TheJulia | JayF: so I think the ironic side test needs to set a shard key on a node and then also wire nova-compute to use that shard key | 00:51 |
---|---|---|
JayF | ack | 01:53 |
rpittau | good morning ironic! o/ | 06:54 |
opendevreview | Riccardo Pittau proposed openstack/bifrost master: Revert "Pin systemd to 252-15 for CentOS Stream 9" https://review.opendev.org/c/openstack/bifrost/+/891014 | 06:58 |
rpittau | considering that the systemd pkg with the fix is not synced yet, we should merge https://review.opendev.org/c/openstack/bifrost/+/890319 to unblock bifrost CI at least in the master branch | 06:59 |
dtantsur | rpittau: fair enough :( | 09:32 |
dtantsur | let's probably not backport it though | 09:32 |
mmalchuk | morning Ironic o/ | 09:54 |
opendevreview | Alex Welsh proposed openstack/bifrost master: Replace Cirros references with generic descriptors https://review.opendev.org/c/openstack/bifrost/+/883943 | 10:41 |
iurygregory | good morning Ironic | 11:35 |
rpittau | dtantsur: yeah, I'll keep the patches in stable branches blocked | 11:59 |
opendevreview | Riccardo Pittau proposed openstack/ironic master: [DNM][trivial] Remove todo note that was already done https://review.opendev.org/c/openstack/ironic/+/890912 | 12:10 |
opendevreview | Merged openstack/bifrost master: Pin systemd to 252-15 for CentOS Stream 9 https://review.opendev.org/c/openstack/bifrost/+/890319 | 12:51 |
opendevreview | Riccardo Pittau proposed openstack/ironic master: [DNM][trivial] Remove todo note that was already done https://review.opendev.org/c/openstack/ironic/+/890912 | 13:14 |
adam-metal3 | Hi, | 13:15 |
adam-metal3 | This command produces a non zero exit code and makes my IPA skip multipath devices when it is looking | 13:15 |
adam-metal3 | for suitable root device (as expected, IPA thinks that there is no multipath support). | 13:15 |
adam-metal3 | In reality there is multipath support and multipathd is already running, but for some reason contrarary to | 13:15 |
adam-metal3 | the commnetsm, multipathd command does not exit with 0 when the daemon is already running. | 13:15 |
adam-metal3 | The command: https://opendev.org/openstack/ironic-python-agent/src/branch/master/ironic_python_agent/hardware.py#L216 | 13:15 |
adam-metal3 | My version of multipath-tools is 0.9.0 | 13:15 |
adam-metal3 | I have removed the line in question in my tests, and everything started working without any issue, now IPA was able | 13:15 |
adam-metal3 | to deploy the user image to the multipath device. | 13:15 |
adam-metal3 | My questions are: | 13:15 |
adam-metal3 | Could we just dismiss the return value of this command? right after this command is executed multipath -ll is also run, | 13:15 |
adam-metal3 | isn't that enough to verify that something is wrong and then exclude the use of multipath devices? | 13:15 |
TheJulia | just considering multipath -ll is likely fine | 13:20 |
TheJulia | worrisome that you don't get a zero exit code, i wonder if that was changed in your distribution/version :( | 13:21 |
rpittau | adam-metal3: what's your distro ? | 13:21 |
adam-metal3 | SLES | 13:21 |
adam-metal3 | 15 SP4 | 13:21 |
rpittau | I remember we had issues with something | 13:21 |
rpittau | ah | 13:21 |
rpittau | so CS9 has 0.8.7 | 13:23 |
rpittau | ubuntu focal has 0.8.3 and jammy has 0.8.8 | 13:23 |
rpittau | smells like regression to me | 13:23 |
rpittau | or maybe a fix :D | 13:23 |
rpittau | debian bullseye has 0.8.5, but bookworm has 0.9.4 mmmmm | 13:24 |
rpittau | zigo maybe experienced some issues with multipathd in bookworm based ipa ? | 13:24 |
zigo | What package ? | 13:25 |
rpittau | multipath-tools | 13:25 |
zigo | I have no experience with multipath, so I can't tell. | 13:25 |
rpittau | ok, just curious if you saw any issue when running ipa with multipath | 13:25 |
rpittau | TheJulia: do you think we could remove that or maybe add a try-except? | 13:27 |
rpittau | adam-metal3: do you know why it's exiting with a non zero code? any error in the logs? | 13:27 |
TheJulia | we likely could just remove the check, although I think the purpose was to make sure it was already running | 13:28 |
rpittau | yeah, exactly | 13:28 |
rpittau | I remember we had some issues over there | 13:28 |
TheJulia | yeah, if the ramdisk is auto-starting it on boot, that is fine, I don't think we inherently expect that on some distros as well | 13:29 |
TheJulia | so maybe try/except is the right path | 13:29 |
opendevreview | Merged openstack/ironic stable/2023.1: Fix ks_template property to be processed only for anaconda deploy https://review.opendev.org/c/openstack/ironic/+/889360 | 13:30 |
adam-metal3 | rpittau, just a very minimal error message that I was able to reproduce via executing the command manually : "process is already running failed to create pidfile" | 13:32 |
adam-metal3 | I actually see it in the IPA log also | 13:32 |
TheJulia | so this could just be rooted in the OS starts by default or not | 13:32 |
rpittau | adam-metal3: thanks, interesting, I wonder if they changed the logic thre there, before it was just exiting with 0 | 13:32 |
TheJulia | *but* it was once in the docs and we were able to reproduce it on debian/centos that you could run it again | 13:33 |
rpittau | mmm yep | 13:33 |
TheJulia | a try/except seems like the right path in my brain | 13:33 |
rpittau | I think we should track this in launchpad also | 13:33 |
adam-metal3 | I can create the launchpad issue and the change later today or tomorrow it that is okay for you | 13:34 |
rpittau | adam-metal3: would you mind opening a bug in launchpad? would help a lot | 13:34 |
adam-metal3 | :D | 13:34 |
rpittau | :D | 13:34 |
rpittau | thanks! | 13:34 |
rpittau | and thanks TheJulia ! :) | 13:34 |
adam-metal3 | thank you folks ! I will be signing off have to run som errands, bye | 13:35 |
rpittau | bye adam-metal3 | 13:36 |
opendevreview | Merged openstack/ironic-python-agent stable/2023.1: Log the number of bytes downloaded https://review.opendev.org/c/openstack/ironic-python-agent/+/890696 | 13:40 |
TheJulia | out of curiosity, does anyone know if metal3 supports asking ironic to do a software raid config? | 14:59 |
JayF | If there's anyone who has some free time today to take a look at something with me, please let me know. I'm working on a bug. | 14:59 |
dtantsur | TheJulia: it does | 15:01 |
dtantsur | but this is the least tested part of metal3 :) | 15:02 |
TheJulia | dtantsur: ack, thanks | 15:02 |
TheJulia | JayF: what sort of bug? | 15:02 |
JayF | Dm | 15:03 |
TheJulia | ack, thanks | 15:03 |
opendevreview | Merged openstack/bifrost master: Create the log file for the disk-image-create command https://review.opendev.org/c/openstack/bifrost/+/822895 | 15:28 |
JayF | fungi: you have a few minutes to chat with TheJulia and I? | 15:53 |
rpittau | good night! o/ | 15:59 |
TheJulia | o/ | 16:01 |
TheJulia | the base token has been on context for 11 years | 16:02 |
JayF | it's crazy how old some of the code we're working on is | 16:03 |
mmalchuk | rpittau thanks for merge my change | 19:03 |
mmalchuk | rpittau is it possible merge backports till xena? I will propose backports | 19:04 |
TheJulia | merge yes, release I don't think so | 20:09 |
JayF | TheJulia: so one fun thing, oslo.messaging doesn't dep on oslo.context | 20:38 |
JayF | TheJulia: which would mean getting a fresh context and populating it would likely be too painful to backport | 20:39 |
JayF | TheJulia: I also assume e.g. just safe_cxt=object() and setting things won't work ... | 20:39 |
JayF | ...or will it... hmmm | 20:39 |
mmalchuk | good night Ironic o/ | 20:45 |
* mmalchuk have unstable internet | 20:45 | |
mmalchuk | TheJulia did you answer me? I didn't see my question in logs(( | 20:46 |
JayF | 20:09:25 TheJulia | merge yes, release I don't think so | 20:46 |
JayF | is all you missed | 20:46 |
JayF | these conversations usually work better in gerrit, fwiw | 20:47 |
mmalchuk | I remember exactly asking the question) | 20:47 |
JayF | ah, just got missed in the answer | 20:47 |
mmalchuk | so, I can propose the backports, all them will be merged but not released? | 20:48 |
* mmalchuk confused | 20:48 | |
JayF | So two things separate here: | 20:48 |
JayF | 1) Backport policy has to be met, which for bifrost I'm not sure what that looks like in practice. If in Ironic your change would be on the borderline. | 20:49 |
JayF | 2) Can we release it? Ironic can only perform releases of items still supported; basically any release made in the last 18 months | 20:49 |
JayF | so we can merge to stable/xena; but it will not be released to pypi/etc because we no longer release *anything* from that branch | 20:49 |
mmalchuk | all them mostly for CI, the extra log file | 20:49 |
JayF | yeah but you have to consider if there was a user running, say, yoga bifrost | 20:49 |
JayF | are you 100% sure every user wants this new behavior? | 20:50 |
JayF | and that it won't break existing, long running workflows | 20:50 |
JayF | those are the kinda questions you have to think about in stable/ that you don't for master changes | 20:50 |
mmalchuk | there no new, but can be used for debug purposes | 20:50 |
mmalchuk | alright, wiil prepare changes.... lets see | 20:50 |
mmalchuk | JayF too complicated) | 20:51 |
opendevreview | Maksim Malchuk proposed openstack/bifrost stable/2023.1: Create the log file for the disk-image-create command https://review.opendev.org/c/openstack/bifrost/+/891124 | 20:51 |
JayF | re: my context note; apparently ctxt at that point in oslo.messaging is a dict | 20:53 |
JayF | so it should be simple | 20:53 |
JayF | I already have evidence it's working | 20:53 |
opendevreview | Maksim Malchuk proposed openstack/bifrost stable/zed: Create the log file for the disk-image-create command https://review.opendev.org/c/openstack/bifrost/+/891125 | 20:55 |
opendevreview | Maksim Malchuk proposed openstack/bifrost stable/yoga: Create the log file for the disk-image-create command https://review.opendev.org/c/openstack/bifrost/+/891127 | 20:58 |
TheJulia | mmalchuk: tl;dr yes we can backport if the branch is still there, but xena we cannot release | 20:59 |
TheJulia | JayF: thanks for bringing more clarity | 21:00 |
mmalchuk | TheJulia thanks for clarification. but Xena is not EOL yet. please. | 21:00 |
TheJulia | JayF: awesome w/r/t context | 21:00 |
mmalchuk | Xena would be last | 21:00 |
JayF | mmalchuk: it's against openstack policy; there's literally nothing anyone outside of releases team and TC can do about it | 21:00 |
JayF | mmalchuk: and in fact; we're going to make this policy stricter soon, and rename branches at that 18 month point to "unmaintained/codename" instead of "stable/codename" so that folks like you can see at first glance what is "really supported" and what is "just a branch with fixes" | 21:01 |
mmalchuk | EM not EOL am I wrong? | 21:01 |
JayF | EM is, by policy, forbidden from being released | 21:01 |
JayF | it's only just a place for people to share patches in a branch | 21:01 |
JayF | and deploy them from git | 21:01 |
mmalchuk | oh. lets merge, but not release | 21:01 |
JayF | that's why we are renaming it to "unsupported" soon, to reflect what's actually happening | 21:01 |
JayF | yeah exactly | 21:02 |
mmalchuk | I'm ok, I need this downstream) even if branch dropped, I will build from tag | 21:02 |
* mmalchuk happy) | 21:02 | |
opendevreview | Maksim Malchuk proposed openstack/bifrost stable/xena: Create the log file for the disk-image-create command https://review.opendev.org/c/openstack/bifrost/+/891134 | 21:04 |
mmalchuk | folks, please review and merge: https://review.opendev.org/q/topic:disk-image-create-log | 21:09 |
mmalchuk | thanks a lot | 21:10 |
opendevreview | Merged openstack/python-ironicclient master: tox: Remove basepython https://review.opendev.org/c/openstack/python-ironicclient/+/890317 | 22:13 |
opendevreview | Merged openstack/python-ironicclient master: These should be indented 4 spaces to match the other things in this block https://review.opendev.org/c/openstack/python-ironicclient/+/794691 | 22:14 |
TheJulia | iurygregory: doing a python-ironicclient change for firmware_interface ? | 22:15 |
opendevreview | Merged openstack/ironic master: tox: Remove basepython https://review.opendev.org/c/openstack/ironic/+/890311 | 22:21 |
iurygregory | TheJulia, planing to handle in the next cycle probably .-. | 22:29 |
JayF | it's hard for us to sell to operators it doesn't exist if we don't have client support by release | 22:30 |
JayF | can we maybe try to slip it in towards the end of the cycle? | 22:30 |
JayF | er I mean, it's hard for us to sell it to operator if the client support doesn't exist (yet) | 22:30 |
iurygregory | yeah, I understand that operators would like to have the client support to check the info we stored in the DB | 22:33 |
iurygregory | thinking a bit here, we can probably expand to already show the resource I think | 22:38 |
JayF | yeah I'm thinking changes will be trivial | 22:39 |
iurygregory | I will try to poke a bit tomorrow | 22:39 |
JayF | tomorrow is long before the actual-it-is-done release deadline :) | 22:39 |
opendevreview | Julia Kreger proposed openstack/python-ironicclient master: Add Service Steps to client https://review.opendev.org/c/openstack/python-ironicclient/+/891140 | 23:40 |
opendevreview | Julia Kreger proposed openstack/ironic-python-agent master: WIP: Add get_service_steps logic to the agent https://review.opendev.org/c/openstack/ironic-python-agent/+/890864 | 23:56 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!