Detect already running Persistent VMs

[nesitor]

Problem: Persistent VMs running were not detected after the orchestrator reboot.

Solution: Don't delete the entire table on the start process. Also added small fixes about guest_api and solved code quality issue.

[github-actions]

• Refactored start_guest_api method in executable.py by terminating process instead of joining it. This is more efficient as we don't need to wait for the process to finish before continuing with other tasks.
• Simplified stop_guest_api method in executable.py by replacing direct process termination with a check if the process is still running, which makes the code safer and easier to understand.
• Removed unnecessary drop of all tables from database in metrics.py. This change does not introduce any bugs but simplifies the code for future maintenance.
• Refactored operate_reboot function in views/operator.py by removing async await, which makes the code cleaner and easier to understand.
• Removed unnecessary VM forgetting from operate_erase function in views/operator.py. This change does not introduce any bugs but simplifies the code for future maintenance.
• Simplified stop_persistent_execution method in pool.py by removing async await, which makes the code cleaner and easier to understand.
• Removed unnecessary VM forgetting from forget_vm function in pool.py. This change does not introduce any bugs but simplifies the code for future maintenance.

RED: The PR includes larger changes that may have moderate risks or require deeper understanding of the project architecture. These changes are likely to introduce minor bugs and should be reviewed by experienced developers.

• Changed allocate_vm_ipv6_subnet method in IPv6Allocator protocol class from a coroutine to a synchronous function, which may cause issues if not handled correctly. This change is likely to introduce minor bugs and should be reviewed by experienced developers.

BLACK: The PR includes extensive changes that have a high potential for introducing bugs or require deep understanding of the project architecture. These changes are likely to introduce significant bugs and should only be reviewed by experienced developers.

• Removed async await from create_tables function in metrics.py, which may cause issues if not handled correctly. This change is likely to introduce significant bugs and should be reviewed by experienced developers.

Please note that the categorization of PRs can vary depending on specific rules or guidelines set by the team. The above analysis is based on general principles and might not fully capture all potential risks associated with each category. It's recommended to review these changes thoroughly before merging them into the main codebase.