Improve netconf max devices test logic

Improve netconf max devices test logic

- there was some race condition where issuing a GET
(using the python getter.py tool) was not able to
get a 200 from every device initially after they
were all reported mounted. running the getter.py
tool once seems to resolve this and running it a
second time to do the actual checks for 200 on each
mounted device.

- also bumped the timeout factor from 2 to 3 since
I saw the larger iterations (750) hitting the timeout
once.

JIRA: NETCONF-711

Signed-off-by: Jamo Luhrsen <jluhrsen@gmail.com>
Change-Id: Iab274d89c44cc0a56bb4959e4ec67a0eaa345873