mongos need restart after rs.remove

when we remove a replica vom a sharded replica set, then the mongos hang up and need restart to work properly.

command failed [getlasterror]: { "serverUsed" : "localhost/127.0.0.1:27017" , "ok" : 0.0 , "errmsg" : "could not get last error from sha
rd replset01/vnode02.local:27015,vnode03.local:27015 :: caused by :: DBClientBase::findN: transport error:

Is thrown from the application.