[18:54:28] <ebernhardson>	 I've just redployed a pod in codfw and it's getting `UnknownHostException: flink-zk2001.codfw.wmnet: Temporary failure in name resolution`.  But from a deploy host i can ping (-4 and -6) zk. Any ideas?
[18:58:26] <ebernhardson>	 it's not isolated to zk2001, separate retries have mentioned zk2002 and zk2003
[19:27:43] <ebernhardson>	 from playing around in the shells(call gethostbyname from perl), it looks like both a working and broken container are using 10.192.72.3 as the dns server, but on the broken one requests are timing out (firewalled?) 
[19:35:44] <inflatador>	 ^^ looking at this, best guess is we need to update for calico network policies
[20:18:13] <inflatador>	 per above, ebernhardson was able to get it to work by redeploying. When the original pod `flink-app-producer-b65b89bf5-5hff9` was scheduled on `wikikube-worker2043.codfw.wmnet ` , it had no networking. Just a heads-up in case anyone wants to take a look. I don't see anything obviously wrong w/that host
[20:18:59] <inflatador>	 just doing a cursory glance at https://wikitech.wikimedia.org/wiki/Calico
[20:23:17] <akosiaris>	 Sigh, something has gone wrong with the reimaging then. I ll have look tomorrow, but in the meantime could you cordon it please inflatador?
[20:23:42] <inflatador>	 akosiaris ACK, will do
[20:23:50] <akosiaris>	 Thanks!
[20:23:58] <inflatador>	 akosiaris do we have a cook-book for this or is it just the regular kubectl cmd?
[20:24:40] <inflatador>	 ah, looks like a cook-book https://wikitech.wikimedia.org/wiki/Kubernetes/Administration
[20:25:39] <akosiaris>	 Yeah there is a cookbook, best run that one.
[20:28:31] <inflatador>	 akosiaris cool, just ran the cook-book and it looks to have worked correctly