[06:16:18] <ryankemper>	 Running to an appt in the morning, won't be in for retro
[09:50:20] <dcausse>	 lunch
[10:43:24] <gehel>	 Lunch 2
[13:17:28] <inflatador>	 <o/
[13:26:33] <gehel>	 !issync
[13:26:34] <ircservserv-wm>	 Syncing #wikimedia-search (requested by gehel)
[13:26:35] <ircservserv-wm>	 No updates for #wikimedia-search
[13:32:00] <dcausse>	 o/
[13:52:19] <dcausse>	 going offline
[13:52:33] <inflatador>	 so soon we part! Have a great rest of the day
[14:48:55] <ebernhardson>	 \o
[14:51:15] <inflatador>	 <o/
[16:00:27] <inflatador>	 workout, back in ~40
[16:47:12] <inflatador>	 back
[17:17:27] * ebernhardson waffles between hacky conversion of pattern -> simple string match, and updating event streams to accept pattern based http routes
[17:33:32] <inflatador>	 lunch, back in ~30
[17:38:19] * ebernhardson sighs...somehow eventutilities jars don't have any SCM information in their poms uploaded to archiva...wonder where that lives
[18:26:41] <inflatador>	 ebernhardson do we need to redeploy cirrus-streaming-updater to get changes from https://gerrit.wikimedia.org/r/c/operations/deployment-charts/+/965256/ ? I see it's in failed state 
[18:29:52] <gehel>	 ebernhardson: SCM infos are in eventutilities-parent
[18:46:18] <ebernhardson>	 inflatador: the failed state is unrelated, it's because our http routes don't make it into event utilities schema loading, so it tries to talk to meta directly (and gets firewalled)
[18:46:22] <ebernhardson>	 inflatador: workin up patches now :)
[18:48:20] <ebernhardson>	 gehel: oh indeed it did make it there, thanks!
[18:50:54] * ebernhardson just now realizes eventutilities wraps the http client we would normally create into a final BasicHttpClient class, which can't be extended or implemented from elsewhere :S
[18:57:33] <inflatador>	 medical appointment, back in ~2h
[19:27:22] <ryankemper>	 inflatador: gehel and I merged those wdqs graph split patches. we did have one issue with a cyclical dependency but we resolved that: https://gerrit.wikimedia.org/r/c/operations/puppet/+/965560
[19:27:50] <ryankemper>	 current status is the clouddumps servers are refusing access so there's probably one last thing for us to figure out on the nfs side of things
[19:28:30] <ryankemper>	 ^ https://phabricator.wikimedia.org/P52919 for that
[19:29:48] <ebernhardson>	 iirc nfs had a specific firewall we had to look at before, but i could be misremembering
[19:36:50] <inflatador>	 did y'all run puppet on the dumps servers?
[19:44:30] <ryankemper>	 yeah
[20:01:39] * ebernhardson just can't decide what is appropriate here...we have a CustomRoutePlanner that does what we need...but it turns out apache re-namespaced their http utils and broke a variety of BC for v5, and event utils uses v4 while we use v5...
[20:02:01] <ebernhardson>	 i don't really want to bring in v4 as a dep just so we can implement this thing twice
[20:30:13] * ebernhardson also wonders if we are supposed to be replacing the `HttpHost target` argument, there is also a `HttpHost proxy` argument that might make sense...
[20:40:25] <inflatador>	 back
[20:48:56] <inflatador>	 NFS stuff is fixed...clouddumps servers needed an `exportfs -r` , which forces NFS to regenerate its exports
[21:11:57] <inflatador>	 ryankemper I started a data reload for each of the graph split hosts...these are running in a tmux window under my user on cumin1001
[21:12:24] <ryankemper>	 ack