| *** ChanServ changes topic to "https://opendev.org/airship || https://wiki.openstack.org/wiki/Airship || https://review.opendev.org/#/q/projects:airship+status:open+NOT+label:Verified%253D-1+NOT+label:Workflow%253D-1+NOT+message:DNM+NOT+message:WIP" | 00:34 | |
| -openstackstatus- NOTICE: The Gerrit service at review.opendev.org is back up and running; for outage details see analysis here: http://lists.opendev.org/pipermail/service-announce/2020-October/000011.html | 00:34 | |
| airship-irc-bot | <kk6740> Yay! | 00:34 |
|---|---|---|
| *** sreejithp has joined #airshipit | 00:56 | |
| *** sreejithp has quit IRC | 01:03 | |
| *** sreejithp has joined #airshipit | 01:20 | |
| *** sreejithp has quit IRC | 01:28 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #airshipit | 04:33 | |
| *** sreejithp has joined #airshipit | 05:24 | |
| *** sreejithp has quit IRC | 05:28 | |
| *** awander has quit IRC | 05:36 | |
| *** airship-irc-bot has quit IRC | 06:40 | |
| *** airship-irc-bot has joined #airshipit | 06:40 | |
| *** sreejithp has joined #airshipit | 07:25 | |
| *** sreejithp has quit IRC | 07:31 | |
| *** avolkov has joined #airshipit | 08:07 | |
| *** roman_g has joined #airshipit | 08:10 | |
| *** sreejithp has joined #airshipit | 09:27 | |
| *** sreejithp has quit IRC | 09:32 | |
| *** SRao has quit IRC | 10:04 | |
| airship-irc-bot | <alexander.hughes> hey team, our very own @mattmceuen is presenting "Own your YAML: extending Kustomize via Plugins" in 15 minutes! be sure to check it out, https://summit.openinfra.dev/a/event/24671 | 13:31 |
| *** jamesgu has quit IRC | 13:33 | |
| mattmceuen | darn... just now realized that I somehow edited out a couple of my own slides during video editing :) #promove | 14:02 |
| airship-irc-bot | <alexander.hughes> the live demo? | 14:04 |
| mattmceuen | No, the part where I introduce plugins - in the final product, I just kinda jump into them without context | 14:05 |
| airship-irc-bot | <alexander.hughes> ah I missed that, only thing that stood out was "Live Demo" slide coming up for a second before your closing remarks | 14:05 |
| airship-irc-bot | <alexander.hughes> great presentation though, really helps me to understand kustomize better! thanks for putting this together | 14:06 |
| mattmceuen | thanks! | 14:06 |
| *** marst has joined #airshipit | 14:26 | |
| mattmceuen | easy one-liner review request: https://review.opendev.org/#/c/759066/ | 14:32 |
| airship-irc-bot | <alexander.hughes> setting records on reviews this morning | 14:33 |
| mattmceuen | This bumps the timeout of the site manifest validation to an hour, which times out sometimes now that we've added provider-specific site definitions | 14:33 |
| airship-irc-bot | <ih616h> beat me to it lol | 14:33 |
| mattmceuen | @alexander.hughes and @ih616h going for the gold | 14:33 |
| airship-irc-bot | <dwalt> mattmceuen said the magic words | 14:33 |
| mattmceuen | was it "one liner" or "broken gate"? ;-) | 14:34 |
| airship-irc-bot | <alexander.hughes> you had me at "easy" | 14:34 |
| airship-irc-bot | <dwalt> and "easy" :slightly_smiling_face: | 14:34 |
| mattmceuen | lol | 14:34 |
| mattmceuen | will remember this trick | 14:34 |
| airship-irc-bot | <alexander.hughes> but really the commit message was great, made for an easy review | 14:34 |
| mattmceuen | after all I do need one more +2 on the *easy easy easy* network catalogue changeset https://review.opendev.org/#/c/749611/ | 14:35 |
| airship-irc-bot | <alexander.hughes> triple easy, seems like a trap | 14:35 |
| * mattmceuen it's a trap | 14:35 | |
| airship-irc-bot | <dwalt> hahaha. you got me | 14:35 |
| * mattmceuen and thank you Stas | 14:36 | |
| airship-irc-bot | <kk6740> looks like fishing for reviewers :slightly_smiling_face: | 14:36 |
| airship-irc-bot | <kk6740> you put a bait with 1 liner | 14:36 |
| airship-irc-bot | <alexander.hughes> like fishing with dynamite, he has us all roped in now | 14:36 |
| airship-irc-bot | <kk6740> yea | 14:36 |
| mattmceuen | yes now that I have your attention, here's three hundred lines of YAML! | 14:36 |
| airship-irc-bot | <alexander.hughes> hahaha | 14:36 |
| airship-irc-bot | <alexander.hughes> will get to this when I get a break from meetings in ~1hr | 14:37 |
| mattmceuen | tytytyty | 14:37 |
| mattmceuen | Heads up, @alexander.hughes' talk on Airship Security is in 15min! | 14:45 |
| airship-irc-bot | <alexander.hughes> https://summit.openinfra.dev/a/event/24575 | 14:45 |
| *** SRao has joined #airshipit | 14:57 | |
| *** marst has quit IRC | 15:27 | |
| *** SRao has quit IRC | 15:28 | |
| mattmceuen | great talk @alexander.hughes, and very comprehensive | 15:28 |
| airship-irc-bot | <alexander.hughes> thanks Matt! | 15:28 |
| airship-irc-bot | <alexander.hughes> another presentation, starting right now, by Prakash and Vivek you all should check out: https://summit.openinfra.dev/a/event/24585 | 16:01 |
| airship-irc-bot | <alexander.hughes> Machine Learning at Edge Cloud | 16:01 |
| mattmceuen | +1 | 16:01 |
| *** portdirect has quit IRC | 16:19 | |
| *** portdirect has joined #airshipit | 16:20 | |
| *** portdirect has quit IRC | 16:21 | |
| *** portdirect has joined #airshipit | 16:21 | |
| mattmceuen | Hey all, question around airshipctl connectivity requirements -- the full list of connectivity needs between airshipctl and a remote site that it's deploying/managing/upgrading are: | 16:59 |
| mattmceuen | 1. redfish to all hosts (ephemeral might be the only hard requirement, but you may want to e.g. power cycle the others) | 16:59 |
| mattmceuen | 2. API server connectivity to the api server endpoint | 17:00 |
| mattmceuen | (#2 for both ephemeral and target clusters) | 17:00 |
| mattmceuen | Is there anything I'm forgetting? | 17:00 |
| airship-irc-bot | <kk6740> that’s all it seems when working with bmh envs | 17:06 |
| mattmceuen | Awesome - I couldn't come up with anything else either. Thanks @kk6740 | 17:06 |
| airship-irc-bot | <kk6740> i think there is some work around different public cloud providers, where they need to have to connect to public cloud endpoints | 17:07 |
| mattmceuen | Good point -- sorta takes the place of the redfish connectivity I suppose | 17:07 |
| mattmceuen | Different question :) when I try to `make` the AirshipUI project, I get angular dependency errors; can someone more familiar with angular/airshipui please point me in the right direction? http://paste.openstack.org/show/799261/ | 17:08 |
| *** marst has joined #airshipit | 17:19 | |
| airship-irc-bot | <mattmceuen> @mf4192 @andrew.schiefelbein does that^ error ring a bell? | 17:38 |
| airship-irc-bot | <mf4192> checking now... | 17:42 |
| airship-irc-bot | <se6518> Hello, need one more review from the core team: https://review.opendev.org/755085 | 17:44 |
| airship-irc-bot | <mf4192> @mattmceuen that's really odd, it should install everything it needs. I haven't run into this before | 17:48 |
| airship-irc-bot | <mf4192> I'll try building on a fresh vm to see if I can replicate this error | 17:48 |
| mattmceuen | Weird! Thanks - if it builds for you on a fresh VM then it's probably an issue on my side. I'll try to do the same when I get a chance, and see if I get better luck | 17:49 |
| airship-irc-bot | <mf4192> new build worked just fine | 18:01 |
| airship-irc-bot | <mf4192> if this isn't a fresh clone, you might try blowing away the airshipui/tools/node-v12.16.3 directory and letting all of that get reinstalled by make | 18:02 |
| mattmceuen | @mf4192 yep that sure fixed it :) | 18:22 |
| mattmceuen | thanks! | 18:23 |
| airship-irc-bot | <mf4192> nice, glad to hear it | 18:23 |
| airship-irc-bot | <mf4192> everything else that's broken is by design ;) | 18:23 |
| mattmceuen | lol | 18:23 |
| mattmceuen | thought you were going to say "everything else that's broken is airshipctl" | 18:23 |
| airship-irc-bot | <mf4192> lol | 18:23 |
| airship-irc-bot | <alexander.hughes> :eyes: | 18:24 |
| marst | Hello. Trivial question: is it recommended to run airship in a production-like environment? | 18:27 |
| marst | *possible | 18:27 |
| mattmceuen | Airship 1: yes. Airship 2: not quite yet, but soon :) | 18:27 |
| marst | great! thank you! :) | 18:27 |
| mattmceuen | sure thing, any time | 18:27 |
| marst | are there any hw requirements for airship2? reference architecture? | 18:27 |
| marst | I can start budgeting and buying equipment in the meantime ) | 18:28 |
| mattmceuen | Depends on what you want to run -- airship 2 runs for the most part outside the cluster itself, and orchestrates/manages things inside the cluster | 18:29 |
| mattmceuen | So the minimum requirements would be enough to run a kubernetes cluster -- pretty minimal | 18:29 |
| mattmceuen | But if you want to run telco workloads on top of that, that's where you would e.g. want to consult the CNTT recommendations for server architecture | 18:29 |
| marst | yes, the idea is to follow CNTT/OPNFV guidelines. | 18:30 |
| marst | and build something as close as possible to their RA model | 18:30 |
| marst | thank you! I'll get in touch with EdgeWG. :) | 18:30 |
| airship-irc-bot | <mattmceuen> @james.gu do you have handy a good reference for CNTT RA hardware recommendations? | 18:30 |
| airship-irc-bot | <mattmceuen> Some of the other things you may want to take into consideration @marst are: | 18:31 |
| airship-irc-bot | <mattmceuen> 1. are you planning on VNF or CNF workloads? | 18:31 |
| airship-irc-bot | <mattmceuen> 2. if CNF, do you want to run them on kubernetes inside VMs? (some CNFs don't play nicely when deployed into the same k8s cluster) | 18:32 |
| airship-irc-bot | <mattmceuen> 3. how soon do you want to be in production? If timeline allows, we'd definitely suggest aligning to Airship 2, since that's where future work will focus (along with its many advantages over Airship 1) | 18:33 |
| marst | thank you!! 1. both VNF & CNF 2. Most likely yes, k8s should be inside VMs. For baremetal CNFs there's a great paper from Intel: https://builders.intel.com/docs/networkbuilders/container-bare-metal-for-2nd-generation-intel-xeon-scalable-processor.pdf | 18:41 |
| marst | 3. Q1 next year, it's unlikely we'll have HW sooner than that. | 18:41 |
| marst | Definitely would love to run airship2. :) | 18:42 |
| marst | If you check page 9 from Intel's BMRA: 90% sure airship2 should be able to run on the same HW. | 18:44 |
| *** jamesgu has joined #airshipit | 18:52 | |
| mattmceuen | marst: thanks for sharing that, will give it a read! Agree - same HW should be just fine | 19:05 |
| marst | mattmceuen: thanks to you! Can't wait to try airship 2. | 19:09 |
| airship-irc-bot | <james.gu> @mattmceuen @marst we don't have hardware recommendation for Airship 2 in CNTT yet. | 19:16 |
| airship-irc-bot | <mattmceuen> Sure, I thought CNTT might have some "agnostic" hardware recommendations though - is that not the case? | 19:17 |
| airship-irc-bot | <james.gu> Fro Airship 1, a reference site is https://wiki.opnfv.org/display/pharos/Intel+Pod15 | 19:18 |
| airship-irc-bot | <mattmceuen> ah, gotcha - maybe that's what I was thinking of | 19:20 |
| airship-irc-bot | <james.gu> Marst, the BMRA hardware requirement seems to be pretty reasonable for Airship 1 and 2. the only concern I have is around the number of disks. The ceph cluster deployed by Airship 1 (can't speak too much yet for Airship 2) for OpenStack requires additional disks, so minimum one additional 1TB disk on top of the primary OS disks as listed on page 9, though you may find one disk per node may be too low number of Ceph OSD's for a | 19:23 |
| airship-irc-bot | production environment. | 19:23 |
| marst | james.gu: thanks a lot! Do you think we should buy a bunch of extra disks or add 3 dedicated storage nodes for ceph/rook? | 19:25 |
| marst | I'm not sure if I can get away with both. :)) | 19:26 |
| airship-irc-bot | <james.gu> marst, are you starting with Airship 2? | 19:28 |
| marst | james.gu: yes. | 19:30 |
| mattmceuen | that was what I recommended james.gu, since they're targeting Q1 next year lab entry. If they needed to be earlier than that I'd have suggested Airship1 | 19:30 |
| mattmceuen | we're targeting A2 GA by sometime 1Q as well | 19:31 |
| airship-irc-bot | <james.gu> For the CNTT reference lab, we are adding extra disks on the controller nodes. @mattmceuen do you have different recommendation for production env? | 19:34 |
| marst | james.gu: just on controller nodes? Or compute nodes too? | 19:43 |
| airship-irc-bot | <james.gu> stand corrected, both controller and compute | 19:43 |
| marst | james.gu: thank you! | 19:54 |
| airship-irc-bot | <james.gu> btw, for Airship1, I believe the charts currently expect ceph disks are on the controller/compute nodes. | 19:54 |
| *** muhaha has joined #airshipit | 19:57 | |
| marst | thanks a lot! y'all are awesome. :) | 20:03 |
| airship-irc-bot | <mattmceuen> Hey @kk6740 , we're planning on making `kind: Phase` 's config.documentEntrypoint be "site agnostic", right? So that instead of having e.g. `manifests/site/test-site/ephemeral/initinfra`, it would just have `ephemeral/initinfra`, and therefore wouldn't require patching / redefinition for each site -- right? | 20:22 |
| airship-irc-bot | <kk6740> let me show u the patchset | 20:23 |
| airship-irc-bot | <kk6740> that does this | 20:23 |
| airship-irc-bot | <kk6740> https://review.opendev.org/#/c/757019/ | 20:24 |
| airship-irc-bot | <kk6740> this allows us to prefix the path | 20:25 |
| airship-irc-bot | <kk6740> and generalize it, in metadata file | 20:25 |
| airship-irc-bot | <mattmceuen> Perfect :slightly_smiling_face: | 20:25 |
| airship-irc-bot | <mattmceuen> thank you | 20:25 |
| airship-irc-bot | <mattmceuen> that looks great | 20:25 |
| airship-irc-bot | <kk6740> https://review.opendev.org/#/c/757019/13/manifests/phases/phases.yaml | 20:25 |
| airship-irc-bot | <kk6740> https://review.opendev.org/#/c/757019/13/manifests/metadata.yaml | 20:25 |
| airship-irc-bot | <kk6740> :+1: | 20:26 |
| airship-irc-bot | <sirajudeen.yasin> so one *metadata.yaml per site ?* | 20:26 |
| airship-irc-bot | <mattmceuen> yeah | 20:26 |
| airship-irc-bot | <sirajudeen.yasin> thanks @mattmceuen | 20:26 |
| airship-irc-bot | <mattmceuen> better one metadata per site than one massive kind:Phase patch-fest per site :slightly_smiling_face: | 20:26 |
| airship-irc-bot | <sirajudeen.yasin> :+1: | 20:27 |
| airship-irc-bot | <mattmceuen> Hey @kk6740 I +2'd it, but then changed my mind. I realized it didn't have a descriptive enough commit message; we should make sure breaking changes in particular have some good context so that other people can fix and rebase | 20:36 |
| airship-irc-bot | <mattmceuen> this bit someone earlier this week, so just want to avoid it if possible | 20:36 |
| airship-irc-bot | <kk6740> its not braking :slightly_smiling_face: | 20:37 |
| airship-irc-bot | <kk6740> its optional :slightly_smiling_face: | 20:37 |
| airship-irc-bot | <kk6740> if u dont want to use that prefix, just dont use it | 20:37 |
| airship-irc-bot | <kk6740> and everything is going to work the same way as before | 20:37 |
| airship-irc-bot | <ih616h> not to ruffle any feathers, but I'm not sure how that precludes good commit messages. That commit has no documentation nor any real description of what it changes in the commit message | 20:40 |
| airship-irc-bot | <ih616h> ah, it is linked to https://github.com/airshipit/airshipctl/issues/356, which at least provides some context | 20:41 |
| airship-irc-bot | <kk6740> well message could be more descriptive of course | 20:41 |
| airship-irc-bot | <kk6740> something i missed in my review | 20:42 |
| airship-irc-bot | <mattmceuen> ok I feel better about it if it's not a breaking change, I didn't pick up on that | 21:01 |
| airship-irc-bot | <kk6740> and there is a link to an issue as well :slightly_smiling_face: | 21:01 |
| airship-irc-bot | <mattmceuen> But agree, I think we should try for informative commit messages (which I missed in my review as well :slightly_smiling_face: ) | 21:02 |
| *** muhaha has quit IRC | 21:30 | |
| *** avolkov has quit IRC | 21:57 | |
| *** roman_g has quit IRC | 22:00 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!