> I have to download 3 million of pure HTML files
> without reading links, here the problems:
You have a list of 3 million urls? Otherwise your statement makes no sense.

> 1) I have to resume download without restart
> scanning and check. How?If you canceled or killed the program, there is no
resume.
If you paused, unpause(resume) there is no rescan.

> 2) I need fast proxy. Where?Not an HTT question. You're still limited by
your connection, proxy is irrelevant.