[10:05:08] <dcausse>	 lunch
[13:09:38] <inflatador>	 <o/
[13:10:20] <dcausse>	 o/
[14:46:35] <dcausse>	 https://zookeeper.apache.org/releases.html says: "ZooKeeper clients from the 3.5 and 3.6 branches are fully compatible with 3.7 servers."
[14:46:43] <dcausse>	 and " ZooKeeper 3.7.0 clients are compatible with 3.5 and 3.6 servers as long as you are not using new APIs not present these versions."
[14:49:23] <dcausse>	 if kafka 1.1.0 ships with zookeeper client 3.4 then it might not work
[14:51:12] <dcausse>	 it also says: clients from 3.4 and 3.5 branch are fully compatible with 3.6 servers
[14:51:29] <dcausse>	 so perhaps installing zookeeper 3.6 could make flink and kafka happy?
[15:08:23] <inflatador>	 hmmm
[15:16:17] <inflatador>	 OK, new phab task is up https://phabricator.wikimedia.org/T341137
[16:08:14] <inflatador>	 dr appointment, back in ~1h
[16:59:57] <dcausse>	 going offline
[17:41:26] <inflatador>	 back
[18:07:15] <inflatador>	 lunch, back in ~1h
[19:12:36] <inflatador>	 sorry, been back
[21:56:12] * ebernhardson tries retries gitlab-ci again ... getting more and more out of disk space errors :( Maybe there is something we could change about ci to need less disk space, but not sure what
[21:58:18] <inflatador>	 ebernhardson I remember mutante pointed me to a phab task where we can report the disk space errors to releng, but I can't find it at the moment...might ask him once he comes back online
[22:00:50] <ebernhardson>	 i can find it, but i suppose i'm not optimistic about our ticket prioritization process :)
[22:04:18] <inflatador>	 From what I heard from mutante, the gitlab stuff is in cloud (public? WMCS?) so it "should" be a matter of just provisioning more storage. Take that with a grain of salt of course
[22:05:14] <ebernhardson>	 yea it is curious to have an isssue with such a small component, it seems we often criminally underspec things. Like sure, maybe CI should be able to work with a 30GB cache. But does a 300GB cache cost that much more on spinning rust?
[22:06:03] <ebernhardson>	 err, s/cache/scratch space/ or whatever :) I guess I'm assuming the random stuff left behind between runs is caches of some sort