When the SQL Server service (re)starts, it will always drop and recreate the tempdb database. Based on the error messages, you did something that made this impossible.

The error messages indicate that tempdb (or at least one of its files - tempdog.ldf) wants to be allocated on the G: drive. This fails with an error message that suggests that there iss not enough disk space. There are numerous scenarioss imaginable that may have caused this, but I guess you actually want a solution right now.

Here is a step by step guide: https://www.xtivia.com/start-sql-server-lost-tempdb-data-files/[/url].

After following that guide, once your instance is back up and running, try to figure out the actual required size of tempdb for your instance, and determine on which drives they should go; ensure that those drives will always have enough room for both the regular size and unexpected growth. Create multiple files for the data (all with equal size and autogrow paramteres) and a single file for the log, all big enough that they normally never need to grow; choose an autogrow setting that finds the middle ground between numerous tiny allocations or one single way-too-big allocation if growth is ever needed. Set trace flags 1117 and 1118 flags to optimize allocation and growth (see https://www.brentozar.com/archive/2014/06/trace-flags-1117-1118-tempdb-configuration/[/url]). Then restart the instance once more to make all these changes take effect.

And then ensure that nobody else ever touches your drives anymore.