Graceful shutdown if disk is full

[ligi]

this is a follow up of #20845

we should periodically monitor the free space on the disk we are operating on and if we fall under a certain limit we should gracefully shutdown the node to prevent database corruption

[howjmay]

Is this a good first issue?

[holiman]

I think this is a fairly good first issue. The feature should:

• Every N seconds, check the disk where the datadir resides,
• If available space is < some limit, maybe 500Mb, then start printing warnings

[howjmay]

Let me take a look, and try to implement it!

[howjmay]

I think this is a fairly good first issue. The feature should:

• Every N seconds, check the disk where the datadir resides,
• If available space is < some limit, maybe 500Mb, then start printing warnings

@holiman So does this be two stage operation? Once the free space is under stage 1 limit then geth will start to print warnings. If the free space is keeping dropping until the stage 2 limit, then geth will graceful shutdown

[vyrwu]

I made a pull request that implements this: #21884.

Would appreciate some review comments and guidance, I'm new to this repository and still learning, thx. ✌️

[alex347]

Please review a new implementation: #22103

[holiman]

Fixed via #22103