Fix support for infinite ?master_timeout

[DaveCTurner]

Specifying ?master_timeout=-1 on an API which performs a cluster state
update means that the cluster state update task will never time out
while waiting in the pending tasks queue. However this parameter is also
re-used in a few places where a timeout of -1 means something else,
typically to timeout immediately. This commit fixes those places so that
?master_timeout=-1 consistently means to wait forever.

[github-actions]

Documentation preview:

• ✨ Changed pages

[elasticsearchmachine]

Pinging @elastic/es-distributed (Team:Distributed)

[elasticsearchmachine]

Hi @DaveCTurner, I've created a changelog YAML for you.

[volodk85]

Shouldn't we compare with -1 explicitly here and everywhere down below?
Since user can specify any negative number in theory. Or doc should say that any negative number works as infinite timeout (which sounds confusing)

[DaveCTurner]

We only accept -1 when parsing a time value:

elasticsearch/libs/core/src/main/java/org/elasticsearch/core/TimeValue.java

Lines 375 to 384 in 0699c93

	if (value < -1) {
	// -1 is magic, but reject any other negative values
	throw new IllegalArgumentException(
	"failed to parse setting ["
	+ settingName
	+ "] with value ["
	+ initialInput
	+ "] as a time value: negative durations are not supported"
	);
	}

However the convention elsewhere in the code seems to be to compare .millis() < 0. I think we could make this neater for sure, but that's a job for another day. The docs definitely don't need to mention any lenience in this area.

[volodk85]

LGTM