Tell a story.

Tarantool remote requests are handled in a IProto thread, and
on_shutdown triggers are handled in the TX thread.

During shutdown we send a request to the leader, which introduces a race
condition in case we are the leader. If iproto thread shuts down before
the request is sent, it will never be handled.

The fix is simple: in case we are the leader during shutdown, don't send
the request, instead handle the request directly in the TX thread.

To highlight all the logging messages containing the given key do this:
* rust: `tlog::highlight_key(log_key, Some(color))`.
* lua: `picolib.log.highlight_key(log_key, color)`.

To disable highlight for the given key do this:
* rust: `tlog::highlight_key(log_key, None)`.
* lua: `picolib.log.highlight_key(log_key)`.

To disable highlighting of all keys do this:
* rust: `tlog::clear_highlight()`.
* lua: `picolib.log.clear_highlight()`.

Georgy Moshkin authored 2 years ago and Georgy Moshkin committed 2 years ago

Some users may change their "target-dir" cargo setting for example to
make it shared between all projects (may speed up some builds).
It that case our integration tests should be looking for the executable
in the right place.

It's already called `raft_conf_change_loop` in code.

Extend documentation with raft log processing algorithm.

Describe `conf_change_loop` and graceful shutdown logic.

Bring the algorithm description in accordance with actual
implementation. Also provide some comments to the code.