Wednesday, 2021-10-13

holtgreweHi. I'm using kolla/ansible on wallaby. I have trouble with the bifrost IPA image. The image boots into a login prompt but I cannot login. I have setup `etc/kayobe/bifrost.yml` as by the manual with `kolla_bifrost_dib_elements_extra` including `- "devuser"`, and set in `kolla_bifrost_dib_env_vars_extra` the env variables `DIB_DEV_USER_USERNAME: "devuser"` and `DIB_DEV_USER_PASSWORD: "calvin"`. I have subsequently removed /httpboot/deployment_05:43
holtgreweimage.qcow2-new from the bifrost container and rebuilt it.05:43
holtgrewe(actually removed .qcow2 not .qcow-new)05:43
holtgreweHowever, I cannot log into a machine booted with the updated image either and can find no reference to the devuser user in the qcow2 image if I mount it and grep inside of it...05:44
jingvarholtgrewe: there are 2 images - ironic python agent (accepts devuser) and deployment image - accepts user via cloudinit 06:06
holtgrewe@jingvar, that's good to know... so, I would have to remove /httpboot/ipa.* and then call `kayobe seed service deploy` again?06:09
holtgreweI need to get a shell on the machine after booting into IPA discovery. Something appears to be weird with the network (I have an LACP bond setup on net2/3 and while net0/1 are not connected, the machine gets an IP during PXE on net2 and boots into the image but for some reason then cannot configure the net2 interface any more...)06:11
holtgreweI *think* that the switches are setup correctly with lacp fallback but for debugging, I really have to see what the IPA image is doing and what the results are.06:11
jingvarjust change pxe_boot_mac to net2 mac06:12
holtgreweSorry, I don't follow. It successfully boots into the IPA image but then I cannot log into the prompt "localhost".06:14
holtgreweBut then network manager takes over and cannot take up the interface.06:14
jingvarironic set ipmi to boot from first nic - and create  /httpboot/pxelinux.cfg/ a file like aa-bb-cc-...06:14
jingvarif node can't boot from interface with this mac it will try boot from second nic, and comes to ironic with different mac and will booted with /httpboot/pxelinux.cfg/default 06:16
holtgreweI might be missing something here. I have just setup the seed machine and I'm using manual ipmitools to have the machine pxe boot and I want it to discover. Yes, it boots with /httpboot/pxelinux.cfg/default06:17
holtgreweSo I'm now seeing "CentOS Linux 8\nKernel...\n....localhost login:"06:17
*** amoralej|off is now known as amoralej06:19
jingvarDo you make a barematal host into ironic with the ipmi creds06:19
holtgreweBut I cannot login with devuser:calvin although I set the user and password for devuser into kolla_bifrost_dib_env_vars_extra and I have devuser in kolla_bifrost_dib_env_vars_extra. I have the following in /etc/bifrost/dib.yml in the bifrost-deploy container.06:20
holtgrewedib_env_vars:06:20
holtgrewe  DIB_DEV_USER_PASSWORD: calvin06:20
holtgrewe(of course that's just an excerpt).06:20
jingvarIPA don't configure bond etc, just boot and collect hardware info and send it to ironic, depends on settings it can't wait command from ironic or shutdown the node06:20
holtgreweyes, yes, it successfully booted into the IPA image06:20
holtgrewebut I now need to figure out what's going on wrong inside the IPA image06:21
holtgrewebut I cannot loginto the machine while it's in the IPA image06:21
holtgrewein other words: what's the root password and/or how do I get a user with known password into the IPA image?06:21
jingvarussaly there is a lot of magic06:22
jingvarby default no user06:22
jingvarlet me check06:22
jingvarDo you have kayobe-config/etc/kolla/.. deuser ?06:24
holtgreweSorry, what exactly?06:27
jingvarkayobe render it config to kolla config 06:29
jingvarsome times i faced with that variables from kayobe don't delivered to kolla06:30
holtgrewe`grep -B 3 devuser etc/kolla/config/bifrost/dib.yml`  => "dib_env_vars:\n    DIB_CLOUD_INIT_DATASOURCES: ConfigDrive\n    DIB_DEV_USER_PASSWORD: calvin\n    DIB_DEV_USER_USERNAME: devuser"06:30
holtgreweYes, and that is also put into the bifrost-deploy container on the host.06:30
holtgreweThe file `/etc/bifrost/dib.yml` inside of the bifrost container also has the `dib_env_vars` with `DIB_DEV_USER_USERNAME`.06:32
holtgreweIs the devuser and password baked into the ipa.initramfs image?06:32
jingvarI think yes06:33
jingvarkolla_bifrost_dib_elements_extra ?06:34
holtgreweInside of the bifrost contrainer in /etc/bifrost/dib.yml I have dib_elements: disable-selinux enable-serial-console vm devuser dynamic-login cloud-init-datasources.06:35
jingvarI think it is ok06:36
holtgreweHm, OK... let me extract the initramfs image and look inside.06:36
jingvarI've added devuser and redeploy seed services06:36
jingvarDo you added devuser in first run or after issues? probably it don't rebuild image because it already exist?06:39
holtgreweI removed the image and then did the `seed service deploy`. But I'm happy to do this again.06:40
jingvarTASK [Bootstrap bifrost (this may take several minutes)] *06:41
holtgreweSame here, and I already havea n ipa.initramfs from 8:44 today. BUT the SHA256 sum is `0d5d1e1b3ad7425aca0eeddecf9a8497e7af0fc2754fefd4a686bc3d66ca30fe  ipa-centos8-stable-wallaby.initramfs` both for the old and the new one.06:46
holtgreweSo apparently it is just downloaded and nothing is baked in.06:46
holtgreweOh... https://docs.openstack.org/kayobe/latest/configuration/reference/ironic-python-agent.html#example-configuring-a-development-user-account06:48
holtgrewed'oh!06:48
holtgreweI have to go to ipa.yaml not bifrost.yaml06:48
jingvar:)06:48
holtgrewe... and I just got new glasses. How could that happen? Thanks and sorry for taking your time.06:49
jingvarI'm not sure, bifrost should brings settings to ipa06:50
holtgreweI'm re-runnign the seed deployment now.06:50
holtgreweOK, still the same sha256 sum.06:53
holtgrewe`ipa_build_images: true` that might help.06:53
jingvarI'v got the same - can't login06:55
jingvaripa_build_images: true` - maybe, but i think it just download base image and install python agent06:56
holtgreweOK, service deploy does not build the image. Maybe `seed container image build`.06:58
jingvarmaybe just unpack initramfs, chroot, add user, pack and update sha06:58
holtgrewe:-D06:59
holtgreweNow I'm seeing `/bin/sh -c ELEMENTS_PATH=/opt/kayobe/src/ironic-python-agent-builder/dib DIB_RELEASE=8-stream DIB_REPOLOCATION_ironic_agent=https://opendev.org/openstack/ironic-python-agent DIB_REPOREF_ironic_agent=stable/wa07:04
holtgrewellaby DIB_DEV_USER_USERNAME=devuser DIB_DEV_USER_PASSWORD=calvin DIB_DEV_USER_PWDLESS_SUDO=yes /opt/kayobe/venvs/kayobe/bin/python /home/stack/.ansible/tmp/ansible-tmp-1634108498.7438307-37998-244734892192366/AnsiballZ_command.py`  in ps, so it's rebuilding.07:04
jingvarok07:06
holtgreweThe command was `kayobe seed deployment image build` as written here https://docs.openstack.org/kayobe/ussuri/configuration/ironic-python-agent.html. What do you know, all very well documented I just have trouble finding it.07:07
jingvarI don't know , seed service deploy - I see output DIB_DEV_USER_USERNAME\": \"devuser\", - and belive that it applyed07:11
jingvarTASK [bifrost-create-dib-image : Initiate ramdisk build]   \"skip_reason\": \"Conditional result was False\"",07:13
jingvarohh07:13
alecorps7hello07:24
holtgreweaaand I have a login!07:28
opendevreviewMichal Nasiadka proposed openstack/kolla master: cinder-volume/ubuntu: add lsscsi and nvme  https://review.opendev.org/c/openstack/kolla/+/81361107:32
opendevreviewMichal Nasiadka proposed openstack/kolla stable/wallaby: gnocchi: update to 4.4.1  https://review.opendev.org/c/openstack/kolla/+/81375207:34
mnasiadkamorning07:34
jingvarmorning07:43
jingvarholtgrewe: how?07:43
holtgreweMy custom built IPA image is served and I can now login with devuser:calvin.07:44
holtgreweAnd I could "fix" my problem by taking down and up the port on the switches... Dell switches are weird sometimes.07:44
opendevreviewVerification of a change to openstack/kolla master failed: docs: Move Kayobe RC1 after Kolla & Kolla Ansible RC1  https://review.opendev.org/c/openstack/kolla/+/81298207:44
opendevreviewBoxiang Zhu proposed openstack/kolla master: Skyline: Add skyline dockerfile  https://review.opendev.org/c/openstack/kolla/+/81079608:19
opendevreviewBoxiang Zhu proposed openstack/kolla-ansible master: Skyline: Add skyline role  https://review.opendev.org/c/openstack/kolla-ansible/+/81056608:20
opendevreviewMaksim Malchuk proposed openstack/kayobe master: Add the missing documentation for mariadb recovery  https://review.opendev.org/c/openstack/kayobe/+/81378108:26
opendevreviewMaksim Malchuk proposed openstack/kayobe master: Add the missing documentation for mariadb recovery  https://review.opendev.org/c/openstack/kayobe/+/81378108:26
opendevreviewMark Goddard proposed openstack/kayobe master: Add support for Ironic inspection through DHCP-relay  https://review.opendev.org/c/openstack/kayobe/+/81326608:53
opendevreviewBoxiang Zhu proposed openstack/kolla-ansible master: Skyline: Add skyline role  https://review.opendev.org/c/openstack/kolla-ansible/+/81056608:56
opendevreviewMerged openstack/kolla master: docs: Move Kayobe RC1 after Kolla & Kolla Ansible RC1  https://review.opendev.org/c/openstack/kolla/+/81298209:02
opendevreviewMark Goddard proposed openstack/kayobe master: Remove stale config dump from seed hypervisor host configure  https://review.opendev.org/c/openstack/kayobe/+/81378609:14
holtgreweSo I have one interface up on the node running the IPA image. What should happen next?09:21
opendevreviewMichal Nasiadka proposed openstack/kolla-ansible master: Tidy up renos before Xena release  https://review.opendev.org/c/openstack/kolla-ansible/+/81378809:21
opendevreviewBoxiang Zhu proposed openstack/kolla master: Skyline: Add skyline dockerfile  https://review.opendev.org/c/openstack/kolla/+/81079609:25
holtgreweThe nodes are now visible in bifrost via baremetal node list... However, `kayobe overcloud introspection data save --output-dir share/introspection-data --output-format YAML` fails.09:27
holtgreweOK... I guess I have to fix the validate results.09:29
opendevreviewVerification of a change to openstack/kolla-ansible master failed: Add support for Ironic inspection through DHCP-relay  https://review.opendev.org/c/openstack/kolla-ansible/+/81326809:32
opendevreviewVerification of a change to openstack/kolla stable/ussuri failed: neutron: Fix update-alternatives on CentOS  https://review.opendev.org/c/openstack/kolla/+/81330509:36
mnasiadkayoctozepto: https://review.opendev.org/c/openstack/kolla-ansible/+/813788 - and then I'll post change to cut RC1s09:40
opendevreviewVerification of a change to openstack/kolla-ansible stable/wallaby failed: Add check and diff options to kolla-ansible  https://review.opendev.org/c/openstack/kolla-ansible/+/81295910:04
holtgreweHm, is there any way to tell kayobe that I want to have bifrost detect with the idrac rather than the ipmi driver?10:25
holtgreweAh, it's probably inspector_discovery_enroll_node_driver10:26
holtgreweWhich kayobe command rebuilds the ironic.conf fiel?11:48
jingvarseed service deploy, or dive into playbooks 11:54
*** amoralej is now known as amoralej|lunch12:02
mgoddardjingvar: it's nice to see that you're on the kayobe learning curve and are now able to give back by helping others12:47
holtgreweI can't seem to get my nodes over the verifying step. Here is the `baremetal node validate stack-1` output. https://pastebin.com/z5LbWejT13:02
opendevreviewMerged openstack/kolla-ansible stable/wallaby: Add check and diff options to kolla-ansible  https://review.opendev.org/c/openstack/kolla-ansible/+/81295913:02
holtgreweI think the deploy one is not evil by the docs.13:03
*** amoralej|lunch is now known as amoralej13:04
opendevreviewMichal Arbet proposed openstack/kolla-ansible stable/wallaby: Trivial fix shebang in keystone's fernet-node-sync.sh.j2  https://review.opendev.org/c/openstack/kolla-ansible/+/81381613:08
opendevreviewMichal Arbet proposed openstack/kolla-ansible stable/wallaby: Trivial fix shebang in keystone's fernet-node-sync.sh.j2  https://review.opendev.org/c/openstack/kolla-ansible/+/81381613:12
holtgreweWill a message `Driver idrac does not support console (disabled or not implemented).` prevent a node from being provisioned by ironic?13:18
jingvarmgoddard: :)13:23
jingvari have no experince with idrac13:29
holtgrewe@jingvar, have you ever seen something similar? when switching to IPMI I see different flavours of problems.13:29
holtgreweooooh13:30
holtgreweI see now. I'm looking into /var/log/kolla/ironic/*13:30
holtgrewe*sigh* I have a separate OOB network and need to add firewall rule exceptions13:31
yoctozeptomnasiadka: reviewed, needs quick fixes14:06
holtgreweWhen I'm configuring mdadm software arrays - how do I get this into provisioning? `kayobe seed service deploy` or do I have to rebuild the image?14:51
holtgreweAh, software raid needs to be done directly in bifrost, apparently.14:59
priteauholtgrewe: kayobe only supports mdadm at `host configure` time on non-root devices14:59
mnasiadka#startmeeting kolla15:00
opendevmeetMeeting started Wed Oct 13 15:00:13 2021 UTC and is due to finish in 60 minutes.  The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot.15:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:00
opendevmeetThe meeting name has been set to 'kolla'15:00
priteauIt may be possible to do software raid on root FS via bifrost15:00
mnasiadka#topic rollcall15:00
holtgrewe@priteau, thanks.15:00
mgoddard\o15:00
mnasiadkao/15:00
fricklero/15:01
priteauo/15:01
opendevreviewMichal Nasiadka proposed openstack/kolla-ansible master: Tidy up renos before Xena release  https://review.opendev.org/c/openstack/kolla-ansible/+/81378815:01
opendevreviewMichal Nasiadka proposed openstack/kolla-ansible master: Tidy up renos before Xena release  https://review.opendev.org/c/openstack/kolla-ansible/+/81378815:02
yoctozeptoo/15:02
alecorpshello15:03
mnasiadka#topic agenda15:04
mnasiadka* Announcements15:04
mnasiadka** Ussuri goes to Extended Maintenance in 1 month (2021-11-12) http://lists.openstack.org/pipermail/openstack-discuss/2021-October/025276.html15:04
mnasiadka* Review action items from the last meeting15:04
mnasiadka* CI status15:04
mnasiadka* Release tasks15:04
mnasiadka* Xena cycle planning15:04
mnasiadka* Yoga PTG planning15:04
mnasiadka* Open discussion15:04
mnasiadka#topic Announcements15:04
mnasiadkaUssuri goes to Extended Maintenance in 1 month (2021-11-12)15:04
mnasiadka#link http://lists.openstack.org/pipermail/openstack-discuss/2021-October/025276.html15:04
mnasiadkaSo it will join Rocky++ club15:05
mnasiadkaUsing the opportunity - do we want to EOL any branches?15:06
mnasiadkamgoddard, yoctozepto?15:07
mgoddardlast eol was queens15:07
mgoddardso rocky could be a candidate15:07
kevko+115:07
mnasiadkaYes, that sounds like a good approach.15:08
mnasiadka#topic Review action items from the last meeting15:09
mnasiadkamnasiadka to run through the release checklist15:09
mnasiadkain progress - kolla-ansible tidy renos patch being reviewed, then I'll propose rc1 for Kolla and Kolla-Ansible15:10
mnasiadkaand then move on with Kayobe as agreed15:10
mnasiadkamgoddard to update release docs to reflect Kayobe should be branched/cut RC1 after we have Kolla/Kolla-Ansible RC1s15:10
mnasiadkaThat's also done15:10
mnasiadka#topic CI status15:10
mnasiadkayoctozepto: how are we with those Galera failures?15:11
yoctozeptomnasiadka: no progress; frickler to hold a node to see if service restart helps15:11
mnasiadkaok then15:11
mnasiadkaany other CI issues we know of?15:11
yoctozeptoseemingly it works more often than not outside of CI15:12
yoctozeptobut can break still15:12
yoctozeptoI guess15:12
mnasiadkaas always15:12
yoctozeptoyup15:12
mnasiadkabut it's not a release blocker as I understand15:12
fricklerah, I have a set of 3 held nodes, but didn't get to look at them yet15:12
frickleryoctozepto: remind me of your ssh key, then I'll get you access15:13
mnasiadkaok, let's move on15:13
mnasiadka#topic Release tasks15:13
mnasiadkaThat's taken care of,15:14
mnasiadka#topic Xena cycle planning15:14
mnasiadkaThat's too late for that :-)15:14
mnasiadka#topic Yoga PTG planning15:14
mnasiadka#link https://etherpad.opendev.org/p/kolla-yoga-ptg15:14
mnasiadkaPlease post your topics, PTG is next week15:14
mgoddardlast chance to discuss before the PTG15:15
mgoddardhttps://etherpad.opendev.org/p/kolla-yoga-ptg15:16
yoctozeptofrickler: https://api.github.com/users/yoctozepto/keys15:16
alecorpshello, are we in open discussion topics ?15:17
mgoddardhi alecorps, not yet15:17
alecorpsok sorry15:17
mgoddardnp15:17
mgoddardare there any topics which need more details?15:17
mnasiadkaok then, I guess no pre-PTG discussion items ;-)15:17
mgoddardI should fill in more info about podman15:19
mgoddardalso, is there any reading we need to do before the PTG?15:20
mgoddarde.g. review the podman patch15:20
mnasiadkaAround multiple environments - can we have a summary what is there now, and what are next steps?15:20
mgoddardI've added myself a todo15:21
mgoddardHas anyone looked at the podman patch?15:22
yoctozeptoI did15:22
yoctozeptoI talked with ohorecny as well15:22
mnasiadkaI added also a todo to populate Xena PTG Actions15:22
yoctozeptobut not checked recently15:22
yoctozeptomy take is to make it work with the current ansible code and do the heavier lifting behind the scenes :-)15:23
mgoddardare there comments in the patch that explain what that means?15:23
mnasiadkaDo I see that correct it doesn't use systemd units for running container images?15:25
fricklerwow, that's some monster patch15:26
opendevreviewMerged openstack/kolla stable/ussuri: neutron: Fix update-alternatives on CentOS  https://review.opendev.org/c/openstack/kolla/+/81330515:26
yoctozeptomgoddard: yes, there were there15:26
mgoddardok, I need to read15:26
yoctozeptofrickler: yup, /me trying to make it simpler15:26
yoctozeptomgoddard: ++15:26
yoctozeptomnasiadka: did not check recently, might not be doing that yet15:26
yoctozeptoit would be good to review it before Tuesday then15:27
yoctozeptowe are targetting Tuesday, right?15:27
mnasiadkawell, we should analyze that patch - and maybe agree on the best approach forward (e.g. if we want to use systemd, other decisions) during the PTG15:27
mnasiadkaSo yes, it's the time to add more descriptions to each item this week15:29
mnasiadkaDo we have any additional topics around the PTG - or should we move to open discussion?15:30
yoctozeptoI don't; I will be mostly offering my knowledge and insight on the proposed topics this time, not having many of my own15:31
mnasiadkaI will probably find some that might be interesting but not picked up in previous cycles (as ,,backup'' to discuss if we'll feel it makes sense)15:33
kevkoadded some comments also :P 15:34
mgoddardkevko: will you attend the PTG?15:35
kevkomgoddard: probably yes 15:36
mgoddardkevko: good. Please add your name to the etherpad15:36
mgoddardkevko: it's a good place to discuss features that you're working on or would like to work on15:36
kevkomgoddard: but I am not sure, it depends on my internet connection ..i will be on vacation 15:36
mgoddardah15:36
mgoddardwell don't ruin your holiday for it ;)15:37
kevkomgoddard: but probably i will 15:37
mnasiadkakevko: ruin or attend? :)15:37
kevkomgoddard: well, i really need to move on with my patches ...I am still rebasing and keeping them live ..because we are using them from stein ...15:37
kevkomgoddard: especially logging, proxysql, mariadb-arbitrator 15:38
kevkoattend :D 15:38
kevko(and working very well :/)15:38
kevkoalready open discussion ? 15:38
mnasiadkamgoddard: https://review.opendev.org/c/openstack/kolla-ansible/+/813788 - and I'll post rc1 patches to releases repo15:38
mnasiadkaok then, let's move to open discussion15:39
mnasiadka#topic Open discussion15:39
mnasiadkaalecorps: you had something?15:39
alecorpsyes15:39
alecorpsI'm looking for two reviews for the two belows gerrit reviews :15:39
alecorpshttps://review.opendev.org/c/openstack/kolla-ansible/+/80740415:39
alecorpshttps://review.opendev.org/c/openstack/kolla-ansible/+/80876015:39
alecorpsmaybe you can help me ?15:39
mnasiadkaUnfortunately we're beyond the point we could review them and merge them in Xena15:41
yoctozeptobut we are known to backport small features15:42
yoctozeptojust need to wait until we stabilise15:42
alecorpsok15:42
alecorpsI just want to be sure the pr will not be abandonned15:43
kevkoI would like to ask why there is such big preasure to remove binary packages ..15:44
mnasiadkaalecorps: if you could add them on the whiteboard on the end of Xena priorities list - we'll remember to review that15:44
mnasiadkaalecorps: https://etherpad.opendev.org/p/KollaWhiteBoard15:44
alecorpsok I will15:44
alecorpsthanks15:44
mnasiadkakevko: it's not pressure, but an open discussion on the PTG15:45
kevkoi know there are problems with it ... but i think option is to drop all meta packages and just install source code (from a package)15:45
kevkomnasiadka: yeah, but it is discussion from version to version :) 15:45
mnasiadkayes, we've had that discussion last time - we're just a limited group with many things to maintain15:45
kevkoI am DD in debian, so if found some bug fixed in master ..i am patching it as fast as I can ..also CVEs and such .. when i have source i can't do it ..15:46
mnasiadkaand most people use source, because most openstack projects are not really good at posting minor releases15:47
kevkoi can for example use some git with my source ..but i can't handle requirements for example ..they are installed from pip15:47
kevkoif it's binary ..i just fix it with patch in debian .. (or in local apt repo for example ..)15:47
mnasiadkakevko: in Debian it might be simple for you, but there are are distributions - where it's really simple to do it source15:48
kevkowell, centos has something similar nope ? 15:48
kevkoi mean ..patches for packages 15:48
yoctozeptokevko: what you describe as needed is even simpler for me to do with source; we might perhaps want to focus on making kolla have better flows and guidelines for your needs15:49
yoctozeptobut otherwise source are *more* flexible15:49
yoctozeptonot less15:49
frickleris there a way to patch things that are installed from pypi? like the recent suds-jurko for example?15:50
mnasiadkakevko: most people that use binary rely on RDO/UCA/Debian OpenStack to deliver packages, and at least RDO and UCA post new OpenStack packages versions only when there's a minor release - which doesn't happen often for some projects (but bugs are fixed in stable/xena for example)15:50
mnasiadkabut yes, I agree with yoctozepto - we should gather stories/scenarios from users and help them with describing a workflow15:51
opendevreviewMerged openstack/kolla-ansible master: Add support for Ironic inspection through DHCP-relay  https://review.opendev.org/c/openstack/kolla-ansible/+/81326815:51
mnasiadkaand fixing what doesn't work + updating docs with how to handle different situations15:51
opendevreviewMerged openstack/kolla-ansible master: Tidy up renos before Xena release  https://review.opendev.org/c/openstack/kolla-ansible/+/81378815:51
kevkoyoctozepto: ok, consider that some service, binary, library, whatever has dependency to some library and its upper-constraints is X.X.X ... but X.X.X is buggy ..and fix is in Y.Y.Y ..but that service, binary, library is not compatible with Y.Y.Y because it is major change ...but buggy thing is only in one line for example ... how i can fullfill this with source ? 15:51
kevkoand library Y.Y.Y is non-openstack ..and author will not fix X.X.X+1 version 15:52
yoctozeptokevko: for that we need a special workflow in kolla and I would prefer to work on that so that all users benefit15:53
yoctozeptofrickler: I think kevko's issue matches your question ^15:53
frickleryes15:54
mgoddardyou could fork the code, and install from git15:54
kevkoin package you just take a diff from Y.Y.Y version (even if whole code is completly different..you just pick one line diff ..add to package ..rebuild image )15:54
yoctozeptomgoddard: might be tricky with arbitrary deps15:54
yoctozeptogood discussion topic nonetheless15:55
yoctozeptolet's add this for the PTG15:55
alecorps@mnasiadka: added in the whiteboard15:55
alecorpshope it's ok for you15:55
kevkomgoddard: yeah, and this is the way, add possibility to override any source library,app whatever15:55
mnasiadkaalecorps: thanks, will check and adapt if needed15:55
kevkomgoddard: but you have to implement whole process of deliver that forked git ..what about dependencies  ? .. c libraries, gcc and others if there will be some source which will need it ?  in my example this is solved on deb/centos builder machine 15:57
kevkoalso, is there some statistics how many users are using binary/source ? 15:58
kevkobecause I think most binary users are not downloading images from kolla officially ..they just building them .. as me 15:59
mnasiadkathe numbers on Docker Hub are similar, but that might also mean the only user is CI :)16:00
mnasiadkaok, let's finish the meeting16:00
yoctozeptolet's discuss during the ptg16:00
mnasiadka#endmeeting16:00
opendevmeetMeeting ended Wed Oct 13 16:00:46 2021 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)16:00
opendevmeetMinutes:        https://meetings.opendev.org/meetings/kolla/2021/kolla.2021-10-13-15.00.html16:00
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/kolla/2021/kolla.2021-10-13-15.00.txt16:00
opendevmeetLog:            https://meetings.opendev.org/meetings/kolla/2021/kolla.2021-10-13-15.00.log.html16:00
yoctozeptothanks mnasiadka16:00
* yoctozepto off16:00
kevkothank you guys 16:01
alecorpsthanks16:10
holtgreweHow do I set the boot device in kayobe without RAID?16:19
priteauholtgrewe: Do you mean select which disk is used for writing the root image?16:21
holtgreweYes, do I put `inspector_rule_root_hint_init` into group_vars/controllers/inspector, for example?16:21
*** amoralej is now known as amoralej|off16:22
holtgreweOr is this something that I set directly in bifrost/ironic via baremetal node set?16:28
priteauYou can do either. Inspector rule will only work on next inspection, while manual set can be done at any time of course16:29
holtgreweThanks. Sorry for the dumb questions, I am still building my mental model of all of this and kayobe -> kolla -> a bunch of openstack is somewhat challenging for me still16:31
holtgreweI now do `kayobe overcloud provision`. This makes the machine go it IPA and boot via DHCP. They get an IP hat is not the same as I have in `etc/kayobe/network-allocation.yml`. Thus the kayobe command fails because it cannot SSH into the machines. What am I doing wrong here?17:00
yoctozeptomnasiadka, mgoddard: could we start ptg with kolla topics instead of general? I will have to evacuate myself onto TC meeting in the 2nd half17:08
yoctozepto(on Monday)17:09
jingvarholtgrewe: there are 2 parts of pxe subnet17:48
jingvar1.  for ironic dhcp - unknow nodes 2. for  known nodes 17:50
jingvarbifrost-deploy)[root@kayobe bifrost-base-archive-stable-victoria]# less /etc/dnsmasq.conf  dhcp-range=10.228.0.128,10.228.0.254,12h17:59
jingvarnetwork-allocation have reocrds from first part of  10.228.0.0/2418:00
opendevreviewMaksim Malchuk proposed openstack/kolla-ansible stable/wallaby: Add support for Ironic inspection through DHCP-relay  https://review.opendev.org/c/openstack/kolla-ansible/+/81389020:03
opendevreviewMaksim Malchuk proposed openstack/kolla-ansible stable/victoria: Add support for Ironic inspection through DHCP-relay  https://review.opendev.org/c/openstack/kolla-ansible/+/81389120:04
opendevreviewMaksim Malchuk proposed openstack/kolla-ansible stable/ussuri: Add support for Ironic inspection through DHCP-relay  https://review.opendev.org/c/openstack/kolla-ansible/+/81389220:04
-opendevstatus- NOTICE: Both Gerrit and Zuul services are being restarted briefly for minor updates, and should return to service momentarily; all previously running builds will be reenqueued once Zuul is fully started again22:50

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!