has joined #openstack-ironic | 10:00 | |
| dtantsur | rpittau: I'm afraid we'll have to drop xenial | 10:01 |
| dtantsur | iurygregory: if they qualify as bug fixes, we should | 10:01 |
| iurygregory | dtantsur, https://review.opendev.org/#/c/686382/ doesn't have a release note, thoughts? | 10:04 |
| patchbot | patch 686382 - ironic-python-agent - Set rd.md.uuid kernel parameter when deploying on ... (MERGED) - 8 patch sets | 10:04 |
| *** mkrai_ has joined #openstack-ironic | 10:08 | |
| *** dsneddon has quit IRC | 10:09 | |
| *** openstackgerrit has quit IRC | 10:12 | |
| etingof | TheJulia, re vbmc nuisance: I do not see if the root cause has been reported. the best thing to do here would be to update this bifrost job to get it running vbmcd from systemd -- https://github.com/openstack/ironic/blob/master/devstack/lib/ironic#L794 | 10:29 |
| etingof | so we'd avoid the legacy way of invoking vbmcd (via vbmc on server response timeout) and get all the logs in place | 10:29 |
| etingof | I can probably take care of that if needed | 10:30 |
| *** openstackgerrit has joined #openstack-ironic | 10:37 | |
| openstackgerrit | Merged openstack/bifrost stable/train: Use upper-constraints for ironicclient, sushy and DIB https://review.opendev.org/703086 | 10:37 |
| *** priteau has quit IRC | 10:41 | |
| *** dsneddon has joined #openstack-ironic | 10:42 | |
| dtantsur | iurygregory: looks an appropriate backport to me. arne_wiebalck? | 10:43 |
| *** goldyfruit has joined #openstack-ironic | 10:47 | |
| *** dsneddon has quit IRC | 10:47 | |
| *** mkrai_ has quit IRC | 10:50 | |
| *** mkrai__ has joined #openstack-ironic | 10:50 | |
| arne_wiebalck | dtantsur: iurygregory the md uuid patch? | 10:53 |
| dtantsur | arne_wiebalck: yep | 10:53 |
| arne_wiebalck | give me a second to refresh my memory ... | 10:54 |
| arne_wiebalck | dtantsur: wouldn't that introduce new functionality? | 10:55 |
| arne_wiebalck | dtantsur: it should not break anything, though | 10:55 |
| arne_wiebalck | dtantsur: *should* ;) | 10:55 |
| dtantsur | :) | 10:56 |
| dtantsur | given that it's a very fresh feature with a lot of rough edges, I'm inclined to let small improvements in | 10:56 |
| iurygregory | well I can backport my uefi fix without, I was just wondering if we want or not XD | 10:59 |
| * iurygregory will backport to train and stein | 11:00 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-python-agent stable/train: Set rd.md.uuid kernel parameter when deploying on software raid https://review.opendev.org/703584 | 11:00 |
| iurygregory | after this one I can backport Julia's patch and then mine | 11:01 |
| arne_wiebalck | dtantsur: I agree. | 11:01 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Start installing DIB from Zuul-cloned sources in the CI https://review.opendev.org/703585 | 11:05 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-python-agent stable/train: Don't try to put a bootloader in place when bootable https://review.opendev.org/703588 | 11:07 |
| *** goldyfruit has quit IRC | 11:10 | |
| *** rpittau is now known as rpittau|bbl | 11:15 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/bifrost master: Vagrant: run ansible-playbook from venv https://review.opendev.org/702885 | 11:18 |
| *** dsneddon has joined #openstack-ironic | 11:22 | |
| *** dsneddon has quit IRC | 11:27 | |
| openstackgerrit | Merged openstack/bifrost master: Python3/Ansible conditional quarks https://review.opendev.org/701590 | 11:28 |
| openstackgerrit | Merged openstack/bifrost master: Hopefully final venv fixes https://review.opendev.org/701846 | 11:28 |
| openstackgerrit | Merged openstack/bifrost master: Remove CI credential setup https://review.opendev.org/702107 | 11:29 |
| *** paddyoneill has joined #openstack-ironic | 11:32 | |
| dtantsur | TheJulia, rpittau|bbl, now that we have a patch fixing the bifrost CI, let's make at least the Bionic job voting before we do *anything* else? | 11:37 |
| *** Lucas_Gray has joined #openstack-ironic | 11:40 | |
| openstackgerrit | Merged openstack/bifrost master: Vagrant: run ansible-playbook from venv https://review.opendev.org/702885 | 11:44 |
| openstackgerrit | Merged openstack/bifrost master: Add a listen-address to the dnsmasq bind https://review.opendev.org/703314 | 11:45 |
| openstackgerrit | Merged openstack/bifrost master: Try shorthand format for console https://review.opendev.org/703522 | 11:45 |
| rpittau|bbl | dtantsur: yep, was thinking the same, I'm out for lunch but I'll check that as soon as I'm back | 11:57 |
| rpittau|bbl | dtantsur, tonyb[m] before I forget, I contacted Travis support for the hardware repo as the ci has an issue for sziviani PR | 11:59 |
| *** dsneddon has joined #openstack-ironic | 12:00 | |
| dtantsur | thx | 12:00 |
| *** dtantsur is now known as dtantsur|bbl | 12:01 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: DNM follow up to root device hints in instance_info https://review.opendev.org/703595 | 12:02 |
| *** dsneddon has quit IRC | 12:05 | |
| tonyb[m] | rpittau|bbl: Thanks | 12:06 |
| *** priteau has joined #openstack-ironic | 12:06 | |
| *** aedc has quit IRC | 12:09 | |
| *** aedc has joined #openstack-ironic | 12:09 | |
| *** yolanda has joined #openstack-ironic | 12:10 | |
| *** mkrai__ has quit IRC | 12:10 | |
| *** yolanda has quit IRC | 12:10 | |
| *** yolanda has joined #openstack-ironic | 12:11 | |
| *** ociuhandu has joined #openstack-ironic | 12:12 | |
| *** ociuhandu has quit IRC | 12:13 | |
| *** Lucas_Gray has quit IRC | 12:13 | |
| *** tkajinam has joined #openstack-ironic | 12:14 | |
| *** Lucas_Gray has joined #openstack-ironic | 12:17 | |
| *** aedc has quit IRC | 12:20 | |
| *** aedc has joined #openstack-ironic | 12:20 | |
| *** mkrai_ has joined #openstack-ironic | 12:25 | |
| *** rcernin has quit IRC | 12:28 | |
| *** dsneddon has joined #openstack-ironic | 12:31 | |
| *** dsneddon has quit IRC | 12:35 | |
| *** afasano has quit IRC | 12:47 | |
| *** afasano has joined #openstack-ironic | 12:48 | |
| *** bfournie has quit IRC | 12:51 | |
| *** ociuhandu has joined #openstack-ironic | 12:52 | |
| iurygregory | stable/train is broken for ipa http://paste.openstack.org/show/788638/ | 12:59 |
| *** ociuhandu has quit IRC | 13:01 | |
| iurygregory | according to infra it should be fixed already | 13:04 |
| * iurygregory trying recheck | 13:04 | |
| *** priteau has quit IRC | 13:05 | |
| *** rpittau|bbl is now known as rpittau | 13:05 | |
| *** mkrai_ has quit IRC | 13:06 | |
| *** dsneddon has joined #openstack-ironic | 13:07 | |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Make bionic job voting again https://review.opendev.org/703609 | 13:13 |
| iurygregory | Make bifrost great again ^ | 13:16 |
| rpittau | iurygregory: I didn't want to be *that* explicit, but yes :D | 13:16 |
| *** dsneddon has quit IRC | 13:17 | |
| iurygregory | rpittau, I was able to get the reference =P | 13:17 |
| *** whoami-rajat_ has joined #openstack-ironic | 13:28 | |
| *** jdandrea has joined #openstack-ironic | 13:32 | |
| *** iurygregory has quit IRC | 13:35 | |
| *** Lucas_Gray has quit IRC | 13:42 | |
| *** ociuhandu has joined #openstack-ironic | 13:47 | |
| *** dsneddon has joined #openstack-ironic | 13:48 | |
| *** gkaratae has joined #openstack-ironic | 13:50 | |
| *** dsneddon has quit IRC | 13:53 | |
| *** ociuhandu has quit IRC | 13:53 | |
| *** khansa has quit IRC | 13:54 | |
| *** bfournie has joined #openstack-ironic | 13:55 | |
| gkaratae | Hello guys! Have someone tried to deploy Ironic standalone within Bifrost on systems such as RHEL7 or Fedora 30 recently? In both of the cases I encounter a pretty bunch of problems such as python 2/3 incompatibility or packages non-existence (e.g. qemu-kvm-tools on Fedora 30). For now, I more likely to think that the problems are not on my side and I'd like to hear others' opinion :) | 13:57 |
| *** dtantsur|bbl is now known as dtantsur | 14:01 | |
| *** rloo has joined #openstack-ironic | 14:04 | |
| *** iurygregory has joined #openstack-ironic | 14:05 | |
| rpittau | gkaratae: hi! We dropped support for Python 2 during this cycle, are you installing from source ? | 14:05 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Fix changes leftovers https://review.opendev.org/703620 | 14:06 |
| dtantsur | gkaratae: tl;dr it's broken for the very reasons you mention. for Train you need to pull https://review.opendev.org/#/c/703086/ | 14:07 |
| patchbot | patch 703086 - bifrost (stable/train) - Use upper-constraints for ironicclient, sushy and DIB (MERGED) - 5 patch sets | 14:07 |
| dtantsur | qemu-kvm-tools may be a bug we're not aware of, please see if you still have it once you figure out the python 2 thing | 14:08 |
| dtantsur | CentOS 7 is a system that we do test againt (up to and including Train), not sure about Fedora | 14:08 |
| openstackgerrit | Doug Hellmann proposed openstack/ironic-inspector master: link introspection API docs to client docs https://review.opendev.org/703622 | 14:12 |
| *** mkrai_ has joined #openstack-ironic | 14:13 | |
| rpittau | gkaratae: I'm pretty sure there's no qemu-kvm-tools package for Fedora since at least Fedora 25 | 14:19 |
| rpittau | maybe it's time to review the Fedora support :) | 14:19 |
| *** gkaratae has quit IRC | 14:20 | |
| etingof | is anyone fluent in Glean lurking here? - https://review.opendev.org/703623 | 14:21 |
| patchbot | patch 703623 - opendev/glean - Fix a handful of bugs in config-drive processing - 1 patch set | 14:21 |
| *** tzumainn has joined #openstack-ironic | 14:24 | |
| *** dsneddon has joined #openstack-ironic | 14:26 | |
| *** dsneddon has quit IRC | 14:31 | |
| *** frickler has joined #openstack-ironic | 14:32 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Document introspection data and collectors https://review.opendev.org/703628 | 14:40 |
| dtantsur | etingof: it may be fungi or clarkb or other infra folks | 14:46 |
| *** goldyfruit has joined #openstack-ironic | 14:46 | |
| fungi | etingof: what's the glean question? | 14:50 |
| fungi | i may or not be able to answer, but can certainly ask around | 14:50 |
| fungi | oh, i see, the patch you just pushed earlier today | 14:50 |
| fungi | ~30 minutes ago i guess | 14:51 |
| etingof | fungi, yeah, the patch is the question | 14:52 |
| fungi | etingof: lgtm | 14:53 |
| fungi | i expect we just don't have enough test coverage for vlan interfaces | 14:53 |
| fungi | thanks for the fixes! | 14:53 |
| etingof | fungi, 10x! however, that's just a heat-up! the real meat of the matter is this short an sweet schema -- https://review.opendev.org/#/c/703133/ | 14:53 |
| patchbot | patch 703133 - nova - Add JSON schema for network_data.json - 3 patch sets | 14:53 |
| fungi | oh neat | 14:54 |
| etingof | fungi, would it make sense to use it somehow in Glean? | 14:54 |
| fungi | i think so, though of course it would depend on the deployed environment including that feature | 14:55 |
| etingof | hmmm? | 14:55 |
| fungi | oh wait | 14:55 |
| etingof | so one way of using this schema would be to validate passed network_data.json right before parsing it inside Glean... | 14:55 |
| fungi | you mean embedding that schema in glean, not having nova serve a copy alongside the metadata in teh configdrive | 14:55 |
| etingof | to fail reliably | 14:55 |
| etingof | I am not sure we should rely on Nova here | 14:56 |
| fungi | s/here// ;) | 14:56 |
| etingof | fungi++ | 14:56 |
| fungi | (i kid, nova is remarkably reliable) | 14:56 |
| etingof | indeed | 14:56 |
| dtantsur | rpioso: hey, are you aware of this? https://bugzilla.redhat.com/show_bug.cgi?id=1793206 | 14:57 |
| openstack | bugzilla.redhat.com bug 1793206 in openstack-ironic "On Dell 14th Gen PowerEdge, iDRAC firmware starting with 3.34.34.34 disables RAW commands." [Unspecified,Assigned] - Assigned to dtantsur | 14:57 |
| etingof | so either we can pass network_data through the schema in Glean to fail early and soundly | 14:57 |
| *** k_mouza has quit IRC | 14:57 | |
| fungi | etingof: but yeah, seems like a reasonable addition to me. obviously i'm just one maintainer. it may make sense to discuss in #openstack-infra if you're looking for additional buy-in before submitting a change to add it... or you could just push up the change and get input through gerrit. up to you | 14:57 |
| etingof | and/or we can validate network_data in ironic before even passing to Glean/ramdisk | 14:57 |
| etingof | fungi, I am half-way through the change | 14:58 |
| fungi | wfm ;) | 14:58 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Set Ansible default version to 2.8 https://review.opendev.org/703633 | 14:58 |
| *** k_mouza has joined #openstack-ironic | 15:04 | |
| *** dsneddon has joined #openstack-ironic | 15:04 | |
| *** dsneddon has quit IRC | 15:10 | |
| *** AJaeger has joined #openstack-ironic | 15:10 | |
| *** tkajinam has quit IRC | 15:16 | |
| openstackgerrit | khansa proposed openstack/sushy-cli master: Add power command to sushy cli https://review.opendev.org/702546 | 15:22 |
| *** khansa has joined #openstack-ironic | 15:22 | |
| *** ociuhandu has joined #openstack-ironic | 15:30 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Follow up to root device hints in instance_info https://review.opendev.org/703595 | 15:30 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Expose collector and hardware manager names via introspection data https://review.opendev.org/703641 | 15:31 |
| *** dsneddon has joined #openstack-ironic | 15:37 | |
| *** hjensas has quit IRC | 15:39 | |
| *** Lucas_Gray has joined #openstack-ironic | 15:42 | |
| *** dsneddon has quit IRC | 15:42 | |
| *** ociuhandu has quit IRC | 15:51 | |
| clarkb | etingof: fungi shouldn't those schemas be defined in nova instead? | 15:52 |
| clarkb | documenting this in glean is just a good way for gleans documentation to be out of date when nova makes changes | 15:52 |
| fungi | clarkb: they *should* but sort of like service types integration in openstacksdk there's a massive lag waiting for things like that to wind up in deployed environments | 15:53 |
| clarkb | fungi: but we are talking about docs right? there shouldn't be lag for that | 15:53 |
| dtantsur | clarkb: nova is not the only thing producing these, nor is it the thing consuming them | 15:53 |
| clarkb | dtantsur: no but it defines it | 15:54 |
| dtantsur | it could just as well be something like API SIG or os_network_data project | 15:54 |
| fungi | clarkb: oh, i see, you mean in documentation not actually programmatically applied by nova to vet its own metadata | 15:54 |
| clarkb | everything else has copied nova essentially | 15:54 |
| clarkb | fungi: that is all this change to glean is doing | 15:54 |
| clarkb | if openstack wants to define the schema globally outside of nova thats fine too. But I don't think glean is the appropriate location | 15:54 |
| etingof | perhaps one way would be to wait for Nova to merge and publish the schema, then just link Glean to it... | 15:55 |
| clarkb | etingof: yes linking to the schema in glean would be fine | 15:55 |
| fungi | but i agree glean is not the place to establish and publish a metadata standard. having glean warn when metadata doesn't validate is more what i was expecting | 15:55 |
| dtantsur | clarkb: ideally, glean should have a schema that it accepts (which may not equal the nova's schema) | 15:55 |
| etingof | the other thought is to have Glean actually using the schema to validate the input | 15:55 |
| etingof | to harden it up | 15:56 |
| clarkb | I am not sure it can | 15:56 |
| clarkb | since it accepts rax config drive data | 15:56 |
| etingof | but that would pull the whole lot of obscure dependencies | 15:56 |
| clarkb | it could warn if there are inconsistencies I suppose | 15:56 |
| etingof | yeah, or crash ;) | 15:56 |
| clarkb | no crashing would be bad | 15:56 |
| clarkb | since it runs on clouds that may change the schema | 15:56 |
| clarkb | and glean doesn't actually care about 90% of the input data | 15:57 |
| clarkb | so if that other 90% is "wrong" meh | 15:57 |
| etingof | hmmm, I think Glean consumes 90% of network_data.json | 15:57 |
| fungi | the dependencies bit is something i hadn't thought about though. if json schema validation isn't easy to do with just python stdlib then it may not be the greatest idea. glean exists mostly to avoid the dependency hell cloud-init brings | 15:57 |
| *** mbeierl has quit IRC | 15:57 | |
| clarkb | etingof: in the best case it may, but often all you need is glean to run enough to configure dhcp and it ignores the network data | 15:57 |
| *** jtomasek has quit IRC | 15:58 | |
| clarkb | etingof: that then allows you to debug why the rest of glean failed | 15:58 |
| etingof | fungi, exactly! no, this battery is not included | 15:58 |
| clarkb | crashing instead would be terrible for users | 15:58 |
| *** TxGirlGeek has joined #openstack-ironic | 15:58 | |
| clarkb | glean should always attempt to do as much work as possible to make debugging possible | 15:59 |
| etingof | so the use case we are mostly interested in is network_data.json validation whenever it comes from a messy human | 16:00 |
| *** mbeierl has joined #openstack-ironic | 16:00 | |
| dtantsur | .. on the ironic side, largely | 16:00 |
| clarkb | why isn't ironic providing it? | 16:01 |
| etingof | ironic is not always aware of networking bits | 16:01 |
| dtantsur | ironic must and will provide it, yes | 16:01 |
| *** iurygregory has quit IRC | 16:01 | |
| dtantsur | I guess, etingof is researching the additional hardening | 16:01 |
| etingof | well, not quite... there are two school of thoughts on this | 16:02 |
| etingof | one is to have ironic enforcing well-formed network_data whenever it comes from the operator | 16:02 |
| dtantsur | and one if them is about to use his core rights to enforce some sanity in this discussion..... | 16:02 |
| etingof | the other is to oblige the operator to provide a well-formed network_data | 16:03 |
| etingof | either way, ensuring well-formed network_data involves schema-based validation | 16:03 |
| clarkb | fwiw I'm not against validation in glean but I agree we should avoid adding deps and the default behavior should be to warn not crash (maybe allow user to opt into crashing instead) | 16:03 |
| dtantsur | etingof: ignore the 2nd option. it is not going to happen. | 16:03 |
| etingof | so the school of thoughts sometimes clash with each other, but that's good | 16:04 |
| clarkb | (and people should set up working dhcp regardless of other network config as a recovery fallback) | 16:04 |
| clarkb | even if its on an unrouted network that you can plug a laptop into | 16:04 |
| dtantsur | clarkb: working DHCP is what we're not going to have | 16:04 |
| clarkb | dtantsur: I understand that isn't your production config but it should be there in the background as a fallback | 16:05 |
| dtantsur | the task at hand is to provide networking information where DHCP is not desired or possible (think, edge) | 16:05 |
| etingof | schema validation seems impossible without external deps | 16:05 |
| clarkb | etingof: not impossible, but more painful | 16:05 |
| dtantsur | clarkb: it may be there or may be not. if it is, it's probably going to be wrong enough to fail everything anyway. this complicates our case and makes me so strict about ironic-side validation. | 16:06 |
| clarkb | dtantsur: the idea is you'd only use it if configuration fails anyway | 16:06 |
| clarkb | dtantsur: then you plug your laptop into the l2 network, console in and figure out why it broke | 16:06 |
| etingof | clarkb, there is the official validation utility (jsonschema) supposedly coming from the same people that maintain the json schema spec | 16:06 |
| dtantsur | Edge deployment, the L2 network is in a mobile station in the woods | 16:06 |
| etingof | clarkb, I am not sure we'd be able to implement our own to avoid extra deps | 16:06 |
| dtantsur | (I'm not exactly arguing, but rather explaining the use case we have) | 16:07 |
| clarkb | dtantsur: and eventually if you can't boot a working network someone is driving out there to find out why | 16:07 |
| dtantsur | yep | 16:07 |
| dtantsur | My point here is that early validation is quite critical since the failure model is quite bad. | 16:08 |
| etingof | now days remote BMC access should solve booting problems... | 16:08 |
| dtantsur | I don't think validation in glean buys us anything, although I can see how it can be overall useful | 16:08 |
| clarkb | dtantsur: ya I think validation in glean would largely be an additional debug output | 16:08 |
| dtantsur | etingof: we only support it for IPMI, and IPMI serial console over WAN is sadness | 16:08 |
| clarkb | glean can assert it got bad input if somehow it got that far | 16:08 |
| dtantsur | but we're talking about redfish anyway, sooo.. | 16:09 |
| dtantsur | the jsonschema python package has a few requirements itself. not sure if you want it by default.. | 16:10 |
| etingof | dtantsur, IPMI over WAN is sadness | 16:10 |
| etingof | dtantsur, try to install it to resolve all the deps | 16:10 |
| * dtantsur does | 16:11 | |
| dtantsur | attrs-19.3.0 importlib-metadata-1.4.0 jsonschema-3.2.0 more-itertools-8.1.0 pyrsistent-0.15.7 six-1.14.0 zipp-2.0.0 | 16:11 |
| dtantsur | (and setuptools if for some reason it's not already there) | 16:12 |
| * etingof has never encountered these except six | 16:12 | |
| dtantsur | me neither | 16:12 |
| etingof | so I suppose depending on so many things goes against Glean design goals | 16:13 |
| *** openstackgerrit has quit IRC | 16:13 | |
| *** mbeierl has quit IRC | 16:13 | |
| etingof | but dtantsur will get used to these packages once we have schema validation built into ironic ;) | 16:14 |
| dtantsur | etingof: we already have (for RAID) | 16:14 |
| dtantsur | and will have more for the transition from WSME, I guess | 16:15 |
| *** mbeierl has joined #openstack-ironic | 16:16 | |
| etingof | do you mean we have them all in ironic indirectly already? | 16:16 |
| dtantsur | ironic depends on jsonschema quite directly | 16:16 |
| dtantsur | we're using it for RAID config format validation | 16:16 |
| etingof | ah, I see | 16:17 |
| *** dsneddon has joined #openstack-ironic | 16:17 | |
| *** k_mouza has quit IRC | 16:20 | |
| *** mbeierl has quit IRC | 16:20 | |
| *** dsneddon has quit IRC | 16:22 | |
| *** ociuhandu has joined #openstack-ironic | 16:23 | |
| *** k_mouza has joined #openstack-ironic | 16:23 | |
| *** jtomasek has joined #openstack-ironic | 16:25 | |
| *** k_mouza has quit IRC | 16:25 | |
| *** ociuhandu has quit IRC | 16:28 | |
| *** k_mouza has joined #openstack-ironic | 16:28 | |
| *** openstackgerrit has joined #openstack-ironic | 16:31 | |
| openstackgerrit | Ilya Etingof proposed openstack/sushy master: Fix 'None' field value processing https://review.opendev.org/703652 | 16:31 |
| etingof | yolanda, this fix ^ should get you around boot-override-target=None problem | 16:33 |
| etingof | but might break some unfortunate OEM as well | 16:33 |
| *** ociuhandu has joined #openstack-ironic | 16:36 | |
| *** gyee has joined #openstack-ironic | 16:39 | |
| *** iurygregory has joined #openstack-ironic | 16:40 | |
| *** bnemec has quit IRC | 16:48 | |
| *** goldyfruit has quit IRC | 16:51 | |
| *** dsneddon has joined #openstack-ironic | 16:53 | |
| *** dsneddon has quit IRC | 16:58 | |
| *** mkrai_ has quit IRC | 17:00 | |
| *** hjensas has joined #openstack-ironic | 17:00 | |
| *** tesseract has quit IRC | 17:01 | |
| *** TxGirlGeek has quit IRC | 17:02 | |
| rpioso | Good morning, ironicers :-) | 17:03 |
| rpittau | good morning rpioso :) | 17:04 |
| rpittau | and good night as well! o/ | 17:04 |
| *** rpittau is now known as rpittau|afk | 17:04 | |
| *** ricolin has quit IRC | 17:04 | |
| *** lucasagomes has quit IRC | 17:05 | |
| *** ricolin has joined #openstack-ironic | 17:05 | |
| rpioso | rpittau|afk: Good day :-) | 17:05 |
| etingof | Earth is a sphere indeed | 17:06 |
| sziviani | Good night :) | 17:06 |
| rpioso | etingof, TheJulia: I'm continuing to test vmedia against a PowerEdge 14G server. I've validated the test environs network config and completed some baseline testing against a dib CentOS ramdisk. | 17:07 |
| etingof | rpioso++ | 17:08 |
| rpioso | etingof, TheJulia: Deployment via iPXE boot with dib CentOS ramdisk succeeds, which is a good start. | 17:08 |
| rpioso | etingof, TheJulia: Deployment via Redfish vmedia boot with dib CentOS ramdsik fails. I believe that ramdisk is too big. Trying to boot the ironic hatched ISO from the iDRAC web UI virtual system console symptomatically fails in the same manner. | 17:09 |
| etingof | that can have something to do with EFI boot loader - they way how it does memory management... | 17:10 |
| *** mkrai_ has joined #openstack-ironic | 17:11 | |
| etingof | so may be trying legacy bios ISO would be another experiment to run | 17:11 |
| etingof | i.e. isolinux-based | 17:12 |
| rpioso | etingof, TheJulia: I'm moving on to try the tinyipa Tiny Core ramdisk. Based on my earlier testing, I believe both deployment approaches will fail, perhaps because it doesn't have the needed NIC drivers. | 17:12 |
| etingof | you should be able to assess drivers situation from console... | 17:13 |
| rpioso | etingof: I thought of that, but would like to get UEFI to succeed first. Does devstack default to isolinux-based when the boot mode is configured to legacy BIOS in local.conf? | 17:13 |
| rpioso | etingof: How so? | 17:14 |
| rpioso | etingof: Also, what NIC h/w did you have success with? | 17:14 |
| *** jtomasek has quit IRC | 17:14 | |
| etingof | iirc isolinux is installed unconditionally. so if you switch ironic node boot mode to bios, ironic will bake isolinux iso | 17:15 |
| etingof | regardless of local.conf | 17:16 |
| openstackgerrit | Merged openstack/ironic stable/train: Fix entry paths for cleaning and deployment https://review.opendev.org/703337 | 17:16 |
| *** mkrai_ has quit IRC | 17:16 | |
| etingof | once tinycore is booted, can you use web-ui to get on the console? | 17:16 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-python-agent stable/train: Search for efi partition https://review.opendev.org/703380 | 17:17 |
| etingof | I do not remember NIC hardware already - EBRAINOVERFLOW | 17:17 |
| rpioso | etingof: Yes, I can. When I last did that, none of the network interface, except the loopback interface, had an IPv4 address. | 17:17 |
| rpioso | etingof: Can you access the server's iDRAC. It'll gladly refresh your memory. | 17:18 |
| etingof | that is arguably more likely to be caused by missing dhcp responder than missing linux drivers | 17:18 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Set Ansible default version to 2.8 https://review.opendev.org/703633 | 17:18 |
| rpioso | etingof: The iDRAC is very powerful ;-) | 17:18 |
| rpioso | etingof: But, but, the DHCP responder works with dib CentOS. | 17:19 |
| rpioso | etingof: Why would that be different? | 17:19 |
| etingof | yeah, puzzling | 17:19 |
| etingof | it's Intel(R) 10GbE 4P X710 rNDC | 17:20 |
| *** khansa has quit IRC | 17:21 | |
| * rpioso checks his | 17:21 | |
| etingof | if you have unnumbered NICs on the system, chances are the drivers are in place | 17:21 |
| etingof | so you can re-run dhcpmasq may be with some debugging to see how it ends up | 17:22 |
| etingof | you can also re-run startup scripts by hand to repeat the exact boot sequence | 17:22 |
| rpioso | etingof: What do you mean by "unnumbered NICs on the system"? | 17:23 |
| * rpioso is unfamiliar with the startup scripts | 17:23 | |
| etingof | ip link reports something besides lo, but (as you say) those interfaces are not assigned L3 addresses | 17:23 |
| etingof | iirc tinycore runs a predefined script somewhere in /opt | 17:25 |
| etingof | boot.sh or something like that | 17:25 |
| etingof | may be it helps to run this script from the console under bash -x to see how it goes | 17:26 |
| *** jtomasek has joined #openstack-ironic | 17:26 | |
| * rpioso compares his NIC hw to etingof's | 17:26 | |
| *** TxGirlGeek has joined #openstack-ironic | 17:33 | |
| *** dsneddon has joined #openstack-ironic | 17:34 | |
| *** rloo has quit IRC | 17:34 | |
| *** rloo has joined #openstack-ironic | 17:34 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-python-agent stable/train: Avoid grub2-install when on UEFI boot mode https://review.opendev.org/703381 | 17:36 |
| *** rloo has quit IRC | 17:36 | |
| *** rloo has joined #openstack-ironic | 17:37 | |
| rpioso | etingof: My R740xd has an Intel(R) 2P X710/2P I350 rNDC. A 1 Gbps I350 port is connected to the provisioning network, which has been used for PXE booting and delivering the ISO overt https. A collleague has informed me that port's driver is different from the 10 Gbps X710's. | 17:38 |
| *** dsneddon has quit IRC | 17:39 | |
| *** rloo has quit IRC | 17:40 | |
| rpioso | etingof: I'll run the Tiny Core tests against that existing network config to confirm the test results. If they fail as I believe they did before, we'll change the port to more closely match your config. | 17:40 |
| *** rloo has joined #openstack-ironic | 17:40 | |
| etingof | the other option is to try coreos with legacy bios... | 17:41 |
| rpioso | etingof: I'm thinking if that should fail, too, then off to legacy BIOS :-( , re-running the startup scripts, etc. | 17:42 |
| rpioso | etingof: Wdyt of that plan? | 17:43 |
| rpioso | etingof: What config is needed to get coreos used with legacy boot? | 17:44 |
| etingof | the plan looks good to me | 17:44 |
| rpioso | \o/ | 17:44 |
| etingof | I think flipping ironic node boot mode to bios should suffice, hopefully | 17:45 |
| *** rloo has quit IRC | 17:45 | |
| dtantsur | s/coreos/centos/ ? | 17:45 |
| *** rloo has joined #openstack-ironic | 17:45 | |
| rpioso | etingof: Are coreos and isolinux synonyms? | 17:46 |
| etingof | nooo | 17:46 |
| etingof | sorry, centos indeed | 17:46 |
| etingof | isolinix is a boot loader | 17:46 |
| rpioso | Whew! Thx dtantsur and etingof! | 17:46 |
| etingof | too much of containers | 17:47 |
| rpioso | etingof: Enough tech speak ;-) | 17:48 |
| * rpioso goes to eat yesterday's lunch. Finally! | 17:50 | |
| *** iurygregory has quit IRC | 17:51 | |
| rpioso | etingof: Thanks, again. | 17:51 |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: DIB: replace targetcli with target-restore on RHEL/CentOS 8 https://review.opendev.org/701719 | 17:51 |
| *** yolanda has quit IRC | 17:52 | |
| *** k_mouza has quit IRC | 17:52 | |
| *** yolanda has joined #openstack-ironic | 17:54 | |
| openstackgerrit | Doug Hellmann proposed openstack/ironic-inspector master: link introspection API docs to client docs https://review.opendev.org/703622 | 17:56 |
| *** alexmcleod has quit IRC | 17:56 | |
| *** dougsz has quit IRC | 17:57 | |
| *** TxGirlGeek has quit IRC | 18:01 | |
| *** dsneddon has joined #openstack-ironic | 18:08 | |
| *** dtantsur is now known as dtantsur|afk | 18:11 | |
| dtantsur|afk | g'night | 18:11 |
| *** jtomasek has quit IRC | 18:13 | |
| *** ociuhandu_ has joined #openstack-ironic | 18:13 | |
| *** dsneddon has quit IRC | 18:14 | |
| *** Lucas_Gray has quit IRC | 18:15 | |
| dtantsur|afk | rpittau|afk: when you're back: we need to replace the bifrost job on ironic: s/xenial/bionic/ (ditto on inspector) | 18:15 |
| *** ociuhandu has quit IRC | 18:17 | |
| *** TxGirlGeek has joined #openstack-ironic | 18:17 | |
| *** ociuhandu_ has quit IRC | 18:18 | |
| *** k_mouza has joined #openstack-ironic | 18:21 | |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Use bionic job for bifrost integration https://review.opendev.org/703665 | 18:45 |
| *** dsneddon has joined #openstack-ironic | 18:45 | |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic-inspector master: Use bionic job for bifrost integration https://review.opendev.org/703668 | 18:46 |
| *** dsneddon has quit IRC | 18:50 | |
| *** afasano has quit IRC | 18:53 | |
| *** TxGirlGeek has quit IRC | 18:56 | |
| *** TxGirlGeek has joined #openstack-ironic | 19:06 | |
| *** mbeierl has joined #openstack-ironic | 19:08 | |
| *** TxGirlGeek has quit IRC | 19:13 | |
| *** dustinc|PTO is now known as dustinc | 19:13 | |
| *** TxGirlGeek has joined #openstack-ironic | 19:14 | |
| *** dsneddon has joined #openstack-ironic | 19:15 | |
| *** gkaratae has joined #openstack-ironic | 19:17 | |
| *** TxGirlGeek has quit IRC | 19:22 | |
| *** paddyoneill is now known as paddyoneill|afk | 19:26 | |
| *** TxGirlGeek has joined #openstack-ironic | 19:26 | |
| *** afasano has joined #openstack-ironic | 19:29 | |
| *** mbeierl has quit IRC | 19:30 | |
| *** TxGirlGeek has quit IRC | 19:31 | |
| *** TxGirlGe_ has joined #openstack-ironic | 19:31 | |
| *** iurygregory has joined #openstack-ironic | 19:33 | |
| *** k_mouza has quit IRC | 20:09 | |
| *** jtomasek has joined #openstack-ironic | 20:10 | |
| *** Lucas_Gray has joined #openstack-ironic | 20:26 | |
| *** yolanda has quit IRC | 20:29 | |
| *** yolanda has joined #openstack-ironic | 20:30 | |
| *** mbeierl has joined #openstack-ironic | 20:32 | |
| *** rcernin has joined #openstack-ironic | 20:42 | |
| *** yolanda has quit IRC | 20:46 | |
| *** gkaratae has quit IRC | 20:48 | |
| *** yolanda has joined #openstack-ironic | 20:50 | |
| openstackgerrit | Steve Baker proposed openstack/ironic master: DNM Demonstrate mixing pecan and flask https://review.opendev.org/703695 | 20:54 |
| *** whoami-rajat_ has quit IRC | 20:58 | |
| *** mbeierl has quit IRC | 21:14 | |
| *** mbeierl has joined #openstack-ironic | 21:18 | |
| *** mbeierl has quit IRC | 21:23 | |
| *** mbeierl has joined #openstack-ironic | 21:25 | |
| *** seba has quit IRC | 21:29 | |
| *** mbeierl has quit IRC | 21:29 | |
| *** seba has joined #openstack-ironic | 21:37 | |
| *** jtomasek has quit IRC | 21:38 | |
| *** mbeierl has joined #openstack-ironic | 21:38 | |
| *** mbeierl has quit IRC | 21:39 | |
| *** mbeierl has joined #openstack-ironic | 21:39 | |
| *** TxGirlGe_ has quit IRC | 21:54 | |
| *** TxGirlGeek has joined #openstack-ironic | 21:56 | |
| *** mbeierl has quit IRC | 22:06 | |
| *** mbeierl has joined #openstack-ironic | 22:11 | |
| *** mbeierl has quit IRC | 22:18 | |
| *** bfournie has quit IRC | 22:20 | |
| *** afasano has quit IRC | 22:26 | |
| *** hwoarang has quit IRC | 22:28 | |
| *** hwoarang has joined #openstack-ironic | 22:30 | |
| *** k_mouza has joined #openstack-ironic | 22:31 | |
| *** mbeierl has joined #openstack-ironic | 22:32 | |
| *** mbeierl has quit IRC | 22:37 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Keystone: Add make /etc/keystone/fernet-keys https://review.opendev.org/703529 | 22:51 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Sync pip_install files https://review.opendev.org/703502 | 22:51 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: DNM Enable remaining ubuntu jobs https://review.opendev.org/701851 | 22:52 |
| *** mbeierl has joined #openstack-ironic | 22:52 | |
| *** mbeierl has quit IRC | 22:57 | |
| *** k_mouza has quit IRC | 22:58 | |
| *** k_mouza has joined #openstack-ironic | 22:58 | |
| *** k_mouza has quit IRC | 23:03 | |
| *** Lucas_Gray has quit IRC | 23:08 | |
| openstackgerrit | Merged openstack/bifrost master: Make bionic job voting again https://review.opendev.org/703609 | 23:09 |
| *** Lucas_Gray has joined #openstack-ironic | 23:09 | |
| *** bfournie has joined #openstack-ironic | 23:20 | |
| *** jdandrea has quit IRC | 23:25 | |
| *** dsneddon has quit IRC | 23:39 | |
| *** k_mouza has joined #openstack-ironic | 23:40 | |
| *** k_mouza has joined #openstack-ironic | 23:41 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!