ajayhn · September 15, 2016 18:14
diff --git a/ifmap-server-error-handling b/ifmap-server-error-handling
 1. timeout when publishing to ifmap-server (mapclient.call())
   when timeout happens keep retrying
 2. ifmap-server code needs to inspect healthcheck object publish and if it misses n of them, exit() OR
     config-nodemgr keeps reading healthcheck object from ifmap-server and if it can't read in n secs or value is stale, kill ifmap-server
 3. control-node keeps reading healthcheck object from ifmap and it sees stale, move to a new ifmap-server
 4. ifmap-publish-greenlet enqueue finds q-full, no change from current
	1. timeout when publishing to ifmap-server (mapclient.call())
	when timeout happens keep retrying
	2. ifmap-server code needs to inspect healthcheck object publish and if it misses n of them, exit() OR
	config-nodemgr keeps reading healthcheck object from ifmap-server and if it can't read in n secs or value is stale, kill ifmap-server
	3. control-node keeps reading healthcheck object from ifmap and it sees stale, move to a new ifmap-server
	4. ifmap-publish-greenlet enqueue finds q-full, no change from current