7313: crunch-job reports an error when a task doesn't record state.

[arvados.git] / sdk / cli / bin / crunch-job

diff --git a/sdk/cli/bin/crunch-job b/sdk/cli/bin/crunch-job
index f78824bf2f14c468a946160335537b4fd66b7702..28da66d0ca74296ef8475a5d96c0cb6fb3d41764 100755 (executable)
--- a/sdk/cli/bin/crunch-job
+++ b/sdk/cli/bin/crunch-job
@@ -1182,6 +1182,9 @@ sub reapchildren
 
   if (!defined $task_success) {
     # task did not indicate one way or the other --> fail
+    Log($jobstepid, sprintf(
+          "ERROR: Task process exited %d, but never updated its task record to indicate success and record its output.",
+          exit_status_s($childstatus)));
     $Jobstep->{'arvados_task'}->{success} = 0;
     $Jobstep->{'arvados_task'}->save;
     $task_success = 0;
@@ -1698,20 +1701,24 @@ sub log_writer_finish()
 
   close($log_pipe_in);
 
+  my $logger_failed = 0;
   my $read_result = log_writer_read_output(120);
   if ($read_result == -1) {
+    $logger_failed = -1;
     Log (undef, "timed out reading from 'arv-put'");
   } elsif ($read_result != 0) {
+    $logger_failed = -2;
     Log(undef, "failed to read arv-put log manifest to EOF");
   }
 
   waitpid($log_pipe_pid, 0);
   if ($?) {
+    $logger_failed ||= $?;
     Log(undef, "log_writer_finish: arv-put exited " . exit_status_s($?))
   }
 
   close($log_pipe_out);
-  my $arv_put_output = $log_pipe_out_buf;
+  my $arv_put_output = $logger_failed ? undef : $log_pipe_out_buf;
   $log_pipe_pid = $log_pipe_in = $log_pipe_out = $log_pipe_out_buf =
       $log_pipe_out_select = undef;
 
@@ -1777,13 +1784,13 @@ sub save_meta
   my $justcheckpoint = shift; # false if this will be the last meta saved
   return if $justcheckpoint;  # checkpointing is not relevant post-Warehouse.pm
   return unless log_writer_is_active();
+  my $log_manifest = log_writer_finish();
+  return unless defined($log_manifest);
 
-  my $log_manifest = "";
   if ($Job->{log}) {
     my $prev_log_coll = api_call("collections/get", uuid => $Job->{log});
-    $log_manifest .= $prev_log_coll->{manifest_text};
+    $log_manifest = $prev_log_coll->{manifest_text} . $log_manifest;
   }
-  $log_manifest .= log_writer_finish();
 
   my $log_coll = api_call(
     "collections/create", ensure_unique_name => 1, collection => {