| kubajj | Good morning everyone o/ | 08:19 |
|---|---|---|
| mmalchuk | morning | 08:20 |
| ravlew | Good morning ironic | 09:33 |
| ravlew | I'm looking for second core review for Bifrost log rotation bug fix: https://review.opendev.org/c/openstack/bifrost/+/893195 | 09:34 |
| janders | Good morning Ironic o/ | 11:14 |
| janders | recently we ran into an issue with managing iLO5 BMCs with Ironic running in FIPS compliant mode. TLS versions on iLO5 (even with latest firmware) seem too low. I created a bug for this: https://bugs.launchpad.net/ironic/+bug/2034191 | 11:15 |
| janders | I wonder if we have anyone from HP here who could look into this with us? Thanks in advance. | 11:15 |
| iurygregory | good morning Ironic | 11:38 |
| masghar | Good morning ironic o/ | 11:39 |
| masghar | I have a general question: In ironic, why do we have a separate node ID and node UUID? I am guessing node ID was added first and node UUID was a later addition, and we probably didn't want to break old database keys (so we didn't remove node ID). But what is special about a UUID (I mean why add it later)? | 11:41 |
| iurygregory | uuid will help to unique identify the resources you are looking for, it will make easier to search logs etc (probably something like this) | 12:24 |
| TheJulia | masghar: the node's id value is the primary key in the database,but not meant to ever be exposed to the user | 13:06 |
| TheJulia | I backported the fedora removal to the first branch, since we need to nuke it across branches it seems. A relatively quick review for stable cores https://review.opendev.org/c/openstack/bifrost/+/893674 If in agreement, I'll backport similiarly | 13:09 |
| iurygregory | +2 | 13:37 |
| TheJulia | since it just job config, we can likely workflow it as well, but we can also wait | 13:40 |
| * TheJulia wonders how many people are still on PTO/vacation or spending time in a nice padded room | 13:40 | |
| iurygregory | yeah, a lot of folks on PTO :D | 13:41 |
| iurygregory | I will +W, don't think JayF will kill me for doing this :D | 13:41 |
| TheJulia | stevebaker[m]: JayF: looks like ngs is failing due to etcd :\ | 13:48 |
| zigo | Hi there! Can we have a release for proliantutils, containing the latest commit, that switches to lextudio ? | 14:00 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent stable/2023.1: preserve/handle config drives on 4k block devices https://review.opendev.org/c/openstack/ironic-python-agent/+/893736 | 14:01 |
| zigo | With the current form of things, it's impossible to release in a coherent way ... | 14:01 |
| zigo | The Debian package python3-pysnmp4 will conflict with python3-pysnmp-lextudio, so I can't install the build-dependencies ... | 14:01 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent stable/zed: preserve/handle config drives on 4k block devices https://review.opendev.org/c/openstack/ironic-python-agent/+/893558 | 14:01 |
| iurygregory | zigo, we need the hpe folks to release if I remember correctly | 14:02 |
| zigo | :/ | 14:03 |
| zigo | And they're not doing it?!? | 14:03 |
| zigo | I will package the tip of master, adding a tag myself then. | 14:03 |
| iurygregory | let me check who is in the release config for it | 14:04 |
| zigo | Cheers. | 14:06 |
| iurygregory | ok, there is no config there, so I think the HPE folks are responsible for doing it | 14:06 |
| TheJulia | yeah, it is all manual tagging for them | 14:09 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent stable/yoga: preserve/handle config drives on 4k block devices https://review.opendev.org/c/openstack/ironic-python-agent/+/893738 | 14:09 |
| zigo | Have you guys made sure nobody in OpenStack continues to use the legacy pyasn1 / pysnmp ? | 14:11 |
| zigo | It'd be otherwise not co-installable ... | 14:11 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent stable/xena: preserve/handle config drives on 4k block devices https://review.opendev.org/c/openstack/ironic-python-agent/+/893740 | 14:12 |
| iurygregory | I see some references when searching https://codesearch.opendev.org/?q=pyasn1&i=nope&literal=nope&files=&excludeFiles=&repos= | 14:14 |
| TheJulia | x/networking-hpe, vitrage look like the straglers | 14:14 |
| TheJulia | vitrage likely needs a TC review and possible de-listing | 14:17 |
| masghar | (makes sense, thanks TheJulia (in reference to my question)) | 14:28 |
| opendevreview | Merged openstack/bifrost stable/2023.1: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893674 | 15:23 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/zed: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893753 | 15:50 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/yoga: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893754 | 15:55 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/zed: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893753 | 15:55 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/yoga: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893754 | 15:56 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/xena: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893755 | 15:58 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/wallaby: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893756 | 15:59 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/victoria: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893757 | 16:00 |
| opendevreview | Julia Kreger proposed openstack/bifrost stable/ussuri: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893760 | 16:02 |
| opendevreview | Julia Kreger proposed openstack/bifrost bugfix/12.0: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893781 | 16:12 |
| opendevreview | Julia Kreger proposed openstack/bifrost bugfix/11.1: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893783 | 16:13 |
| opendevreview | Julia Kreger proposed openstack/bifrost bugfix/11.0: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893784 | 16:14 |
| opendevreview | Julia Kreger proposed openstack/bifrost bugfix/10.1: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893785 | 16:16 |
| opendevreview | Julia Kreger proposed openstack/bifrost bugfix/10.0: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893786 | 16:17 |
| opendevreview | Julia Kreger proposed openstack/bifrost bugfix/8.3: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893787 | 16:18 |
| opendevreview | Julia Kreger proposed openstack/bifrost bugfix/8.2: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893788 | 16:18 |
| TheJulia | that was... painful | 16:18 |
| JayF | Do I need to email HPE team and ask for proliantutils release? | 16:19 |
| TheJulia | likely yes | 16:20 |
| iurygregory | yeah | 16:20 |
| iurygregory | ++ | 16:20 |
| JayF | do they need to do anything w/r/t requirements/constraints/etc | 16:21 |
| TheJulia | I've approved/workflowed the anchient bifrost butfix branch changes where the CI config is already noop'ed/disabled | 16:21 |
| JayF | or literally just "tag a new release for this release"? | 16:21 |
| TheJulia | if they need to update constraints, they needed to make that call at some point before a release, we have no way of knowing (i.e. they have to evaluate it) | 16:21 |
| iurygregory | yeah, likely we would bump the version in driver-requirements | 16:22 |
| JayF | I don't trust they'll do that but it looks like it should be clean anyway | 16:22 |
| JayF | we'll bump th... yes | 16:22 |
| TheJulia | yeah, driver-reqs is independent | 16:22 |
| TheJulia | should likely do so yeah | 16:22 |
| TheJulia | since pysnmp | 16:23 |
| JayF | it's emailed | 16:23 |
| iurygregory | zigo, fyi ^ | 16:23 |
| TheJulia | question on https://review.opendev.org/c/openstack/ironic/+/889750 | 16:26 |
| JayF | basically is what I wanted | 16:28 |
| JayF | API takes names, we store UUIDs | 16:28 |
| JayF | just like we do if you change a node generally accessing it by name | 16:29 |
| TheJulia | k | 16:33 |
| opendevreview | Merged openstack/bifrost master: Replace Cirros references with generic descriptors https://review.opendev.org/c/openstack/bifrost/+/883943 | 17:05 |
| opendevreview | Merged openstack/bifrost bugfix/10.0: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893786 | 17:05 |
| opendevreview | Merged openstack/bifrost bugfix/10.1: Remove Fedora from the CI https://review.opendev.org/c/openstack/bifrost/+/893785 | 17:05 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Enable OVN CI https://review.opendev.org/c/openstack/ironic/+/885087 | 17:42 |
| JayF | https://review.opendev.org/c/openstack/nova/+/844120 I really, really like the idea of this. | 17:47 |
| opendevreview | Julia Kreger proposed openstack/ironic master: DB: Only re-query for a lock holder if we cannot lock https://review.opendev.org/c/openstack/ironic/+/893532 | 17:48 |
| JayF | Assuming nova lands it, I wouldn't hate us to go the same direction (using black for python styling) | 17:48 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Enable OVN CI https://review.opendev.org/c/openstack/ironic/+/885087 | 17:50 |
| TheJulia | yeouch, that much change is backport nightmares | 17:52 |
| TheJulia | I was going to backport the 4k fix to wallaby then saw I'd have to patch ironic-lib instead and noped out. | 17:59 |
| TheJulia | wheeeee "Image streamed onto device /dev/vda in 197.2619252204895 seconds for 2483482173 bytes. Server originaly reported 2483486720." happened in CI | 18:11 |
| TheJulia | ... I suspect stalled transport 197 seconds is a long time | 18:12 |
| JayF | it's mildly impressive it held on | 18:12 |
| TheJulia | yeah | 18:17 |
| TheJulia | so old bifrost branches, any objection if I just turn off jobs, in particular older bifrost bugfix branches already carry this, but thinking of moving this up through all bugfix branches on it | 18:28 |
| JayF | define: "this" ? | 18:32 |
| JayF | I've lost context | 18:32 |
| TheJulia | removal of fedora from CI | 18:33 |
| JayF | I thought that was a situation where we didn't have much choice | 18:33 |
| TheJulia | and bifrost branches starting in bugfix/10.1 | 18:33 |
| JayF | TheJulia: your comment re: OVN only being the dhcp server change, I'm not sure I really fully understood what that meant until now. | 19:18 |
| JayF | So it's probably safe to say that 95%+ of the complication in that PR is just CI silliness, and not representative of how OVN looks IRL with Ironic ... which matches up with all the other conversations we have had | 19:18 |
| JayF | I was just very, very confused | 19:18 |
| TheJulia | Configuration will vary, the MTU handling and and TFTP NAT stuffs is an issue, it is what it is | 19:59 |
| TheJulia | *likely* operators are not actually using it in bmaas configs, or are not making a big deal out of it | 19:59 |
| JayF | zigo: TheJulia: FYI; hpe indicated to me they have scheduled a release for Friday | 20:53 |
| JayF | zigo: TheJulia: for x/proliantutils | 20:54 |
| *** dking is now known as Guest1963 | 21:50 | |
| *** Guest1963 is now known as dking | 21:52 | |
| hjensas | TheJulia: any idea why https://paste.opendev.org/show/b92EJWG3vTNqyl6o2AnF/ is not working on Stream 9? (It's not installing ipmitool either, which is in bindep.txt as "default" i.e not bound to dpkg/rpm) | 22:00 |
| JayF | hjensas: what's the output? | 22:07 |
| JayF | hjensas: usually when bindep runs it will spit an error if there is one, then it stops processing the file | 22:07 |
| JayF | hjensas: so what you describe sounds to me like an err somewhere in the middle of the file causing everything below it (and that package) to get missed | 22:07 |
| * JayF has been troubleshooting this kinda err w/Debian Bullseye CI jobs all last week | 22:07 | |
| TheJulia | so the issue is the labeling default | 22:41 |
| TheJulia | it likely needs to go | 22:41 |
| hjensas | TheJulia: ah, it's failing xinetd | 22:58 |
| hjensas | 2023-09-05 22:05:41.600 | No match for argument: xinetd | 22:58 |
| hjensas | 2023-09-05 22:05:41.612 | Error: Unable to find a match: xinetd | 22:58 |
| opendevreview | Harald Jensås proposed openstack/ironic master: devstack bindep - [platform:rpm] https://review.opendev.org/c/openstack/ironic/+/893813 | 23:15 |
| opendevreview | Harald Jensås proposed openstack/ironic master: devstack - support tftp systemd socket activation https://review.opendev.org/c/openstack/ironic/+/893814 | 23:15 |
| opendevreview | Harald Jensås proposed openstack/ironic master: devstack - replace xinetd with systemd tftp socket https://review.opendev.org/c/openstack/ironic/+/893815 | 23:15 |
| JayF | oooh hjensas I'm going to like these patches | 23:36 |
| JayF | hjensas: left a review, something minor/specific on the bindep update, but for the systemd-activated tftp, is there a good reason we don't flip debian/ubuntu to that method, too? | 23:40 |
| JayF | hjensas: I have an ubuntu-based devstack up if there's any specific output I could put there to assist (e.g. if you need to ensure paths/filenames are the same and don't have an ubuntu box at hand) | 23:41 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!