solving Maximum runtime exceeded problem

[feri]

hi
at first sry for my bad english

i have 11000 rrds for processing and when network goes bad i get time exceeded from poller
i think by my slef many processes solve this problem
i wrote my script to run spines from remote ssh and for example divergence 24 process to my machines, each one run spines and all the things goes well (multipoller was bugi for me and i didn't like it)
goes well means total polling time become better :d
i didn't read all of the codes for polling mechanism and didnt read one line of spine code, just know what going on and my script goes in spine path and act like that, no code of cacti changes at all, script just pass first and last ids into remote spines
but today network broke again and i see something very surprising that think i dont really know what goings on spines!
poller.php cant kill remote spines, and i see its about 1 hour spines are there and dont return any output in poller_output and seems they stop until get data query from some devices that cant reply, but why spine doesnot skip after number of trys!!!? :-?

do i describe situation good? :))
thanks