Re: Trouble getting a core dump from clamd
On Tue, January 19, 2010 12:35, Chuck Swiger wrote: Hi-- On Jan 19, 2010, at 10:04 AM, Doug Poland wrote: No joy. # sysctl kern.corefile=/var/crash/clamd.core # /usr/local/etc/rc.d/clamav-clamd start Starting clamav_clamd. Segmentation fault # ll /var/crash total 2 -rw-r--r-- 1 root wheel 5 Apr 10 2005 minfree Please see man 5 core; in particular, clamd changes userid: By default, a process that changes user or group credentials whether real or effective will not create a corefile. This behaviour can be changed to generate a core dump by setting the sysctl(8) variable kern.sugid_coredump to 1. ...so you will need to toggle that sysctl also. I'd suggest doing something more like: mkdir /cores chmod 1777 /cores sysctl kern.sugid_coredump=1 sysctl kern.corefile=/cores/%N.%P.core ...but adjust the path to be under /var or /usr if free diskspace makes doing so necessary. Thanks for the suggestion and the pointer to man 5 core. I implemented your suggestions but still get no core dump. Very strange... -- Regards, Doug ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
Hi-- On Jan 20, 2010, at 12:46 PM, Doug Poland wrote: Thanks for the suggestion and the pointer to man 5 core. I implemented your suggestions but still get no core dump. Very strange... Check your default resource limits (shell startup files /etc/login.conf) and see whether coredumpsize is set to 0, maybe? If you build and run: % cat crash.c #include stdlib.h main() { char *bad = NULL; bad[0] = '1'; } % cc -o crash crash.c % ./crash [1]16550 segmentation fault (core dumped) ./crash % ls /cores crash.16550.core ...do you get this trivial program to dump core? Regards, -- -Chuck ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
On Wed, January 20, 2010 15:05, Chuck Swiger wrote: Hi-- On Jan 20, 2010, at 12:46 PM, Doug Poland wrote: Thanks for the suggestion and the pointer to man 5 core. I implemented your suggestions but still get no core dump. Very strange... Check your default resource limits (shell startup files /etc/login.conf) and see whether coredumpsize is set to 0, maybe? Nope, nothin' % set | grep core | egrep -v grep % env | grep core | egrep -v grep % grep core /etc/login.c* | egrep -v # /etc/login.conf::coredumpsize=unlimited:\ If you build and run: % cat crash.c #include stdlib.h main() { char *bad = NULL; bad[0] = '1'; } % cc -o crash crash.c % ./crash [1]16550 segmentation fault (core dumped) ./crash % ls /cores crash.16550.core ...do you get this trivial program to dump core? No, not on the system in question. Just tried it on a plain-vanilla 8.0-RELEASE box I just built and I do get the the core dump there. % echo $SHELL /bin/tcsh % uname -a FreeBSD hostname.domain.com 7.2-RELEASE-p3 FreeBSD 7.2-RELEASE-p3 #0: Sat Sep 12 11:15:05 CDT 2009 r...@hostname.domain.com:/usr/obj/usr/src/sys/GENERIC-SERVER i386 % cat /sys/i386/conf/GENERIC-SERVER ident GENERIC-SERVER include GENERIC options INCLUDE_CONFIG_FILE # Include this file in kernel options QUOTA % sysctl -a | grep core kern.corefile: /var/coredumps/%U/%N.core kern.nodump_coredump: 0 kern.coredump: 1 kern.sugid_coredump: 1 debug.elf32_legacy_coredump: 1 % touch /var/coredumps/file % ll !$ ll /var/coredumps/file -rw-r--r-- 1 doug wheel 0 Jan 20 15:13 /var/coredumps/file -- Regards, Doug ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
Hi-- On Jan 20, 2010, at 1:24 PM, Doug Poland wrote: % sysctl -a | grep core kern.corefile: /var/coredumps/%U/%N.core Does /var/coredumps/doug and/or /var/coredumps/clamav exist and have appropriate permissions (or be 1777 like /tmp to avoid a problem with that)...? -- -Chuck ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
On Wed, 20 Jan 2010 15:24:51 -0600 Doug Poland wrote: % sysctl -a | grep core kern.corefile: /var/coredumps/%U/%N.core As CORE(5) says %U is a UID... % touch /var/coredumps/file % ll !$ ll /var/coredumps/file -rw-r--r-- 1 doug wheel 0 Jan 20 15:13 /var/coredumps/file ... so you should have a directory /var/coredumps/UID owned by (well, at least have permissions to write by) this UID. -- WBR, Boris Samorodov (bsam) Research Engineer, http://www.ipt.ru Telephone Internet SP FreeBSD Committer, http://www.FreeBSD.org The Power To Serve ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
On 2010-01-20 17:11, Boris Samorodov wrote: On Wed, 20 Jan 2010 15:24:51 -0600 Doug Poland wrote: % sysctl -a | grep core kern.corefile: /var/coredumps/%U/%N.core As CORE(5) says %U is a UID... % touch /var/coredumps/file % ll !$ ll /var/coredumps/file -rw-r--r-- 1 doug wheel 0 Jan 20 15:13 /var/coredumps/file ... so you should have a directory /var/coredumps/UID owned by (well, at least have permissions to write by) this UID. That was it, thanks for the keen vision! -- Regards, Doug ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Trouble getting a core dump from clamd
Hello, I'm running 7.2-RELEASE-p3 i386 and am having an issue getting a core dump from a program that is seg faulting. Last night, inexplicably, clamd started to seg fault. I was trying to obtain a core dump for further analysis but no .core file can be found. I've tried the following in a tcsh: # sysctl -a | grep core kern.corefile: %N.core kern.nodump_coredump: 0 kern.coredump: 1 kern.sugid_coredump: 0 debug.elf32_legacy_coredump: 0 # set ulimit=-c # /usr/local/sbin/clamd /usr/local/etc/clamd.conf Segmentation fault # tail /var/log/messages Jan 19 10:16:19 hostname kernel: pid 42315 (clamd), uid 106: exited on signal 11 But I get no .core file. Suggestions? -- Regards, Doug ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
2010/1/19 Doug Poland d...@polands.org Hello, I'm running 7.2-RELEASE-p3 i386 and am having an issue getting a core dump from a program that is seg faulting. Last night, inexplicably, clamd started to seg fault. I was trying to obtain a core dump for further analysis but no .core file can be found. I've tried the following in a tcsh: # sysctl -a | grep core kern.corefile: %N.core kern.nodump_coredump: 0 kern.coredump: 1 kern.sugid_coredump: 0 debug.elf32_legacy_coredump: 0 # set ulimit=-c # /usr/local/sbin/clamd /usr/local/etc/clamd.conf Segmentation fault # tail /var/log/messages Jan 19 10:16:19 hostname kernel: pid 42315 (clamd), uid 106: exited on signal 11 But I get no .core file. Suggestions? -- Regards, Doug ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org set a path in the sysctl variable kern.corefile. WIll make the core file easier to find if one is generated. Generally much tider as well then having core files littered all over the system ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
On Tue, January 19, 2010 11:10, krad wrote: 2010/1/19 Doug Poland d...@polands.org Hello, I'm running 7.2-RELEASE-p3 i386 and am having an issue getting a core dump from a program that is seg faulting. set a path in the sysctl variable kern.corefile. WIll make the core file easier to find if one is generated. Generally much tider as well then having core files littered all over the system No joy. # sysctl kern.corefile=/var/crash/clamd.core # /usr/local/etc/rc.d/clamav-clamd start Starting clamav_clamd. Segmentation fault # ll /var/crash total 2 -rw-r--r-- 1 root wheel 5 Apr 10 2005 minfree -- Regards, Doug ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Trouble getting a core dump from clamd
Hi-- On Jan 19, 2010, at 10:04 AM, Doug Poland wrote: No joy. # sysctl kern.corefile=/var/crash/clamd.core # /usr/local/etc/rc.d/clamav-clamd start Starting clamav_clamd. Segmentation fault # ll /var/crash total 2 -rw-r--r-- 1 root wheel 5 Apr 10 2005 minfree Please see man 5 core; in particular, clamd changes userid: By default, a process that changes user or group credentials whether real or effective will not create a corefile. This behaviour can be changed to generate a core dump by setting the sysctl(8) variable kern.sugid_coredump to 1. ...so you will need to toggle that sysctl also. I'd suggest doing something more like: mkdir /cores chmod 1777 /cores sysctl kern.sugid_coredump=1 sysctl kern.corefile=/cores/%N.%P.core ...but adjust the path to be under /var or /usr if free diskspace makes doing so necessary. Regards, -- -Chuck ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org