Patch already submitted to JDEV but no modification in actual code although it seems serious ? And I want to be the first in the contrib area download ... Argh ! It seems mtq.c has a little problem ... so all the threads of the pool become busy very quickly : here is a little patch : if we have found a not busy thread to work ... you could also do this just before the for(n = 0; n < MTQ_THREADS; n++) if(mtq__master->all[n]->busy == 0) { mp = mtq__master->all[n]->mp; mtq__master->all[n]->busy = 1; break; } feel free to contact me : rod@idealx.com PS: I hate firewall ...