[10:11:08] <dcausse>	 lunch
[12:43:10] <inflatador>	 <o/
[12:49:13] <dcausse>	 o/
[12:51:59] <pfischer>	 o/
[13:50:44] <dcausse>	 errand
[14:54:08] <inflatador>	 back
[15:27:36] <pfischer>	 dcausse: regarding the delete (and to some extend the reconcile operation): Since I cannot determine which subgraphs the (deleted) entity has been assigned to (I only have the entity ID): Do we simply send a regular delete patch + delete stubs patch to *all* subgraphs?
[15:28:38] <dcausse>	 pfischer: yes a delete is going to broadcast a DeleteOp everywhere
[15:28:59] <dcausse>	 just finished some tests, pushing them
[15:29:25] <pfischer>	 thanks! I cherry-picked you signature changes and adapted my code accordingly
[15:31:53] <dcausse>	 pfischer: this is what I have in mind: https://gerrit.wikimedia.org/r/c/wikidata/query/rdf/+/1047015/2/streaming-updater-producer/src/test/scala/org/wikidata/query/rdf/updater/SubgraphDiffUnitTest.scala#124
[15:32:50] <dcausse>	 beware that there might be some mistakes in the tests tho
[15:34:09] <dcausse>	 for deleting stubs I realized that it might be a whole lot easier to re-use a DeleteOp rather than crafting a custom Patch
[15:35:38] <dcausse>	 but was about to add more edge cases... so perhaps there might still be a need to have custom Patch for deleting stubs
[15:38:12] <pfischer>	 Jep, just implemented the delete reusing the DeleteOp
[16:04:15] <pfischer>	 dcausse: added a patchset with the updated signatures + delete implementation to my CR
[16:04:40] <dcausse>	 pfischer: thanks! will rebase mine
[16:05:45] <pfischer>	 I’ll look into your reconcile comment tomorrow, I’m out for today.
[16:38:26] <dcausse>	 dinner
[17:19:59] <inflatador>	 lunch, back in ~30
[17:47:19] <inflatador>	 back
[17:53:15] <ryankemper>	 ebernhardson: inflatador: I can’t make pairing at the normal time jfyi
[17:58:08] <inflatador>	 ryankemper ACK
[18:34:18] <gehel>	 pfischer: if you are around: https://meet.google.com/iti-ahax-yme (JVM meeting)
[18:35:20] <inflatador>	 ebernhardson I got in late to the mtg...I'm just doing kubernetes stuff, no big deal either way if you wanna join or not
[18:36:32] <ebernhardson>	 inflatador: sec
[18:55:53] <ebernhardson>	 inflatador: helm3 template cirrus-streaming-updater-producer-eqiad charts/flink-app --debug -f charts/flink-app/values.yaml -f .fixtures/general-eqiad.yaml -f helmfile.d/services/cirrus-streaming-up dater/values.yaml.gotmpl -f helmfile.d/services/cirrus-streaming-updater/values-main.yaml -f helmfile.d/services/cirrus-streaming-updater/values-staging.yaml  -f helmfile.d/services/cirr
[18:55:55] <ebernhardson>	 us-streaming-updater/values-producer.yaml  -f helmfile.d/services/cirrus-streaming-updater/values-producer-eqiad.yaml --set "app.job.args[2]=200" | less␠
[22:37:17] <inflatador>	 ebernhardson was indeed a FW issue with postgres, have a puppet patch up if you or ryankemper have a chance to take a look: https://gerrit.wikimedia.org/r/c/operations/puppet/+/1047189
[22:45:50] <ryankemper>	 inflatador: Looks reasonable at first glance. Gave my +1
[22:46:47] <inflatador>	 ryankemper thanks, I've been discussing with Ben on slack so should not be any surprises
[22:52:57] <inflatador>	 OK, that did it! The airflow pod connected postgres and created the DB tables