Add Grace period to HAProxy at startup
'HealthCheckGracePeriodSeconds' defaults to 0, meaning
the task might get killed several time just because the container takes
some seconds to start. This uselessly delays the stack creation.
Change-Id: I33808b23fbaf6ff606fba90649f1f055857efd2b
diff --git a/dual-primary/cf-service-lb.yml b/dual-primary/cf-service-lb.yml
index 0865f04..1d49ffe 100644
--- a/dual-primary/cf-service-lb.yml
+++ b/dual-primary/cf-service-lb.yml
@@ -82,6 +82,10 @@
Description: HAPRoxy Syslog Sidecar service name
Type: String
Default: gerrit-haproxy-sidecar
+ HealthCheckGracePeriodSeconds:
+ Description: HAProxy Healtcheck Grace Period
+ Type: Number
+ Default: 60
Resources:
LBService:
@@ -95,6 +99,7 @@
!Join [':', [!Ref 'ClusterStackName', 'ClusterName']]
DesiredCount: !Ref DesiredCount
TaskDefinition: !Ref TaskDefinition
+ HealthCheckGracePeriodSeconds: !Ref HealthCheckGracePeriodSeconds
LoadBalancers:
- ContainerName: !Ref LBServiceName
ContainerPort: !Ref HTTPContainerPort
