Event timeout increases with every client loop

[olliewalsh]

The logic for the proc based client is ok:

Set timeout to 5 initially:

salt/salt/client/__init__.py

Line 638 in 055beeb

inc_timeout = timeout

Set start to time.time():

salt/salt/client/__init__.py

Line 642 in 055beeb

start = int(time.time())

If time.time() > start + timeout we have timed out:

salt/salt/client/__init__.py

Line 687 in 055beeb

if int(time.time()) > start + timeout:

If minions returned increase timeout by inc_timeout (5s)

salt/salt/client/__init__.py

Line 700 in 055beeb

timeout += inc_timeout

However for the event based client the event system is polled with a wait time that increases with each loop:

salt/salt/client/__init__.py

Line 749 in 055beeb

raw = self.event.get_event(timeout, jid)

[olliewalsh]

Related to pull req #8581 and issue #7157 (and a number of related issues)

[basepi]

Good detective work! Thanks for your continued work on these timeout issues, @olliewalsh!

[cachedout]

@olliewalsh Is this among the issues that was fixed with your recent round of changes? No worries if you're not sure, just wanted to check with you before I spent the time to try and reproduce. Thanks again!

[olliewalsh]

Yea, this is fixed too