We have an Umbraco installation that takes around a minute to boot. Sometimes Umbraco will successfully boot. But some of the time Umbraco will fail to boot with the “failed to boot” screen and the logs will point to database time-outs as the reason.
However, even after the database is verifiably warmed up and radiply accessible from Umbraco (using a diagnostic page outside of Umbraco), Umbraco still shows the “failed to boot” screen immediately on page refresh, and the logs will still point to database time-outs as the reason. Only now these time-outs are happening immediately.
So it seems Umbraco is “caching” the failure to boot, including the reason for boot failure. Is this correct? Is there a way to turn off this “caching” behavior, so that subsequent page refreshs would result in a successful boot?
If not, is there a way to increase the time-outs during boot, specifically the database time-out? To stop the boot failures occuring in the first place.
Or is there a different way to try to get Umbraco to boot after a “failed to boot”?
The Umbraco connection string is verified to work, and sometimes Umbraco will boot after a restart.
The installation is load-balanced admin/front setup running on Azure Web Apps. We’ve never had any problems with this type of setup before.
The same solution works in the production environment. This is the test environment with rougly the same server sizing.
The only thing special about this Umbraco solution is that it takes long to start, probably because of custom start-up code or similar.
Umbraco 8.13.
This is a companion discussion topic for the original entry at https://our.umbraco.com/forum/106389-umbraco-fails-to-boot-even-after-database-connection-is-warmed-up-it-seems-the-failture-to-boot-is-cached