[06:52:41] Dentist appt tmrw, no retro for me [09:55:13] Lunch [12:36:47] pfischer: I see that https://gerrit.wikimedia.org/r/c/search/cirrus-streaming-updater/+/832379 is finally merged! Congratulation on your first major patch being merged. And thanks for your patience through all those iterations! [13:03:07] o/ [13:44:59] Taking a break with Oscar. Back for the retro [15:02:08] ejoseph, ryankemper: retrospective is starting: https://meet.google.com/eki-rafx-cxi [16:11:39] going offline [17:35:02] * ebernhardson wonders why reindexing happens inside SpecificAliasVisitor [17:35:21] s/Visitor/Validator/ [17:40:14] bit of a mess in here that makes it awkward to delete a fresh index that failed creation, but make sure we don't delete an index that has potentially been made live [17:42:13] i guess we go roundabout and ask elastic if the index has live aliases... [18:30:14] ryankemper: I'll be 5' late [18:30:23] gehel: ack [18:48:44] meh... PhanImpossibleCondition Impossible attempt to cast false of type false to truthy [19:25:33] maybe i'm going about this wrong way...instead i suppose drop fatalError from ConfigUtils and have it return StatusValue's everywhere [19:32:46] (trying to make maintenance scripts not exit from random places) [19:33:19] looks like some exciting master eligible stuff is happening! [19:36:08] nice! [19:37:56] maybe I should disable my bug for 'elastic' while I'm slumming it in ServiceOps ;P [19:40:24] i dunno, service ops is pretty OP. You might be slumming it over here :P [19:42:44] ryankemper: looks like we have a config check that we did not tell that the configuration has changed [19:43:43] Oh no, I'm definitely increasing the slum quotient wherever I go ;) [19:44:33] ryankemper: I'll let you silence this check in Icinga if it takes more than 1h to update the settings [19:44:48] and on that note, I'm walking away from the keyboard for real! [19:44:56] gehel: sounds good! [20:47:13] * ebernhardson has some memory that underscores in indexBaseName can cause difficulties, and we have arbcom__content_123456 indices. I wonder why that doesn't cause issues