[08:23:43] wdqs reload failed on 1009 (journal corruption), still running on 1010 (https://grafana-rw.wikimedia.org/d/000000489/wikidata-query-service?orgId=1&var-cluster_name=wdqs-test&from=1675191710223&to=1676276548657) [08:23:53] :/ [10:55:31] Lunch [11:05:41] lunch 2 [13:55:15] o/ [14:00:15] dcausse that sucks! Is it OK if I restart 1009 or would you like me to copy down any logs or anything? [14:01:49] inflatador: we can try to restart, I doubt there's anything new we can learn from wdqs1009 logs [14:11:20] ACK, looks like I need to clean up some more of the reuse_munge stuff...working on patch now [14:13:54] https://gerrit.wikimedia.org/r/c/operations/cookbooks/+/888707 [14:18:03] inflatador: I've a question for you about a running cookbook on cumin1001 [14:18:26] (not urgent at all, can wait if you're busy) [14:18:59] volans sure, we can talk now. Probably related to the patch I just pushed [14:20:48] I was wondering why there is a cookbook running since Feb 2 that seems to be run from a non-stardard location [14:23:32] It's WDQS data reload, those take weeks...so that would explain the length of the run https://phabricator.wikimedia.org/T323096 [14:24:00] but why is not run as a normal cookbook from the official repo? [14:26:20] I was testing it before merging the cookbook to prod...it worked so I just left it running from the test version [14:26:44] I see various things "wrong" with that setuo [14:26:46] *setup [14:27:28] We're in a weird state with this data reload. Are there any other cookbooks that run constantly for weeks? I feel like cookbooks might not be the best approach for this [14:28:05] the problems I see are not related with its duration fwiw [14:30:26] Sorry I just realized I have a dr appointment in 15 m. I'll hit you up when I get back if you like [14:30:38] sure, ttyl [15:38:26] pfischer: for when you're back: there are training options for Flink on https://www.ververica.com/training (not super clear on the dates, I suspect you'll need to register to get the dates of the next training). Let me know if you're interested. [15:44:00] volans I'm back but still not feeling great. OK if we talk tomorrow? [15:54:16] no prob at all, take care! [19:12:06] seems ok, although i often don't notice until i realize i'm not in the invite-only channels. As you've done, a reconnect often works it out [19:17:11] dcausse: gehel: If forking a java project from github to our gitlab, do we think it's important to do things like switch over to discovery-parent-pom, setup CI, etc? In particular i'm pulling over the two packages mjolnir uses for feature selection. the total extent of spark3 migration is super trivial, but upstream is no longer maintaining the project [19:17:18] example migration: https://gitlab.wikimedia.org/repos/search-platform/spark-infotheoretic-feature-selection/-/commit/4f5b45b3ffd5c13ad2f335f30ebe8d369fb6431e [19:18:31] tempted to configure the maven release plugin, make a release to archiva, and call it good enough [19:26:54] ebernhardson: sounds good to me [19:27:26] using the parent pom I guess you'll have to either make the code happy with our checks or simply ignore all our rules [19:28:10] I don't think it makes sense to migrate to our parent pom, unless we want to take over that project entirely [19:29:41] ok sounds good [20:42:41] thanks for restarting dcausse , sorry I didn't grab that earlier [20:43:22] np! I was around so just did it [22:05:41] heading out slightly early, see ya tomorrow!