Hi, Valgrind results:
output: # valgrind /usr/sbin/smbd ------------------- ==18143== Memcheck, a memory error detector. ==18143== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al. ==18143== Using LibVEX rev 1854, a library for dynamic binary translation. ==18143== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP. ==18143== Using valgrind-3.3.1-Debian, a dynamic binary instrumentation framework. ==18143== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al. ==18143== For more details, rerun with: -v ==18143== ==18143== ==18143== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 19 from 2) ==18143== malloc/free: in use at exit: 353,195 bytes in 395 blocks. ==18143== malloc/free: 1,005 allocs, 610 frees, 411,722 bytes allocated. ==18143== For counts of detected errors, rerun with: -v ==18143== searching for pointers to 395 not-freed blocks. ==18143== checked 670,752 bytes. ==18143== ==18143== LEAK SUMMARY: ==18143== definitely lost: 935 bytes in 7 blocks. ==18143== possibly lost: 301 bytes in 3 blocks. ==18143== still reachable: 351,959 bytes in 385 blocks. ==18143== suppressed: 0 bytes in 0 blocks. ==18143== Rerun with --leak-check=full to see details of leaked memory. SERVER:~# ==18144== ==18144== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 19 from 2) ==18144== malloc/free: in use at exit: 353,509 bytes in 400 blocks. ==18144== malloc/free: 1,031 allocs, 631 frees, 413,922 bytes allocated. ==18144== For counts of detected errors, rerun with: -v ==18144== searching for pointers to 400 not-freed blocks. ==18144== checked 671,104 bytes. ==18144== ==18144== LEAK SUMMARY: ==18144== definitely lost: 935 bytes in 7 blocks. ==18144== possibly lost: 301 bytes in 3 blocks. ==18144== still reachable: 352,273 bytes in 390 blocks. ==18144== suppressed: 0 bytes in 0 blocks. ==18144== Rerun with --leak-check=full to see details of leaked memory. ------------------- output: # valgrind --leak-check=full /usr/sbin/smbd ------------------- ==18247== Memcheck, a memory error detector. ==18247== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al. ==18247== Using LibVEX rev 1854, a library for dynamic binary translation. ==18247== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP. ==18247== Using valgrind-3.3.1-Debian, a dynamic binary instrumentation framework. ==18247== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al. ==18247== For more details, rerun with: -v ==18247== ==18247== ==18247== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 19 from 2) ==18247== malloc/free: in use at exit: 353,195 bytes in 395 blocks. ==18247== malloc/free: 1,005 allocs, 610 frees, 411,722 bytes allocated. ==18247== For counts of detected errors, rerun with: -v ==18247== searching for pointers to 395 not-freed blocks. ==18247== checked 670,752 bytes. ==18247== ==18247== 15 bytes in 1 blocks are definitely lost in loss record 2 of 30 ==18247== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18247== by 0x7917D71: strdup (in /lib/libc-2.7.so) ==18247== by 0x5A6D7B: string_set (in /usr/sbin/smbd) ==18247== by 0x478820: (within /usr/sbin/smbd) ==18247== by 0x47BA1D: lp_load_ex (in /usr/sbin/smbd) ==18247== by 0x47C328: lp_load_initial_only (in /usr/sbin/smbd) ==18247== by 0x46FE57: main (in /usr/sbin/smbd) ==18247== ==18247== ==18247== 85 bytes in 1 blocks are possibly lost in loss record 6 of 30 ==18247== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18247== by 0x748F80A: talloc_strdup (in /usr/lib/libtalloc.so.1.2.0) ==18247== by 0x5AEEDB: get_myname (in /usr/sbin/smbd) ==18247== by 0x5AEF00: myhostname (in /usr/sbin/smbd) ==18247== by 0x4787A9: (within /usr/sbin/smbd) ==18247== by 0x47BA1D: lp_load_ex (in /usr/sbin/smbd) ==18247== by 0x47C328: lp_load_initial_only (in /usr/sbin/smbd) ==18247== by 0x46FE57: main (in /usr/sbin/smbd) ==18247== ==18247== ==18247== 216 bytes in 2 blocks are possibly lost in loss record 12 of 30 ==18247== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18247== by 0x74910F3: _talloc_zero (in /usr/lib/libtalloc.so.1.2.0) ==18247== by 0x5BD8FA: event_context_init (in /usr/sbin/smbd) ==18247== by 0x46F4AB: smbd_event_context (in /usr/sbin/smbd) ==18247== by 0x46F4DF: smbd_messaging_context (in /usr/sbin/smbd) ==18247== by 0x46FEB1: main (in /usr/sbin/smbd) ==18247== ==18247== ==18247== 920 (648 direct, 272 indirect) bytes in 3 blocks are definitely lost in loss record 18 of 30 ==18247== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18247== by 0x7491575: _talloc_array (in /usr/lib/libtalloc.so.1.2.0) ==18247== by 0x5A6776: str_list_make (in /usr/sbin/smbd) ==18247== by 0x478DD1: (within /usr/sbin/smbd) ==18247== by 0x47BA1D: lp_load_ex (in /usr/sbin/smbd) ==18247== by 0x47C328: lp_load_initial_only (in /usr/sbin/smbd) ==18247== by 0x46FE57: main (in /usr/sbin/smbd) ==18247== ==18247== LEAK SUMMARY: ==18247== definitely lost: 663 bytes in 4 blocks. ==18247== indirectly lost: 272 bytes in 3 blocks. ==18247== possibly lost: 301 bytes in 3 blocks. ==18247== still reachable: 351,959 bytes in 385 blocks. ==18247== suppressed: 0 bytes in 0 blocks. ==18247== Reachable blocks (those to which a pointer was found) are not shown. ==18247== To see them, rerun with: --leak-check=full --show-reachable=yes jano:~# ==18249== ==18249== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 19 from 2) ==18249== malloc/free: in use at exit: 353,509 bytes in 400 blocks. ==18249== malloc/free: 1,031 allocs, 631 frees, 413,922 bytes allocated. ==18249== For counts of detected errors, rerun with: -v ==18249== searching for pointers to 400 not-freed blocks. ==18249== checked 671,104 bytes. ==18249== ==18249== 15 bytes in 1 blocks are definitely lost in loss record 2 of 30 ==18249== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18249== by 0x7917D71: strdup (in /lib/libc-2.7.so) ==18249== by 0x5A6D7B: string_set (in /usr/sbin/smbd) ==18249== by 0x478820: (within /usr/sbin/smbd) ==18249== by 0x47BA1D: lp_load_ex (in /usr/sbin/smbd) ==18249== by 0x47C328: lp_load_initial_only (in /usr/sbin/smbd) ==18249== by 0x46FE57: main (in /usr/sbin/smbd) ==18249== ==18249== ==18249== 85 bytes in 1 blocks are possibly lost in loss record 6 of 30 ==18249== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18249== by 0x748F80A: talloc_strdup (in /usr/lib/libtalloc.so.1.2.0) ==18249== by 0x5AEEDB: get_myname (in /usr/sbin/smbd) ==18249== by 0x5AEF00: myhostname (in /usr/sbin/smbd) ==18249== by 0x4787A9: (within /usr/sbin/smbd) ==18249== by 0x47BA1D: lp_load_ex (in /usr/sbin/smbd) ==18249== by 0x47C328: lp_load_initial_only (in /usr/sbin/smbd) ==18249== by 0x46FE57: main (in /usr/sbin/smbd) ==18249== ==18249== ==18249== 216 bytes in 2 blocks are possibly lost in loss record 12 of 30 ==18249== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18249== by 0x74910F3: _talloc_zero (in /usr/lib/libtalloc.so.1.2.0) ==18249== by 0x5BD8FA: event_context_init (in /usr/sbin/smbd) ==18249== by 0x46F4AB: smbd_event_context (in /usr/sbin/smbd) ==18249== by 0x46F4DF: smbd_messaging_context (in /usr/sbin/smbd) ==18249== by 0x46FEB1: main (in /usr/sbin/smbd) ==18249== ==18249== ==18249== 920 (648 direct, 272 indirect) bytes in 3 blocks are definitely lost in loss record 18 of 30 ==18249== at 0x4C2360E: malloc (vg_replace_malloc.c:207) ==18249== by 0x7491575: _talloc_array (in /usr/lib/libtalloc.so.1.2.0) ==18249== by 0x5A6776: str_list_make (in /usr/sbin/smbd) ==18249== by 0x478DD1: (within /usr/sbin/smbd) ==18249== by 0x47BA1D: lp_load_ex (in /usr/sbin/smbd) ==18249== by 0x47C328: lp_load_initial_only (in /usr/sbin/smbd) ==18249== by 0x46FE57: main (in /usr/sbin/smbd) ==18249== ==18249== LEAK SUMMARY: ==18249== definitely lost: 663 bytes in 4 blocks. ==18249== indirectly lost: 272 bytes in 3 blocks. ==18249== possibly lost: 301 bytes in 3 blocks. ==18249== still reachable: 352,273 bytes in 390 blocks. ==18249== suppressed: 0 bytes in 0 blocks. ==18249== Reachable blocks (those to which a pointer was found) are not shown. ==18249== To see them, rerun with: --leak-check=full --show-reachable=yes ------------------- Thanks 2009/7/27 Steve Langasek <vor...@debian.org> > On Mon, Jul 27, 2009 at 12:03:08PM +0200, klxout wrote: > > Package: samba > > Version:3.2.5-4lenny6 > > Severity: important > > > We have a production server with a PDC and BDC using as a domain with > > Ldap. Since weeks ago we have detected a high use of memory in PDC > > server, smbd processes start to increase memory usage and then crash > > ldap. It seems that can work correctly about one week then crash ldap. > > server has 1 GB of RAM > > Could you please run smbd under valgrind, to see where the memory leaks are > happening? > > -- > Steve Langasek Give me a lever long enough and a Free OS > Debian Developer to set it on, and I can move the world. > Ubuntu Developer http://www.debian.org/ > slanga...@ubuntu.com vor...@debian.org >