debug

Author: samtkaplan

src/Schedulers.jl

@@ -615,7 +615,9 @@ function loop(eloop::ElasticLoop, journal, journal_task_callback, tsk_map, tsk_r
             if haskey(Distributed.map_pid_wrkr, bad_pid)
                 @debug "bad pid=$bad_pid is known to Distributed, calling rmprocs"
                 try
+                    @debug "calling rmprocs with bad_pid=$bad_pid"
                     rmprocs(bad_pid)
+                    @debug "done calling rmprocs with bad_pid=$pad_pid"
                 catch e
                     @warn "unable to remove bad woker with pid=$bad_pid"
                     logerror(e)
@@ -624,14 +626,17 @@ function loop(eloop::ElasticLoop, journal, journal_task_callback, tsk_map, tsk_r
                 @debug "bad pid=$bad_pid is not known to Distributed, calling kill"
                 try
                     # required for cluster managers that require clean-up when the julia process on a worker dies:
+                    @debug "calling kill with bad_pid=$bad_pid"
                     Distributed.kill(wrkrs[bad_pid].manager, bad_pid, wrkrs[bad_pid].config)
+                    @debug "done calling kill with bad_pid=$bad_pid"
                 catch e
                     @warn "unable to kill bad worker with pid=$bad_pid"
                     logerror(e)
                 end
             end
             haskey(wrkrs, bad_pid) && delete!(wrkrs, bad_pid)
         end
+        @debug "...done removing pad pids."
 
         # check to see if we need to add or remove machines
         δ,n_remaining_tasks = 0,0