| vanou | good morning ironic | 00:27 |
|---|---|---|
| jlvillal | TheJulia, Thanks. That's confusing for my little brain :D | 00:34 |
| vanou | JayF: How about communication with OpenStack vul team? | 02:41 |
| kaloyank | morning everyone o/ | 06:21 |
| kaloyank | dtantsur: yes, it looks like it's working, I tested in last night | 06:22 |
| rpittau | good morning ironic! o/ | 08:56 |
| opendevreview | Harald Jensås proposed openstack/metalsmith stable/yoga: list_instances - cache allocations https://review.opendev.org/c/openstack/metalsmith/+/876671 | 09:05 |
| opendevreview | Harald Jensås proposed openstack/metalsmith stable/xena: list_instances - cache allocations https://review.opendev.org/c/openstack/metalsmith/+/876672 | 09:06 |
| opendevreview | Harald Jensås proposed openstack/metalsmith stable/wallaby: list_instances - cache allocations https://review.opendev.org/c/openstack/metalsmith/+/876673 | 09:06 |
| dtantsur | kaloyank: nice, thanks for confirming! | 09:16 |
| kaloyank | dtantsur: Thanks for looking into my case | 09:17 |
| kaloyank | on a side note, do you need more information for the missing PXE boot parameters on the Neutron port when fast-track deployment fails? | 09:18 |
| dtantsur | a report in storyboard with ironic-conductor logs (if possible) could be a good start | 09:22 |
| dtantsur | I cannot promise to jump on it right away, unfortunately - too many other things to do | 09:22 |
| kaloyank | dtanstur: I'll see to it | 09:36 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Do not recalculate checksum if disk_format is not changed https://review.opendev.org/c/openstack/ironic/+/876595 | 10:57 |
| dtantsur | kaloyank: the final version of the fix ^^ | 10:57 |
| dtantsur | I had to adjust the code for non-glance images | 10:58 |
| kaloyank | Thanks, I'll update the patch | 11:17 |
| kaloyank | OK, another question: When cleaning a node with fast-track, why do we remove the cleaning port from Neutron after the node moves to "available" state? The node is still powered on and still has an IP assigned by Neutron | 11:32 |
| kaloyank | Logs https://pastebin.com/XmxXALDa ^^ | 11:33 |
| iurygregory | morning Ironic | 11:47 |
| kubajj | morning iurygregory | 12:51 |
| timeu | quit | 13:13 |
| mnasiadka | TheJulia: I was able to run the upgrade through to Yoga based on your patch - thanks again :) | 13:25 |
| dtantsur | kaloyank: it's a bit of a grey area. Theoretically, it's wrong to keep the cleaning port when cleaning is not running. | 13:41 |
| dtantsur | Worst case, you're going to exhaust the cleaning subnet with ports that correspond to available nodes | 13:42 |
| dtantsur | There is also a question of when to delete it. | 13:43 |
| rpittau | if any core has a moment please check a small fix for bifrost CI https://review.opendev.org/c/openstack/bifrost/+/872634 | 13:43 |
| kaloyank | dtantsur: well since I share the cleaning subnet with instances running, sometimes there are IP conflicts because Neutron hands out IP addresses that are already assigned to alive IPAs | 13:47 |
| kaloyank | I was thinking that it may be useful to "switch" from the cleaning network to the provisioning network and leave it as it if it's the same Neutron subnet | 13:48 |
| dtantsur | Switching means re-creating the port, so it may receive a different IP anyway? | 13:49 |
| kaloyank | it most certainly will, I'm not exactly familiar but the IPA should update it's callback URL once it has the new IP address | 13:50 |
| dtantsur | updating callback URL is explicitly forbidden (I think the reason was to minimize a possibility of hijacking the agent - TheJulia will remember). | 13:56 |
| dtantsur | you'll need to reboot the node using the ironic API. at which point the value of fast-track is questionable.. | 13:56 |
| kaloyank | I see.. So it only makes sense to retain the port if the cleaning subnet is the same as the provisioning? | 14:00 |
| kaloyank | and callback URL is the URL the agent is available at or I'm mistaken? | 14:00 |
| TheJulia | good morning | 14:01 |
| kaloyank | o/ | 14:03 |
| opendevreview | Merged openstack/ironic bugfix/21.2: Fixes for tox 4.0 https://review.opendev.org/c/openstack/ironic/+/876409 | 14:10 |
| TheJulia | mnasiadka: awesome! That is the best news! | 14:10 |
| TheJulia | Yeah, it was to prevent a man in the middle attack | 14:11 |
| opendevreview | Merged openstack/ironic bugfix/21.2: Update .gitreview for bugfix/21.2 https://review.opendev.org/c/openstack/ironic/+/867826 | 14:12 |
| TheJulia | I guess code could be added around port changes to permit updating the callback URL | 14:12 |
| TheJulia | One thing I have been thinking about is supporting multiple callback urls | 14:12 |
| TheJulia | ... Since most physical machines have multiple interfaces in various networks | 14:13 |
| TheJulia | kaloyank: I think your base question is worthy of discussion during the PTG, would you be willing to join and discuss ? | 14:17 |
| kaloyank | TheJulia: I'd love to, two questions: What's a PTG and when is it held? | 14:19 |
| TheJulia | kaloyank: https://openinfra.dev/ptg/ | 14:19 |
| TheJulia | March 27-31st, virtual this time around | 14:19 |
| TheJulia | Our etherpad of discussion topics https://etherpad.opendev.org/p/ironic-bobcat-ptg | 14:20 |
| TheJulia | kaloyank: there is a similar topic down by line 72 on that etherpad. Honestly I think the topic your looking for is more along the lines of "can we make fast track work with neutron integrated usage" | 14:25 |
| kaloyank | TheJulia: found it, I'll read thoroughly in a bit :) | 14:26 |
| TheJulia | kaloyank: I added thoughts, please feel free to add/change what I wrote | 14:31 |
| dtantsur | meanwhile, https://review.opendev.org/c/openstack/ironic/+/876595 is ready for review | 14:31 |
| TheJulia | dtantsur: sure, I just need to send a board related email first which will take me a few minutes | 14:33 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876174 | 14:54 |
| TheJulia | Always fun when you get to go dig through bylaws first thing in the morning | 14:57 |
| TheJulia | dtantsur: reviewed, +2 although two comments one likely just informational for another reviewer | 15:03 |
| dtantsur | thx! | 15:05 |
| opendevreview | Merged openstack/bifrost master: Finally fix jinja[spacing] https://review.opendev.org/c/openstack/bifrost/+/872634 | 15:22 |
| kaloyank | TheJulia: I updated the issue description | 15:27 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Fix online upgrades for Bios/Traits https://review.opendev.org/c/openstack/ironic/+/876626 | 16:51 |
| rpittau | good night! o/ | 16:58 |
| kaloyank | good night everyone o/ | 17:42 |
| opendevreview | Julia Kreger proposed openstack/ironic master: add default conductor group capability https://review.opendev.org/c/openstack/ironic/+/855705 | 18:06 |
| jlvillal | Maybe dumb question. If I do an `lsblk` should I see a partition for the configdrive on a deployed system? | 18:08 |
| jlvillal | Real reason I'm asking. My SSH key isn't getting added to the system and I'm trying to figure out why... | 18:10 |
| JayF | jlvillal: I believe so, yes. | 18:19 |
| JayF | jlvillal: that's how cloud-init works, it looks for it and mounts it up and gets the data out | 18:19 |
| JayF | jlvillal: so you can totes interact w/config drive using stnadard linux block tools | 18:20 |
| jlvillal | JayF, Thanks. I don't seem to see one. Thus why I am confused... But would make sense why the SSH keys aren't being added. | 18:20 |
| JayF | I'd check conductor logs and see if it appears to be picking up a config drive at all | 18:28 |
| jlvillal | Thanks | 18:29 |
| JayF | iurygregory: rpittau: TheJulia: Is there any reason you all know of I should not approve the releases to cut 2023.1 branches for all Ironic stuff right now? | 19:07 |
| JayF | I guess 876626 would be good to get in | 19:08 |
| JayF | although we could backport it | 19:08 |
| JayF | I just +2'd 876626, someone could land it and I'll update that change req | 19:09 |
| * iurygregory looks | 19:11 | |
| iurygregory | let's get 876626 in | 19:12 |
| kubajj | JayF, TheJulia: sorry for bothering you, but have you managed to have a look at the email? there is still no rush, but don't want to forget about it and leave it to the last mine. | 19:13 |
| iurygregory | at least we save some CI resources =) | 19:13 |
| JayF | kubajj: you should be bothering me, yikes, I'm so sorry | 19:13 |
| JayF | iurygregory: you landing it? | 19:13 |
| JayF | kubajj: I'm going to go take care of it right now | 19:13 |
| iurygregory | just +W 876626 | 19:13 |
| iurygregory | we can probably land https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/876361 and update https://review.opendev.org/c/openstack/releases/+/876611 | 19:17 |
| iurygregory | https://review.opendev.org/c/openstack/releases/+/876571 we are good to go | 19:18 |
| iurygregory | for bifrost I saw some patches, so we can probably wait and update I would say | 19:19 |
| iurygregory | for ironic we need to wait and update virtualpdu I think? | 19:24 |
| TheJulia | kubajj: sorry, forgot. Things have been chaos :( | 19:26 |
| TheJulia | JayF: afaik no as long as the commit hash lines up with something not weeks ago | 19:27 |
| jlvillal | JayF, I figured out why `lsblk` didn't show the partition. I had sshed into the wrong system. Duh! :( But now to figure out why the SSH keys didn't get added to the authorized_keys file... | 19:34 |
| JayF | jlvillal: did the correct one at least have the configdrive? | 19:42 |
| jlvillal | JayF, Yep. | 19:42 |
| JayF | if that has your key, it's a cloud-init issue | 19:42 |
| JayF | which is an entire world of pain I'm lucky enough not to have scoped into my life | 19:43 |
| JayF | although if it breaks when looking for a BM configdrive, it is still my problem :| | 19:43 |
| jlvillal | Hehe. Yeah. I've done it before and I always get confused. Since I only dig into it like once a year and forget everything the next time I need to do something. | 19:43 |
| stevebaker[m] | hey, does anybody know the package which provides tgt on recent centos? It hasn't been scsi-target-utils since centos7 | 20:10 |
| JayF | in Gentoo, it's literally sys-block/tgt | 20:11 |
| JayF | maybe it split into it's own package in RH-land? | 20:11 |
| stevebaker[m] | oh, its not a thing since centos8. I'm including the ramdisk element, which ironic-python-agent-builder doesn't actually use | 20:23 |
| JayF | if we want https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/876361 in antelope, someone needs to W+1 it nowish... TheJulia you have a sec? | 20:38 |
| JayF | or stevebaker[m] | 20:38 |
| JayF | I approved all release requests, some needed a small update, except Ironic (waiting for 876626 to land) and tempest plugin (hoping someone looks at dtantsur's change so we can get it in pre-branch) | 20:39 |
| stevebaker[m] | done | 20:40 |
| JayF | thanks steve! | 20:42 |
| kubajj | JayF: thanks | 20:45 |
| kubajj | TheJulia: no worries | 20:45 |
| opendevreview | Merged openstack/ironic master: Fix online upgrades for Bios/Traits https://review.opendev.org/c/openstack/ironic/+/876626 | 20:57 |
| opendevreview | Merged openstack/ironic master: Add missing include for inventory API reference https://review.opendev.org/c/openstack/ironic/+/874946 | 20:57 |
| JayF | Aight, updating the ironic one | 21:02 |
| JayF | we're going to have a stack! of opens! | 21:02 |
| JayF | aight, Ironic release signed off on | 21:05 |
| JayF | just gotta watch 876361 and get tempest | 21:05 |
| JayF | and we've stacked an open | 21:05 |
| JayF | dtantsur: rpittau: I'm trying to get https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/876361 in before https://review.opendev.org/c/openstack/releases/+/876611 lands. It's about to fail the gate. Can one or both of you own ensuring it 1) gets in and the releases PR is updated+approved or 2) is broken and we release without it | 21:59 |
| iurygregory | JayF, also for ironic, we need the updates in the code to cut the release | 22:00 |
| JayF | iurygregory: what do you mean? | 22:00 |
| opendevreview | Verification of a change to openstack/ironic-tempest-plugin master failed: Split the ramdisk ISO test https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/876361 | 22:00 |
| JayF | iurygregory: I did the release notes prelude; what else needs to be done? | 22:01 |
| iurygregory | https://docs.openstack.org/ironic/latest/contributor/releasing.html#things-to-do-after-releasing | 22:01 |
| * JayF didn't recall doing anything extra last release, but you might have JFDI | 22:01 | |
| JayF | iurygregory: that's after, yeah? not before? | 22:01 |
| iurygregory | ohhh right (in the past we did the release mappings before cutting it I think) | 22:03 |
| JayF | honestly, you wanna help out with taking care of that? | 22:03 |
| JayF | otherwise I'll probably end up doing it on an airplane lol | 22:03 |
| * TheJulia waits for unit tests to spin making a "whirrrrrrrrrrrrrrrrl" sound | 22:23 | |
| JayF | TheJulia: what's the thing I need? I saw the -1 on the releases pr | 22:30 |
| TheJulia | oh jebus, the 2023.1 crap breaks it | 22:31 |
| TheJulia | https://www.irccloud.com/pastebin/huYQoMUH/ | 22:31 |
| JayF | TheJulia: we also need to update that comment on Ln544, we need to worry about SLURP now too | 22:33 |
| TheJulia | so since antelope is not the actual named release anymore, it is 2023.1, I guess we justdrop the test? | 22:33 |
| TheJulia | keep yoga around? | 22:34 |
| JayF | I gotta be honest, I know what this code is for in theory but I don't know how to exercise it | 22:34 |
| JayF | TheJulia: I know that we'll need to keep zed when we cut 2023.2; I'm not sure about yoga now | 22:34 |
| JayF | TheJulia: the simpler answer for us as a project is probably to just leave the extra at all times | 22:34 |
| TheJulia | yeah | 22:34 |
| TheJulia | agree | 22:34 |
| JayF | so is this what gets called if, e.g., ironic-dbsync 2023.1 ? | 22:35 |
| JayF | like I'm trying to determine where the release version slug gets input from a user | 22:35 |
| TheJulia | so, it is version mapping for objects if they need translation or not | 22:35 |
| TheJulia | retesting.... | 22:37 |
| JayF | we could fix the sorting | 22:37 |
| JayF | https://github.com/openstack/ironic/blob/master/ironic/common/release_mappings.py#L554 could have a specified list of openstack versions that get sorted in first | 22:38 |
| JayF | idk; I trust you to handle this I'm going to point my brain elsewhere unless you need it | 22:38 |
| TheJulia | k | 22:38 |
| TheJulia | AssertionError: Versions (21, 1) and (21, 4) are not sequential | 22:38 |
| TheJulia | ... oh | 22:44 |
| TheJulia | I see what is going on | 22:44 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Update release mappings for 21.4 release https://review.opendev.org/c/openstack/ironic/+/876786 | 22:52 |
| iurygregory | now we will add 2023.1 and Antelope? O.o | 22:52 |
| TheJulia | well, so technically it needs to match out to the branch name | 22:53 |
| TheJulia | but people are going to get things mixed up no matter what | 22:53 |
| TheJulia | so better to do both I think | 22:53 |
| TheJulia | jayf ^ | 22:54 |
| JayF | I don't object | 22:54 |
| JayF | but again, I don't fully understand when/where an operator would input this slug | 22:54 |
| TheJulia | ironic.conf to pin out versions | 22:54 |
| TheJulia | one minute | 22:55 |
| JayF | yeah, so you pin the objects to the previous release | 22:55 |
| JayF | start upgrading | 22:55 |
| JayF | remove the pin, yeah? | 22:55 |
| JayF | for rolling upgrades | 22:55 |
| TheJulia | yeah | 22:55 |
| JayF | you shook the dust free | 22:55 |
| TheJulia | pin_release_version | 22:55 |
| iurygregory | I'm ok with this approach =) | 22:56 |
| TheJulia | Then +2+A's seem in order if there are no issues | 22:57 |
| JayF | in this same vein; it doesn't matter if the release_mappings are in place correctly for bugfix releases | 22:57 |
| JayF | because the cat is already out of the proverbial bag there | 22:57 |
| TheJulia | you can run the version just fine | 22:57 |
| JayF | but we should still be doing it | 22:57 |
| TheJulia | it is when you go to upgrade | 22:57 |
| JayF | specifically a rolling upgrade | 22:57 |
| TheJulia | any upgrade, the new version needs to know about the mappings | 22:58 |
| TheJulia | the issue is we break the master branch immediately upon release if we don't do the mappings before the overall cycle stable branch | 22:58 |
| JayF | I'm saying, in your change, I suspect e.g. the node wasn't that version when 21.2 was cut | 22:58 |
| TheJulia | JayF: I literally went to the bugfix branches and copied out the "master" state | 22:58 |
| JayF | I shoulda known you'd leave no i undotted | 22:59 |
| JayF | iurygregory: you wanna +2A https://review.opendev.org/c/openstack/ironic/+/876786 | 22:59 |
| * TheJulia thinks that is a compliment :) | 22:59 | |
| JayF | well you're literally cleaning up details that I think, with my PTL hat on, I was responsible for | 23:00 |
| JayF | having a higher attention to detail than me is not super difficult :D lol | 23:00 |
| iurygregory | JayF, doing now =) | 23:00 |
| * TheJulia shrugs, it is a team effort | 23:00 | |
| JayF | sure, but I wish there was a J in team :P | 23:00 |
| TheJulia | ... Unfortuantely I don't have a grumpy cat to live stream at you for that ;) | 23:01 |
| JayF | I have a grumpy dog and a grumpy cat, my AI made them :D | 23:01 |
| TheJulia | anyhow, I think I'm going to call it a day, my brain is hello | 23:01 |
| JayF | you got to meet them yesterday lol | 23:01 |
| TheJulia | hehehehe | 23:01 |
| TheJulia | s/hello/jello/ | 23:01 |
| JayF | case-in-point | 23:02 |
| JayF | have a nice day :D see you tomorrow or Thurs :D | 23:02 |
| TheJulia | o/ | 23:03 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!