| stevebaker[m] | ianw: hey what do you think of this suggestion? https://review.opendev.org/c/openstack/diskimage-builder/+/826976/3/diskimage_builder/elements/bootloader/finalise.d/50-bootloader#160 | 00:52 |
|---|---|---|
| ianw | stevebaker[m]: tbh i'm not sure that doesn't make a more complex path even more complex ... i know clarkb also had some trouble getting his head around it so i think we just need a path that if someone is reading top-to-bottom they can understand roughly what's going on where | 01:06 |
| ianw | if we have a bootloader-element, i feel like it could do what it needs to do in that if statement, then later just do | 01:06 |
| ianw | if [[bootloader-stevebaker]]; then # <insert comment here> exit 0; fi | 01:07 |
| ianw | for mine, that is clear enough, and people aren't like to break things by moving things around on it, etc. | 01:08 |
| ianw | i think i've tried about 3 times to clean the old grub paths in that file, that would help too | 01:08 |
| stevebaker[m] | ianw: so to clarify the problem, overcloud-full is a partition image, so it depends on the grub2 element. Due to bad centos9 grub config defaults (and a need for correct config in general), overcloud-full needs to have its grub config regenerated as for 50-bootloader#160 onwards. Can you describe what change could allow this? | 01:12 |
| ianw | what's the bit it *doesn't* need to do then? | 01:13 |
| stevebaker[m] | any kind of grub-install, because in a partition image there is no device to install to | 01:15 |
| ianw | is there a reason why we can't shuffle the grub-install after all the config-file editing we do? | 01:16 |
| ianw | that's probably what you're getting at with a separate element. but maybe we can just exit 0 before doing that? | 01:17 |
| stevebaker[m] | yeah, just move the grub-install to the end, exit out first if there is nothing to install to? | 01:18 |
| ianw | yep | 01:25 |
| ianw | if we could do it in three steps of | 01:25 |
| ianw | 1) removing the old grub paths | 01:25 |
| ianw | 2) moving it to the end | 01:25 |
| ianw | 3) adding on the early exit | 01:25 |
| ianw | that seems ideal and a good way to not introduce more than one problem in a change :) | 01:25 |
| stevebaker[m] | ok, I'll have a crack. Meanwhile I did this https://review.opendev.org/c/openstack/diskimage-builder/+/828617 | 01:31 |
| stevebaker[m] | ianw: oh, what is 1) the old grub paths? | 01:31 |
| ianw | i feel like there's still old grub1 paths in there? | 01:32 |
| ianw | https://review.opendev.org/c/openstack/diskimage-builder/+/790878 | 01:32 |
| ianw | type thing | 01:32 |
| stevebaker[m] | ianw: that change looks good as it is, can you rebase it so I can review it? | 01:37 |
| ianw | yeah i think things just kept moving around on it. i'll rebase it today | 01:40 |
| stevebaker[m] | ok thanks | 01:43 |
| opendevreview | Steve Baker proposed openstack/diskimage-builder master: Replace kpartx with qemu-nbd in extract-image https://review.opendev.org/c/openstack/diskimage-builder/+/828617 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Futher bootloader cleanups https://review.opendev.org/c/openstack/diskimage-builder/+/790878 | 03:09 |
| ianw | stevebaker[m] : i *think* i merged that right, but wants a careful eye | 03:10 |
| * stevebaker[m] looks | 03:10 | |
| stevebaker[m] | ianw: just one bit missing I think | 03:15 |
| ianw | stevebaker[m]: hrm, does anything use /boot/grub (! /boot/grub2) | 03:21 |
| stevebaker[m] | I have assumed some grub2 distros use /boot/grub as their path, but I'm not sure which right now. One of the quirks of everyone having their own fork | 03:22 |
| stevebaker[m] | I can follow up on that tomorrow, I'm about to disappear | 03:23 |
| ianw | stevebaker[m]: np. it came in with -> https://review.opendev.org/c/openstack/diskimage-builder/+/814866 | 03:23 |
| ianw | prometheanfire: ^ is /boot/grub still true? | 03:23 |
| ianw | i'll put it back for now | 03:23 |
| prometheanfire | iirc, yes, let me see... | 03:26 |
| prometheanfire | https://wiki.gentoo.org/wiki/GRUB2#Main_configuration_file has some paths, including /boot/grub | 03:27 |
| ianw | ok, thanks | 03:35 |
| stevebaker[m] | Something to follow up later, but I don't think bootloader ppc64 is tested and ppc IBM jobs broke on centos-8 functest removal | 03:52 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Futher bootloader cleanups https://review.opendev.org/c/openstack/diskimage-builder/+/790878 | 04:09 |
| ianw | yeah we spent quite a while with tonyb working on that, but i think it's time has passed | 04:14 |
| ianw | 2022-02-10 04:16:17.921800 | centos-8 | "msg": "The centos-8 node type is no longer supported. You need to switch this job to use a centos-8-stream node type." | 04:16 |
| ianw | that's mmore like it | 04:16 |
| rpittau | stevebaker[m], TheJulia, ianw, please have a look at https://review.opendev.org/c/openstack/diskimage-builder/+/828266 when you got a minute, that will fix dhcp-all-interfaces for systemd with netifs predictable names scheme only, thanks! | 10:32 |
| *** priteau_ is now known as priteau | 16:11 | |
| *** priteau is now known as priteau_ | 16:48 | |
| *** priteau_ is now known as priteau | 16:48 | |
| opendevreview | Merged openstack/diskimage-builder master: Futher bootloader cleanups https://review.opendev.org/c/openstack/diskimage-builder/+/790878 | 21:02 |
| *** prometheanfire is now known as Guest2 | 23:49 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!