| *** seongsoocho has joined #openstack-ironic | 00:31 | |
| seongsoocho | join openinfra-summit | 00:32 |
|---|---|---|
| janders | good morning/afternoon Ironic o/ | 00:42 |
| janders | the great IRC migration question: is it now okay to leave the old channels? | 00:43 |
| janders | I keep looking at the old ones out of a habit and miss out on what's going on here | 00:44 |
| *** rloo has quit IRC | 00:46 | |
| stevebaker | janders: I have just left. A few will remain behind to redirect folk here | 00:56 |
| janders | thank you stevebaker, leaving now too | 00:58 |
| opendevreview | Merged openstack/ironic-python-agent stable/train: Prevent broken partition image UEFI deploys https://review.opendev.org/c/openstack/ironic-python-agent/+/782885 | 01:09 |
| opendevreview | Merged openstack/ironic-python-agent stable/victoria: Limit qemu-img execution arenas https://review.opendev.org/c/openstack/ironic-python-agent/+/794030 | 01:09 |
| *** lmcgann has quit IRC | 01:11 | |
| *** tzumainn has quit IRC | 01:31 | |
| *** paras333 has quit IRC | 02:07 | |
| *** paras333 has joined #openstack-ironic | 02:12 | |
| *** paras333 has quit IRC | 02:20 | |
| opendevreview | Merged openstack/ironic-python-agent stable/train: Increase the memory limit for qemu-img https://review.opendev.org/c/openstack/ironic-python-agent/+/793931 | 02:29 |
| opendevreview | Merged openstack/ironic-python-agent stable/train: Limit qemu-img execution arenas https://review.opendev.org/c/openstack/ironic-python-agent/+/794034 | 02:30 |
| opendevreview | Merged openstack/ironic-python-agent stable/train: Add fstab pointer to EFI partition https://review.opendev.org/c/openstack/ironic-python-agent/+/782886 | 02:30 |
| *** sdanni_ has quit IRC | 02:37 | |
| *** stendulker has joined #openstack-ironic | 04:52 | |
| *** stendulker_ has joined #openstack-ironic | 05:00 | |
| *** stendulker has quit IRC | 05:00 | |
| arne_wiebalck | Good morning, ironic! | 05:42 |
| arne_wiebalck | iurygregory: thanks for merging the ESP backports! | 05:44 |
| *** stendulker has joined #openstack-ironic | 06:24 | |
| *** stendulker_ has quit IRC | 06:24 | |
| iurygregory | good morning janders arne_wiebalck and Ironic o/ | 06:30 |
| iurygregory | arne_wiebalck, yw =) | 06:30 |
| iurygregory | if you have some time today to review some patches it would help since we need to start the release process later today =) | 06:30 |
| arne_wiebalck | They are all in the weekly prio list, I guess? | 06:32 |
| iurygregory | yeah =) | 06:36 |
| opendevreview | Verification of a change to openstack/ironic-python-agent failed: Increase the memory limit for qemu-img https://review.opendev.org/c/openstack/ironic-python-agent/+/793930 | 07:10 |
| opendevreview | Verification of a change to openstack/ironic-python-agent failed: Limit qemu-img execution arenas https://review.opendev.org/c/openstack/ironic-python-agent/+/794033 | 07:10 |
| iurygregory | arne_wiebalck, re your comment in https://review.opendev.org/c/openstack/ironic-python-agent/+/790471/4/ironic_python_agent/extensions/standby.py#142 | 07:13 |
| iurygregory | I think dtantsur said it was optional because in _write_image is optional | 07:14 |
| iurygregory | L228 | 07:14 |
| arne_wiebalck | iurygregory: still, it is either a copy/paste error in the comment or the signature needs adaptation | 07:45 |
| iurygregory | arne_wiebalck, yeah =) agree | 07:46 |
| *** lucasagomes has joined #openstack-ironic | 08:14 | |
| *** lucasagomes has quit IRC | 08:19 | |
| *** derekh has joined #openstack-ironic | 08:21 | |
| *** lucasagomes has joined #openstack-ironic | 08:26 | |
| *** lucasagomes has quit IRC | 08:27 | |
| *** lucasagomes has joined #openstack-ironic | 08:56 | |
| *** lucasagomes has quit IRC | 08:58 | |
| *** Luzi has joined #openstack-ironic | 09:10 | |
| *** mgoddard_ has joined #openstack-ironic | 09:14 | |
| arne_wiebalck | TheJulia: your -1 on https://review.opendev.org/c/openstack/ironic/+/768009 can be lifted, no? | 09:16 |
| dtantsur | good morning ironic | 09:16 |
| iurygregory | morning dtantsur | 09:16 |
| *** mgoddard_ is now known as mgoddard | 09:18 | |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-python-agent master: Stop accepting duplicated configdrive https://review.opendev.org/c/openstack/ironic-python-agent/+/790471 | 09:19 |
| dtantsur | arne_wiebalck: ^^ | 09:19 |
| iurygregory | dtantsur, tks! | 09:20 |
| dtantsur | hmm, I wonder why smartctl doesn't log anything: https://zuul.opendev.org/t/openstack/build/be59c1e3400c4f04bccf709f7d1ecfe7/log/controller/logs/ironic-bm-logs/node-1_no_ansi_2021-06-02-07:45:29_log.txt#2506 | 09:22 |
| iurygregory | *magic* | 09:22 |
| iurygregory | :D | 09:22 |
| dtantsur | actually, it seems like failed commands no longer log stdout/err | 09:22 |
| dtantsur | oh well, s/no longer // | 09:24 |
| *** lucasagomes has joined #openstack-ironic | 09:34 | |
| *** stendulker has quit IRC | 09:35 | |
| *** stendulker has joined #openstack-ironic | 09:36 | |
| *** lucasagomes has quit IRC | 09:37 | |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-lib master: utils.execute: log stdout and stderr even on failure https://review.opendev.org/c/openstack/ironic-lib/+/794171 | 09:37 |
| dtantsur | here goes ^^^ | 09:37 |
| iurygregory | funny =) | 09:38 |
| iurygregory | dtantsur, L98 shouldn't be if stdout? | 09:39 |
| dtantsur | iurygregory: nope, log_stdout is a flag to execute() | 09:40 |
| iurygregory | oh ok I just saw =) | 09:40 |
| iurygregory | sorry =) | 09:40 |
| *** frigo has joined #openstack-ironic | 09:40 | |
| dtantsur | no worries, it confused me as well at first | 09:41 |
| dtantsur | any opinions on inserting a line break after the colon? | 09:41 |
| *** lucasagomes has joined #openstack-ironic | 09:42 | |
| *** lucasagomes has quit IRC | 09:44 | |
| iurygregory | I don't have a strong option for this one hehe | 09:44 |
| iurygregory | do we need ironic-lib released with this and a bump in ironic for the release? | 09:45 |
| dtantsur | we don't bump versions for bug fixes | 09:45 |
| arne_wiebalck | dtantsur: re configdrive, thanks! | 09:54 |
| arne_wiebalck | dtantsur: re ironic-lib, on failure I think it would be more useful to increase the log level beyond debug | 09:55 |
| dtantsur | arne_wiebalck: we don't know if the failure is expected or not | 09:55 |
| arne_wiebalck | dtantsur: ? | 09:55 |
| dtantsur | we use the try-and-catch approach very often | 09:55 |
| arne_wiebalck | oh | 09:55 |
| arne_wiebalck | hmm | 09:56 |
| arne_wiebalck | ok | 09:56 |
| arne_wiebalck | I would still think that in the majority of cases you probably try to run sth you would like to succeed, no? | 09:57 |
| iurygregory | this would be the happy path (but we are talking about software + hardware) :D | 09:57 |
| arne_wiebalck | iurygregory speaks the truth | 09:58 |
| arne_wiebalck | and this is supporting my point, actually :) | 09:58 |
| arne_wiebalck | the default may be the wrong way round | 09:59 |
| dtantsur | this needs an addition to the function's signature, which is beyond the scope of this change at the very least | 10:00 |
| dtantsur | and honestly, I'd rather have the caller do a meaningful logging | 10:00 |
| arne_wiebalck | I agree | 10:00 |
| arne_wiebalck | to the first point :) | 10:01 |
| arne_wiebalck | seems like in most cases the caller does not even capture the returned (stdout, stderr) | 10:05 |
| dtantsur | in many cases they're not interesting | 10:06 |
| arne_wiebalck | if the command succeeds | 10:06 |
| arne_wiebalck | anyway, seems to be ok the way it is in the majority of cases | 10:07 |
| arne_wiebalck | I am just thinking we could probably save a "enable debug and retry" cycle in some failure scenarios | 10:08 |
| arne_wiebalck | but, yeah, beyond this patch for sure | 10:09 |
| mnaser | good morning | 10:24 |
| iurygregory | morning mnaser | 10:24 |
| opendevreview | Mark Goddard proposed openstack/bifrost stable/wallaby: Install DIB dependencies in bifrost-ironic-install role https://review.opendev.org/c/openstack/bifrost/+/794176 | 10:27 |
| opendevreview | Merged openstack/ironic-python-agent stable/ussuri: Increase the memory limit for qemu-img https://review.opendev.org/c/openstack/ironic-python-agent/+/793930 | 10:27 |
| opendevreview | Mark Goddard proposed openstack/bifrost stable/victoria: Install DIB dependencies in bifrost-ironic-install role https://review.opendev.org/c/openstack/bifrost/+/794177 | 10:31 |
| *** lucasagomes has joined #openstack-ironic | 10:32 | |
| opendevreview | Mark Goddard proposed openstack/bifrost stable/victoria: Install DIB dependencies in bifrost-ironic-install role https://review.opendev.org/c/openstack/bifrost/+/794177 | 10:32 |
| *** lucasagomes has quit IRC | 10:47 | |
| *** lucasagomes has joined #openstack-ironic | 10:51 | |
| *** jelabarre-rh has quit IRC | 11:01 | |
| *** paras333 has joined #openstack-ironic | 11:32 | |
| *** lucasagomes has quit IRC | 11:35 | |
| *** lucasagomes has joined #openstack-ironic | 11:37 | |
| *** owalsh has quit IRC | 11:39 | |
| *** stendulker has quit IRC | 11:48 | |
| *** lucasagomes has quit IRC | 11:55 | |
| *** lucasagomes has joined #openstack-ironic | 12:01 | |
| *** tzumainn has joined #openstack-ironic | 12:32 | |
| opendevreview | Merged openstack/ironic-python-agent master: Stop accepting duplicated configdrive https://review.opendev.org/c/openstack/ironic-python-agent/+/790471 | 12:37 |
| iurygregory | yay! | 12:38 |
| *** lmcgann has joined #openstack-ironic | 12:57 | |
| *** timeu has joined #openstack-ironic | 13:00 | |
| cenne | :) | 13:11 |
| cenne | \o/ | 13:11 |
| TheJulia | good morning | 13:12 |
| *** rloo has joined #openstack-ironic | 13:13 | |
| trandles | TheJulia: I missed your earlier question about image format. In OSP 16.1 it's still a flat partition image | 13:14 |
| trandles | chapter 24 of the director guide is about how to create a full disk image from the flat partition image provided | 13:15 |
| TheJulia | Ahh that makes sense then because your not network booting that partition image | 13:19 |
| * TheJulia suspects we have cleanup in the image verification code + destructions | 13:20 | |
| *** Luzi has quit IRC | 13:21 | |
| trandles | TheJulia: that's why I was slightly confused by my deployed image not working. I had updated the IPA ramdisk on the directory and the deploy ramdisk in glance, but it still wasn't working. I hadn't thought about the ramdisk in /boot/ in the image. | 13:30 |
| trandles | Everything saw all the drives just fine until the actual boot from local disk... | 13:31 |
| trandles | I also did see that support for these devices is being added back into 8.4. I'm tempted to open a support case asking that the support be backported to the overcloud image for OSP 16 | 13:32 |
| TheJulia | so I think the fact we ask for a kernel/ramdisk is more historical in nature from network booting | 13:34 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: Set download stable branch for IPA https://review.opendev.org/c/openstack/ironic/+/794224 | 13:34 |
| *** sdanni_ has joined #openstack-ironic | 13:34 | |
| TheJulia | trandles: so you'll want osp 16.2 once we release it | 13:34 |
| *** sdanni_ has quit IRC | 13:35 | |
| TheJulia | we can't cross rhel releases, at least on the osp side of the house | 13:35 |
| *** sdanni has joined #openstack-ironic | 13:35 | |
| trandles | TheJulia: will it rebase to whatever the current RHEL8 minor version is at the time? | 13:35 |
| TheJulia | and what I mean by that is get driers in to rhel 8.3 from 8.4 | 13:35 |
| TheJulia | 16.1 afaik stays on rhel 8.3, but sometimes if people ask these things change | 13:36 |
| trandles | 16.1 is still on 8.2 | 13:36 |
| TheJulia | but 8.4 has caused us a bit of pain downstream right now | 13:36 |
| TheJulia | hmm, maybe it skips 8.3, I dunno | 13:36 |
| trandles | it's been a bumpy road getting OSP installed but now that it is, I really like it, moreso than the other non-OSP installs I had done in the past | 13:37 |
| *** lucasagomes has quit IRC | 13:38 | |
| opendevreview | Julia Kreger proposed openstack/ironic stable/victoria: Set IPA download branch to stable/victoria for victoria https://review.opendev.org/c/openstack/ironic/+/794227 | 13:39 |
| TheJulia | wow gerrit is slow this morning | 13:39 |
| TheJulia | if I can get a quick review on https://review.opendev.org/c/openstack/ironic/+/794224 it would be appreciated. Stable/wallaby CI is presently broken on downloading the wrong IPA images which prevents us from landing backports to ironic as some jobs are failing with iscsi extension not being found | 13:41 |
| *** lucasagomes has joined #openstack-ironic | 13:48 | |
| arne_wiebalck | I have two Ironic 101 questions: a) the "node not found" hook triggered upon inspection does not mean Ironic does not have the node in its DB, it only means Ironic is surprised that introspection data has arrived since it is not aware of an ongoing introspection (like during auto-discovery), is that correct? | 13:53 |
| *** tosky has joined #openstack-ironic | 13:54 | |
| iurygregory | Gerrit is always slow for me :D | 13:58 |
| iurygregory | +30s to provide feedback/send a patch | 13:58 |
| TheJulia | arne_wiebalck: node not found hook should be the node is a surprise we didn't find it | 13:58 |
| TheJulia | node not found shouldn't fire otherwise, if my memory is correct | 13:59 |
| iurygregory | TheJulia, why stable/victoria instead of stable/wallaby? | 13:59 |
| arne_wiebalck | TheJulia: we did not find it as in "it is not in the DB" or we did not find it as in "we did not tell to inspect" ? | 13:59 |
| iurygregory | or we need the previous IPA stable version? | 13:59 |
| TheJulia | iurygregory: both need it, both were set to master, just different jobs and underlying job settings it looks like | 14:00 |
| iurygregory | ack | 14:00 |
| TheJulia | wallaby was set to download, not build on a couple jobs | 14:00 |
| TheJulia | I'm surprised it took this long to show up, but also could have been just pulling from a cache or a mirror for a while :\ | 14:00 |
| *** kkillsfirst has joined #openstack-ironic | 14:00 | |
| *** paras333 has quit IRC | 14:08 | |
| *** paras333 has joined #openstack-ironic | 14:08 | |
| dtantsur | good morning TheJulia | 14:21 |
| dtantsur | TheJulia: https://review.opendev.org/c/openstack/ironic/+/794224/1/devstack/lib/ironic is against wallaby but says victoria | 14:22 |
| TheJulia | doh | 14:22 |
| iurygregory | I think I wasn't clear when I asked :D | 14:22 |
| dtantsur | ah, I haven't noticed that it was already being discussed | 14:23 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: Set download stable branch for IPA https://review.opendev.org/c/openstack/ironic/+/794224 | 14:23 |
| TheJulia | oh, iurygregory didn't realize that is what you were asking | 14:23 |
| iurygregory | no worries, I was in a hurry to join a meeting when I asked :D | 14:23 |
| TheJulia | I thought you asked why because I posted patches for both branches | 14:23 |
| iurygregory | the meeting just finished | 14:23 |
| iurygregory | I'm finishing the updates in release notes and release mappings for ironic | 14:24 |
| iurygregory | will push in a few =) | 14:24 |
| TheJulia | ack | 14:24 |
| dtantsur | approved both | 14:24 |
| iurygregory | the flash :D | 14:25 |
| * TheJulia needs to wake up before doing the next set of backports | 14:25 | |
| dtantsur | heh, it wasn't a long reading :) | 14:25 |
| iurygregory | yup =) | 14:25 |
| *** frigo has quit IRC | 14:35 | |
| *** opendevreview has quit IRC | 14:38 | |
| *** stendulker has joined #openstack-ironic | 14:47 | |
| *** edleafe has joined #openstack-ironic | 14:48 | |
| *** owalsh has joined #openstack-ironic | 14:48 | |
| *** edleafe has left #openstack-ironic | 14:48 | |
| *** edleafe has joined #openstack-ironic | 14:48 | |
| *** edleafe has left #openstack-ironic | 14:48 | |
| *** Gue______ is now known as jamesdenton | 14:51 | |
| *** paras333 has quit IRC | 15:05 | |
| *** paras333 has joined #openstack-ironic | 15:05 | |
| arne_wiebalck | coming back to my "node not found" hook question: I think it really is a "did not expect introspection data from this node" hook, isn't it? | 15:07 |
| dtantsur | pretty much | 15:08 |
| arne_wiebalck | dtantsur: thanks | 15:09 |
| *** opendevreview has joined #openstack-ironic | 15:09 | |
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic master: Prepare 18.0 and clean up release note https://review.opendev.org/c/openstack/ironic/+/794248 | 15:09 |
| arne_wiebalck | that brings me to my second 101 question :) | 15:09 |
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic master: Prepare 18.0 and clean up release note https://review.opendev.org/c/openstack/ironic/+/794248 | 15:09 |
| iurygregory | LOL | 15:09 |
| dtantsur | arne_wiebalck: that must be 102 then! | 15:09 |
| arne_wiebalck | heh | 15:09 |
| iurygregory | 2 patches wtf git review O.o | 15:10 |
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic master: Prepare 18.0 and clean up release note https://review.opendev.org/c/openstack/ironic/+/794250 | 15:10 |
| iurygregory | LOL | 15:10 |
| iurygregory | O.o | 15:10 |
| dtantsur | moar patches! | 15:10 |
| iurygregory | my git review is drunk | 15:10 |
| iurygregory | WTF | 15:10 |
| dtantsur | 18.0 is such an awesome release that we'll prepare it twice | 15:10 |
| iurygregory | dtantsur, because of iscsi :D | 15:11 |
| dtantsur | iscsi-- | 15:11 |
| arne_wiebalck | what triggers the IPA to do an inspection and send the data, or does it always inspect? | 15:11 |
| dtantsur | arne_wiebalck: always when an inspector URL is present | 15:11 |
| arne_wiebalck | aha! | 15:11 |
| iurygregory | https://review.opendev.org/c/openstack/ironic/+/794250 will be the valid one =) | 15:11 |
| arne_wiebalck | dtantsur: this is what I was about to ask | 15:12 |
| dtantsur | but I'm faster! | 15:12 |
| arne_wiebalck | dtantsur: we moved to pass both URLs all the time | 15:12 |
| arne_wiebalck | :-D | 15:12 |
| arne_wiebalck | dtantsur: and now get node not found all the time | 15:13 |
| dtantsur | iurygregory: the release mapping part is good, I'll let someone English-speaking comment on the notes (that be TheJulia I guess) | 15:13 |
| iurygregory | dtantsur, ack =) | 15:14 |
| iurygregory | I did some check to make some release notes more simple and fixed a few typos only | 15:14 |
| iurygregory | I'm wondering if we want to create a bugfix for inspector we only have http://paste.openstack.org/show/806278/ | 15:19 |
| opendevreview | Alexander Evseev proposed openstack/bifrost master: Drop external tftp service in favor to use dnsmasq's one https://review.opendev.org/c/openstack/bifrost/+/782531 | 15:19 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-inspector stable/wallaby: Trivial: better guard against malformed data in extra_hardware https://review.opendev.org/c/openstack/ironic-inspector/+/794183 | 15:20 |
| dtantsur | if we release wallaby with ^^, then we can probably skip inspector | 15:20 |
| dtantsur | since it's the only meaningful change there | 15:20 |
| iurygregory | ++ | 15:20 |
| iurygregory | I'm also planning in push releases for wallaby victoria ussuri | 15:21 |
| dtantsur | nice! | 15:21 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-python-agent master: Check the target root device with smartctl https://review.opendev.org/c/openstack/ironic-python-agent/+/793953 | 15:24 |
| dtantsur | had to rebase unfortunately ^^^ | 15:24 |
| iurygregory | do we consider ipab a service? =) asking because we need to update https://docs.openstack.org/ironic/latest/contributor/releasing.html#independent :D | 15:32 |
| dtantsur | iurygregory: mmmmmmm, a good question. do they have something like "other"? :) | 15:37 |
| dtantsur | oh, and we need to remove sushy-cli | 15:37 |
| iurygregory | yeah =) | 15:38 |
| * dtantsur has no idea where to put IPA-b | 15:38 | |
| iurygregory | other sounds like a good idea =) | 15:40 |
| iurygregory | I think we can remove molteniron... | 15:40 |
| dtantsur | probably? | 15:40 |
| TheJulia | iurygregory: thanks for the other releases :) | 15:56 |
| iurygregory | TheJulia, np! | 15:56 |
| *** stendulker has quit IRC | 16:07 | |
| dtantsur | folks, I'm out this week, talk to you on Monday! | 16:07 |
| arne_wiebalck | bye dtantsur o/ | 16:07 |
| *** lucasagomes has quit IRC | 16:13 | |
| *** stendulker has joined #openstack-ironic | 16:17 | |
| *** ipilcher has joined #openstack-ironic | 16:30 | |
| *** stendulker has quit IRC | 16:34 | |
| ipilcher | Greetings all! Would this be a proper place for questions about pyghmi? | 16:35 |
| TheJulia | ipilcher: Well, you can ask, but we likely would end up pointing you in the direction of the maintainer. Whats up? | 16:41 |
| trandles | TheJulia: closing the loop on my deployment drama. For some reason I still couldn't get the ramdisk to auto-load the driver. I loaded it manually from the dracut emergency shell, continued the boot, fixed up the ramdisk using dracut, reinstalled grub, and rebooted successfully. The whole time ironic said the node was happily active. I don't know what could/should be done | 16:42 |
| trandles | from ironic's perspective to know the node was actually in a bad state. | 16:42 |
| TheJulia | Well, ironic has no agent running on that machine after reboot, so it can't really know the state | 16:44 |
| TheJulia | so.. hmm | 16:44 |
| trandles | yeah, my thoughts too... | 16:44 |
| TheJulia | when you say updated the image, how did you update the ramdisk in it? | 16:44 |
| TheJulia | We've talked and discussed an in-workload agent for kexecing to cleaning, but huge security implications there along with everything else | 16:45 |
| trandles | There's a longer version of this but the tl;dr is first I used virt-copy-in to put the kmod RPM in the image (it comes from ElRepo) and then virt-customize to dnf install it. When I finally booted successfully the driver was indeed installed fine, but the initramfs hadn't been updated. | 16:46 |
| trandles | Second, I ended up copying-out the initramfs, unwinding it (skipcpio, pax, etc.) and using rpm2cpio to just put the kmod extras driver in manually. | 16:47 |
| trandles | That's when I was finally able to load it from the dracut command line using insmod, the device and drive were seen, and I could continue booting. | 16:48 |
| *** tosky has quit IRC | 16:48 | |
| trandles | (sorry, I created the new initramfs and copy-in to the overcloud image) | 16:49 |
| trandles | I also tried to use rd.driver.pre=megaraid_sas on the kernel commandline in grub but that didn't work either. I had to insmod it by hand from the dracut command line. | 16:50 |
| opendevreview | Merged openstack/ironic stable/victoria: Set IPA download branch to stable/victoria for victoria https://review.opendev.org/c/openstack/ironic/+/794227 | 16:57 |
| TheJulia | hmm | 16:58 |
| TheJulia | running a quick errand | 16:59 |
| *** bfournie has quit IRC | 16:59 | |
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic-python-agent master: Clean-up releasenotes for 8.0.0 release https://review.opendev.org/c/openstack/ironic-python-agent/+/794277 | 16:59 |
| ipilcher | TheJulia: I'd like to write some sort of IPMI proxy that will provide LDAP-based auth for the Dell systems in our lab. pyghmi seems like it could be a good starting point, but I haven't been able to figure it out by simply reading the code. I'm also not sure if it's even *possible*. (Does an IPMI "client" ever send the actual password, or does it only send a hash?) | 17:04 |
| ipilcher | (So basically vbmc, with with actual BMCs on the back end.) | 17:05 |
| *** derekh has quit IRC | 17:06 | |
| opendevreview | Merged openstack/ironic-python-agent stable/ussuri: Limit qemu-img execution arenas https://review.opendev.org/c/openstack/ironic-python-agent/+/794033 | 17:07 |
| arne_wiebalck | bye everyone o/ | 17:08 |
| iurygregory | bye arne_wiebalck | 17:12 |
| iurygregory | I'm wondering if we need bugfix branch for ipab also since we have for ipa | 17:12 |
| TheJulia | ipilcher: so I think it only sends the hash, so integrating it in with another authentication system might be problematic. I *think* one can do a null hash, but then your basically sending stuff in the clear aiui | 17:32 |
| TheJulia | ipilcher: and of course, only the hash because the underlying transport is udp packets | 17:32 |
| TheJulia | tzumainn: you might be interested in this topic ipilcher is looking at | 17:33 |
| ipilcher | :-( That would explain why iDRACs won't do LDAP auth for IPMI. | 17:33 |
| TheJulia | yeah, ipmi is basically evil. | 17:33 |
| iurygregory | bifrost didn't require any release note clean-up so I've pushed 11.0.0 https://review.opendev.org/c/openstack/releases/+/794320 since we have removed the support for iscsi | 17:34 |
| ipilcher | I guess I should start looking at Redfish. I wonder if it's actually usable on iDRAC 8. | 17:35 |
| TheJulia | ipilcher: it is usable, at least the base functionality | 17:35 |
| ipilcher | That's encouraging, at least. | 17:35 |
| TheJulia | The extended functionality, aiui, has been chosen not to be implemented on the idrac8s | 17:36 |
| TheJulia | what that truly is, I dunno, but people have used just redfish with idrac 8s to do power/boot control for a while. | 17:36 |
| ipilcher | Power/boot control is all we need. (Pretty much what Ironic needs.) | 17:37 |
| TheJulia | \o/ | 17:37 |
| TheJulia | tzumainn has experssed interest in building a proxy infront of ironic that speaks $somebmcprotocol | 17:37 |
| TheJulia | For those following along on improving api performance. Today's benchmark of master was 205 nodes/second. Up from like 86 nodes/second at the beginning of the cycle. | 18:17 |
| iurygregory | WOOOT | 18:17 |
| TheJulia | now... to get it to ~460 nodes/second | 18:17 |
| opendevreview | Merged openstack/ironic stable/wallaby: Set download stable branch for IPA https://review.opendev.org/c/openstack/ironic/+/794224 | 18:20 |
| trandles | TheJulia: is that nodes queried or images served or ?? | 18:23 |
| tzumainn | ipilcher, ah, yes, I've played with this a little, mostly with sushy_tools / redfish though | 18:23 |
| trandles | Whatever it is, ~205 nodes/second sounds exciting just in general ;) | 18:23 |
| TheJulia | trandles: nodes returned via the api when doing chunked requests of 1000 | 18:24 |
| trandles | gotcha | 18:24 |
| TheJulia | trandles: I have a data set with ~113k nodes that I generated locally to simulate some of the large database issues some of the larger operators have experiencedd | 18:25 |
| *** jelabarre-rh has joined #openstack-ironic | 18:25 | |
| trandles | If we adopt ironic for stateless cluster provisioning we'll probably fall in that category. We have clusters in the 10k-20k node range right now. | 18:26 |
| *** jelabarre-rh has quit IRC | 18:27 | |
| *** jelabarre-rh has joined #openstack-ironic | 18:27 | |
| TheJulia | yeah | 18:28 |
| TheJulia | so 10k is not horrible really database wise, it is 40k when the database really starts crying | 18:28 |
| TheJulia | *but* a lot of that has to do with indexes, db hit patterns, and how various actions get translated into db queries | 18:29 |
| ipilcher | tzumainn - It looks like I'll have to use Redfish in order to get LDAP auth working, as it seems that the IPMI protocol doesn't send the password (which is needed for an LDAP bind). A quick look indicates that my iDRACs probably have adequate Redfish support, so I can hit them directly, which means I don't need any sort of intermediate proxy. | 18:29 |
| trandles | TheJulia: I've been meaning to ask if the inspector can be configured to use something other than swift to store all the inspection data | 18:30 |
| ipilcher | Basically, I had been thinking that the iDRAC inability to use LDAP auth for IPMI was an iDRAC limitaion, but it looks like it's really a limitation of IPMI itself. | 18:30 |
| *** kkillsfirst has quit IRC | 18:31 | |
| TheJulia | some additional indexes made a decently measurable difference alone. Trying to improve the big painful query point which is also an issue with syncing with Nova. The other aspect is going to be conductors and if you want power sync... or not. | 18:31 |
| TheJulia | trandles: webdav, but I don't think we fed it credentials. I guess you could also point it at ceph object store as long as it is swift compatible | 18:32 |
| TheJulia | of course, patches are always welcome | 18:32 |
| trandles | We really like the inventory inspector gives us. We have to track things like DIMM replacements so being able to inspect and harvest all the serial numbers is really nice. Saving that data in something like a DB to make it more useful for reporting, etc. might be another project we (LANL) pursues. | 18:33 |
| trandles | It might also just be simpler to write our own tooling to take it from the object store and pipe it into splunk. | 18:34 |
| tzumainn | ipilcher, ah, okay! | 18:35 |
| trandles | But splunk is likely where we would send it ultimately. Thinking out loud, all of our splunk ingestion flows through rabbitmq or kafka so maybe it would be worth having it go directly there during inspection and bypass the object store entirely. | 18:35 |
| TheJulia | so post processing wise, I don't think we grok or have a concept of doign that, but it sounds feasible | 18:38 |
| TheJulia | NobodyCam: this might tickle your brain a little | 18:38 |
| NobodyCam | Good Morning Ironic folks! | 18:39 |
| * NobodyCam reads scroll back | 18:39 | |
| NobodyCam | so inspection logs direct to Splunk or Kibana | 18:42 |
| iurygregory | morning NobodyCam o/ | 18:44 |
| TheJulia | does it seem usefu? | 18:48 |
| TheJulia | useful | 18:50 |
| *** bfournie has joined #openstack-ironic | 18:51 | |
| *** kkillsfirst has joined #openstack-ironic | 19:05 | |
| *** tosky has joined #openstack-ironic | 19:23 | |
| *** paras333 has quit IRC | 20:15 | |
| iurygregory | hey stevebaker, if you have a few minutes today can you review https://review.opendev.org/c/openstack/ironic-python-agent/+/794277 and https://review.opendev.org/c/openstack/ironic/+/794250 ? We need this two to cut the bugfix branches for ironic and ipa tomorrow =) | 20:17 |
| TheJulia | I... in theory, could just approve the later one | 20:18 |
| TheJulia | I did approve the first one | 20:19 |
| iurygregory | TheJulia, if you give the +2 in the ipa change it would help hehe | 20:19 |
| TheJulia | iurygregory: refresh it ;) | 20:20 |
| iurygregory | woot | 20:20 |
| iurygregory | :O | 20:20 |
| iurygregory | ty :D | 20:20 |
| kkillsfirst | Hello, I have read the "Getting Started" and "Developer's Guide" on OpenDev. I have used the sandbox repo to go through the Gerrit workflow. Where/what should I be looking at to start making contributions? | 20:21 |
| TheJulia | kkillsfirst: so I'd download like the ironic repository, run `git review -s`, and go from there. Or ironic-python-agent. If there was something in one of the documents that you want to try and change to verify you've got the workflow down, ironic does hold the vast majority of the documentation. | 20:23 |
| kkillsfirst | TheJulia: Thank you. | 20:24 |
| iurygregory | you can probably push a change to remove sushy-cli molteniron from https://docs.openstack.org/ironic/latest/contributor/releasing.html#independent https://docs.openstack.org/ironic/latest/contributor/releasing.html#client-libraries | 20:24 |
| TheJulia | ++ | 20:24 |
| * TheJulia glares at "ironic.common.exception.InsufficentMemory: Available memory at 279.66796875, Insufficent as 1024 is required to proceed at this time" | 20:24 | |
| iurygregory | you will find the file in the ironic repository under doc/source/contributor/ , after editing make sure you run tox -e docs to see the change is ok, after doing this you can push the change and we will be happy to review | 20:26 |
| TheJulia | iurygregory: are you going to prepare a blog post for ironicbaremetal.org? | 20:26 |
| TheJulia | or is that my [honor|?curse?]? | 20:27 |
| iurygregory | for the release you mean or I'm missing something? =) | 20:27 |
| TheJulia | for 18.0's release | 20:27 |
| iurygregory | yup it's on my list to update things in the repo after we have the release | 20:28 |
| TheJulia | k | 20:28 |
| *** paras333 has joined #openstack-ironic | 20:29 | |
| opendevreview | Merged openstack/ironic-python-agent master: Clean-up releasenotes for 8.0.0 release https://review.opendev.org/c/openstack/ironic-python-agent/+/794277 | 20:30 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Set stage for objects to handle selected field lists. https://review.opendev.org/c/openstack/ironic/+/792275 | 20:32 |
| *** rloo has quit IRC | 20:36 | |
| *** rloo has joined #openstack-ironic | 20:36 | |
| *** paras333 has quit IRC | 21:00 | |
| *** jamesdenton has quit IRC | 21:21 | |
| *** Gues_____ has joined #openstack-ironic | 21:23 | |
| stevebaker | iurygregory: done | 21:56 |
| *** lmcgann has quit IRC | 21:57 | |
| iurygregory | stevebaker, ty! | 22:04 |
| *** jelabarre-rh has quit IRC | 22:20 | |
| *** jelabarre-rh has joined #openstack-ironic | 22:20 | |
| opendevreview | Kam Killsfirst proposed openstack/ironic master: Changes made to release documentation: https://review.opendev.org/c/openstack/ironic/+/794346 | 22:43 |
| *** paras333 has joined #openstack-ironic | 23:05 | |
| TheJulia | kkillsfirst: \o/ | 23:10 |
| TheJulia | kkillsfirst: approved | 23:13 |
| *** kkillsfirst has quit IRC | 23:20 | |
| *** tosky has quit IRC | 23:25 | |
| opendevreview | Merged openstack/ironic master: Prepare 18.0 and clean up release note https://review.opendev.org/c/openstack/ironic/+/794250 | 23:30 |
| opendevreview | Merged openstack/ironic master: Changes made to release documentation: https://review.opendev.org/c/openstack/ironic/+/794346 | 23:32 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!