[09:30:31] <volans>	 inflatador: I've left some comments on https://phabricator.wikimedia.org/T321605#8344805, hope that helps
[10:46:42] <dcausse>	 lunch
[13:01:27] <inflatador>	 volans oh wow! Thanks for the context on that
[13:02:42] <inflatador>	 <o/
[13:03:17] <volans>	 no problem :)
[13:03:44] <dcausse>	 o/
[13:10:42] <inflatador>	 dcausse I keep forgetting to ask you, on https://phabricator.wikimedia.org/T321491 how do you want to try flink operator? as  Deployment or Jobmanager? Or both? 
[13:12:37] <dcausse>	 inflatador: what do you mean by jobmanager vs deployment? (never looked into this k8s operator so I might miss some context)
[13:15:48] <inflatador>	 dcausse sorry, was under some mistaken impressions. Looking at https://wikitech.wikimedia.org/wiki/Wikidata_Query_Service/Flink_On_Kubernetes#Using_a_flink-operator and thinking you wrote it
[13:16:15] <inflatador>	 no wait, you did...maybe I'm just not using the right words
[13:16:42] <inflatador>	 specifically, if you want to run flink application cluster as jobmanager or deployment. 
[13:16:48] <inflatador>	 or evaluate both
[13:17:12] <dcausse>	 perhaps you mean "session cluster" vs "application cluster" ?
[13:18:09] <inflatador>	 Actually I meant jobmanager as k8s deployment or job resource, sorry
[13:18:15] <dcausse>	 ah ok
[13:18:30] <dcausse>	 I think that part will be chosen by the operator itself I suppose
[13:19:04] <dcausse>	 we should not have a choice (it might become an internal implementation detail of the k8s operator)
[13:20:55] <dcausse>	 because there will be a new type of k8s resource (FlinkDeployment) if I'm getting this right: https://nightlies.apache.org/flink/flink-kubernetes-operator-docs-main/docs/custom-resource/job-management/
[13:22:44] <inflatador>	 Thanks, I'm pretty ignorant of this whole thing. Was reading Modena's email and it sounded to me like we're currently using session cluster but wanted to use application cluster too
[13:23:05] <dcausse>	 yes we do but these are different things
[13:23:43] <dcausse>	 "Job" and "Deployment" are k8s resources 
[13:24:09] <dcausse>	 "application cluster" and "session cluster" are different kind of flink setup
[13:24:11] <inflatador>	 Right, that I understand. I was just thinking that the application cluster had different ways to deploy
[13:24:27] <inflatador>	 Like as a job or deployment
[13:24:55] <inflatador>	 and we needed to evaluate those two
[13:25:18] <dcausse>	 this part should be taken care of by the k8s operator
[13:25:28] <dcausse>	 it's only when you deploy this manually that you have these options
[13:25:57] <inflatador>	 Ah, that was the piece I was missing. Sorry for my ignorance here
[13:26:04] <dcausse>	 no worries!
[13:26:20] <inflatador>	 As far as the operator, we can test with the newest flink operator from the nightlies
[13:26:35] <inflatador>	 Or whatever you prefer
[13:26:43] <dcausse>	 I have no preferences
[13:26:59] <dcausse>	 if it's "installable" that's already great tbh! :)
[14:46:22] <mpham>	 feeling a little sore after my booster yesterday and have a few small errands to do, so will skip wed meeting today. you can ping me if there's something relevant though
[14:47:18] <dcausse>	 mpham: take care!
[16:02:46] <inflatador>	 workout, back in ~30
[16:46:17] <inflatador>	 back
[18:32:27] <inflatador>	 voting/errand, back in ~1h
[18:56:34] <inflatador>	 back...voting lines too long, will have to try again some other time
[19:27:20] <ebernhardson>	 mail-in ftw :)
[19:32:05] <inflatador>	 Yeah, I wish! The library I normally vote at has about 1/2 the parking space it needs...I'll probably end up hitting one of the bigger places this weekend
[21:20:39] <inflatador>	 ryankemper I started the data-transfer cookbook on cumin2002 in a tmux window. wcqs2003 is in progress , if it finishes cleanly only 1003 will be left
[21:21:01] <inflatador>	 https://phabricator.wikimedia.org/T316236#8347630 also noted here
[21:21:17] <ryankemper>	 ack, thanks
[21:30:52] <inflatador>	 ryankemper NP. xfer from 1001 -> 2003 failed for the 3rd time, so I'm giving up for today