Esri / arcgis-gitops

GitHub Actions workflows for ArcGIS Enterprise deployment and operation
Apache License 2.0
5 stars 1 forks source link

Upgrade from 11.2 to 11.3 fails for arcgis-server-linux template #81

Closed pbobov closed 1 month ago

pbobov commented 3 months ago

Upgrade from 11.2 to 11.3 fails for arcgis-server-linux template at setting server's system properties with "Failed to clear REST cache on one or more server machines. Please check the log for more details." error.

 TASK [Set ArcGIS Server system properties]
│ *************************************
│ fatal: [ip-10-0-64-166.us-west-2.compute.internal]: FAILED! => {"changed":
│ false, "msg": "REST service error: code=500, message='Failed to clear REST
│ cache on one or more server machines. Please check the log for more
│ details.', details=None", "response": ""}

The server log has records:

WARNING | Failed to clear REST cache on one or more server machines. {"status":"error","messages":["Could not find resource or operation 'clearRestCache' on the system."],"code":404} 

and

WARNING | The machine 'IP-10-0-64-166.US-WEST-2.COMPUTE.INTERNAL' was not reachable while performing this operation. The server machine will be notified of this change when it comes back online. If ArcGIS Server is going to continue to be unavailable, remove the machine from the site.

This may have something to do with the system properties set to:

{
  "machineSuspendThreshold": 60,
  "suspendedMachineUnregisterThreshold": 60
}

or that the site uses hostnames. Though the machines hostnames resolve on all the EC2 instances.

pbobov commented 1 month ago

The issue was fixed by https://github.com/Esri/arcgis-gitops/pull/107

The problem was that updating ArcGIS Server system properties does not work is some server machines in the site are not available. The fix makes the upgrade process skip updating the system properties is they are not changed.