| *** penick has quit IRC | 00:08 | |
| *** diurnalist has joined #openstack-ironic | 00:13 | |
| *** penick has joined #openstack-ironic | 00:14 | |
| *** rloo has joined #openstack-ironic | 00:17 | |
| *** rloo has quit IRC | 00:17 | |
| *** penick has quit IRC | 00:30 | |
| *** diurnalist has quit IRC | 01:03 | |
| *** ricolin has joined #openstack-ironic | 01:15 | |
| *** penick has joined #openstack-ironic | 01:25 | |
| *** penick has quit IRC | 01:33 | |
| *** zzzeek has quit IRC | 01:54 | |
| *** yedongcan has joined #openstack-ironic | 01:56 | |
| *** zzzeek has joined #openstack-ironic | 01:57 | |
| *** rcernin has quit IRC | 02:26 | |
| *** rcernin has joined #openstack-ironic | 02:35 | |
| *** mkrai has joined #openstack-ironic | 02:56 | |
| *** tzumainn has quit IRC | 02:58 | |
| *** rcernin has quit IRC | 03:04 | |
| *** rh-jlabarre has quit IRC | 03:15 | |
| *** rcernin has joined #openstack-ironic | 03:15 | |
| *** Lucas_Gray has quit IRC | 03:19 | |
| *** rcernin has quit IRC | 03:19 | |
| *** rcernin has joined #openstack-ironic | 03:19 | |
| *** diurnalist has joined #openstack-ironic | 03:50 | |
| *** mkrai has quit IRC | 03:52 | |
| *** mkrai has joined #openstack-ironic | 03:53 | |
| *** rcernin has quit IRC | 03:55 | |
| *** rcernin has joined #openstack-ironic | 03:56 | |
| *** yedongcan has quit IRC | 04:09 | |
| *** mkrai has quit IRC | 04:32 | |
| *** mkrai has joined #openstack-ironic | 04:33 | |
| openstackgerrit | Richard G. Pioso proposed openstack/ironic master: Correct Redfish boot once fallback conditional https://review.opendev.org/737892 | 04:34 |
|---|---|---|
| *** rcernin has quit IRC | 04:45 | |
| *** rcernin has joined #openstack-ironic | 04:46 | |
| *** akahat|pto is now known as akahat | 05:10 | |
| *** diurnalist has quit IRC | 05:32 | |
| *** yedongcan has joined #openstack-ironic | 05:43 | |
| arne_wiebalck | Good morning, ironic! | 05:59 |
| *** mnasiadka_ is now known as mnasiadka | 06:42 | |
| openstackgerrit | Verification of a change to openstack/ironic failed: DRAC: Fix a failure to create virtual disk https://review.opendev.org/730725 | 06:59 |
| ftarasenko | Good morning! | 07:00 |
| ftarasenko | arne_wiebalck: Arne, still having problems with software raid. Maybe, you've seen the same? Grub2 install fails with "Installing for i386-pc platform.\ngrub2-install: error: install device isn't specified.\n" | 07:01 |
| ftarasenko | Full log is here https://pastebin.com/qiWF5EFe | 07:01 |
| rpittau | oh Great Cthulhu! The UEFI jobs work with ubuntu focal and the native ovmf package! https://review.opendev.org/716889 | 07:07 |
| patchbot | patch 716889 - ironic - Fix uefi jobs with native ubuntu ovmf package - 49 patch sets | 07:07 |
| rpittau | good morning ironic! Happy Friday! o/ | 07:07 |
| arne_wiebalck | ftarasenko: Hi! I am in a meeting now, will have a look a little later. | 07:12 |
| iurygregory | good morning arne_wiebalck rpittau and Ironic o/ | 07:13 |
| rpittau | hey iurygregory arne_wiebalck :) | 07:13 |
| *** dougsz has joined #openstack-ironic | 07:16 | |
| *** mkrai has quit IRC | 07:16 | |
| *** rcernin has quit IRC | 07:24 | |
| openstackgerrit | Riccardo Pittau proposed openstack/sushy-tools master: Use victoria tests https://review.opendev.org/739173 | 07:34 |
| iurygregory | i was working on the py38 fix .-. but rpittau is faster XD | 07:39 |
| rpittau | I was changing to victoria, ran the py38 tests, and had a deja-vu | 07:42 |
| rpittau | then I realized it was a new error! | 07:42 |
| openstackgerrit | vinay kumar muddu proposed openstack/ironic master: [WIP] Isolate the ISO creation logic from redfish https://review.opendev.org/739174 | 07:44 |
| rpittau | ftarasenko: that's a limitation of ironic-python-agent as you can see here: https://opendev.org/openstack/ironic-python-agent/src/branch/master/ironic_python_agent/extensions/image.py#L561 | 07:48 |
| rpittau | probably something we should fix :) | 07:48 |
| ftarasenko | rpittau: my mistake, I think, already checking that. I have server with DUAL boot mode by default, but need UEFI for both ramdisk and target. Thank you | 07:50 |
| rpittau | ah ok! cool :) | 07:50 |
| rpittau | if anyone has a minute we should really merge this https://review.opendev.org/738438 | 07:51 |
| patchbot | patch 738438 - ironic-python-agent-builder - Disable EPEL on CentOS 8 by default - 2 patch sets | 07:51 |
| iurygregory | rpittau, no worries =) | 07:54 |
| iurygregory | rpittau, looking now | 07:54 |
| *** ociuhandu has joined #openstack-ironic | 08:02 | |
| *** lucasagomes has joined #openstack-ironic | 08:09 | |
| *** dtantsur|afk is now known as dtantsur | 08:11 | |
| dtantsur | morning ironic | 08:11 |
| *** priteau has joined #openstack-ironic | 08:14 | |
| arne_wiebalck | ftarasenko: no, I have not seen this yet | 08:14 |
| arne_wiebalck | ftarasenko: but to be fair, we do not have UEFI with RAID at scale yet | 08:15 |
| arne_wiebalck | ftarasenko: I only tested if it works, we will have more in the coming 8 weeks or so | 08:15 |
| arne_wiebalck | ftarasenko: from now on, all nodes will be (or are planned to be) UEFI | 08:15 |
| ftarasenko | arne_wiebalck: it looks like problem with dual boot, already fixed, moving on. Be back in hour or two if it works or not | 08:16 |
| ftarasenko | Thank you! | 08:16 |
| arne_wiebalck | ftarasenko: dual boot | 08:16 |
| arne_wiebalck | ? | 08:16 |
| arne_wiebalck | ftarasenko: oh, the IPA is booted in legacy mode while the image will require UEFI? | 08:17 |
| ftarasenko | arne_wiebalck: dual boot means server, capable of booting to BIOS and UEFI simultaneously | 08:19 |
| ftarasenko | arne_wiebalck: I've booted IPA in UEFI and have target image set to UEFI, but for some reason it failed. I forced UEFI in server settings and got another error (missing package) which I'm trying to solve right now | 08:21 |
| arne_wiebalck | ftarasenko: server settings means in the BIOS? | 08:22 |
| ftarasenko | arne_wiebalck: yes. BIOS boot mode in Supermicro has 3 options - bios, uefi and dual. | 08:22 |
| arne_wiebalck | ftarasenko: do you know what dual means? it tries both? in some order? | 08:23 |
| rpittau | ftarasenko: out of curiosity, which package? | 08:23 |
| *** mkrai has joined #openstack-ironic | 08:25 | |
| ftarasenko | arne_wiebalck: it tries both, not sure which order. on support sites I've seen that Windows could not be installed in DUAL mode, need to force UEFI, so it's not just simple fallback | 08:25 |
| ftarasenko | rpittau: efibootmgr, missed it for unknown reason( | 08:25 |
| arne_wiebalck | ftarasenko: ok, thanks! | 08:26 |
| rpittau | thanks :) | 08:26 |
| arne_wiebalck | ftarasenko: yes, we needed to add this as well | 08:26 |
| arne_wiebalck | ftarasenko: efibootmgr | 08:26 |
| arne_wiebalck | ftarasenko: but I think this was only required to install the legacy boot loader when booted in UEFI mode | 08:27 |
| arne_wiebalck | ftarasenko: no, scrap this | 08:27 |
| ftarasenko | arne_wiebalck: I've thought that I've added it yesterday, but it was on another project | 08:28 |
| arne_wiebalck | ftarasenko: it was required to install the legacy bootloader when booting a UEFI capable image in legacy mode | 08:28 |
| rpittau | ftarasenko: you're building your own image, right? Did you check ipa-builder? | 08:29 |
| arne_wiebalck | ftarasenko: anyway, good luck and let us know how it goes, this is all still newish territory | 08:29 |
| ftarasenko | arne_wiebalck: we have efibootmg calls from here to install bootloader. is it only for legacy mode? https://opendev.org/openstack/ironic-python-agent/src/branch/master/ironic_python_agent/extensions/image.py#L707 | 08:30 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Limit Inspection->Lookup->Heartbeat lag https://review.opendev.org/739120 | 08:31 |
| dtantsur | TheJulia: some enhancements on top of ^^ I hope you don't mind | 08:31 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Switch the inspector job to tinyIPA until DIB is fixed https://review.opendev.org/739178 | 08:32 |
| ftarasenko | rpittau: this is required for target image, I get error form chroot | 08:32 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Limit Inspection->Lookup->Heartbeat lag https://review.opendev.org/739120 | 08:32 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Revert "Switch the inspector job to tinyIPA until DIB is fixed" https://review.opendev.org/739180 | 08:33 |
| *** rachit7 has joined #openstack-ironic | 08:33 | |
| ftarasenko | arne_wiebalck: rpittau: looks like grub2-install calls for efibootmgr http://paste.openstack.org/show/795520/ | 08:33 |
| rpittau | ftarasenko: not sure I get it, you get error from chroot in ipa-builder? | 08:33 |
| rpittau | ah ok | 08:33 |
| rpittau | nvm | 08:33 |
| rpittau | yeah you need efibootmgr, it's installed in the centos8 image provided by us from ipa-builder | 08:34 |
| arne_wiebalck | ftarasenko: that is possible ... it was maybe installed to make the image s/w RAID and UEFI capable | 08:35 |
| arne_wiebalck | ftarasenko: I have seen this error when we were building our image | 08:35 |
| ftarasenko | rpittau: but ipa-builder is only for provision, and I get this error while deploy chroots to target image. So target should also have this package inside it | 08:35 |
| arne_wiebalck | ftarasenko: right, this should be in the user image | 08:36 |
| rpittau | ftarasenko: sorry, my bad, I didn't read it was for the target image | 08:36 |
| dtantsur | mgoddard: morning! I'm going to request a sushy-tools release if you don't mind | 08:37 |
| mgoddard | dtantsur: sure, go for it | 08:38 |
| *** k_mouza has joined #openstack-ironic | 08:39 | |
| *** rachit7 has quit IRC | 08:40 | |
| openstackgerrit | Merged openstack/ironic-specs master: Amend the in-band deploy steps spec https://review.opendev.org/738991 | 08:43 |
| *** Lucas_Gray has joined #openstack-ironic | 08:52 | |
| *** derekh has joined #openstack-ironic | 08:53 | |
| dtantsur | TheJulia: now another thing is bothering me: now that we start heartbeating pretty quickly, there is a risk that a lock is still held on the node | 08:56 |
| iurygregory | good morning dtantsur o/ | 08:57 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent stable/ussuri: Increase the ESP partition size to 550 MiB when using software RAID https://review.opendev.org/739184 | 08:57 |
| *** Lucas_Gray has quit IRC | 09:05 | |
| *** Lucas_Gray has joined #openstack-ironic | 09:09 | |
| *** yedongcan has quit IRC | 09:13 | |
| *** sshnaidm|afk is now known as sshnaidm|off | 09:15 | |
| *** mkrai has quit IRC | 09:15 | |
| *** mkrai has joined #openstack-ironic | 09:16 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/networking-generic-switch master: [DNM] Adding more debug to ngs to test grenade https://review.opendev.org/739186 | 09:17 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic master: Native zuulv3 grenade multinode multitenant https://review.opendev.org/705030 | 09:19 |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: Disable EPEL on CentOS 8 by default https://review.opendev.org/738438 | 09:20 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Log when a node should be fast-track-able but it's not https://review.opendev.org/739189 | 09:22 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic-python-agent master: Revert jobs to dib https://review.opendev.org/738442 | 09:28 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic-python-agent-builder master: Build tinyipa on focal https://review.opendev.org/725799 | 09:28 |
| *** ricolin has quit IRC | 09:33 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up nodes in DELETING on conductor restart https://review.opendev.org/739192 | 09:42 |
| *** tkajinam has quit IRC | 09:46 | |
| *** k_mouza has quit IRC | 09:47 | |
| *** rcernin has joined #openstack-ironic | 09:51 | |
| *** rcernin has quit IRC | 09:51 | |
| *** k_mouza has joined #openstack-ironic | 09:58 | |
| ftarasenko | arne_wiebalck: dtantsur: Thank you so much for your support, I've managed to boot server with UEFI & Software RAID. One more question, how target sets boot order for UEFI? I still have pxe boot first after provision. | 10:36 |
| dtantsur | ftarasenko: usually using efibootmgr (unless arne_wiebalck can correct me) | 10:37 |
| dtantsur | we also set the desired boot device out-of-band (via idrac in your case) | 10:38 |
| ftarasenko | but what additional settings do I need to set in Ironic? I have only 'boot_option': 'local' in my node's settings | 10:38 |
| arne_wiebalck | dtantsur: I think we do not use efibootmgr directly with software RAID | 10:45 |
| *** sshnaidm_ has joined #openstack-ironic | 10:47 | |
| arne_wiebalck | and from what I remember the RAID code does not touch the order, not sure how this is handled tbh | 10:48 |
| *** sshnaidm|off has quit IRC | 10:48 | |
| iurygregory | we dont use efibootmgr with SW Raid | 10:48 |
| iurygregory | =) | 10:48 |
| iurygregory | you wanted grub2 afaik | 10:49 |
| arne_wiebalck | since grub2 was tested and working :) | 10:50 |
| arne_wiebalck | but how is the boot order controlled? | 10:50 |
| iurygregory | the answer will always be *magic* | 10:50 |
| iurygregory | =P | 10:51 |
| iurygregory | sorry it's friday | 10:51 |
| arne_wiebalck | (I can't tell from our deployment since we *always* PXE boot) | 10:51 |
| * arne_wiebalck breaks for lunch | 10:52 | |
| dtantsur | arne_wiebalck, iurygregory, why exactly don't we use efibootmgr with sw RAID? | 10:59 |
| iurygregory | dtantsur, well https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/extensions/image.py#L735 | 11:14 |
| iurygregory | because of this | 11:14 |
| dtantsur | I guess we need to fix the FIXME two lines above :) | 11:15 |
| dtantsur | but I'm curious why we did that | 11:15 |
| iurygregory | we had a discussion about that | 11:15 |
| iurygregory | I think we added that on arne_wiebalck patch to fix SW Raid | 11:16 |
| iurygregory | dtantsur, https://review.opendev.org/#/c/696156/ | 11:16 |
| patchbot | patch 696156 - ironic-python-agent - Software RAID: Add UEFI support (MERGED) - 11 patch sets | 11:16 |
| iurygregory | maybe we need to specify something that would allow grub2 or not | 11:18 |
| *** priteau has quit IRC | 11:19 | |
| dtantsur | we can start with efibootmgr, then try grub? | 11:19 |
| dtantsur | although it doesn't explain why the boot device is not changed in idrac | 11:19 |
| iurygregory | that would make sense to me, arne_wiebalck wdyt? | 11:19 |
| iurygregory | if it fails with efibootmgr we would go for grub2, but I do think that arne_wiebalck wants to keep grub2 first for their case (?)... | 11:21 |
| arne_wiebalck | no, from what I remember, the UEFI s/w RAID patch had been lingering for a while, testing is not easy, and the patch set with grub2 was tested and working | 11:22 |
| arne_wiebalck | so, I simply did not want to redo things at the time | 11:23 |
| arne_wiebalck | we were also running out of time before ussuri | 11:23 |
| iurygregory | ohh | 11:24 |
| arne_wiebalck | so we agreed to -- once all is confirmed to be working -- to take a step back and restructure the code | 11:24 |
| iurygregory | (we probably forgot to add the FIXME XD) | 11:24 |
| arne_wiebalck | since it has heavily grown in the past 12 months | 11:25 |
| arne_wiebalck | and there are lots of turns now | 11:25 |
| *** hjensas has quit IRC | 11:25 | |
| arne_wiebalck | s/grub/efibootmgr/ is only a part of this clean up I think | 11:25 |
| *** mkrai has quit IRC | 11:26 | |
| iurygregory | so it would be fine to just go with efibootmgr (and maybe add a fallback to try grub2 if it fails | 11:26 |
| arne_wiebalck | I think your change to rely on efobootmgr came after the UEFI/grub patch, but got merged first; then we decided to basically have an "if" to have the (working) UEFI patch be merged as well | 11:28 |
| arne_wiebalck | efibootmgr | 11:28 |
| *** diurnalist has joined #openstack-ironic | 11:29 | |
| arne_wiebalck | I don't think there is a technical reason why this should not work with efibootmgr, the main issue here is testing to make sure it does ... | 11:29 |
| iurygregory | yeah | 11:29 |
| arne_wiebalck | as the current code seems to be working (we have at least 3 or 4 deployments which tried this by now) | 11:30 |
| * arne_wiebalck sounds like an operator :-D | 11:31 | |
| iurygregory | arne_wiebalck, you are | 11:31 |
| iurygregory | even if you only do things for testing XD | 11:31 |
| arne_wiebalck | iurygregory: true, but I meant in the "dear devs, don't put my deployment at risk!" sense | 11:32 |
| iurygregory | arne_wiebalck, gotcha | 11:32 |
| ftarasenko | So, do I correctly understand that I can remove one line and efibootmgr will set boot order. Or put me an error, of course. And the only reason for this check is lack of testing? https://github.com/openstack/ironic-python-agent/blob/ba3caa6c644460a76518e6a0b1977b03516bc323/ironic_python_agent/extensions/image.py#L735 | 11:33 |
| arne_wiebalck | we should certainly move fwd and clean this up, my point is that this requires resources to test this carefully | 11:33 |
| *** diurnalist has quit IRC | 11:33 | |
| arne_wiebalck | ftarasenko: I am not sure this will be enough, since the bootloader will not be installed if Ironic finds a bootloader | 11:34 |
| arne_wiebalck | ftarasenko: and we need to reinstall the bootlader for s/w RAID | 11:34 |
| iurygregory | arne_wiebalck, correct and +1 | 11:35 |
| arne_wiebalck | ftarasenko: how about you give it a try (if you have time), and we take it from there? iurygregory is the one who implemented most of the new logic around efibootmgr | 11:37 |
| ftarasenko | arne_wiebalck: sure, I'll try today and be back with the results. | 11:37 |
| iurygregory | I think he would probably need to change the ipa code to remove the md_device check | 11:37 |
| dtantsur | iurygregory: we can do it a bit more sophisticated and try both | 11:39 |
| iurygregory | dtantsur, yeah | 11:39 |
| iurygregory | the idea of "if it fails" we try grub2 right? | 11:40 |
| arne_wiebalck | we will also need to install on all holder disks | 11:42 |
| arne_wiebalck | in _manage_uefi I guess | 11:42 |
| arne_wiebalck | basically, the logic of _install_grub2 needs to be translated | 11:43 |
| iurygregory | for the case of SW | 11:43 |
| arne_wiebalck | yes | 11:43 |
| iurygregory | so we move there adding the check md | 11:43 |
| *** uzumaki has joined #openstack-ironic | 11:57 | |
| *** derekh has quit IRC | 12:02 | |
| uzumaki | Hello ironic! o/ | 12:09 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: [WIP] Software RAID: don't try to set local_gb to MAX https://review.opendev.org/739216 | 12:09 |
| dtantsur | arne_wiebalck: any ideas how to handle ^^^ better? | 12:09 |
| dtantsur | hi uzumaki | 12:09 |
| uzumaki | how's it going dtantsur ? o/ | 12:10 |
| dtantsur | it's Friday :) slowly approaching the weekend | 12:10 |
| uzumaki | Ah! I see, for the ones in the states, it might be the 4th of july holiday today, I assume? | 12:10 |
| dtantsur | uzumaki: yep, I assume we won't see them | 12:11 |
| uzumaki | I see.. I wonder where iurygregory, rpittau and arne_wiebalck are.. | 12:11 |
| dtantsur | all here | 12:12 |
| rpittau | physically? :) | 12:12 |
| dtantsur | ah, physically? | 12:12 |
| dtantsur | Czechia, France and Switzerland accordingly | 12:12 |
| dtantsur | (unless rpittau finally got to Italy) | 12:13 |
| uzumaki | dtantsur, that was my point, yes, they don't get to disappear | 12:13 |
| rpittau | I guess arne_wiebalck is literally an 'edge' case :D | 12:13 |
| uzumaki | rpittau, xD | 12:13 |
| rpittau | heh unfortunately not before Septemver :/ | 12:13 |
| uzumaki | well, I do hope things get better, Italy's your hometown, yes? rpittau | 12:14 |
| rpittau | yeah, I'm originally from Sardegna | 12:14 |
| uzumaki | rpittau, cool! | 12:14 |
| rpittau | oh look https://review.opendev.org/725799 is green again :P | 12:15 |
| patchbot | patch 725799 - ironic-python-agent-builder - Build tinyipa on focal - 11 patch sets | 12:15 |
| uzumaki | great! I wonder why the tempest is in such a tempest on the CI | 12:16 |
| dtantsur | heh | 12:16 |
| rpittau | never had more appropriate name chosen... | 12:16 |
| uzumaki | ikr? It's the irony that kills me xD | 12:17 |
| rpittau | :) | 12:17 |
| uzumaki | so, any special weekend plans? rpittau dtantsur iurygregory arne_wiebalck | 12:17 |
| dtantsur | nothing special, I only want to sleep | 12:18 |
| dtantsur | maybe some forest walk if the weather allows | 12:18 |
| uzumaki | dtantsur, forest walks sounds sooooooo aesthetic <3 | 12:19 |
| dtantsur | and relaxing! | 12:19 |
| uzumaki | dtantsur, sleep? Oh, I forget humans have to sleep | 12:19 |
| dtantsur | yeeah, a stupid limitation | 12:19 |
| uzumaki | yes, often when I start making stupid mistakes on the keyboard, reminds me I haven't slept in ages, drop on the bed right away xD | 12:19 |
| * arne_wiebalck is an French/Swiss edge case | 12:19 | |
| uzumaki | oui arne_wiebalck, nous sommes d'accord | 12:20 |
| rpittau | dtantsur: not sure you saw this :) https://review.opendev.org/716889 | 12:21 |
| patchbot | patch 716889 - ironic - Fix uefi jobs with native ubuntu ovmf package - 49 patch sets | 12:21 |
| rpittau | I mean, not sure you saw it's actually passing :D | 12:21 |
| dtantsur | OH | 12:23 |
| uzumaki | 49 patchets? gosh, what do you have in there? the new world order? xD rpittau | 12:23 |
| rpittau | uzumaki: that's nothing, that's an issue going on since 1 year and a half | 12:24 |
| rpittau | Multiple patches, multiple bugs | 12:25 |
| uzumaki | my my! | 12:25 |
| iurygregory | hey uzumaki o/ | 12:25 |
| dtantsur | rpittau: a few questions inline | 12:25 |
| * iurygregory was in a meeting | 12:25 | |
| uzumaki | hola iurygregory o/ | 12:25 |
| dtantsur | uzumaki: I've seen a patch going to 100+ patchsets | 12:25 |
| dtantsur | and then gerrit became so slow on it, that it was re-uploaded and had a few dozens more | 12:25 |
| uzumaki | of course you have dtantsur, the evil designer of the new world order xD | 12:26 |
| iurygregory | uzumaki, about special plans I don't have any for the weekend (but I have for monday since it's holiday)XD | 12:26 |
| dtantsur | hey, it wasn't MY patch! :D | 12:26 |
| uzumaki | iurygregory, oh? what's on moday? | 12:26 |
| iurygregory | uzumaki, karaoke =) | 12:26 |
| uzumaki | dtantsur, xD I guess I'll never know | 12:26 |
| iurygregory | dtantsur, 100+ sounds like what will take for the grenade multinode.... | 12:26 |
| uzumaki | iurygregory, you have a national karaoke holiday in cz? I'm coming right over.. | 12:26 |
| iurygregory | uzumaki, noooo (we have holiday on monday Saints Cyril and Methodius) | 12:27 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Software RAID: don't try to set local_gb to MAX https://review.opendev.org/739216 | 12:27 |
| arne_wiebalck | dtantsur: hmm, no ... getting the actual size back is probably tricky ... the scheduling breaks with 'MAX'? | 12:27 |
| uzumaki | iurygregory, I see, well, either way it's a long weekend, so have a good one! | 12:27 |
| dtantsur | arne_wiebalck: you just cannot set that, it does not pass validation | 12:27 |
| iurygregory | and in the past I had to go every monday to a karaoke pub | 12:27 |
| dtantsur | my patch above at least prevents this value from being set | 12:27 |
| iurygregory | and I will go on monday XD | 12:27 |
| arne_wiebalck | dtantsur: which validation? | 12:29 |
| dtantsur | arne_wiebalck: InvalidParameterValue: The following properties for node b6c3528e-f838-447a-ba11-7f77a87a05d1 should be non-negative integers, but provided values are: local_gb=MAX | 12:31 |
| dtantsur | coming from https://opendev.org/openstack/ironic/src/branch/master/ironic/objects/node.py#L182 | 12:32 |
| arne_wiebalck | dtantsur: this is validated when being set in the post creation hook? | 12:32 |
| dtantsur | yep | 12:32 |
| arne_wiebalck | is this new? | 12:32 |
| dtantsur | arne_wiebalck: nope, it has been there forever | 12:33 |
| arne_wiebalck | b/c we allow for MAX for RAID since the beginning, no? | 12:33 |
| * arne_wiebalck is confused | 12:33 | |
| dtantsur | for RAID - yes, but not in properties['local_gb'] | 12:33 |
| arne_wiebalck | right, but does that mean all RAIDs created with MAX raised this validation error? | 12:34 |
| dtantsur | do you set is_root_volume when building RAID? | 12:34 |
| dtantsur | setting local_gb is only triggered when is_root_volume is set | 12:34 |
| openstackgerrit | Merged openstack/ironic-python-agent stable/ussuri: Increase the ESP partition size to 550 MiB when using software RAID https://review.opendev.org/739184 | 12:34 |
| arne_wiebalck | in the target_raid_config you mean? | 12:35 |
| dtantsur | yup | 12:35 |
| arne_wiebalck | nope | 12:35 |
| arne_wiebalck | :) | 12:35 |
| * arne_wiebalck is unconfused | 12:36 | |
| arne_wiebalck | I guess there is no easy way to set this to some sane value ... but this is true for hardware RAID as well, no? | 12:37 |
| dtantsur | yeah | 12:38 |
| dtantsur | I don't remember if hw RAID supports MAX though | 12:38 |
| arne_wiebalck | I think it does. | 12:39 |
| arne_wiebalck | I don't remember to have change the initial validation at least and the docs have an example. | 12:39 |
| arne_wiebalck | *changed | 12:40 |
| *** mkrai has joined #openstack-ironic | 12:45 | |
| *** mkrai has quit IRC | 13:11 | |
| *** hjensas has joined #openstack-ironic | 13:13 | |
| *** uzumaki has quit IRC | 13:18 | |
| TheJulia | dtantsur: I was thinking about that as well, it is definitely a risk. Then again it is fine if ipa heartbeats fail and are not processed. And conductor side ops should have a lock on them the entire time... So I kind of suspect it maybe something we would just have to explore to see. The only real fix in that case is separate locking and callback url storage from the node object, yet that also seems like over | 13:24 |
| TheJulia | optimization of the interaction for additional db transactions for every single operation. | 13:24 |
| iurygregory | good morning TheJulia =) | 13:25 |
| dtantsur | TheJulia: morning! missed heartbeats are fine, but in case of fast-track it would cause a reboot | 13:25 |
| dtantsur | because the next heartbeat will likely come too late | 13:25 |
| TheJulia | dtantsur: I think one hardware raid driver does, another doesn't if memory serves | 13:25 |
| TheJulia | dtantsur: the conductor side window would need to be sufficient to account for that by being larger than the heartbeat interval. | 13:27 |
| dtantsur | there is not window on the conductor side, it boils down to how quickly cleaning/deploy is requested after inspection | 13:27 |
| dtantsur | it's around 30 seconds for metal3 | 13:27 |
| TheJulia | We must be thinking different things because I don't understand why your saying that unless the time window was stripped from the conductor. | 13:28 |
| TheJulia | Let's talk next week about it | 13:29 |
| TheJulia | Anyway, it is a holiday here. I should... Well... I don't know what. But it will likely involve kerbals | 13:29 |
| rpittau | kerbals... I miss it :/ | 13:31 |
| TheJulia | 1.10 has been released.... | 13:31 |
| TheJulia | Err, although my laptop apparently never suspended last nighj | 13:32 |
| dtantsur | TheJulia: enjoy! | 13:53 |
| openstackgerrit | wu.chunyang proposed openstack/python-ironicclient master: Remove pypy https://review.opendev.org/739235 | 13:55 |
| openstackgerrit | Merged openstack/ironic master: DRAC: Fix a failure to create virtual disk https://review.opendev.org/730725 | 14:10 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Follow up of fix uefi jobs with ovmf native ubuntu package https://review.opendev.org/739244 | 14:14 |
| openstackgerrit | wu.chunyang proposed openstack/python-ironicclient master: Remove pypy https://review.opendev.org/739235 | 14:28 |
| *** hjensas has quit IRC | 14:47 | |
| *** diurnalist has joined #openstack-ironic | 14:49 | |
| *** Goneri has joined #openstack-ironic | 14:54 | |
| *** diurnalist has quit IRC | 15:04 | |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic-python-agent master: Refactor part of image module https://review.opendev.org/739251 | 15:05 |
| *** markmcclain has quit IRC | 15:06 | |
| *** markmcclain has joined #openstack-ironic | 15:08 | |
| *** sshnaidm_ is now known as sshnaidm|off | 15:11 | |
| *** dtantsur is now known as dtantsur|afk | 15:20 | |
| dtantsur|afk | have a great weekend | 15:20 |
| *** bdodd_ has joined #openstack-ironic | 15:22 | |
| *** bdodd has quit IRC | 15:25 | |
| *** dhellmann has quit IRC | 15:25 | |
| *** dhellmann has joined #openstack-ironic | 15:26 | |
| *** uzumaki has joined #openstack-ironic | 15:35 | |
| *** hjensas has joined #openstack-ironic | 15:46 | |
| *** lucasagomes has quit IRC | 16:00 | |
| *** k_mouza has quit IRC | 16:02 | |
| *** ociuhandu_ has joined #openstack-ironic | 16:13 | |
| *** ociuhandu has quit IRC | 16:16 | |
| *** ociuhandu_ has quit IRC | 16:17 | |
| openstackgerrit | Merged openstack/ironic master: Fix uefi jobs with native ubuntu ovmf package https://review.opendev.org/716889 | 16:26 |
| uzumaki | and.... it's a merge! \o// | 16:28 |
| uzumaki | <confetti> <horns> | 16:28 |
| *** Lucas_Gray has quit IRC | 16:30 | |
| *** dougsz has quit IRC | 16:32 | |
| rpittau | <pizza> <beer> | 16:59 |
| *** ociuhandu has joined #openstack-ironic | 18:09 | |
| *** diurnalist has joined #openstack-ironic | 18:29 | |
| *** diurnalist has quit IRC | 18:46 | |
| *** penick has joined #openstack-ironic | 19:06 | |
| *** penick has quit IRC | 19:24 | |
| openstackgerrit | Verification of a change to openstack/ironic-python-agent-builder failed: Build tinyipa on focal https://review.opendev.org/725799 | 19:58 |
| *** ociuhandu has quit IRC | 20:29 | |
| *** ociuhandu has joined #openstack-ironic | 20:29 | |
| *** ociuhandu has quit IRC | 20:34 | |
| *** ociuhandu has joined #openstack-ironic | 20:40 | |
| *** diurnalist has joined #openstack-ironic | 21:22 | |
| *** penick has joined #openstack-ironic | 21:22 | |
| *** penick has quit IRC | 21:26 | |
| *** diurnalist has quit IRC | 21:36 | |
| *** dking has left #openstack-ironic | 21:51 | |
| *** bdodd_ has quit IRC | 22:41 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!