[09:24:34] <pfischer>	 o/ dcausse: I’ve upgraded the RDF tools to spark 3 https://gerrit.wikimedia.org/r/c/wikidata/query/rdf/+/882691. However, I’d like to test its runtime behaviour before merging anything. I’m afraid that some overridden versions do not apply for the actual runtime dependencies.
[09:26:23] <dcausse>	 pfischer: o/ I'll take a look, for testing I'll have to recollect what jobs are in there to know how to run them :)
[09:27:02] <pfischer>	 Alright, thank you!
[09:27:43] <dcausse>	 pfischer: your patch also updates the stream-updater I don't think we should change it in this same patch
[09:27:54] <dcausse>	 upgrading flink is done via https://gerrit.wikimedia.org/r/c/wikidata/query/rdf/+/879822
[09:32:48] <dcausse>	 ho I think it needs to be rebased, I think Erik has made plenty of changes to the way we build scala
[09:36:52] <dcausse>	 oh perhaps not, got confused by the presence of https://repo.artima.com/releases which I think Erik got rid of
[09:37:08] <dcausse>	 might just be that we forgot to remove  the repository
[10:41:13] <dcausse>	 hm... seems like the the data-reload script is running twice on wdqs1009
[10:41:30] <dcausse>	 inflatador: ^
[10:44:51] <dcausse>	 running only once on wdqs1010
[10:46:31] <dcausse>	 they're fetching entities/20230102/wikidata-20230102-all-BETA.ttl.bz2 which is too old, they should fetch 20230116/wikidata-20230116-all-BETA.ttl.bz2
[11:06:29] <gehel>	 Lunch
[11:07:41] <dcausse>	 pfischer: some ideas for testing rdf-spark-tools P43307
[11:08:22] <dcausse>	 full link https://phabricator.wikimedia.org/P43307
[11:08:39] <pfischer>	 dcausse: Thanks, looking into it.
[11:09:19] <gehel>	 Lunch 
[11:09:42] <gehel>	 (for real this time)
[11:10:04] <dcausse>	 lunch 2
[13:55:41] <gehel>	 inflatador: I've moved our 1:1 30' later (conflict with the Quarter in Review). Let's try to use that time for ITC (I'll do my best to be ready by then)
[14:16:52] <inflatador>	 <o/
[14:17:06] <inflatador>	 dcausse will check into that now
[14:17:13] <dcausse>	 o/
[14:17:15] <dcausse>	 thanks!
[14:21:52] <inflatador>	 np, thanks for catching that. I restarted the reloads, should be good now but LMK if you notice anything amiss
[14:24:54] <dcausse>	 inflatador: will take a look, curious to know how the dump gets picked-up tho? I thought that the cookbook would follow the symlink?
[14:25:36] <inflatador>	 dcausse I hard-coded it because I was suspicious what would happen if the dump rolled over when it was still munging and forgot to put it back ;(
[14:25:52] <dcausse>	 oh ok makes sense
[14:26:15] <inflatador>	 I'll get a patch up shortly, even if it's a problem the script should be able to work around it without hardcoding
[14:26:46] <gehel>	 dcausse, inflatador, ryankemper, pfischer: I've sent you the invite to the k8s Special Interest Group later today. I'm not exactly sure what the scope is. I won't be able to be there.
[14:27:10] <inflatador>	 gehel thanks, I'm planning on going
[14:27:19] <gehel>	 It probably does not make sense to have the 4 of you joining, so please check on who is going and who's not.
[14:28:33] <dcausse>	 inflatador: thanks for attending
[14:29:54] <inflatador>	 NP, if anyone else wants to go instead of me, we can work it out with a game of rochambeau
[14:34:56] <dcausse>	 til "game of rochambeau" is pierre-feuille-ciseaux (commonly named chifoumi around here)
[14:39:25] <inflatador>	 That's funny, the French-sounding name that isn't actually used in France
[14:40:14] <gehel>	 Around here, it would be "feuille-cailloux-ciseaux", or also chifoumi
[14:41:58] <dcausse>	 we also got the "well" extension but this one is bit like cheating, the "well" only loses against the paper
[14:45:09] <gehel>	 trivia: The Interplanetary Chifoumi Championship is organized in Lausanne every now and then...
[14:45:33] <gehel>	 https://www.lausannebondyblog.ch/championnat-interplanetaire-de-feuille-caillou-ciseaux-clap-de-fin/
[14:46:05] <gehel>	 Or used to, those funny days might be over
[14:46:10] <gehel>	 https://www.20min.ch/fr/story/une-competition-insolite-a-lausanne-985018706794
[14:46:55] <dcausse>	 so close to you, you'd better start your training asap
[14:49:12] <gehel>	 Nah, I'm not into extreme sports. The risk of injury is too high.
[14:51:23] <inflatador>	 Don't want to get "scissor finger"
[14:52:00] <inflatador>	 Hmm, something funky is happening on wdqs1009 and 1010 , but not 2010. Might have something to do with the deploy yesterday
[14:53:20] <dcausse>	 inflatador: lemme know if I can help
[14:55:24] <inflatador>	 dcausse thanks, it looks like something wrong with the script invocation https://phabricator.wikimedia.org/P43313
[14:57:39] <inflatador>	 looks munge.sh is different between working/non-working
[14:58:44] <dcausse>	 not sure I remember that we we updated the munge script recently
[15:00:28] <inflatador>	 seems to work if I copy the old munge.sh over
[15:03:07] <dcausse>	 hm indeed  it changed https://gerrit.wikimedia.org/r/c/wikidata/query/deploy/+/879645/1/munge.sh
[15:04:11] <dcausse>	 oh I see
[15:04:20] <dcausse>	 it's to solve the symlink issue
[15:05:13] <inflatador>	 oh yeah, sure enough. Damn
[15:05:34] <dcausse>	 I guess it has to be done in python within the cookbook with pathlib and Path('dumpfile').resolve() instead
[15:05:52] <inflatador>	 Sounds reasonable
[15:05:57] <inflatador>	 I'll try and get a patch up today
[15:06:01] <dcausse>	 thanks!
[15:07:00] <dcausse>	 we should revert the patch to munge.sh it'll never work in our context (it'll always receive "-" to specify stdin)
[15:10:26] <inflatador>	 Cool, will work on that too
[15:14:55] <inflatador>	 revert is up here , not sure if it needs a +1 or next steps https://gerrit.wikimedia.org/r/c/wikidata/query/rdf/+/883208
[15:44:06] <ebernhardson>	 \o
[15:44:12] <inflatador>	 <o/
[15:47:45] <dcausse>	 o/
[15:58:26] <inflatador>	 workout, back in ~40
[16:31:59] <gehel>	 early cooking / dinner. Back later for a long evening!
[16:45:43] <inflatador>	 back
[17:46:20] <inflatador>	 quick break, back in ~15
[18:15:12] <inflatador>	 back
[18:33:09] <dcausse>	 inflatador: re https://gerrit.wikimedia.org/r/c/operations/puppet/+/882748 namespaces created on the "main" entry are usable in the dse-k8s?
[18:34:27] <dcausse>	 I understood that we still want to create a namespace there with the same name "rdf-streaming-updater" under dse-k8s
[18:36:31] <inflatador>	 dcausse not sure, I was just going off what Ben did to create the "stream-enrichment-poc" namespace
[18:36:47] <inflatador>	 https://gerrit.wikimedia.org/r/c/operations/puppet/+/854007/
[18:37:41] <dcausse>	 inflatador: yes I think we want a namespace in the dse-k8s for our test of the streaming-updater with the flink-k8s-operator
[18:38:23] <dcausse>	 that namespace can be "rdf-streaming-updater" apparently, it won't clash with the one created on the wikikube cluster
[18:38:39] <inflatador>	 dcausse ah OK, that was my screw-up, I put it in the wrong place
[18:38:56] <inflatador>	 I'll fix it after my lunch, back in ~45
[18:39:07] <dcausse>	 sure, no rush :)
[19:07:15] <dcausse>	 dinner
[19:22:15] <inflatador>	 back
[19:28:59] * ebernhardson can't seem to figure out why something is importable from the REPL, but when loaded through pytest/airflow get a ModuleNotFoundError