Fix the issue widget model issue
So the problem is happening intermittently
because resources and traversal starts up parallel
sometimes they would be ready at the same time
and sometimes they are not but the HAProxy deployment
needs to wait until resources, traversal and graphadmin
is ready before it starts to be in the ready state.
Also modifying the readiness probe to make the
echo request to /aai/util/echo rather than rely
on the port check to get a better indicator of haproxy readiness
Issue-ID: AAI-1759
Change-Id: I847abd7fd0375da59254656f5891ed14547058ab
Signed-off-by: Kajur, Harish (vk250x) <vk250x@att.com>
diff --git a/kubernetes/aai/templates/deployment.yaml b/kubernetes/aai/templates/deployment.yaml
index d32db17..3f16e25 100644
--- a/kubernetes/aai/templates/deployment.yaml
+++ b/kubernetes/aai/templates/deployment.yaml
@@ -39,7 +39,11 @@
- /root/ready.py
args:
- --container-name
+ - aai-resources
+ - --container-name
- aai-traversal
+ - --container-name
+ - aai-graphadmin
env:
- name: NAMESPACE
valueFrom:
@@ -75,8 +79,17 @@
periodSeconds: {{ .Values.liveness.periodSeconds }}
{{ end -}}
readinessProbe:
- tcpSocket:
+ httpGet:
+ path: /aai/util/echo
port: {{ .Values.service.internalPort2 }}
+ scheme: HTTPS
+ httpHeaders:
+ - name: X-FromAppId
+ value: OOM_ReadinessCheck
+ - name: X-TransactionId
+ value: {{ uuidv4 }}
+ - name: Accept
+ value: application/json
initialDelaySeconds: {{ .Values.readiness.initialDelaySeconds }}
periodSeconds: {{ .Values.readiness.periodSeconds }}
resources: