Submitted by: @AlexPeshkoff

Under heavy load classic server, actively working at the same time with events, gets locked.

That happens due to processing of lock manager and event signals in different order in different instances of fb_inet_server. One locked mutex for events and tries to lock mutex for lock manager, another locked mutex for lock manager and tries lo lock mutex for events. I.e. we have typical deadlock.

To avoid it we need to learn to inhibit signals not only from main code, but from AST's too.