Is there some explanation for the panic in pwnam_r when a session reconnects after a while? This is an exagerated example. The user leaves on a Thursday, comes back next Tuesday - bang! My Samba 3.0.24 runs under a SuSE 10.1. The passwd files are used for authentication. [2007/03/15 17:27:44, 1] smbd/service.c:close_cnum(1150) pmn12 (172.24.204.113) closed connection to service PRIMA [2007/03/20 11:08:31, 0] lib/fault.c:fault_report(41) ===========================================================[2007/03/20 11:08:31, 0] lib/fault.c:fault_report(42) INTERNAL ERROR: Signal 11 in pid 17526 (3.0.24-0.1.49-1172-SUSE-CODE10) Please read the Trouble-Shooting section of the Samba3-HOWTO [2007/03/20 11:08:31, 0] lib/fault.c:fault_report(44) From: http://www.samba.org/samba/docs/Samba3-HOWTO.pdf [2007/03/20 11:08:31, 0] lib/fault.c:fault_report(45) ==========================================================[2007/03/20 11:08:31, 0] lib/util.c:smb_panic(1599) PANIC (pid 17526): internal error [2007/03/20 11:08:31, 0] lib/util.c:log_stack_trace(1706) BACKTRACE: 30 stack frames: #0 /usr/sbin/smbd(log_stack_trace+0x1c) [0x555555752bcc] #1 /usr/sbin/smbd(smb_panic+0x43) [0x555555752cb3] #2 /usr/sbin/smbd [0x555555741212] #3 /lib64/libc.so.6 [0x2b36fc6e4b20] #4 /lib64/libc.so.6 [0x2b36fc7a258f] #5 /lib64/libc.so.6 [0x2b36fc7a0857] #6 /lib64/libc.so.6 [0x2b36fc7a0bf6] #7 /lib64/libc.so.6(getpwnam_r+0x138) [0x2b36fc745c78] #8 /lib64/libc.so.6(getpwnam+0x70) [0x2b36fc7455a0] #9 /usr/sbin/smbd(getpwnam_alloc+0x4f) [0x555555745c2f] #10 /usr/sbin/smbd [0x55555574566f] #11 /usr/sbin/smbd [0x55555574566f] #12 /usr/sbin/smbd [0x55555574566f] #13 /usr/sbin/smbd(Get_Pwnam_alloc+0x2a2) [0x555555745942] #14 /usr/sbin/smbd(smb_getpwnam+0x7d) [0x55555578f81d] #15 /usr/sbin/smbd(make_server_info_info3+0x199) [0x55555578fd79] #16 /usr/sbin/smbd [0x55555578e507] #17 /usr/sbin/smbd [0x55555578ea56] #18 /usr/sbin/smbd [0x55555578cc1e] #19 /usr/sbin/smbd [0x555555789e5f] #20 /usr/sbin/smbd [0x555555793ec6] #21 /usr/sbin/smbd [0x555555655f7e] #22 /usr/sbin/smbd(ntlmssp_update+0x253) [0x5555556547c3] #23 /usr/sbin/smbd(auth_ntlmssp_update+0x2a) [0x555555793bda] #24 /usr/sbin/smbd(reply_sesssetup_and_X+0x1651) [0x5555555f06c1] #25 /usr/sbin/smbd [0x555555617c22] #26 /usr/sbin/smbd(smbd_process+0x720) [0x555555618c00] #27 /usr/sbin/smbd(main+0xa0b) [0x5555557e594b] #28 /lib64/libc.so.6(__libc_start_main+0xf4) [0x2b36fc6d2154] #29 /usr/sbin/smbd [0x5555555b22a9] [2007/03/20 11:08:32, 0] lib/fault.c:dump_core(173) dumping core in /var/log/samba/cores/smbd [2007/03/20 12:04:29, 0] lib/fault.c:fault_report(41)
On Sun, Mar 25, 2007 at 10:57:14PM +0200, Dragan Krnic wrote:> Is there some explanation for the panic in pwnam_r when a session > reconnects after a while? > This is an exagerated example. The user leaves on a Thursday, > comes back next Tuesday - bang! > > My Samba 3.0.24 runs under a SuSE 10.1. > The passwd files are used for authentication.Is this in any way reproducable at will? You can for example force a reconnect by killing the smbd that is responsible for a particular client. You can find out the pid of that smbd by running "smbstatus". Then any subsequent access to that server from the client will trigger a reconnect. If that is reproducable, could you please install the debuginfo package that comes with your binary package. I am very sure I am able to fix that bug If I can get a valgrind log of that panic. Look at https://bugzilla.samba.org/ bug 4434 for information how to get the valgrind log. Thanks, Volker -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://lists.samba.org/archive/samba/attachments/20070325/e7355fae/attachment.bin
Skipped content of type multipart/mixed-------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part Url : http://lists.samba.org/archive/samba/attachments/20070326/a519c7cc/attachment.bin
Andrew, patch was missing, can you resend, please ? Guenther Andrew Bartlett wrote:> On Sun, 2007-03-25 at 23:18 +0200, Volker Lendecke wrote: >> On Sun, Mar 25, 2007 at 10:57:14PM +0200, Dragan Krnic wrote: >>> Is there some explanation for the panic in pwnam_r when a session >>> reconnects after a while? >>> This is an exagerated example. The user leaves on a Thursday, >>> comes back next Tuesday - bang! >>> >>> My Samba 3.0.24 runs under a SuSE 10.1. >>> The passwd files are used for authentication. >> Is this in any way reproducable at will? You can for example >> force a reconnect by killing the smbd that is responsible >> for a particular client. You can find out the pid of that >> smbd by running "smbstatus". Then any subsequent access to >> that server from the client will trigger a reconnect. >> >> If that is reproducable, could you please install the >> debuginfo package that comes with your binary package. >> >> I am very sure I am able to fix that bug If I can get a >> valgrind log of that panic. Look at >> https://bugzilla.samba.org/ bug 4434 for information how to >> get the valgrind log. > > As a random idea, I worked on, but never committed, this (attached) > patch for what I suspected to be some issues here. It might (or might > not) be the same bug, but I never got a confirmation from that reporter. > > Andrew Bartlett > >-- G?nther Deschner GPG-ID: 8EE11688 Red Hat gdeschner@redhat.com Samba Team gd@samba.org