From 34a092a7f4a8bdd089d9e55d02f2259f86ab54dd Mon Sep 17 00:00:00 2001
From: Dmitrii Kabanov <dk370c@att.com>
Date: Thu, 11 Oct 2018 16:02:02 -0700
Subject: [PATCH] Add startingDeadlineSeconds field to cronJobs

This patch set adds "startingDeadlineSeconds" field to cronJobs.
When the field is not set, the controller counts how many missed
jobs occured from the last scheduled time till now. And if it happends
more than 100 time the job will not be scheduled. To avoid this
the "startingDeadlineSeconds" field should be set to sufficient period
of time. In this case the controller counts how many missed jobs occured
during this period of time. The value of the field should be less than
time (in seconds) needed for running >100 jobs (according to schedule).

Change-Id: I3bf7c7077b55ca5a3421052bd0b59b70c9bbcf24
---
 cinder/templates/cron-job-cinder-volume-usage-audit.yaml | 3 +++
 cinder/values.yaml                                       | 1 +
 heat/templates/cron-job-engine-cleaner.yaml              | 3 +++
 heat/values.yaml                                         | 1 +
 nova/templates/cron-job-cell-setup.yaml                  | 3 +++
 nova/templates/cron-job-service-cleaner.yaml             | 3 +++
 nova/values.yaml                                         | 2 ++
 7 files changed, 16 insertions(+)

diff --git a/cinder/templates/cron-job-cinder-volume-usage-audit.yaml b/cinder/templates/cron-job-cinder-volume-usage-audit.yaml
index e8727147d1..a15d064f27 100644
--- a/cinder/templates/cron-job-cinder-volume-usage-audit.yaml
+++ b/cinder/templates/cron-job-cinder-volume-usage-audit.yaml
@@ -33,6 +33,9 @@ spec:
   schedule: {{ .Values.jobs.volume_usage_audit.cron | quote }}
   successfulJobsHistoryLimit: {{ .Values.jobs.volume_usage_audit.history.success }}
   failedJobsHistoryLimit: {{ .Values.jobs.volume_usage_audit.history.failed }}
+  {{- if .Values.jobs.volume_usage_audit.starting_deadline }}
+  startingDeadlineSeconds: {{ .Values.jobs.volume_usage_audit.starting_deadline }}
+  {{- end }}
   concurrencyPolicy: Forbid
   jobTemplate:
     metadata:
diff --git a/cinder/values.yaml b/cinder/values.yaml
index e225130bd3..368b1610b8 100644
--- a/cinder/values.yaml
+++ b/cinder/values.yaml
@@ -71,6 +71,7 @@ images:
 jobs:
   volume_usage_audit:
     cron: "*/5 * * * *"
+    starting_deadline: 600
     history:
       success: 3
       failed: 1
diff --git a/heat/templates/cron-job-engine-cleaner.yaml b/heat/templates/cron-job-engine-cleaner.yaml
index 87b4ba2ca9..c63159fe8b 100644
--- a/heat/templates/cron-job-engine-cleaner.yaml
+++ b/heat/templates/cron-job-engine-cleaner.yaml
@@ -33,6 +33,9 @@ spec:
   schedule: {{ .Values.jobs.engine_cleaner.cron | quote }}
   successfulJobsHistoryLimit: {{ .Values.jobs.engine_cleaner.history.success }}
   failedJobsHistoryLimit: {{ .Values.jobs.engine_cleaner.history.failed }}
+  {{- if .Values.jobs.engine_cleaner.starting_deadline }}
+  startingDeadlineSeconds: {{ .Values.jobs.engine_cleaner.starting_deadline }}
+  {{- end }}
   concurrencyPolicy: Forbid
   jobTemplate:
     metadata:
diff --git a/heat/values.yaml b/heat/values.yaml
index fd02006bed..1cc6858de2 100644
--- a/heat/values.yaml
+++ b/heat/values.yaml
@@ -67,6 +67,7 @@ images:
 jobs:
   engine_cleaner:
     cron: "*/5 * * * *"
+    starting_deadline: 600
     history:
       success: 3
       failed: 1
diff --git a/nova/templates/cron-job-cell-setup.yaml b/nova/templates/cron-job-cell-setup.yaml
index 66df4dfa51..ee86ca321f 100644
--- a/nova/templates/cron-job-cell-setup.yaml
+++ b/nova/templates/cron-job-cell-setup.yaml
@@ -30,6 +30,9 @@ spec:
   schedule: {{ .Values.jobs.cell_setup.cron | quote }}
   successfulJobsHistoryLimit: {{ .Values.jobs.cell_setup.history.success }}
   failedJobsHistoryLimit: {{ .Values.jobs.cell_setup.history.failed }}
+  {{- if .Values.jobs.cell_setup.starting_deadline }}
+  startingDeadlineSeconds: {{ .Values.jobs.cell_setup.starting_deadline }}
+  {{- end }}
   concurrencyPolicy: Forbid
   jobTemplate:
     metadata:
diff --git a/nova/templates/cron-job-service-cleaner.yaml b/nova/templates/cron-job-service-cleaner.yaml
index 60f5a984bb..47fdad96ad 100644
--- a/nova/templates/cron-job-service-cleaner.yaml
+++ b/nova/templates/cron-job-service-cleaner.yaml
@@ -30,6 +30,9 @@ spec:
   schedule: {{ .Values.jobs.service_cleaner.cron | quote }}
   successfulJobsHistoryLimit: {{ .Values.jobs.service_cleaner.history.success }}
   failedJobsHistoryLimit: {{ .Values.jobs.service_cleaner.history.failed }}
+  {{- if .Values.jobs.service_cleaner.starting_deadline }}
+  startingDeadlineSeconds: {{ .Values.jobs.service_cleaner.starting_deadline }}
+  {{- end }}
   concurrencyPolicy: Forbid
   jobTemplate:
     metadata:
diff --git a/nova/values.yaml b/nova/values.yaml
index 6558182ac4..777469a68c 100644
--- a/nova/values.yaml
+++ b/nova/values.yaml
@@ -100,11 +100,13 @@ jobs:
   # TODO(portdirect): Add a post-start action to nova compute pods that registers themselves.
   cell_setup:
     cron: "0 */1 * * *"
+    starting_deadline: 600
     history:
       success: 3
       failed: 1
   service_cleaner:
     cron: "0 */1 * * *"
+    starting_deadline: 600
     history:
       success: 3
       failed: 1