Hi Mladen, I found it :)
Mladen Turk schrieb:
Rainer Jung wrote:
Mladen Turk wrote:
Rainer Jung wrote:
Does that make sense for you?
Yes it does. The most important is that if
there was no retries set for ajp worker, it gets copied
from lb worker.
In fact we don't copy (this is not new), because the retries for an lb
has a different functional meaning than for an lb sub worker. If an
ajp worker (main worker or lb sub worker) doesn't have a retries set,
it gets the default value JK_RETRIES (=2).
Hmm, I think the elected worker in mod_jk was setting the service's
retries, and then service was setting particular real workers retries.
I'm almost sure each worker inherited the retries from
lb (at least it should).
It was like that in 1.2.15. Then in July 2006, shortly before 1.2.18 you
changed that:
http://svn.apache.org/viewvc?view=rev&revision=420265
From that change on, the retries element in the service struct was no
more used. We set it, but didn't read from it anywhere.
Since retries for an lb worker does something different, than for an lb
ajp member worker, I find the situation to keep them separate fine.
At least my recent commit doesn't change the situation we had since 1.5
years :)
Regards,
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
