Thanks Daniel, I'll give that a go!
I'm surprised no-one has come across this before, is pfstatd not widely
used? I'm looking at using custom snmp traps to gather this information
instead, but that also looks like a non-standard method. What do most
people use for pf performance monitoring?
Cliff.
Daniel Hartmeier wrote:
When the process tries to write to the socket after the connection has
been closed, it gets a SIGPIPE signal. Without custom signal handling,
the default action is to terminate the process, see signal(3).
signal(3).
Basic socket programming issue, the author sucks. Try the patch below ;)
Daniel
Index: pfstatd.c
===================================================================
RCS file: /var/cvs/pfstat/pfstatd/pfstatd.c,v
retrieving revision 1.1.1.1
diff -u -r1.1.1.1 pfstatd.c
--- pfstatd.c 11 Jan 2007 16:01:58 -0000 1.1.1.1
+++ pfstatd.c 26 Mar 2008 17:58:05 -0000
@@ -39,6 +39,7 @@
#include <errno.h>
#include <fcntl.h>
#include <pwd.h>
+#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
@@ -97,6 +98,7 @@
}
if (argc != optind)
usage();
+ signal(SIGPIPE, SIG_IGN);
if ((fdp = open("/dev/pf", O_RDONLY)) < 0) {
fprintf(stderr, "open: /dev/pf: %s\n", strerror(errno));
return (1);
