With the update the behavior is different, but there is still something wrong I think. This is what I did: 1) Run syslog-ng 2) Run syslog_overflow_test.pl 3) The following lines are written to syslog: Mar 25 15:46:02 dockee jbq[12238]: 0 Mar 25 15:46:05 dockee 1 Mar 25 15:46:05 dockee 2 Mar 25 15:46:05 dockee 3 Mar 25 15:46:05 dockee 4 Mar 25 15:46:05 dockee 5 Mar 25 15:46:05 dockee 6 Mar 25 15:46:05 dockee 7 Mar 25 15:46:05 dockee 8 Mar 25 15:46:05 dockee 9 Mar 25 15:46:05 dockee 10 Mar 25 15:46:05 dockee 11 Mar 25 15:46:05 dockee 12 Mar 25 15:46:05 dockee 13 Mar 25 15:46:05 dockee 14 Mar 25 15:46:05 dockee 15 Mar 25 15:46:05 dockee 16 Mar 25 15:46:05 dockee 17 Mar 25 15:46:05 dockee 18 Mar 25 15:46:05 dockee 19 Mar 25 15:46:05 dockee 20 Mar 25 15:46:05 dockee 21 Mar 25 15:46:05 dockee 22 Mar 25 15:46:05 dockee 23 Mar 25 15:46:05 dockee 24 Mar 25 15:46:05 dockee 25 Mar 25 15:46:05 dockee 26 Mar 25 15:46:05 dockee 27 Mar 25 15:46:05 dockee 28 Mar 25 15:46:05 dockee 29 Mar 25 15:46:05 dockee 30 Mar 25 15:46:05 dockee 31 This block of messages is repeated many times during a short period of 3 seconds, but then no output appears anymore, and syslog-ng reports an error: Destination queue full, dropping message; queue_len='8192', mem_fifo_size='8192' If I increase the mem_fifo_size to 32000, syslog-ng is hogging 101% (!) of the CPU, and nothing is logged... Also, I notice that now every number is on its own line. This is not necessarily a problem, but it is different from what we had before, ie all numbers on the same line. Best regards, -- Jean-Baptiste Quenot http://caraldi.com/jbq/blog/