Hi, I yesterday and today encountered core dump in hobbitd_rrd. Here's the backtrace : (gdb) bt #0 0xb7dcd83b in raise () from /lib/tls/libc.so.6 #1 0xb7dcefa2 in abort () from /lib/tls/libc.so.6 #2 0x08057db1 in sigsegv_handler (signum=0) at sig.c:57 #3 <signal handler called> #4 0xb7e1c093 in strchr () from /lib/tls/libc.so.6 #5 0x08050972 in do_ncv_rrd (hostname=0xb7ce10f8 "orbadm-as01p.orb75.cvf", testname=0xb7ce110f "websvcs", msg=0x0, tstamp=1185778194) at do_ncv.c:51 #6 0x080529cb in update_rrd (hostname=0xb7ce10f8 "orbadm-as01p.orb75.cvf", testname=0xb7ce110f "websvcs", msg=0x0,tstamp=1185778194, sender=0x0, ldef=0x0) at do_rrd.c:411 #7 0x08049f69 in main (argc=0, argv=0xbff6f204) at hobbitd_rrd.c:160 (gdb) list 160 update_rrd(hostname, testname, restofmsg, tstamp, sender, ldef); 161 } 162 else if (strncmp(metadata[0], "@@shutdown", 10) == 0) { 163 running = 0; 164 continue; 165 } 166 else if (strncmp(metadata[0], "@@logrotate", 11) == 0) { 167 char *fn = xgetenv("HOBBITCHANNEL_LOGFILENAME"); 168 if (fn && strlen(fn)) { 169 freopen(fn, "a", stdout); My hobbitd_rrd is patched against the splitncv feature. It may be useful to know that the size of the core dump is 51MB. On fresh startup, the process is a few KB large. So I suspect a memory leak in that module, probably introduced by my patch. Alas, I won't be able to fix it soon, so beware. Regards, -- Charles Goyard - charles.goyard at orange-ftgroup.com - (+33) 1 45 38 01 31 Orange Business Services - online multimedia // ingénierie