We have a setup with two identical test environments (A and B) each with one portal, one hosting server, data store and one federated server. We created successfully a backup some days ago, and have tested to import it. The backup is taken in environment A and restored in environment B. The setup from where take the backup (environment A) has aprox 60 services running.
After a lot of testing and scratching our heads we have figured out the following:
We have two test "backup - restore" scenarios (this are all about the federated server not the hosting server)
Scenario 1) A -> B (No services running on env B before restore)
Before restore
Env A (60 Services) , Env B (no services)
After restore
Env A (60 servies), Env B (60 services)
Timing 22 minutes from start to finish. The timing is from the log on these events:
"Import operation started" -> "Import operation completed" and "SERVER READY"
Scenario 2) A -> B (ca 60 services running on env B before restore)
Before restore
Env A (60 Services) , Env B (60 services)
After restore
Env A (60 servies), Env B (60 overwritten services)
Timing 37 minutes from start to finish. The timing is from the log on these events:
"Import operation started" -> "Import operation completed" and "SERVER READY"
Bot scenarios seems to finish restoring of the federated server according to the federated arcgis server logs.
In scenario 1) webgisdr is informed that the federated server restore is finished and webgisdr continues with portal restore and everything finishes - SUCCESS
In scenario 2) webgisdr is NOT informed that the federated server restore is finished and webgisdr continues to wait for ever for the federated server to finish, webgisdr is never notified so the process never finishes - FAIL.
Are there some timeout in play here? Why is webgisdr never notified in scenario 2?
I have been struggling with this several days now, so all help is highly appreciated!
Thanks!
