Recommended Reboot Process for Highly Available Portal for ArcGIS

[AlexStevenson]

What is the recommended reboot process/order for rebooting the primary and standby servers serving Portal for ArcGIS?  We've been experiencing an issue with the portal page not loading after these servers come up from weekly scheduled reboots.  Once we stop the Portal for ArcGIS service on the standby and primary servers, then start up the primary and standby (in that order), the portal page works again.  

The errors from the logs show the following:

[JonathanQuinn]

If your portal is HA, I suggest you stop the standby, stop the primary, start the primary, and start the standby.

The problem is if you stop the standby and stop the primary, but then start the standby, it won't be functional as it can't connect to the database on the primary.  It also won't assume the role of primary as there may be data loss.  For example, you stop the standby and create some data on the primary.  The standby won't get the new data as you've stopped it.  You stop the primary and start the standby.  It can't receive the new data as the primary is not off, so it won't assume the role of primary.

[AlexStevenson]

Thanks, Jonathan.  We made the change to the scheduled reboots and I haven't seen the issue in the logs since then.

[BorjaParesFuente]

Hi Jonathan,

is there any link or web reference that could give datails of the start/stop services order on a AGE HA environment? This will be a very usefull information for a Disaster Recovery Plan.

Regards

Borja