| TheJulia | JayF: physical_network I think needs to match the physical network name as recorded in neutron so networking-baremetal can reconcile it | 00:39 |
|---|---|---|
| TheJulia | rm_work[m]: JayF: https://github.com/openstack/networking-baremetal/blob/master/networking_baremetal/agent/ironic_neutron_agent.py#L238-L239 | 01:22 |
| JayF | I will either review a change by somebody to update that doc with the specifics, or I'll take care of pushing the update in the morning if it hasn't been done | 01:33 |
| rm_work[m] | hmm, yeah that tracks but I feel like it also goes directly to some ovn mapping (which i don't see there? unless I am skimming badly) | 01:49 |
| Amarachi_O | Good morning Ironic! | 05:46 |
| freemanboss[m] | Good morning Amarachi_O: | 06:42 |
| ayo_ | Good morning Amarachi_O | 06:48 |
| Amarachi_O | Please have anyone been able to successfully deploy nodes | 06:48 |
| freemanboss[m] | <Amarachi_O> "Please have anyone been able..." <- Yes | 07:22 |
| Amarachi_O | Please how did you go about it | 07:24 |
| freemanboss[m] | where are you stuck at? | 07:24 |
| rpittau | good morning ironic! o/ | 07:44 |
| rpittau | Hashaam[m]: hi! We're going to reply to you soon, in the meantime I suggest to have a look at bifrost documentation | 07:46 |
| Amarachi_O | freemanboss[m]: I am not stuck really I just want to confirm if the process I went through was the right one, after enrolling the nodes, did you use /bifrost-cli deploy / tmp/baremetal. json | 07:50 |
| Amarachi_O | -e @baremetal-install-env. json or was there another process in between | 07:50 |
| Amarachi_O | freemanboss[m]: and what is the next thing to do after that I didn't see | 07:51 |
| rpittau | Hashaam[m]: I can't actually find your email, did you send it to both masghar and myself? | 07:51 |
| freemanboss[m] | if the baremetal.json is present yes thats the file else use baremetal-inventorfy.json | 07:51 |
| freemanboss[m] | you are using ansible or cli? | 07:52 |
| Amarachi_O | Cli | 07:52 |
| freemanboss[m] | alright | 07:53 |
| Amarachi_O | Are you using cli or ansible | 07:53 |
| freemanboss[m] | I used both | 07:53 |
| freemanboss[m] | The process is simple and direct it's just that it seems hard to get/understand and apply. | 07:54 |
| Amarachi_O | Okay I will try that, so now that you have deployed the hardware what next are we meant to do ? | 07:55 |
| freemanboss[m] | First thing to avoid unnecessary issues is to create an ssh key pairs | 07:56 |
| freemanboss[m] | ssh-keygen -t rsa -b 4096 -f ~/.ssh/id_rsa | 07:56 |
| Amarachi_O | Oh okay I ran into that issue initially but I was able to create the ssh key pairs | 07:57 |
| freemanboss[m] | Then prepare to run ironic in a test mode ./bifrost-cli testenv... (full message at <https://matrix.org/oftc/media/v1/media/download/ASjKlJ0e6dIIpze_Yqsx0hMVaOV9KN5jpOwkXDzSW9COANYZl0mWj6ou5KgMIV92zFc49CQ10Uh7Ahx4nXmz9aNCeWGY7ytgAG1hdHJpeC5vcmcvT2FxeEtDSHhrc1hQVkFIdktYT2dVYkRw>) | 07:58 |
| freemanboss[m] | Amarachi_O: Great good | 07:58 |
| freemanboss[m] | Amarachi_O: Write up check your mail or the previous conversation. | 07:59 |
| freemanboss[m] | After the writeup rpittau: will tell us what next | 07:59 |
| Amarachi_O | Okay no problem thank you so much! | 08:00 |
| freemanboss[m] | Amarachi_O: You're welcome | 08:00 |
| rpittau | freemanboss[m]: don't forget to submit your bifrost experience as second contribution in the outreachy website, when you're ready :) | 08:00 |
| freemanboss[m] | w | 08:01 |
| freemanboss[m] | rpittau: yes I will do that as soon as possible | 08:01 |
| Amarachi_O | rpittau okay are we still to send that via email ? Or just to submit it on the website | 08:02 |
| rpittau | Amarachi_O: what really matters is the submission on the outreachy website | 08:03 |
| freemanboss[m] | adeolaadesina[m]: Yes it's doing the same thing. But I think the /tmp/baremental.json is only created when you use the script/setup.sh. | 08:04 |
| Amarachi_O | rpittau: okay thank you | 08:04 |
| freemanboss[m] | When ./bifrost-cli is used 2 files got created in the working directory which are baremetal-nodes.json and baremetal-inventory.json and also ansible too has its created files too | 08:05 |
| freemanboss[m] | adeolaadesina[m]: Exactly | 08:06 |
| opendevreview | Michal Nasiadka proposed openstack/bifrost unmaintained/2023.1: Switch to unmaintained/2023.1 https://review.opendev.org/c/openstack/bifrost/+/945542 | 08:31 |
| ayo_ | Amarachi_O: what are we using the key pair for ? | 10:18 |
| freemanboss[m] | ayo_: It's to establish the deployment. Ironic needs to connect with the host | 10:20 |
| Amarachi_O | ayo_: If you don't include it in the deployment phase you will run into an error | 10:22 |
| ayo_ | Thank you freemanboss[m] Amarachi_O | 10:32 |
| adavize_ | freemanboss[m] : I'm still stuck at after installing ironic. "./bifrost-cli install --testenv" runs fine but after that I can't seem to get any other thing to work. | 11:21 |
| adavize_ | I've followed the documentation steps thoroughly and I don't know if I am missing anything. I'm trying to run Bifrost on Ubuntu wsl | 11:26 |
| freemanboss[m] | <freemanboss[m]> "Then prepare to run ironic in..." <- > <@freemanboss:matrix.org> Then prepare to run ironic in a test mode ./bifrost-cli testenv... (full message at <https://matrix.org/oftc/media/v1/media/download/AVHjEji28o_vzE9KrrZqUdO9OxDOOjGv_3UV0nkd2-U7ZGffPPyqklXN11-5kRUoodsBL_9hFiWin_EPEoe3aSFCeWGk9XogAG1hdHJpeC5vcmcvdFN6eVdEakNuZmNOSVpycnBOY05IUEZv>) | 11:28 |
| freemanboss[m] | <freemanboss[m]> "First thing to avoid unnecessary..." <- > <@freemanboss:matrix.org> First thing to avoid unnecessary issues is to create an ssh key pairs... (full message at <https://matrix.org/oftc/media/v1/media/download/AWBvpnKzEipS1FZJi9h7Q5xrtBwuqQabyDM8qhI0lTpSsmFam8jl4mzvFvrRW6_N4OVKaWJLwtv1RMcKIA5fFr1CeWGk_9ogAG1hdHJpeC5vcmcvYU5PY09sd1JBWm1zT0NpZ1RPU0R2eGlr>) | 11:29 |
| Amarachi_O | adavize_: what error are you getting | 11:32 |
| ps_adavize | when i run enroll command "TASK [ironic-enroll-dynamic : Wait for the node to finish cleaning] ****************************************************FAILED - RETRYING: [testvm1 -> localhost]: Wait for the node to finish cleaning (45 retries left). | 11:32 |
| ps_adavize | FAILED - RETRYING: [testvm2 -> localhost]: Wait for the node to finish cleaning (45 retries left)." | 11:32 |
| opendevreview | cid proposed openstack/ironic master: docs: Clarify physical_network should be a name https://review.opendev.org/c/openstack/ironic/+/945562 | 11:33 |
| Amarachi_O | ps_adavize: what is the line of code you use to run it | 11:34 |
| adavize_ | Amarachi_0 : "./bifrost-cli enroll baremental-inventory.json" | 11:35 |
| freemanboss[m] | ps_adavize: ./bifrost-cli node list | 11:35 |
| freemanboss[m] | Send the results | 11:35 |
| ps_adavize | freemanboss[m] : https://postimg.cc/njFhZh5C after running the enroll command. | 11:38 |
| freemanboss[m] | freemanboss[m]: > <@freemanboss:matrix.org> ./bifrost-cli node list | 11:39 |
| freemanboss[m] | > Send the results | 11:39 |
| freemanboss[m] | ps_adavize: | 11:39 |
| freemanboss[m] | If you can just follow the processes I highlighted already I don't think you'll face any issue | 11:40 |
| ps_adavize | ./bifrost-cli node list -->>>>>>>> "usage: Bifrost CLI [-h] [--debug] {testenv,install,enroll,deploy} ... | 11:40 |
| ps_adavize | Bifrost CLI: error: argument {testenv,install,enroll,deploy}: invalid choice: 'node' (choose from 'testenv', 'install', 'enroll', 'deploy')" | 11:40 |
| ps_adavize | freemanboss[m] : https://postimg.cc/5jHMzzwf | 11:41 |
| freemanboss[m] | ps_adavize: Ohhh sorry | 11:42 |
| freemanboss[m] | baremetal node list | 11:42 |
| ps_adavize | freemanboss[m] : I actually followed the steps you outlined in the former post | 11:42 |
| freemanboss[m] | ps_adavize: You started each process all over again? | 11:42 |
| ps_adavize | freemanboss[m] : https://postimg.cc/NKBVd2r9 | 11:44 |
| ps_adavize | yes I ran exactly that process | 11:44 |
| ayo_ | Yep they enrolled ps_adavize | 11:45 |
| ps_adavize | ayo_ : what about the errors showing up? saying waiting for node to finish cleaning ? | 11:47 |
| ayo_ | Cleaning up would most likely be for redeployed nodes | 11:47 |
| freemanboss[m] | ps_adavize: You can also validate it to see what's happening | 11:48 |
| ayo_ | Have you deployed before? ps_adavize ? | 11:49 |
| freemanboss[m] | ps_adavize: After the error that's when you change the state to manageable right? Just enroll again since you did that already | 11:50 |
| ps_adavize | freemanboss[m] : https://postimg.cc/vg0fwjPw the validate output | 11:50 |
| ps_adavize | ayo_ i haven't deployed yet | 11:51 |
| freemanboss[m] | ps_adavize: Just restart from the process I highlighted | 11:51 |
| freemanboss[m] | Make sure you follow step by step | 11:51 |
| ps_adavize | freemanboss[m]: ayo_ what system are you using?, I am using ubuntu on wsl | 11:53 |
| ayo_ | Centos | 11:55 |
| freemanboss[m] | ps_adavize: That error I saw there I don't think it's system based you can try enrolling again if you think you want to tackle that issue first | 11:56 |
| freemanboss[m] | I use Ubuntu | 11:56 |
| Amarachi_O | I am using Ubuntu on a vm though but I don't think that would be the issue since you have gone this far | 11:57 |
| freemanboss[m] | ps_adavize: If it's system based error you shouldn't get into enrollment there should have been clear issue during installation | 11:57 |
| freemanboss[m] | Amarachi_O: Exactly | 11:58 |
| ps_adavize | okay | 11:58 |
| ps_adavize | freemanboss[m] : can you run baremetal driver list, let me see what shows up in yours | 11:58 |
| freemanboss[m] | ps_adavize: My system is off | 12:00 |
| ps_adavize | also are you using the stable/2024.2 branch freemanboss[m] Amarachi_0 | 12:08 |
| freemanboss[m] | ps_adavize: I used master but that too will work fine | 12:09 |
| freemanboss[m] | Those are not real issue | 12:09 |
| Amarachi_O | ps_adavize: yes | 12:12 |
| ps_adavize | let me try to delete everything and start again, although i've done that severally | 12:14 |
| freemanboss[m] | ps_adavize: Don't delete anything just start again and be cautious | 12:17 |
| freemanboss[m] | What's the resources you're using? Ram? CPU? | 12:17 |
| ps_adavize | freemanboss[m] that's what I have been doing for 2 days now, 24 gig ram, 2 cores. I also tried on a 16gig, 6core | 12:19 |
| freemanboss[m] | ps_adavize: Alright it's enough resources just follow the processes and make sure you've ssh key pairs | 12:20 |
| ps_adavize | apart from git and python3 is there any other prerequisite that you installed manually? | 12:20 |
| ps_adavize | freemanboss[m] at what point do i create the ssh key pairs | 12:20 |
| Amarachi_O | When you are deploying | 12:21 |
| Amarachi_O | Before you deploy actually | 12:21 |
| ps_adavize | my own problem is still at the enroll command | 12:22 |
| freemanboss[m] | First thing to avoid unnecessary issues is to create an ssh key pairs... (full message at <https://matrix.org/oftc/media/v1/media/download/AThkkQL7LbYNlbrkHvhvx1pmjuG2CuI5iKvm4hyTb2P_HCzVgPZ_rWK1dLLr4-SdvbExakFCLOieFaZUOoQ-qGZCeWGoBL2wAG1hdHJpeC5vcmcvU0psclhnY2lIa2NYaGNaY1RUU2JsZlV4>) | 12:22 |
| freemanboss[m] | After installing git and python3 then changing dir to bifrost | 12:22 |
| freemanboss[m] | Just follow these steps | 12:22 |
| Amarachi_O | freemanboss[m]: thanks for this | 12:23 |
| ps_adavize | Thankyou freemanboss[m] I'm currently trying it out again will get back on the result | 12:27 |
| freemanboss[m] | ps_adavize: Alright no problem | 12:28 |
| Amarachi_O | @freemanboss[m]: please what are the steps for using ansible aswell | 12:34 |
| freemanboss[m] | Amarachi_O: Get that done first or you faced any error? | 12:36 |
| Amarachi_O | Get what done first sorry? | 12:37 |
| freemanboss[m] | When using ansible just follow the docs and the enroll-dynamic.yaml and deploy-dynamic.yml is the substitute for enroll and deploy as in CLI | 12:37 |
| freemanboss[m] | Amarachi_O: Ohh sorry | 12:37 |
| Amarachi_O | Okay thanks I will try that | 12:39 |
| TheJulia | good morning | 13:01 |
| cid | \o | 13:02 |
| cid | freemanboss[m], could you record any instructions in the doc that you encountered didn't work and you had to do it a different way, as well as any omissions, in an etherpad https://etherpad.opendev.org/ so that you can just refer your peers, as well as to help us know how to update the docs later on. This also applies to anyone else that has encountered a problem and later fixed it, they can update the pad. | 13:02 |
| freemanboss[m] | cid: Yeah, I'll do that. | 13:04 |
| cid | Great, \o/. Thanks. | 13:05 |
| freemanboss[m] | cid: my pleasure | 13:08 |
| cardoe | JayF: definitely poke me on some of the network model stuff cause that's where we are battling as well. skrobul will be interested. there's someone else on my team that I'm poking to stay on IRC that's gonna be my point person. | 13:12 |
| TheJulia | So Jay, Myself, and johnthetubaguy had a long chat on ?Monday? and I started drafting a spec regarding some of the network issues specifically as it relates to nova and figuring out/managing how to sort portgroup and bonds, as well as influencing actual device selection. Does that happen to be relating to your interests cardoe ? | 13:17 |
| cardoe | Yes | 13:17 |
| cardoe | actual device selection is the big item for us right now. | 13:18 |
| TheJulia | Checkout https://etherpad.opendev.org/p/VNdieUWeejYChegDwbHh is the high level idea | 13:18 |
| cardoe | So many irons in the fire. | 13:18 |
| TheJulia | indeed | 13:18 |
| cardoe | I really need physical_network on vxlan or another type that's "l2vni". I don't know how the neutron folks want to go with that. | 13:18 |
| TheJulia | cardoe: cross-project session perhaps? | 13:19 |
| cardoe | yeah. except I need open minds / eyes at that session | 13:19 |
| TheJulia | so the question is how do we breakdown that barrier | 13:20 |
| cardoe | I understand that the current view of vxlan in neutron is that it's an overlay on top of bare metal networks. | 13:20 |
| TheJulia | if we invite them to us, I'm not sure we'll get traction but those interested in crossing their project's boundry would be the ones to talk to | 13:20 |
| cardoe | But in my case, it's the underlay. | 13:20 |
| cardoe | So roughly. I've got boxes that have 4 ports of attachment. There's two physically separate fabrics that those ports live on. | 13:21 |
| cardoe | I want all 4 ports to exist in Ironic (and Neutron) because I want to generate cloud-init metadata for them. | 13:22 |
| cardoe | As discussed above (and in the networking doc), ironic selects the port based on physical_network. But vxlan doesn't have a physical_network field. | 13:23 |
| TheJulia | oh | 13:23 |
| TheJulia | heh | 13:23 |
| cardoe | And I'm not sure if my approach is 100% correct but I'll describe it real quick if it helps. | 13:24 |
| cardoe | or I've got a few to sync talk if that would be better and then I'll share the design write up. | 13:24 |
| TheJulia | so, I think you should read the etherpad real quick | 13:25 |
| TheJulia | but give me a couple minutes and I'll start a call so we can chat | 13:25 |
| TheJulia | https://meet.google.com/hrg-utbx-qqd | 13:28 |
| TheJulia | I'll be back in ~30 seconds, grabbing coffee from the maker | 13:28 |
| * TheJulia begins Round two of Caffination | 13:29 | |
| ps_adavize | freemanboss[m] this is the error I'm getting at the end when I try to run the enroll command https://pastebin.com/Na6EpNyF | 13:30 |
| ps_adavize | so I just spin up an ec2 instance on aws and everything worked fine, I don't know if it has something to do with me using wsl? rpittau cid masghar | 13:33 |
| ayo_ | Or you’ll try a vm? | 13:43 |
| Amarachi_O | I think a vm will work fine | 13:43 |
| satoshi | ps_adavis: I don't know the root cause of this error. | 13:43 |
| satoshi | "last_error": "During sync_power_state, max retries exceeded for node 878c3113-0035-5033-9f99-46520b89b56d, node state None does not match expected state 'power off'. Updating DB state to 'None' Switching node to maintenance mode. Error: IPMI call failed: power status. | 13:43 |
| satoshi | Can you check the power state of the node with something like baremetal node show node-name/node-id? | 13:43 |
| ayo_ | I can’t say if there’s a dependency wsl might not be having ps_adavize | 13:44 |
| rpittau | ps_adavize: it's probably wsl | 13:50 |
| opendevreview | Anton proposed openstack/networking-baremetal master: Clean up baremetal agents on node delete https://review.opendev.org/c/openstack/networking-baremetal/+/945500 | 13:52 |
| ps_adavize | so does it mean wsl can't run bifrost? or maybe it's a configuration i need to fix? rpittau | 13:52 |
| rpittau | ps_adavize: I'm not sure how it is now, but I used a lot of wsl in the past, long time ago, I've never got a good experience with complex tasks/operations, always better to have a VM for that | 13:55 |
| satoshi | Just a thought for the bifrost documentation improvement. Adding system resource requirement(or link) to the Bifronst installation page would be nice as many of us are facing errors related to it. | 13:55 |
| ps_adavize | Alright thanks rpittau | 14:31 |
| JayF | ps_adavize: wsl can probably run bifrost; absolutely cannot run bifrost testenv | 14:39 |
| opendevreview | Kaifeng Wang proposed openstack/ironic master: Fixes the issue of flat network provision when porgroup is configured https://review.opendev.org/c/openstack/ironic/+/934433 | 14:45 |
| ps_adavize | that sounds plausible, since ironic got installed and running. JayF any particular reason why wsl can't run testenv? | 14:45 |
| freemanboss[m] | <ps_adavize> "Freeman Boss this is the error I..." <- baremetal node manage testvm1 | 15:04 |
| freemanboss[m] | baremetal node provide testvm1 | 15:04 |
| freemanboss[m] | Same for testvm2 | 15:04 |
| freemanboss[m] | openstack baremetal node list | 15:05 |
| freemanboss[m] | Then send the result of this | 15:05 |
| freemanboss[m] | You're using ansible or cli? | 15:06 |
| koperg[m] | Hi, Looking for recommendations on how-to configure pin_release_version : | 15:08 |
| koperg[m] | https://docs.openstack.org/ironic/latest/configuration/config.html#DEFAULT.pin_release_version | 15:08 |
| koperg[m] | In preparation for a slurp upgrade (N+2) in Epoxy release. | 15:08 |
| hjensas | So, are we supposed to disconnect virtual media and still think the agent is accessible? snippet of logs, https://paste.opendev.org/show/beHv9TjxLPfxL4zqK4ea/ - It is disconnecting vmedia and then trying to run -> {'step': 'tear_down_agent' | 15:12 |
| hjensas | I am using sushy-emulator vmedia on openstack instance, so not like "normal" hardware. I guess the agent would still be running from memory on normal hardware. | 15:17 |
| JayF | adavize_ps: are you the same as ps_advize? If so, the answer to your question is that bifrost testenv uses virtual machines to pretend to be bare metal; I wouldn't trust that to work; or work performantly; inside the hyperv-vm that is a wsl instance | 15:33 |
| adavize_ps | okay I think I understand what you mean JayF, thank you | 15:35 |
| JayF | like, it's *possible* it could work, but it'd be slow and potentially need tweaking | 15:35 |
| JayF | not a route I'd suggest for someone new to the community | 15:35 |
| JayF | We do, however, try to ensure our unit tests will pass in WSL. If you find a unit test that won't pass in WSL; that would be a bug | 15:36 |
| adavize_ps | alright, I'll just install ubuntu Os then so I don't have all that run around, JayF. And I dont think I saw any test fail, the error messages I could understand were mostly related to memory and hardware interaction | 15:38 |
| satoshi | https://etherpad.opendev.org/p/XxXVAdF2rK9P_qLX_VYK | 15:41 |
| satoshi | I'm working on a bug and not sure how to proceed. Would be great if someone can give me a pointer about this. The details are in the link. | 15:41 |
| TheJulia | koperg[m]: so your on Epoxy-2 today, and I guess your intending to do a rolling upgrade over some period of time to newer version? | 15:55 |
| TheJulia | koperg[m]: If the docs are not clear, you should only need to pin if you intend to have a slow rolling upgrade process and have records to create/change in a mixed upgrade state. If your doing the direct jump, you shouldn't need to leverage the pin | 16:05 |
| TheJulia | Entirely depends on what exactly your doing and and that option offers a control to help sort of lock things so you can slowly roll instead of just jump, if that makes sense | 16:06 |
| TheJulia | hjensas: So... ISO contents should come from the ramdisk and operate from ram, yeah, that detach would break it I guess | 16:07 |
| TheJulia | sushy-tools does something under the hood, I just don't remember what | 16:07 |
| koperg[m] | TheJulia: Thank you. That makes more sense to me now! | 16:07 |
| TheJulia | koperg[m]: patches are always welcome to revise text strings and docs to make things more clear :) | 16:08 |
| hjensas | TheJulia: yeah, the sushy-emulator rebuilds the openstack instance on attach/detach - so the agent running in ram is gone as soon as that detach happens. This is metal3+fast-track enabled a well. | 16:13 |
| hjensas | TheJulia: anyhow, I'll try to dig into it more tomorrow. Might be an option to hold-off on sushy-emulator rebuild of the instance until a power cycle/reset/off action happens. | 16:15 |
| TheJulia | Wouldn't be a bad idea | 16:15 |
| TheJulia | Anyway, have a great evening! | 16:15 |
| *** iurygregory_ is now known as iurygregory | 17:26 | |
| iurygregory | TheJulia, hey o/, by any chance you know how can I check the privilege level of the user someone is using in the ipmitool command? | 17:27 |
| TheJulia | iurygregory: an API consumer, or ironic, or someone doing it externally/manually ? | 17:30 |
| TheJulia | iurygregory: asking to help orient/understand the question | 17:32 |
| iurygregory | I think in this case the customer created a custom user, when I checked the logs the ipmitool had ocpadmin as the user in the -U option | 17:32 |
| iurygregory | so I'm thinking they might have provided wrong privilege level or something.. | 17:33 |
| iurygregory | since they cant retrieve the power status (even after providing different cipher suites -C 3 and -C 17) | 17:33 |
| TheJulia | hmmmm | 17:34 |
| iurygregory | so I'm wondering if there is a command they can use to check that.. | 17:34 |
| iurygregory | or maybe the machine is weird and doesn't like normal cipher suites... | 17:35 |
| TheJulia | so, if memory serves, there can be differnet privilege levels, we by default assume ADMINISTRATOR | 17:35 |
| iurygregory | if you want I can share the thread on slack | 17:35 |
| iurygregory | yeah, ironic will issue the command with ADMINISTRATOR | 17:35 |
| TheJulia | that is with -L | 17:35 |
| iurygregory | yup | 17:36 |
| TheJulia | sure | 17:36 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: WIP: Trait based port selection and dynamic portgroups https://review.opendev.org/c/openstack/ironic-specs/+/945642 | 17:47 |
| TheJulia | JayF: per discussion ^ | 17:48 |
| JayF | Would be nice to get https://review.opendev.org/c/openstack/ironic-python-agent/+/941714 + https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/943348 landed for satoshi before his internship is over | 17:50 |
| TheJulia | JayF: I think the dividing line along the networking around the task.driver.network interface allows us to delienate cleanly, just pure standalone users will need a network interface, perhaps mercury or something else | 17:58 |
| TheJulia | Jay: copyright header is wrong unless the year is 2015 and satoshi works for Rackspace I’m stuck in traffic on my phone at the moment so can’t easily add the comment | 18:50 |
| JayF | lolsob | 18:55 |
| JayF | I went ahead and proxied that feedback | 18:57 |
| JayF | I was just looking through some code with cid that /actually was/ written at Rackspace in 2014/2015 lol | 19:01 |
| * JayF was relieved to find that our api app uses real threads to do the work | 19:04 | |
| TheJulia | JayF: no need to lolsob, I just notice things like that | 19:05 |
| JayF | oh, it's just a funny thing to see generally | 19:06 |
| JayF | and one of those "of course the code is good but $openstack-paperwork-reasons means -1" -- which frankly I've been on the giving side of more than once | 19:07 |
| TheJulia | Yeah, sucks. Sometimes I’ve just fixed it | 19:08 |
| JayF | it's all part of what makes openstack openstack | 19:09 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent-builder master: Create Podman DIB Element for container-based cleaning https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/943348 | 19:17 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent master: WIP Ironic Config Intergration for IPA https://review.opendev.org/c/openstack/ironic-python-agent/+/944951 | 19:19 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent master: Add ContainerHardwareManager https://review.opendev.org/c/openstack/ironic-python-agent/+/941714 | 19:21 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic master: WIP Ironic Container Cofig for IPA Integration https://review.opendev.org/c/openstack/ironic/+/944796 | 19:25 |
| satoshi | TheJulia: Thanks for the review. I removed the line. | 19:27 |
| TheJulia | satoshi: another question on the change :) | 21:47 |
| TheJulia | I guess I'm also trying to understand which change is first, and why are they not stacked?! | 21:49 |
| JayF | TheJulia: the IPA-B and ContainerHardwareManager are batch 1; the dynamic config from conductor is batch 2 | 22:05 |
| JayF | trying to ensure he can fully deliver a thing before his fellowship is up, even if the other two are in review a bit longer | 22:05 |
| TheJulia | I get that part, but it looks like the two ipa patches should be stacked together | 22:05 |
| JayF | yeah, you're right I think | 22:06 |
| JayF | but if we land the others it becomes moot, at least | 22:06 |
| opendevreview | Jay Faulkner proposed openstack/ironic-python-agent master: Remove eventlet greendns patching https://review.opendev.org/c/openstack/ironic-python-agent/+/945043 | 22:11 |
| JayF | https://review.opendev.org/q/topic:%22ironic-container-steps%22 I put all of satoshi's outstanding ContainerHWM + config integration patches in this topic | 22:12 |
| TheJulia | to help inform how often I should be looking, when does the session end? | 22:12 |
| TheJulia | session may be the wrong word | 22:12 |
| TheJulia | hmm curious how hard it will be for us to remove eventlet monkey patching entirely from ipa | 22:15 |
| JayF | it breaks things to remove the one remaining bit | 22:15 |
| JayF | cid and I are working on IPA eventlet removal, now that better docs exist and it's not like wandering in the wildernessw | 22:15 |
| JayF | targetting replacing the wsgi server as a required first step | 22:15 |
| JayF | then maybe "really really hope we can just stop patching" is the only remaining step :D | 22:15 |
| TheJulia | Yeah, I glanced at some of the stuff the other day and while a lot, it seems like good progress is being made overall | 22:16 |
| JayF | I can't find my doc on when the cohort ends for mlh :| | 22:16 |
| TheJulia | JayF: can't we just ask satoshi ? ;) | 22:20 |
| JayF | sure; but I like knowing answers to questions I'm supposed to know lol | 22:21 |
| JayF | my usual problem is happening: I know it's in my notes. Which notes? 🤷♂️ | 22:22 |
| TheJulia | In stack 47A, under subheading 3 in row F, of building G | 22:22 |
| JayF | oh hell I left that building in my other laptop | 22:23 |
| TheJulia | wait?! *in* | 22:23 |
| TheJulia | oh my... | 22:23 |
| TheJulia | was that the VR building?! | 22:23 |
| TheJulia | There you go! | 22:23 |
| JayF | alice in kernelspace | 22:23 |
| TheJulia | Ring 0? | 22:23 |
| TheJulia | Speaking of changes, I'd <3 my network simulator stuff to get reviews, its just devstack stuffs for humans to do things | 22:32 |
| satoshi | MLH fellowship ends on April 18. | 22:32 |
| TheJulia | satoshi: Thanks! | 22:32 |
| TheJulia | satoshi: do you happen to understand what I mean by stacking the patches? | 22:32 |
| satoshi | I haven't done that before but I think I can try. | 22:33 |
| JayF | basically you want your branch to have both commits on it, in the right order | 22:34 |
| JayF | and when you run "git review" it'll ask if you mean to create/update two patches | 22:34 |
| TheJulia | Yes, and you can use "git review -x" to place one after the other | 22:34 |
| TheJulia | (if you haven't changed them) | 22:34 |
| JayF | the easiest way to get to this point IME: `git review -d $firstchange`, `git review -x $secondchange`, git review | 22:34 |
| * TheJulia confirms "this is the way" | 22:34 | |
| JayF | (-d = download, -x = cherry pick FWIW) | 22:35 |
| JayF | there's also -X which cherry picks, one puts a "Cherry picked from SHA" message in the commit | 22:35 |
| JayF | and I can never remember which is which without checking manpages | 22:35 |
| TheJulia | I've done it enough that I know -X is with the commit message change | 22:36 |
| JayF | it's like a USB port for me | 22:37 |
| JayF | whichever one I use the first time is not the one I wanted | 22:38 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent master: Ironic Config Intergration for IPA https://review.opendev.org/c/openstack/ironic-python-agent/+/944951 | 22:54 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic master: Ironic Container Cofig for IPA Integration https://review.opendev.org/c/openstack/ironic/+/944796 | 23:00 |
| satoshi | I think I stacked right for IPA-B. I took off WIP from the second batch. | 23:02 |
| TheJulia | IPA-B can't be stacked with the other patches | 23:04 |
| TheJulia | but, I think you had a depends-on tag in the commit message which is good and serves a similar purpose for cross-repo items | 23:04 |
| satoshi | Yeah, Jay asked me to add the tag. | 23:06 |
| satoshi | I don't get this. Do I need to change something? | 23:06 |
| satoshi | > IPA-B can't be stacked with the other patches | 23:06 |
| opendevreview | Merged openstack/ironic-python-agent-builder master: Create Podman DIB Element for container-based cleaning https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/943348 | 23:06 |
| TheJulia | Well, there you go, you don't need the tag now | 23:07 |
| TheJulia | :) | 23:07 |
| TheJulia | nor need to stack it, focus on the IPA patches now, not IPA-B :) | 23:07 |
| TheJulia | you can only stack patches inside of the same repository | 23:07 |
| satoshi | Oh yes sorry I mean IPA. | 23:07 |
| satoshi | I mixed up IPA with IPA-B | 23:08 |
| TheJulia | for things crossing outside of a single repository, we have a notation for commit messages "Depends-On" or "Needed-By" | 23:08 |
| TheJulia | Depends-On is a firm requirement, needed-by is for humans | 23:08 |
| TheJulia | no worries! | 23:08 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!