supervisor: Avoid prematurely pushing some jobs

This is a follow-up to ff615aae541032c647e78d3959d368f595c93e31; it caused us to
submit the first-in-suite and last-in-suite jobs to paddles. Those present has
having 'unknown' status, which will be confusing to users.

Signed-off-by: Zack Cerza <zack@cerza.org>

diff --git a/teuthology/dispatcher/supervisor.py b/teuthology/dispatcher/supervisor.py
index b235d4041009f0618c22780b9a831abb2b7fa5a3..dbcaea512953e247678d2d82ede44b0ce89c7d05 100644 (file)
--- a/teuthology/dispatcher/supervisor.py
+++ b/teuthology/dispatcher/supervisor.py
@@ -42,11 +42,12 @@ def main(args):
     except SkipJob:
         return 0
 
-    report.try_push_job_info({
-        'name': job_config['name'],
-        'job_id': job_config['job_id'],
-        'pid': os.getpid(),
-    })
+    if not (job_config.get('first_in_suite') or job_config.get('last_in_suite')):
+        report.try_push_job_info({
+            'name': job_config['name'],
+            'job_id': job_config['job_id'],
+            'pid': os.getpid(),
+        })
 
     # reimage target machines before running the job
     if 'targets' in job_config: