| janders | good morning Ironic o/ | 00:08 |
|---|---|---|
| janders | uzumaki not sure, sorry... can you point me to the doco you're talking about? I can have a look around for you too | 00:09 |
| uzumaki | janders, morning! o/ | 00:09 |
| uzumaki | lemmi see if I can find that out | 00:09 |
| uzumaki | https://docs.openstack.org/ironic/latest/admin/raid.html#software-raid | 00:09 |
| uzumaki | see bullet 2 | 00:09 |
| * janders is looking at code | 00:10 | |
| janders | uzumaki what hardware/driver are you using? | 00:11 |
| uzumaki | I'm currently thinking of software raid | 00:11 |
| uzumaki | but for the hardware counterpart, it's idrac | 00:11 |
| janders | ACK (and good point BTW, I should have remembered from the link :) | 00:12 |
| uzumaki | to err is human janders | 00:13 |
| *** JamesBenson has joined #openstack-ironic | 00:13 | |
| janders | I don't see an explicit check for this in the code. Unless anyone else has a better idea my recommendation would be test it and see how you go. I wouldn't be surprised if it works, just considered not-so-good practice | 00:15 |
| uzumaki | Hmmm that's what I thought as well.. | 00:16 |
| *** JamesBenson has quit IRC | 00:18 | |
| *** chihhsin has joined #openstack-ironic | 00:37 | |
| *** rcernin_ has joined #openstack-ironic | 00:45 | |
| *** rcernin has quit IRC | 00:47 | |
| *** JamesBenson has joined #openstack-ironic | 00:56 | |
| *** JamesBenson has quit IRC | 01:00 | |
| *** uzumaki has quit IRC | 01:04 | |
| *** ijw__ has quit IRC | 01:20 | |
| *** ijw has joined #openstack-ironic | 01:20 | |
| openstackgerrit | Christopher Dearborn proposed openstack/sushy master: WIP: Allow monitoring progress of a firmware update https://review.opendev.org/745950 | 01:22 |
| *** Qianbiao has joined #openstack-ironic | 01:29 | |
| *** JamesBenson has joined #openstack-ironic | 01:35 | |
| *** JamesBenson has quit IRC | 01:40 | |
| *** JamesBenson has joined #openstack-ironic | 02:46 | |
| *** rcernin_ has quit IRC | 03:19 | |
| *** stevebaker has quit IRC | 03:21 | |
| *** rcernin_ has joined #openstack-ironic | 03:34 | |
| openstackgerrit | Li Yingjun proposed openstack/bifrost master: Variable testing should be converted to bool in if case https://review.opendev.org/749421 | 03:55 |
| *** zaneb has quit IRC | 04:12 | |
| *** zaneb has joined #openstack-ironic | 04:12 | |
| *** cdearborn has quit IRC | 04:25 | |
| *** Qianbiao has quit IRC | 04:40 | |
| *** rh-jelabarre has quit IRC | 04:41 | |
| janders | bfournie dtantsur|afk here are the firewall-cmd commands from my Aussie colleague that I mentioned earlier - these made firewall configuration more appropriate for real baremetal work: http://paste.openstack.org/show/797358/ | 04:47 |
| *** gyee has quit IRC | 05:07 | |
| *** tzumainn has quit IRC | 05:31 | |
| *** hjensas has joined #openstack-ironic | 05:39 | |
| openstackgerrit | paresh sao proposed openstack/ironic master: [WIP] OOB one button secure erase for iLO5 based HPE Proliant servers. https://review.opendev.org/743868 | 05:45 |
| arne_wiebalck | Good morning, ironic! | 05:51 |
| janders | good morning arne_wiebalck | 05:54 |
| arne_wiebalck | hey janders :) | 05:54 |
| *** Qianbiao has joined #openstack-ironic | 05:58 | |
| *** arne_wiebalck has quit IRC | 06:01 | |
| *** arne_wiebalck has joined #openstack-ironic | 06:02 | |
| *** rajinir__ has joined #openstack-ironic | 06:02 | |
| *** rajinir has quit IRC | 06:03 | |
| *** rajinir__ is now known as rajinir | 06:03 | |
| *** JamesBenson has quit IRC | 06:39 | |
| *** rcernin_ has quit IRC | 06:51 | |
| *** rcernin_ has joined #openstack-ironic | 06:54 | |
| iurygregory | good morning arne_wiebalck janders and Ironic! | 06:57 |
| arne_wiebalck | hey iurygregory o/ | 06:57 |
| *** jtomasek has joined #openstack-ironic | 07:00 | |
| openstackgerrit | Ankit Kumar proposed openstack/ironic master: Enhance certificate verification for ilo harware type https://review.opendev.org/743490 | 07:16 |
| janders | hey iurygregory | 07:20 |
| openstackgerrit | paresh sao proposed openstack/ironic master: OOB one button secure erase for iLO5 based HPE Proliant servers. https://review.opendev.org/743868 | 07:24 |
| openstackgerrit | paresh sao proposed openstack/ironic master: OOB one button secure erase for iLO5 based HPE Proliant servers. https://review.opendev.org/743868 | 07:26 |
| *** sshnaidm|afk is now known as sshnaidm | 07:32 | |
| *** rcernin_ has quit IRC | 07:35 | |
| *** tosky has joined #openstack-ironic | 07:38 | |
| rpittau | good morning ironic! o/ | 07:38 |
| iurygregory | morning rpittau o/ | 07:39 |
| rpittau | hey iurygregory :) | 07:39 |
| rpittau | mmmm multinode is still broken ? | 07:41 |
| iurygregory | yes =( | 07:41 |
| iurygregory | I'm looking at it but no luck no sar | 07:42 |
| iurygregory | so far* | 07:42 |
| iurygregory | lol | 07:42 |
| rpittau | and metalsmith too | 07:42 |
| * iurygregory needs to stop trying to talk in slack and irc at same time with different people | 07:42 | |
| iurygregory | yeah | 07:42 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic-python-agent master: Move ironic-standalone-ipa-src to Ubuntu Focal https://review.opendev.org/748213 | 07:46 |
| *** uzumaki has joined #openstack-ironic | 07:49 | |
| *** fmuyassarov has joined #openstack-ironic | 07:53 | |
| *** fmuyassarov has quit IRC | 07:54 | |
| Qianbiao | Hello Ironic. | 07:54 |
| Qianbiao | good morning, iurygregory rpittau and arne_wiebalck> | 07:55 |
| rpittau | hey Qianbiao :) | 07:55 |
| arne_wiebalck | hey Qianbiao o/ | 07:55 |
| Qianbiao | hey 0/ | 07:55 |
| rpittau | there's a new version of setuptools available, I'll try with a recheck | 07:56 |
| iurygregory | morning Qianbiao | 07:56 |
| Qianbiao | morning :) | 07:57 |
| Qianbiao | i need some suggestion for ironic devstack env settings. | 07:59 |
| Qianbiao | currently, i setup env with doc https://docs.openstack.org/ironic/latest/contributor/dev-quickstart.html. | 07:59 |
| Qianbiao | the virt nodes deployed. | 07:59 |
| Qianbiao | And now i want to deploy real baremetals. | 07:59 |
| openstackgerrit | Riccardo Pittau proposed openstack/metalsmith master: [DNM] test ci https://review.opendev.org/749452 | 07:59 |
| Qianbiao | what doc should i try next? | 07:59 |
| Qianbiao | I am stucted with the pxe dhcp network issue for days. | 08:00 |
| *** lucasagomes has joined #openstack-ironic | 08:05 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 08:06 | |
| *** stevebaker has joined #openstack-ironic | 08:17 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic master: [DNM] Test multinode https://review.opendev.org/749459 | 08:19 |
| *** dtantsur|afk is now known as dtantsur | 08:23 | |
| dtantsur | morning ironic | 08:23 |
| iurygregory | morning dtantsur | 08:23 |
| rpittau | hey dtantsur :) | 08:23 |
| stevebaker | dtantsur: good evening | 08:23 |
| dtantsur | janders: it makes sense, but I don't know if we should use internal or create our own zone. I assume our own zone is more flexible, especially in presence of other software on the machine | 08:24 |
| janders | good morning dtantsur | 08:31 |
| janders | dtantsur +1 regarding own zome | 08:32 |
| janders | s/zome/zone | 08:32 |
| janders | I'll be AFK for the next 2 hours and a bit | 08:33 |
| openstackgerrit | Dmitry Tantsur proposed openstack/bifrost master: Create our own firewalld zone and use it on real bare metal https://review.opendev.org/749285 | 08:39 |
| *** rcernin_ has joined #openstack-ironic | 08:39 | |
| dtantsur | bfournie, janders, trandles, updated ^^ | 08:39 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Fix error 601 https://review.opendev.org/749209 | 08:42 |
| openstackgerrit | Dmitry Tantsur proposed openstack/bifrost master: Use packaged iPXE ROMs for openSUSE https://review.opendev.org/747894 | 08:42 |
| *** rcernin_ has quit IRC | 08:46 | |
| rpittau | it's interesting that metalsmith integration started failing between 17 and 19 utc yesterday | 08:53 |
| dtantsur | sigh | 08:53 |
| iurygregory | magic | 08:54 |
| rpittau | I'm truly puzzled | 08:54 |
| rpittau | yep, metalsmith is completely broken | 08:55 |
| rpittau | even with the latest setuptools version | 08:55 |
| *** Nisha_Agarwal has quit IRC | 08:59 | |
| dtantsur | a link handy? | 09:00 |
| rpittau | dtantsur: https://2738d7d5de88b98d3378-26cfc451d4618cee8c0e628e323364bb.ssl.cf2.rackcdn.com/749452/1/check/metalsmith-integration-glance-localboot-centos7/76dbfca/job-output.txt | 09:02 |
| * dtantsur blinks | 09:03 | |
| *** derekh has joined #openstack-ironic | 09:05 | |
| dtantsur | lemme try something | 09:05 |
| openstackgerrit | Dmitry Tantsur proposed openstack/metalsmith master: Add the setuptools workaround https://review.opendev.org/749468 | 09:08 |
| dtantsur | dunno, maybe unrelated, but let's try ^^ | 09:08 |
| rpittau | I saw the change in devstack, I didn't think about applying that there :/ | 09:11 |
| rpittau | the multinode job issue seems more tricky, the instances stopped loading from pxe | 09:12 |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-specs master: Snapshot support https://review.opendev.org/746935 | 09:14 |
| dtantsur | rpittau: the multinode job started passing this morning :) | 09:20 |
| * dtantsur has no clue | 09:20 | |
| * rpittau facepalm and table flipping at the same time | 09:21 | |
| dtantsur | two green runs: https://zuul.openstack.org/builds?job_name=ironic-tempest-ipa-wholedisk-direct-tinyipa-multinode | 09:21 |
| dtantsur | that does not mean it's fixed, of course | 09:21 |
| rpittau | well, yeah, looking at the history though it's a huge improvement! | 09:21 |
| *** zzzeek has quit IRC | 09:32 | |
| *** zzzeek has joined #openstack-ironic | 09:33 | |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Fix error 601 https://review.opendev.org/749209 | 09:35 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Fix error 601 https://review.opendev.org/749209 | 09:35 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Fix error 601 https://review.opendev.org/749209 | 09:37 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Fix error 601 https://review.opendev.org/749209 | 09:38 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Fix error 601 https://review.opendev.org/749209 | 09:41 |
| *** zzzeek has quit IRC | 09:48 | |
| *** zzzeek has joined #openstack-ironic | 09:49 | |
| rpittau | can I please get one more review for https://review.opendev.org/746901 and https://review.opendev.org/747228 ? | 10:05 |
| patchbot | patch 746901 - ironic-python-agent-builder - Build centos8 on centos8 - 4 patch sets | 10:05 |
| patchbot | patch 747228 - ironic-python-agent-builder - Remove old proc before finalising tinyipa - 2 patch sets | 10:05 |
| *** zzzeek has quit IRC | 10:08 | |
| openstackgerrit | Verification of a change to openstack/ironic failed: Handle an older agent with agent_token https://review.opendev.org/745736 | 10:09 |
| *** zzzeek has joined #openstack-ironic | 10:10 | |
| dtantsur | we have a release deadline for ironic-lib and metalsmith this week, please review https://review.opendev.org/749053 and https://review.opendev.org/747641 | 10:18 |
| patchbot | patch 749053 - ironic-lib - Drop the dependency on oslo.service loopingcall - 1 patch set | 10:18 |
| patchbot | patch 747641 - ironic-lib - Move some generic functions from IPA - 2 patch sets | 10:18 |
| openstackgerrit | Akihiro Motoki proposed openstack/ironic-ui master: [goal] Migrate testing to ubuntu focal https://review.opendev.org/744708 | 10:18 |
| *** ijw has quit IRC | 10:19 | |
| *** ijw has joined #openstack-ironic | 10:20 | |
| *** ijw has quit IRC | 10:25 | |
| *** janders has quit IRC | 10:26 | |
| *** yolanda has joined #openstack-ironic | 10:26 | |
| *** janders has joined #openstack-ironic | 10:27 | |
| *** k_mouza has joined #openstack-ironic | 10:32 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 10:33 | |
| *** ijw has joined #openstack-ironic | 10:39 | |
| *** chihhsin has quit IRC | 10:40 | |
| *** rcernin_ has joined #openstack-ironic | 10:42 | |
| *** ijw has quit IRC | 10:44 | |
| *** rcernin_ has quit IRC | 10:47 | |
| dtantsur | yay, https://review.opendev.org/#/c/749468/ worked, please review | 10:49 |
| patchbot | patch 749468 - metalsmith - Add the setuptools workaround - 1 patch set | 10:49 |
| dtantsur | this ^^ fixes metalsmith, ironic and IPA CI hopefully | 10:52 |
| iurygregory | \o/ | 10:56 |
| rpittau | awesome | 10:57 |
| *** zzzeek has quit IRC | 10:57 | |
| *** zzzeek has joined #openstack-ironic | 11:00 | |
| *** Nisha_Agarwal has quit IRC | 11:05 | |
| *** hjensas has quit IRC | 11:06 | |
| *** hjensas has joined #openstack-ironic | 11:07 | |
| *** k_mouza has quit IRC | 11:09 | |
| *** Qianbiao has quit IRC | 11:10 | |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Explicitely do not allocate initial space for virtual volumes https://review.opendev.org/749357 | 11:11 |
| *** ijw has joined #openstack-ironic | 11:11 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic master: Adds few of the security dashboard parameters to capabilities https://review.opendev.org/746416 | 11:12 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic master: Adds few of the security dashboard parameters to capabilities https://review.opendev.org/746416 | 11:13 |
| *** ijw has quit IRC | 11:15 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic master: Adds few of the security dashboard parameters to capabilities https://review.opendev.org/746416 | 11:18 |
| *** k_mouza has joined #openstack-ironic | 11:30 | |
| *** Qianbiao has joined #openstack-ironic | 11:31 | |
| *** k_mouza has quit IRC | 11:31 | |
| *** k_mouza has joined #openstack-ironic | 11:32 | |
| *** k_mouza has quit IRC | 11:34 | |
| *** ijw has joined #openstack-ironic | 11:42 | |
| *** Lucas_Gray has joined #openstack-ironic | 11:47 | |
| *** ijw has quit IRC | 11:47 | |
| *** Goneri has joined #openstack-ironic | 11:49 | |
| *** rh-jelabarre has joined #openstack-ironic | 11:53 | |
| *** Lucas_Gray has quit IRC | 11:57 | |
| *** JamesBenson has joined #openstack-ironic | 12:01 | |
| *** uzumaki has quit IRC | 12:02 | |
| *** Lucas_Gray has joined #openstack-ironic | 12:05 | |
| *** derekh has quit IRC | 12:09 | |
| *** ijw has joined #openstack-ironic | 12:14 | |
| *** ijw has quit IRC | 12:19 | |
| *** Lucas_Gray has quit IRC | 12:30 | |
| *** Lucas_Gray has joined #openstack-ironic | 12:39 | |
| openstackgerrit | Merged openstack/metalsmith master: Add the setuptools workaround https://review.opendev.org/749468 | 12:44 |
| *** ijw has joined #openstack-ironic | 12:46 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/metalsmith stable/ussuri: Add the setuptools workaround https://review.opendev.org/749495 | 12:46 |
| openstackgerrit | Verification of a change to openstack/ironic-lib failed: Drop the dependency on oslo.service loopingcall https://review.opendev.org/749053 | 12:52 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Support caching http:// images locally in the direct deploy https://review.opendev.org/749496 | 13:00 |
| *** derekh has joined #openstack-ironic | 13:03 | |
| TheJulia | good morning | 13:11 |
| *** fdegir has quit IRC | 13:11 | |
| *** maelk has quit IRC | 13:11 | |
| iurygregory | good morning TheJulia =) | 13:12 |
| rpittau | good morning TheJulia :) | 13:12 |
| dtantsur | morning TheJulia | 13:13 |
| rpittau | TheJulia: I'm afraid I was right about the initial allocation for the virtual volumes, we were not allocating any as the behavior of virsh was "faulty", at least looking at the output of the commands | 13:14 |
| openstackgerrit | Ankit Kumar proposed openstack/ironic master: Enhance certificate verification for ilo harware type https://review.opendev.org/743490 | 13:14 |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Follow up on "Fix error 601" https://review.opendev.org/749499 | 13:15 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Support file:/// images for the direct deploy https://review.opendev.org/748379 | 13:15 |
| TheJulia | rpittau: oh joy | 13:16 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Support caching http:// images locally in the direct deploy https://review.opendev.org/749496 | 13:16 |
| * TheJulia needs to go put like... going out in the big blue room cloths to drop her fiesta off at the dealer for repairs | 13:16 | |
| openstackgerrit | Riccardo Pittau proposed openstack/bifrost master: Follow up on "Fix error 601" https://review.opendev.org/749499 | 13:18 |
| TheJulia | dtantsur: your file:/// change lgtm | 13:20 |
| dtantsur | thx! | 13:21 |
| iurygregory | somehow the multinode job is getting happier O.o | 13:23 |
| rpittau | iurygregory: yep | 13:23 |
| dtantsur | yeah.. | 13:23 |
| iurygregory | 11 jobs today 9 green O.o | 13:23 |
| TheJulia | dtantsur: I did put one comment on maybe noting a caveat, but yeah. | 13:26 |
| TheJulia | wut.... | 13:26 |
| TheJulia | w/r/t grenade | 13:26 |
| TheJulia | if nova is no longer being upgraded though.... it is likely passing | 13:27 |
| TheJulia | or the other side is down maybe | 13:27 |
| TheJulia | we have actually seen that before :\ | 13:27 |
| dtantsur | TheJulia: which multi-conductor issue do you have in mind? I'd expect us to re-cache images on rebuild, no? | 13:28 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic-python-agent-builder master: Use tinycore 11 to build tinyipa https://review.opendev.org/711647 | 13:29 |
| TheJulia | dtantsur: if someone only has the file on a single conductor and then requests deploy which is recieved managed and processed by a different conductor | 13:29 |
| dtantsur | TheJulia: ah, this. well, we should document it, but that has always been a problem with file:// images. | 13:29 |
| dtantsur | I'll add this as a work item on https://storyboard.openstack.org/#!/story/2008075 | 13:30 |
| TheJulia | true, yeah | 13:32 |
| * TheJulia looks at the corgi overlord who thinks we're taking him to a friends place because I have shoes on.... | 13:32 | |
| * TheJulia just facepalms | 13:32 | |
| janders | hi TheJulia | 13:34 |
| janders | I'm wrapping up for the night, see you tomorrow Ironic o/ | 13:35 |
| rpittau | bye janders :) | 13:35 |
| janders | ( TheJulia apologies for saying "Hi" and "Bye" at the same time :) ) | 13:35 |
| iurygregory | bye janders | 13:35 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Allow setting image_download_source per node https://review.opendev.org/749504 | 13:41 |
| dtantsur | okay, this should be the last missing bit | 13:41 |
| iurygregory | if anyone has some time today https://review.opendev.org/#/c/689844/ the only missing thing for dhcp-less =) | 13:47 |
| patchbot | patch 689844 - ironic - Add L3 boot section to the docs - 20 patch sets | 13:47 |
| *** tzumainn has joined #openstack-ironic | 13:48 | |
| dtantsur | iurygregory: could we have a CI job as well? :) | 13:49 |
| dtantsur | (+2 to the docs patch) | 13:50 |
| iurygregory | dtantsur, if you want I can give some time to figure out how to setup a job for it =) | 13:50 |
| dtantsur | that would be awesome. it's a non-trivial feature that deserves some coverage. | 13:51 |
| dtantsur | I guess it will be a separate job where you disable DHCP on the neutron subnet we use | 13:51 |
| dtantsur | or we can update one of the existing vmedia jobs? | 13:51 |
| iurygregory | hummm | 13:52 |
| iurygregory | update the vmedia jobs would make sense | 13:52 |
| iurygregory | just need to figure out how to include the simple-init element and generate a valid network_data.json | 13:52 |
| iurygregory | I think that would be the path... | 13:53 |
| dtantsur | oh yeah, a custom DIB image.. | 13:57 |
| iurygregory | yup | 13:58 |
| iurygregory | would be necessary | 13:58 |
| *** sshnaidm is now known as sshnaidm|bbl | 13:58 | |
| dtantsur | iurygregory: maybe update your doc patch to make it more obvious that the IPA images that we build by default (and ship) will NOT work? | 13:58 |
| iurygregory | dtantsur, L 250 says it's necessary to add the element https://review.opendev.org/#/c/689844/20/doc/source/admin/drivers/redfish.rst | 13:59 |
| patchbot | patch 689844 - ironic - Add L3 boot section to the docs - 20 patch sets | 13:59 |
| *** ijw has quit IRC | 14:00 | |
| iurygregory | but also say that the default images we offer won't work right? | 14:00 |
| iurygregory | I'm wondering if would be good to update the ipa-builder docs about it | 14:00 |
| *** Lucas_Gray has quit IRC | 14:05 | |
| *** thiagop has joined #openstack-ironic | 14:07 | |
| *** ijw has joined #openstack-ironic | 14:17 | |
| *** ijw has quit IRC | 14:21 | |
| dtantsur | +1 to both | 14:22 |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: Remove old proc before finalising tinyipa https://review.opendev.org/747228 | 14:23 |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: Build centos8 on centos8 https://review.opendev.org/746901 | 14:23 |
| iurygregory | ok I will update the docs =) | 14:24 |
| iurygregory | warning or note? | 14:24 |
| *** dking has quit IRC | 14:29 | |
| *** Qianbiao has quit IRC | 14:31 | |
| *** Lucas_Gray has joined #openstack-ironic | 14:32 | |
| *** fdegir has joined #openstack-ironic | 14:34 | |
| *** maelk has joined #openstack-ironic | 14:35 | |
| dtantsur | iurygregory: worth a warning, I guess? | 14:37 |
| dtantsur | up to you | 14:37 |
| iurygregory | dtantsur, ack =) | 14:37 |
| *** rcernin_ has joined #openstack-ironic | 14:44 | |
| *** Qianbiao has joined #openstack-ironic | 14:46 | |
| *** ijw has joined #openstack-ironic | 14:48 | |
| *** rcernin_ has quit IRC | 14:49 | |
| TheJulia | okay, I'm back | 14:53 |
| *** ijw has quit IRC | 14:53 | |
| dtantsur | anyone brave enough to review the TLS patch? :) https://review.opendev.org/747921 | 14:56 |
| patchbot | patch 747921 - bifrost - TLS support for API services - 21 patch sets | 14:56 |
| *** thiagop has quit IRC | 14:56 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic master: Add L3 boot section to the docs https://review.opendev.org/689844 | 14:58 |
| *** belmoreira has joined #openstack-ironic | 14:59 | |
| *** thiagop has joined #openstack-ironic | 14:59 | |
| iurygregory | dtantsur, I don't have enough knowledge but I will try =) | 15:01 |
| dtantsur | iurygregory: I didn't have enough knowledge to write it :D I believe you can review it! | 15:01 |
| iurygregory | yay! | 15:01 |
| *** Qianbiao has quit IRC | 15:03 | |
| iurygregory | gotta love gerrit "Code Review - Error / Server Unavailable 0" | 15:18 |
| dtantsur | yeah | 15:19 |
| *** ijw has joined #openstack-ironic | 15:19 | |
| rpittau | dtantsur: I'll give that a try, it's just big :D | 15:20 |
| iurygregory | 38 files | 15:22 |
| iurygregory | I'm going to deploy an env to see how it goes | 15:22 |
| *** ijw has quit IRC | 15:24 | |
| * TheJulia writes scarey warnings because... why not | 15:28 | |
| *** vmud213 has joined #openstack-ironic | 15:28 | |
| openstackgerrit | Verification of a change to openstack/ironic-python-agent failed: If listen_tls is true, enable TLS on wsgi server https://review.opendev.org/747193 | 15:29 |
| *** tosky has quit IRC | 15:31 | |
| rpittau | dtantsur: the logic looks good, just left 1 comment for linters, if you want to correct it now | 15:37 |
| dtantsur | rpittau: I guess it's copy-paste, I can update it quickly | 15:38 |
| rpittau | ok | 15:39 |
| dtantsur | rpittau: although.. it's going to cause a conflict with your lint change anyway :( | 15:40 |
| dtantsur | iurygregory: could you maybe review/approve https://review.opendev.org/#/c/749209/ so that I can rebase? | 15:40 |
| patchbot | patch 749209 - bifrost - Fix error 601 - 8 patch sets | 15:40 |
| iurygregory | dtantsur, sure | 15:42 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: [WIP] Accept and use a TLS certificate from the agent https://review.opendev.org/747136 | 15:42 |
| *** ijw has joined #openstack-ironic | 15:52 | |
| *** belmoreira has quit IRC | 15:53 | |
| openstackgerrit | James Denton proposed openstack/networking-generic-switch master: Support multiple links in link_local_information dict of portgroups https://review.opendev.org/557051 | 15:55 |
| *** ijw has quit IRC | 15:57 | |
| * TheJulia wonders if anything will break with the current patch she is working on | 15:57 | |
| thiagop | trust the gatekeeper?! | 15:58 |
| TheJulia | possibly | 15:59 |
| *** lucasagomes has quit IRC | 16:00 | |
| *** ijw has joined #openstack-ironic | 16:00 | |
| *** ijw has quit IRC | 16:01 | |
| *** ijw has joined #openstack-ironic | 16:01 | |
| openstackgerrit | Merged openstack/metalsmith stable/ussuri: Add the setuptools workaround https://review.opendev.org/749495 | 16:03 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Replace retrying with tenacity https://review.opendev.org/376574 | 16:04 |
| rpittau | slowly tenaciously moving forward with tenacity :) | 16:04 |
| rpittau | fantastic, ironic-standalone-ipa-src is passing on ubuntu focal with the correct disk space \o/ | 16:05 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic-python-agent master: Move ironic-standalone-ipa-src to Ubuntu Focal https://review.opendev.org/748213 | 16:06 |
| TheJulia | \o/ | 16:08 |
| clarkb | rpittau: did you figure out why it used more disk on focal? | 16:09 |
| rpittau | clarkb: yes! well... kinda :) | 16:10 |
| rpittau | basically the behavior of virsh was "wrong" in bionic, it seems, it was initializing the volumes with no initial allocated space, while in focal it looks correct and you need to add allocation 0 | 16:10 |
| rpittau | I'm not sure it's virsh or the translation to the resize command in qemu | 16:11 |
| TheJulia | w/r/t https://zuul.opendev.org/t/openstack/build/16831cc81d0c4c9ebb61ad3823532286 <-- It feels like we can change the tempest test to optionally delete the VIF if is an ironic since ?train? | 16:15 |
| rpittau | mmm yeah | 16:15 |
| clarkb | rpittau: thanks, good to know that the behavior is changing there | 16:15 |
| rpittau | clarkb: no problem, it's kind of fun that last 3 times I talked on #openstack-infra for a problem, I found the fix right after :D | 16:17 |
| openstackgerrit | Merged openstack/bifrost master: Fix error 601 https://review.opendev.org/749209 | 16:21 |
| clarkb | rpittau: sometimes just talking about the problem is all that is needed to figure things out (I am probably guilty of doing this a lot) | 16:22 |
| openstackgerrit | Dmitry Tantsur proposed openstack/bifrost master: TLS support for API services https://review.opendev.org/747921 | 16:25 |
| dtantsur | finally rebased ^^ | 16:25 |
| rpittau | clarkb: I completely agree, not the "guilt" part :) | 16:25 |
| * dtantsur can confirm as well | 16:25 | |
| rpittau | dtantsur: I did a follow up on 601 ansbile lint, not sure if there are conflicts with your patch, if yours merge first it's totally fine | 16:28 |
| dtantsur | shouldn't be, the keystone start was the only place | 16:28 |
| rpittau | ok | 16:28 |
| * rpittau won't let pizza wait | 16:29 | |
| rpittau | goodnight o/ | 16:29 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Remove token-less agent support https://review.opendev.org/749407 | 16:39 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Minor agent version code cleanup https://review.opendev.org/749552 | 16:39 |
| TheJulia | I think bob would appreciate reviews on https://review.opendev.org/#/c/747072 | 16:43 |
| patchbot | patch 747072 - ironic-python-agent - Update the cache if we don't have a root device hint - 6 patch sets | 16:43 |
| *** dking has joined #openstack-ironic | 16:48 | |
| openstackgerrit | vinay kumar muddu proposed openstack/ironic master: Adds ilo-uefi-https boot interface to ilo5 https://review.opendev.org/745501 | 16:50 |
| *** Lucas_Gray has quit IRC | 16:50 | |
| TheJulia | hmm, random job failures are a bit strong today | 16:51 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Minor agent version code cleanup https://review.opendev.org/749552 | 16:52 |
| openstackgerrit | Verification of a change to openstack/ironic-lib failed: Drop the dependency on oslo.service loopingcall https://review.opendev.org/749053 | 16:56 |
| *** dtantsur is now known as dtantsur|afk | 16:57 | |
| dtantsur|afk | o/ | 16:57 |
| *** vmud213 has quit IRC | 16:58 | |
| TheJulia | Anyone interested in mentoring new contributors over the next 3 or so months? | 16:59 |
| TheJulia | iurygregory: interested? | 16:59 |
| iurygregory | TheJulia, I'm interested but need to get approval from my manager probably... | 17:02 |
| TheJulia | k, I'm putting in an application for Boston University, I'll note at we may be able to provide multiple mentors and I'll just put myself on the hook. This fall seems to be mentorship season since I have also been asked to do some other mentorship | 17:03 |
| iurygregory | and we would probably need to define some tasks for the contributors I assume .. | 17:03 |
| TheJulia | Yeah, I have some thoughts around kexec and persistent agents | 17:03 |
| iurygregory | gotcha | 17:04 |
| TheJulia | since that would actually end up being cross-community collaborative in terms of working with operator/scientific communities | 17:04 |
| *** rcernin_ has joined #openstack-ironic | 17:04 | |
| iurygregory | nice | 17:04 |
| iurygregory | I'm mentor in the program of my university in Brazil (next week we will have a person adding the support for volume target on openstacksdk) =) | 17:05 |
| *** derekh has quit IRC | 17:08 | |
| *** rcernin_ has quit IRC | 17:08 | |
| *** mgoddard has quit IRC | 17:12 | |
| openstackgerrit | Merged openstack/ironic master: Handle an older agent with agent_token https://review.opendev.org/745736 | 17:52 |
| TheJulia | woot! | 18:00 |
| openstackgerrit | Julia Kreger proposed openstack/ironic bugfix/15.2: Handle an older agent with agent_token https://review.opendev.org/749563 | 18:00 |
| openstackgerrit | Julia Kreger proposed openstack/ironic bugfix/15.1: Handle an older agent with agent_token https://review.opendev.org/749564 | 18:00 |
| openstackgerrit | Merged openstack/ironic master: Explicitely do not allocate initial space for virtual volumes https://review.opendev.org/749357 | 18:03 |
| openstackgerrit | Christopher Dearborn proposed openstack/sushy master: WIP: Allow monitoring progress of a firmware update https://review.opendev.org/745950 | 18:12 |
| trandles | TheJulia: I saw your message about "mentoring new contributors" and got all excited. I'm not enrolling at BU but would love to have someone to nag with questions so I don't feel like I'm just a general annoyance. | 18:15 |
| TheJulia | lol | 18:15 |
| TheJulia | trandles: you can already nag us :) | 18:15 |
| TheJulia | and no, none of it is an annoyance! | 18:15 |
| trandles | thanks, that helps | 18:15 |
| TheJulia | So how can we help!? | 18:16 |
| trandles | oh that's actually a good transition...I have a question about IPA images in /httpboot/. Does Ironic do a sanity check (compute the sha256 hash) when the kernel and initramfs are transferred? Do I need to provide the hash in /httpboot/ for images I build? | 18:18 |
| TheJulia | so which images, the deploy kernel/ramdisk in the uuid folder. So /httpboot/<uuid>/deploy_[kernel|ramdisk]? or say /httpboot/ipa.[kernel|initramfs] | 18:19 |
| TheJulia | ? | 18:19 |
| trandles | /httpboot/ipa.* | 18:20 |
| trandles | I haven't started playing too much with the deploy_* yet this go around. | 18:21 |
| *** k_mouza has joined #openstack-ironic | 18:32 | |
| openstackgerrit | Merged openstack/ironic-python-agent master: Make WSGI server respect listen_* directives https://review.opendev.org/748767 | 18:35 |
| TheJulia | so those are placed by the playbook and would be replaced upon bifrost re-install | 18:36 |
| TheJulia | ironic doesn't look or check for checksums on those files since generally it is something you've locked up in glance/swift or your system automation is placing it | 18:36 |
| *** k_mouza has quit IRC | 18:36 | |
| * TheJulia looks at a unit test merge conflict and sighs | 18:37 | |
| *** thiagop has quit IRC | 18:46 | |
| *** thiagop has joined #openstack-ironic | 18:46 | |
| *** thiagop has quit IRC | 18:48 | |
| *** thiagop has joined #openstack-ironic | 18:54 | |
| *** ftarasenko has joined #openstack-ironic | 19:08 | |
| *** zzzeek has quit IRC | 19:17 | |
| *** zzzeek has joined #openstack-ironic | 19:18 | |
| *** zzzeek has quit IRC | 19:25 | |
| *** zzzeek has joined #openstack-ironic | 19:27 | |
| openstackgerrit | Merged openstack/ironic master: Suffix the ISO names with '.iso' https://review.opendev.org/748277 | 19:40 |
| openstackgerrit | Verification of a change to openstack/ironic-python-agent failed: If listen_tls is true, enable TLS on wsgi server https://review.opendev.org/747193 | 19:40 |
| TheJulia | ugh | 19:51 |
| *** zzzeek has quit IRC | 19:53 | |
| *** zzzeek has joined #openstack-ironic | 19:56 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic stable/ussuri: Handle an older agent with agent_token https://review.opendev.org/749585 | 19:57 |
| openstackgerrit | Julia Kreger proposed openstack/metalsmith master: Replace the Centos7 image job with a Centos8 UEFI based job https://review.opendev.org/748315 | 20:13 |
| *** gyee has joined #openstack-ironic | 20:17 | |
| openstackgerrit | Julia Kreger proposed openstack/metalsmith master: Replace the Centos7 image job with a Centos8 UEFI based job https://review.opendev.org/748315 | 20:20 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent master: If listen_tls is true, enable TLS on wsgi server https://review.opendev.org/747193 | 20:23 |
| JayF | ^ rebase of previously approved patch | 20:23 |
| JayF | TheJulia: the IPA TLS patch merge was a conflict, not a CI failure fwiw. You are now +1 ugh :p | 20:24 |
| TheJulia | JayF: thanks | 20:25 |
| TheJulia | I was banging my head on a backport when I said ugh, but seems like the feeling is familiar :) | 20:25 |
| JayF | I knew those would probably conflict | 20:27 |
| JayF | but didn't wanna force a merge order | 20:27 |
| JayF | I wish I could teach git just a little python so it knows how to resolve those :D | 20:28 |
| *** ijw has quit IRC | 20:34 | |
| *** ijw has joined #openstack-ironic | 20:35 | |
| *** Lucas_Gray has joined #openstack-ironic | 20:42 | |
| *** jtomasek has quit IRC | 20:42 | |
| openstackgerrit | Merged openstack/ironic-lib master: Drop the dependency on oslo.service loopingcall https://review.opendev.org/749053 | 20:43 |
| *** zzzeek has quit IRC | 20:50 | |
| *** zzzeek has joined #openstack-ironic | 20:51 | |
| *** dustinc has joined #openstack-ironic | 20:58 | |
| *** uzumaki has joined #openstack-ironic | 21:02 | |
| *** rcernin_ has joined #openstack-ironic | 21:06 | |
| *** rcernin_ has quit IRC | 21:12 | |
| *** k_mouza has joined #openstack-ironic | 21:12 | |
| *** ijw has quit IRC | 21:14 | |
| *** ijw has joined #openstack-ironic | 21:14 | |
| TheJulia | clarkb: is there any way to tell zuul to not to retry jobs? just collect the logs from the single failure and return it? | 21:19 |
| clarkb | I don't think so | 21:20 |
| clarkb | well thats not entirely true I guess. Jobs retry under two conditions. The first is if ansible returns an exit code that indicates a network error. The other is if the failure happens in a pre playbook | 21:21 |
| clarkb | in the case of issues with a pre playbook you could move the problematic code to the run playbook | 21:21 |
| *** k_mouza has quit IRC | 21:21 | |
| clarkb | TheJulia: ^ | 21:21 |
| TheJulia | hmm, it is a devstack plugin setup that is timing out | 21:21 |
| *** k_mouza has joined #openstack-ironic | 21:27 | |
| *** Goneri has quit IRC | 21:30 | |
| *** k_mouza has quit IRC | 21:31 | |
| *** uzumaki has quit IRC | 21:39 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent stable/ussuri: Make WSGI server respect listen_* directives https://review.opendev.org/749597 | 21:52 |
| JayF | ^^ backport of that bugfix I wrote to stable/ussuri. | 21:52 |
| JayF | looks like stable/ussuri was the first release it was in | 21:55 |
| TheJulia | +2 :) | 21:55 |
| JayF | is there some bugfix branch that needs to go to, as well? | 21:55 |
| TheJulia | yes | 21:55 |
| JayF | what's the branch and I'll cherry pick to there, too | 21:55 |
| JayF | and/or teach me how to fish :) | 21:56 |
| TheJulia | I've got it | 21:56 |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent bugfix/6.3: Make WSGI server respect listen_* directives https://review.opendev.org/749598 | 21:56 |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent bugfix/6.2: Make WSGI server respect listen_* directives https://review.opendev.org/749599 | 21:56 |
| TheJulia | there it goes | 21:56 |
| JayF | any better place than here to discover those versions? https://docs.openstack.org/releasenotes/ironic-python-agent/unreleased.html | 21:57 |
| TheJulia | JayF: best place is the openstack/releases repo | 21:59 |
| TheJulia | getting you a link | 21:59 |
| JayF | racing you to find it :D | 21:59 |
| *** brtknr has quit IRC | 21:59 | |
| *** SpamapS has quit IRC | 21:59 | |
| TheJulia | https://github.com/openstack/releases/blob/master/deliverables/victoria/ironic-python-agent.yaml | 22:00 |
| JayF | ack, perfect, will know to check there | 22:00 |
| JayF | does anyone run those bugfix ipa versions? | 22:00 |
| *** brtknr has joined #openstack-ironic | 22:00 | |
| *** SpamapS has joined #openstack-ironic | 22:00 | |
| TheJulia | they went back to the beginning of time and recorded all of the commit shas/tags too | 22:00 |
| JayF | or are we just checking a box so we can keep cycle-with-intermediary | 22:00 |
| TheJulia | we have downstream consumption of intermediate releases | 22:01 |
| JayF | nice | 22:01 |
| JayF | I'm pretty excited we'll be able to run stock IPA + the TLS patch (which probably shouldn't be backported due to the eventlet.monkey_patch() addition) | 22:02 |
| clarkb | TheJulia: have a pointer to the thing that is in a retry loop? | 22:02 |
| TheJulia | or at least, we're working towards it, I've not been on that call in a while, but the idea is basically be able to track newer branch for other stuff and let OSP track the longer term stable branch | 22:02 |
| TheJulia | clarkb: sure, one minute | 22:02 |
| * clarkb has dug out of reviews and can take a look now | 22:02 | |
| TheJulia | clarkb: I _think_ it is likely because it is all native zuul playbooks for metalsmith and then devstack with ironic | 22:02 |
| TheJulia | clarkb: 748315,6 in ruunning zuul jobs right now | 22:04 |
| TheJulia | basically I'm trying to swap a job around, but of course I've got a failure occuring so it retires 3 times because it is trying to get to the actual test job | 22:05 |
| TheJulia | I guess that means devstack runs in pre.yaml which... eeeeek | 22:05 |
| clarkb | ya some places do that because the assumption is if you're testing things against a cloud then the cloud is assumed to just spin up successfully | 22:05 |
| TheJulia | yeeeahh.. roles run-devstack are in pre.yaml | 22:06 |
| TheJulia | That ought to be in the job yaml | 22:06 |
| *** k_mouza has joined #openstack-ironic | 22:07 | |
| * TheJulia looks at the time, looks at the spoons, and just suspects it is time for a glass of wine or something | 22:07 | |
| trandles | OMG I would love a beer or wine right now... | 22:07 |
| TheJulia | well, if you have a teleporter.... | 22:07 |
| trandles | I have to be good. I'm on call for production until 17:00 | 22:07 |
| JayF | The last three trips to the store I've made, they've had zero gluten free beer /o\ | 22:07 |
| clarkb | ya its a balance. I want to say "typical" tempest jobs put devstack in run beacuse spinning up a working cloud is part of what is tested | 22:08 |
| clarkb | but for say openstacksdk it just needs a working cloud to run against | 22:08 |
| TheJulia | yeah, I guess I'll put up a separate change tomorrow to rework the jobs so things are not auto-retried when it is not metalsmith directly that is breaking | 22:08 |
| TheJulia | trandles: so...... 51 minutes ;) | 22:09 |
| trandles | TheJulia: indeed | 22:09 |
| * TheJulia goes and takes the corgi overlord out | 22:10 | |
| clarkb | TheJulia: if its just a pre-run timeout I think you can bump the pre-run timeout value | 22:10 |
| clarkb | I'm double checking what the config for that is now | 22:10 |
| *** thiagop has quit IRC | 22:10 | |
| *** k_mouza has quit IRC | 22:11 | |
| clarkb | TheJulia: pre-run and run share the same timeout. Once the job starts running they both have to finish under the normal job timeout value. Then post-timeout applies to post-run only and is a separate value. The split is done so that things that need to happen after a failure can take more time even if the previous playbooks failed | 22:14 |
| TheJulia | k, good to know. | 22:15 |
| *** rcernin_ has joined #openstack-ironic | 22:15 | |
| TheJulia | sadly it is not timeout related. It is legitimately timing out because the VM is not standing up a UEFI VM with the required firmware for some reason. Which is surprising :\ because I thought ubuntu had fixed the packages for that but I guess not :\ | 22:19 |
| clarkb | the linaro cloud we use is all uefi VMs (but arm64) I wonder if they have pointers | 22:20 |
| TheJulia | it specifically has to do with embedded ipxe in the network card firmware | 22:21 |
| TheJulia | or the efi firmware doing something totally funky | 22:26 |
| stevebaker | Hey I've got an appointment in town, I'll be off and online as cafe internet allows | 22:27 |
| rpioso | TheJulia: Regarding the concern about our proposed meta cleaning/deploy step's prospective interaction with software RAID, doesn't the same concern exist if a user utilizes agent RAID to configure software RAID, changes the node's RAID interface to a hardware RAID interface such as idrac-wsman, and configures hardware RAID? And yet again if the software RAID and hardware RAID order is reversed? | 22:27 |
| rpioso | TheJulia: Seems to me the side effect on the node's RAID config on successful step completion may not be robustly implemented in all cases. | 22:29 |
| rpioso | TheJulia: Shouldn't it be add the new software/hardware RAID config to the existing complementary hardware/software RAID config? | 22:31 |
| TheJulia | rpioso: well, that too is a possibility from the example that I think ?dmitry? pointed out on the call. Doing so is entirely up to the operator and maybe they know what they are doing in that case. The larger concern is not an order of operation issue, but a greater consistency and compatibility rabbit hole of sorts. | 22:31 |
| *** stevebaker has quit IRC | 22:33 | |
| TheJulia | I think that focusing on a user changing interfaces and setting/changing target raid configuration between them is kind of totally unrelated to the larger sets of conundrums. That being said, given the nature of specifying disks with hardware raid, if it was done then software raid that seems feasible. The software raid model doesn't quite (yet) have the same support/functionality (afaik, i may be wrong) | 22:33 |
| rpioso | TheJulia: I guess I don't understand the larger set of issues. Please elaborate. | 22:34 |
| TheJulia | if someone were to define, and intentionally swap between two configs and different interfaces, that is really up to them as they have taken explicit actions to apply their desired configuration | 22:34 |
| rpioso | TheJulia: Seems similar applies to using the meta step to configure RAID in ironic RAID schema speak and using the RAID interface, too. | 22:35 |
| rpioso | TheJulia: Note that no such issue applies to BIOS. | 22:36 |
| rpioso | TheJulia: Are there larger issues? | 22:36 |
| TheJulia | Well, in dell's specific case it does not | 22:36 |
| rpioso | TheJulia: Others? | 22:36 |
| TheJulia | let me try and re-form the words | 22:36 |
| TheJulia | Give me like 10 minutes | 22:36 |
| rpioso | TheJulia: Absolutely :-) | 22:37 |
| *** k_mouza has joined #openstack-ironic | 22:41 | |
| TheJulia | so where should I start in this thought process? Asking so I can try and figure out where the context disconnect seems to be | 22:42 |
| TheJulia | rpioso: ^^^ | 22:44 |
| *** k_mouza has quit IRC | 22:47 | |
| *** sshnaidm|bbl is now known as sshnaidm|afk | 22:47 | |
| openstackgerrit | Merged openstack/ironic-lib master: Move some generic functions from IPA https://review.opendev.org/747641 | 22:54 |
| openstackgerrit | Merged openstack/ironic-python-agent master: Refactor and fix documentation https://review.opendev.org/747723 | 22:54 |
| openstackgerrit | Merged openstack/ironic-python-agent master: Document in-band deploy steps and add more docs for custom steps https://review.opendev.org/747753 | 22:54 |
| rpioso | TheJulia: Aside from our addressing the very legitimate concerns expressed about the storage, security, and access to the storage configuration files, which we're now calling "configuration molds", I do not understand what would block the approval of our vendor-independent proposal. | 22:55 |
| TheJulia | Okay | 22:55 |
| * TheJulia tries to figure out a way to word this | 22:56 | |
| rpioso | TheJulia: And we're working on addressing those configuration mold concerns. | 22:56 |
| TheJulia | rpioso: cool | 22:56 |
| TheJulia | so you were on the call and we spent quite a bit of time going back and forth over the operational permutations correct? | 22:56 |
| TheJulia | I realize you were on it, but I'm trying to really get to the heart of the point as to why the community is encouraging you to defer your perception of vendor-independent and translated/cross compatible model as described | 22:58 |
| rpioso | TheJulia: Yes, I was leading the discussion :-) I believe I caught most, if not all of it. Following the meetup, I've further thought about the community's feedback. The example, similar funky interactions identified above came out of that. Seems there's a fundamental issue, which has nothing to do with our proposal. | 23:00 |
| rpioso | TheJulia: Our proposal is just another way to expose it. | 23:00 |
| TheJulia | Actually, the conundrum is more with the size, scope, and single use nature of the proposal. A vendor specific step, is really kind of trivial, where what kind of ended up as circular discussions is when we started digging into the conundrums related to vendor independent espescially with configuration translation/transformation to try and be cross-interface compatible | 23:01 |
| TheJulia | so the consensus and encouragement was more "focus on the first part", and from there it can be iterated upon. Of course different words, but we're really all trying to save you guys headache and pain in the process of getting the feature in by focusing on the first logical chunk | 23:02 |
| rpioso | TheJulia: If our proposal is implemented in a vendor-agnostic manner, ironic consumers, such as Metal3, could consume it when offered by a driver, especially if/when dhellman's capability proposal is implemented. | 23:04 |
| *** rcernin_ has quit IRC | 23:04 | |
| TheJulia | the larger whole, while intriguing and potentially useful, is also kind of equally vague and has many potential traps with what operators could do because out of the gate it is still vendor specific. | 23:04 |
| *** rcernin has joined #openstack-ironic | 23:04 | |
| rpioso | TheJulia: I appreciate that you're trying to save us from headaches and pain, but I need to better understand the source of that. | 23:05 |
| TheJulia | so to your own point, "when offered by a driver". We've had other driver contributors explicitly state they have no intention of doing exactly what you guys are attempting | 23:05 |
| TheJulia | which also makes us more anxious about the larger functionality | 23:05 |
| rpioso | TheJulia: It would be an optional cleaning/deploy step. Nothing vendor-specific about that. | 23:06 |
| rpioso | TheJulia: Their choice. Just like optionally setting boot mode :-) | 23:06 |
| TheJulia | rpioso: if that step was not trying to make compatible configuration representing what was captured, then that is a totally different story | 23:06 |
| TheJulia | rpioso: indeed, but generally there is enough commonality to have a generally supportable pattern there | 23:07 |
| rpioso | TheJulia: What do you mean by "make compatible configuration representing what was captured"? | 23:07 |
| janders | good morning Ironic o/ | 23:07 |
| *** Lucas_Gray has quit IRC | 23:07 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent master: If listen_tls is true, enable TLS on wsgi server https://review.opendev.org/747193 | 23:07 |
| TheJulia | now, if say I go make a generic "save raid and bios settings" step, that could fit with-in what you guys are proposing | 23:08 |
| *** dustinc has quit IRC | 23:08 | |
| JayF | ^^ caught the unit test failure right as I left. Just had to add `use_ssl` to the new unit test added by the host/port bugfix. | 23:08 |
| * JayF out for the day, glhf | 23:08 | |
| TheJulia | where enter the danger zone of sorts is committing to stable interactions across potential and even explicit interface changes | 23:08 |
| TheJulia | g'night JayF | 23:08 |
| JayF | TheJulia: rpioso: I am not an expert by any means on this proposal, but I remember back in the day, we always suggested waiting until two vendors implemented a thing before making it generic | 23:09 |
| JayF | in fact IIRC that was the entire purpose behind vendor_passthru being added | 23:09 |
| TheJulia | rpioso: you folks indicated that if say a raid5 was natively configured in capturing the configuration it would be translated into the target raid configuration | 23:09 |
| JayF | the only thing worse than no abstraction is the wrong one :( | 23:09 |
| TheJulia | JayF: this is true, sadly vendor_passthrough is kind of an undocumented thing for many vendors or the underlying methods are documented or stuff like that because it is vendor specific code area :( | 23:10 |
| TheJulia | JayF: ack, thanks for the test fix | 23:10 |
| rpioso | TheJulia: I don't recall any discussion of RAID 5. | 23:10 |
| JayF | I mean, it's for me :D | 23:10 |
| JayF | have a good one folks; good luck resolving the issues around the giant red [delete it all] button :D | 23:11 |
| *** hjensas has quit IRC | 23:11 | |
| rpioso | TheJulia: We could add such a step "save BIOS and RAID settings" step, but the existing steps should do the right thing by themselves. BIOS already does, RAID not so much. And such a new step could require booting the ramdisk yet again. | 23:12 |
| TheJulia | rpioso: which kind of goes back to I think what dmitry said, which was something along the lines of if it is just saving and then restoring that is one thing, but it is the additional behavior that is where things get complex | 23:13 |
| TheJulia | rpioso: I'm not sure I follow your last statement. Specifically the last sentence | 23:13 |
| rpioso | TheJulia: Obtaining the software RAID configuration from the source of truth requires IPA. | 23:15 |
| TheJulia | in what context are you percieving this as being the case? | 23:15 |
| TheJulia | what interfaces? | 23:15 |
| rpioso | TheJulia: The new "save BIOS and RAID settings" step you suggested. | 23:15 |
| TheJulia | Well, so in your term, the collect the mold step | 23:16 |
| TheJulia | now would the execution of that step do *anything* else other than publish a file? | 23:17 |
| rpioso | TheJulia: AFAIU, iDRAC doesn't know anything about software RAID. | 23:17 |
| rpioso | TheJulia: It cannot configure it :-) | 23:17 |
| rpioso | TheJulia: The collect mold step, export_config, would do nothing other than what you described, publish a file. | 23:18 |
| TheJulia | i'm sure if it did it wouldn't take long for a "Hi, I want a raid 1 with a single hard disk" support case or better yet "raid 5 with 2 disks" | 23:18 |
| rpioso | TheJulia: It presently can't. Don't know if it's even possible. | 23:19 |
| *** k_mouza has joined #openstack-ironic | 23:19 | |
| rpioso | TheJulia: I agree any step shouldn't mess with things it doesn't touch. Hardware RAID shouldn't mess with software RAID config stored on the node and vice versa. | 23:19 |
| TheJulia | so in nothing that your perceiving as the scope of work for the larger feature includes any configuration compatability transformation? | 23:20 |
| rpioso | Please define "configuration compatibility transformation". | 23:20 |
| TheJulia | any part of the step taking any data and touching existing node bios settings or raid configuration fields | 23:21 |
| rpioso | TheJulia: Fix existing, broken RAID config steps' storage of RAID config on the node following completion and all would be good. Of course, the meta step would have to behave correctly, too. | 23:22 |
| rpioso | TheJulia: BIOS is good, because it caches all of the settings on completion. | 23:22 |
| TheJulia | define fix existing | 23:23 |
| *** k_mouza has quit IRC | 23:24 | |
| TheJulia | and what are you percieving as broken exactly about the raid config steps? | 23:24 |
| rpioso | TheJulia: Hardware RAID replaces node's raid_config without considering software RAID config which wasn't affected. | 23:24 |
| rpioso | TheJulia: I suspect vice versa exists, too. | 23:24 |
| rpioso | TheJulia: You and dtantsur seemed to be familiar with this interaction. | 23:25 |
| TheJulia | so yes, it would potentially be out of date but the contract is what was last set throught the raid commit configuration step | 23:25 |
| rpioso | TheJulia: That contract is broken. | 23:25 |
| TheJulia | how so, exactly | 23:25 |
| rpioso | TheJulia: It does not accurately represent the node's RAID configuration. | 23:25 |
| TheJulia | in what case | 23:26 |
| rpioso | TheJulia: I explained it ^^^ | 23:26 |
| TheJulia | *sigh* | 23:26 |
| rpioso | TheJulia: agent RAID followed by any hardware RAID or vice versa. | 23:26 |
| TheJulia | Yes, it would not accurately, and could possibly never acurately represent everything unless it was a multi-demmensional object | 23:27 |
| rpioso | TheJulia: Then resulting raid_config is one or the other, instead of a combination of both software and hardware RAID config. | 23:27 |
| rpioso | TheJulia: Each should know about its own controllers and disks managed by the same. | 23:27 |
| TheJulia | It is becoming clear we need to have another call about this | 23:28 |
| rpioso | TheJulia: Don't mess with volumes controlled by controllers it can't interact with. | 23:28 |
| TheJulia | huh? | 23:28 |
| *** lifeless has quit IRC | 23:29 | |
| *** irclogbot_2 has quit IRC | 23:29 | |
| rpioso | TheJulia: Hardware RAID should not wipe out software RAID configuration on the node and vice versa. It can identify software RAID config by the ID of the RAID controller identified in the logical_volume. | 23:30 |
| TheJulia | There is a fundimental difference in behavior and the meta step reconcilation and application of present state data over the existing fields is problematic, it would be almost easier to just wipe the contents out and say we know nothing about the present state after the step runs, to be totally honest | 23:30 |
| TheJulia | The raid interfaces were never designed around supporting such nested configurations. We as a community don't presently support it, but we know people _DO_ these sorts of things in the field and for hyper specific reasons | 23:31 |
| TheJulia | So it comes down to a cost of supporting such a case for the community, at least in terms of the raid configuration. | 23:32 |
| rpioso | TheJulia: I don't agree. It seems very solvable to me. Key off of the RAID controller ID in the volume's definition. | 23:32 |
| rpioso | TheJulia: No nesting. | 23:32 |
| rpioso | TheJulia: Each logical volume has a RAID controller ID. | 23:32 |
| *** irclogbot_3 has joined #openstack-ironic | 23:33 | |
| TheJulia | Now in terms of the meta step, the sticking point is saving the collected data over the existing fields when configuration is applied, which is only echoed by our discussion several times over now in just the last half hour :) | 23:33 |
| TheJulia | rpioso: in that case, by all means submit a patch and treat it as a bug | 23:33 |
| rpioso | TheJulia: And do you agree that bug should have no bearing on the meta step's acceptability? | 23:36 |
| TheJulia | Does the meta step's saving of the mold to the machine have an intent on changing bios or raid configuration data stored in ironic to represent in ironic's format? | 23:37 |
| rpioso | s/saving/applying/ ? | 23:38 |
| TheJulia | Your proposed sed statement is acceptable | 23:38 |
| TheJulia | but back to your question if it has bearing | 23:39 |
| TheJulia | Will hose fields/data in ironic's db have updates applied | 23:39 |
| rpioso | TheJulia: Yes, when the BIOS and/or RAID config is expressed in ironic terms: ironic RAID schema and BIOS key-value pairs and only then. If it expressed in OEM terms, no, it will not modify the node's saved configuration. | 23:39 |
| rpioso | TheJulia: OEM could, too, but that seems more questionable. | 23:41 |
| TheJulia | Still working on the huge block of text | 23:54 |
| TheJulia | but, agreed on OEMey specific things, but if they are visible or even nested, then maybe!?! The bios interface specifically was modeled around redfish's key/value pair fields without an OEM field so maybe it is not a good idea^(tm). | 23:55 |
| *** k_mouza has joined #openstack-ironic | 23:57 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!