Polymix-4 run starts from the fail point of the previous run

[Alex Rousskov]

On 12/21/18 8:17 PM, Srivattsan s wrote:

> I am using Web-polygraph 4.13 - polymix-4 test to benchmark my caching
> system. I use SSH to login to the remote test-bench. 

These two suggestions are not directly related to your question below,
but they might make your testing experience nicer:

* If you are using ssh to manually start tests, I recommend using
screen(1) or a similar tool to preserve your remote sessions between
logins (and when you lose connectivity to the remote box).

* Ideally, you should reboot the machines before each test and start the
tests without manually logging in. Ssh supports all that, but some
scripting is required to make it work smoothly.


> I noticed that the current state is being stored, but I am not sure if
> web-polygraph restores from the stored state in case of a failed
> previous run. Can someone help me with this issue? Thanks in advance!

Web Polygraph loads saved state only if it is explicitly told to load
that state (via command line options). Polygraph does not know what
happened (or did not happen) during previous tests(s) -- it simply
follows your instructions to load the specified state and "continue" the
test.

Since tests, especially long ones, often fail for reasons outside your
control, and Polygraph itself does not really know what you consider a
usable state as far as test resumption is concerned, Polygraph does not
automatically invalidate saved state when encountering test errors.

If you do not want Polygraph to continue a test (for any reason), then
start a new test without loading the old state instead of telling
Polygraph to continue the test.


HTH,

Alex.