Start master feature test jobs 4 hours earlier

Start the airship master feature tests jobs 4 hours earlier. Jobs
sometimes fail overnight due to webhook failures, which seem related to
high I/O latency in the environment. We're guessing that the high I/O
latency may be caused by CityCloud's disaster recovery snapshots and
hope to avoid the period of their higher latency by starting the jobs
earlier in the evening.

Change-Id: I9e41e862d88742060f13cea6e83eb90ca368c1ea

diff --git a/jjb/airship/job_master_feature_tests_centos.yml b/jjb/airship/job_master_feature_tests_centos.yml
index 80eacced286a6ec52a5aa0f224bc85364dde048f..d2e5fd1899e2011ef5e40a1b9238ca8b2ea832e4 100644 (file)
--- a/jjb/airship/job_master_feature_tests_centos.yml
+++ b/jjb/airship/job_master_feature_tests_centos.yml
@@ -90,7 +90,7 @@
         artifact-num-to-keep: -1
 
     triggers:
-      - timed: "@midnight"
+      - timed: "H 20 * * *"
 
     pipeline-scm:
       scm:

diff --git a/jjb/airship/job_master_feature_tests_ubuntu.yml b/jjb/airship/job_master_feature_tests_ubuntu.yml
index 77092a2a4128623fbd0a9f43ba38e237a8f9cfea..bbfcc94a8381c2fd295aa606db89c6906e8ddabd 100644 (file)
--- a/jjb/airship/job_master_feature_tests_ubuntu.yml
+++ b/jjb/airship/job_master_feature_tests_ubuntu.yml
@@ -90,7 +90,7 @@
         artifact-num-to-keep: -1
 
     triggers:
-      - timed: "@midnight"
+      - timed: "H 20 * * *"
 
     pipeline-scm:
       scm: