When deploying a cloud template with a very long input (greater than 262144 Bytes) set on a custom property of a vSphere VM, after some time the deployment finishes and the resource status reports the message "unable to sync the status of this resource".
In the catalog-service-app log you see errors similar to:
"Failed to sync resource with id: <resource id> of type: 'Cloud.vSphere.Machine' and origin: 'DEPLOYED' with error: 500 Internal Server Error from POST"
Aria Automation 8.16.1
This issue is due to the default size values for both the blueprint-service pod and catalog-service-app being too small for the input size.
Workaround:
- Take a snapshot of all Aria Automation Nodes
- ssh into the vRA instance
- Run the below command, which will open an edit window:
kubectl -n prelude edit deployment tango-blueprint-service-app
- Search for JAVA_OPTS and you will find section similar to the below:
- name: JAVA_OPTS
value: -DHhttps.proxyHost=$(JAVA_PROXY_HOST)
- Add the line
-Dspring.codec.max-in-memory-size=16MB
at the end of the value field
- Save and exit
:wq!
- Still in the ssh connection to the vRA instance
- Run the below command, which will open an edit window:
kubectl -n prelude edit deployment catalog-service-app
- Search for JAVA_OPTS and you will find section similar to the below:
- name: JAVA_OPTS
value: -DHhttps.proxyHost=$(JAVA_PROXY_HOST)
- Add the line
-Dspring.codec.max-in-memory-size=16MB
at the end of the value field
- Save and exit
:wq!
The blueprint-service pod and catalog-service-app will restart and pickup 16MB as the size of memory, which means now it should be able to support the input of several MBs.
Note that the workaround will be wiped out if a platform reboot happens.
To make the workaround permanent after a platform reboot, please follow the kb for Persisting Custom Settings in Helm Charts for VMware Aria Automation
For example, for the catalog-service-app, You would need to modify:
/opt/charts/catalog-service/templates/deployment.yaml