Increase ping delay in WatchdogActorTest to try and reduce spurious test failures...

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

diff --git a/sdk/cli/bin/crunch-job b/sdk/cli/bin/crunch-job
index a2e42d6fac498b1b30437ad71def8bad7e946f3d..e0b27107986a36abd1bd99759d758162c006a042 100755 (executable)
--- a/sdk/cli/bin/crunch-job
+++ b/sdk/cli/bin/crunch-job
@@ -417,18 +417,22 @@ if ($docker_locator = $Job->{docker_image_locator}) {
   Log (undef, "docker image hash is $docker_hash");
   $docker_stream =~ s/^\.//;
   my $docker_install_script = qq{
-if $docker_bin images -q --no-trunc --all | grep -xF \Q$docker_hash\E >/dev/null; then
-    exit 0
+loaded() {
+  id=\$($docker_bin inspect --format="{{.ID}}" \Q$docker_hash\E) || return 1
+  echo "image ID is \$id"
+  [[ \${id} = \Q$docker_hash\E ]]
+}
+if loaded >&2 2>/dev/null; then
+  echo >&2 "image is already present"
+  exit 0
 fi
-declare -a exit_codes=("\${PIPESTATUS[@]}")
-if [ 0 != "\${exit_codes[0]}" ]; then
-   exit "\${exit_codes[0]}"  # `docker images` failed
-elif [ 1 != "\${exit_codes[1]}" ]; then
-   exit "\${exit_codes[1]}"  # `grep` encountered an error
-else
-   # Everything worked fine, but grep didn't find the image on this host.
-   arv-get \Q$docker_locator$docker_stream/$docker_hash.tar\E | $docker_bin load
+echo >&2 "docker image is not present; loading"
+arv-get \Q$docker_locator$docker_stream/$docker_hash.tar\E | $docker_bin load
+if ! loaded >&2; then
+  echo >&2 "`docker load` exited 0, but image is not found (!)"
+  exit 1
 fi
+echo >&2 "image loaded successfully"
 };
 
   my ($exited, $stdout, $stderr) = srun_sync(
@@ -715,6 +719,15 @@ foreach (split (/\n/, $Job->{knobs}))
 {
   Log (undef, "knob " . $_);
 }
+my $resp = api_call(
+  'nodes/list',
+  'filters' => [['hostname', 'in', \@node]],
+  'order' => 'hostname',
+  'limit' => scalar(@node),
+    );
+for my $n (@{$resp->{items}}) {
+  Log(undef, "$n->{hostname} $n->{uuid} ".JSON::encode_json($n->{properties}));
+}
 
 
 
@@ -1593,9 +1606,10 @@ sub create_output_collection
 import arvados
 import sys
 print (arvados.api("v1").collections().
-       create(body={"manifest_text": sys.stdin.read()}).
+       create(body={"manifest_text": sys.stdin.read(),
+                    "owner_uuid": sys.argv[2]}).
        execute(num_retries=int(sys.argv[1]))["portable_data_hash"])
-}, retry_count());
+}, retry_count(), $Job->{owner_uuid});
 
   my $task_idx = -1;
   my $manifest_size = 0;
@@ -1843,6 +1857,7 @@ sub croak
   my ($package, $file, $line) = caller;
   my $message = "@_ at $file line $line\n";
   Log (undef, $message);
+  release_allocation();
   freeze() if @jobstep_todo;
   create_output_collection() if @jobstep_todo;
   cleanup();