Aki,
(Not speaking for Aki)
I understand that salted passwords saved in my database and stronger hash algorithm course that it will require more processor time/power to crack my passwords.
But only when hackers have direct access to my database what means that hackers have access to my passwords hashes (eg. hackers stolen my database).
My Dovecot use passwords saved in database as SHA256 and hackers can use only SMTP, IMAP or POP3 services to try crack it using dictionary attack (I understand that they using plain text dictionaty passwords).
Stronger hash algorithm and salt is useful when hackers have direct access to my database but when they use services as SMTP, IMAP or POP3 to crack passwords only longer and more complicated password can be more secure.
I do not understand this correctly ?
Yes, your understanding is basically correct. However, history gives lots of examples of broken systems that explicitly or implicitly relied on one critical system not failing -- they lacked defense in depth or resilience.
Examples are "this system has no bugs", "my system does not leak hashes", "this algorithm is unbreakable", "we'll never see a CAT5 hurricane", etc. If these critical assumption ever becomes untrue, the foundation of your defense crumbles.
If you narrow your attack definition to only include in-protocol remote brute forcing, then any decent password will take far too long to break that way (esp. with throttling controls that are built-in). Your log files will overflow recording the attempts long before you can expect a password to be cracked. However, you're still susceptable to the qwerty passwords. If this is your *only* line of defense, it is brittle.
A robustly secure system will overlap protection: strong hashes, password compliance systems, brute force countermeasures, file permissions/OS hardening, network origins vetting, anti-DoS measures, etc.
Keep this picture in mind that I found on CLCERT
https://www.clcert.cl/humor/img/weakest-link-road.jpgJoseph Tam <jtam.home@gmail.com>