HTTP retries

[sebb <se...@gmail.com>]

The current JMeter release candidate uses zero retries (default) for
the HC4 implementation, the default for HC3.1 (3) and Java (1 retry, I
think).

We noticed that the HC4 implementation shows occasional errors when
run against some websites, whereas the HC3 does not.
When the HC4 implementation is run with 3 retries, the errors disappear.

Just wondering why the default of 3 was chosen? Is that what other
applications such as browsers do?

Also, I notice that the retry count is available from the
RequestWrapper (getExecCount(), subtract 1) in HC4.

It does not appear to be available in HC3.1 - just checking, is that correct?

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org

Re: HTTP retries

[Oleg Kalnichevski <ol...@apache.org>]

On Fri, 2011-09-23 at 14:26 +0100, sebb wrote:
> The current JMeter release candidate uses zero retries (default) for
> the HC4 implementation, the default for HC3.1 (3) and Java (1 retry, I
> think).
> 
> We noticed that the HC4 implementation shows occasional errors when
> run against some websites, whereas the HC3 does not.
> When the HC4 implementation is run with 3 retries, the errors disappear.
> 
> Just wondering why the default of 3 was chosen? Is that what other
> applications such as browsers do?
> 

Purely by accident.

> Also, I notice that the retry count is available from the
> RequestWrapper (getExecCount(), subtract 1) in HC4.
> 
> It does not appear to be available in HC3.1 - just checking, is that correct?
> 

I believe so. It's been years since I looked at the HC 3.x code, though.

Oleg



---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org