Hello,
We recently went live with our ArcGIS Enterprise install which federated our publicly exposed ArcGIS Server with a Portal machine and designated it as a hosting server. Most of the services on our ArcGIS Server are public except for a few secured services.
Not long after going live, we had a network interruption between the Portal machine and our ArcGIS Server. We almost immediately began receiving calls that our public maps were prompting for a username and password. I found log entries like this in our ArcGIS server log:
<Msg time="2022-05-24T09:55:15,281" type="WARNING" code="6602" source="Admin" process="14020" thread="1" methodName="" machine="gismachine.domain.com" user="" elapsed="" requestID="9c98dd85-e8dd-437b-9850-b27fc27e1075">Failed to return the service configuration 'PublicService.MapServer'. Could not connect to the ArcGIS component at URL 'https://ourportalmachine.domain.com:7443/arcgis/sharing/rest/content/items/8fd724e5f7bd4cdfb53821bcd...'. The ArcGIS component on that machine may not be running or the machine may not be reachable at this time.Error: Connect to ourportalmachine.domain.com:7443 [ourportalmachine.domain.com/xxx.xxx.xxx.xxx] failed: connect timed out</Msg>
Are there any settings such as caching that can be adjusted on the ArcGIS server machine to make it more fault tolerant to connection failures?
Any sort of redundancy such as a highly available configuration is probably off the table due to scarcity of resources at our organization. Thanks for any tips or suggestions!
Did you ever get this resolved? We are experiencing the exact same thing.
Would like to know this to as it is happening at a clients and it has been working fine prior and for a few weeks.
We are seeing similar issues with 11.2 environment. Did this ever get resolved?
Unfortunately no but we've had fewer outages recently to highlight the problem.
We had this issue but it went away after we put all our services in our Portal and then used the Portal connection to feed into the public web map. The login issue went away after that. Hasn't reoccurred since then. Originally we were using the rest services url when the login issue occurred.