Hi,
Has anyone any ideas what is causing my smbd processes to crash? The files
being accessed via samba are on an NFS mounted directory. In the samba log
file I am getting the following errors:
[2004/12/03 16:31:33, 0] smbd/oplock.c:request_oplock_break(1026)
request_oplock_break: no response received to oplock break request to pid
2181
9 on port 53336 for dev = 2c00002, inode = 297939333
for dev = 2c00002, inode = 297939333, tv_sec = 41b09217, tv_usec = 2be1b
[2004/12/03 16:31:33, 0] smbd/open.c:open_mode_check(555)
open_mode_check: exlusive oplock left by process 21819 after break ! For
file
5000_photos/Thumbs.db, dev = 2c00002, inode = 297939333. Deleting it to
continue
...
[2004/12/03 16:31:33, 0] lib/util.c:smb_panic(1055)
PANIC: open_mode_check: Existant process 21819 left active oplock.
and a truss -d for the process which has crashed gives the following:
*** SUID: ruid/euid/suid = 0 / 2053 / 2053 ***
*** SGID: rgid/egid/sgid = 0 / 500 / 500 ***
poll(0xFFBEED08, 2, 32000) (sleeping...)
poll(0xFFBEED08, 2, 32000) = 0
time() = 1102092631
getuid() = 0 [2053]
write(25, " [", 1) = 1
write(25, " 2 0 0 4 / 1 2 / 0 3 1".., 19) = 19
write(25, " , 0", 3) = 3
write(25, " ] ", 2) = 2
write(25, " s m b d / o p l o c k .".., 13) = 13
write(25, " :", 1) = 1
write(25, " r e q u e s t _ o p l o".., 20) = 20
write(25, " (", 1) = 1
write(25, " 1 0 2 6", 4) = 4
write(25, " )\n", 2) = 2
getuid() = 0 [2053]
write(25, " r e q u e s t _ o p".., 133) = 133
getuid() = 0 [2053]
write(25, " f o r d e v = ".., 75) = 75
fcntl(14, F_SETLKW64, 0xFFBEEEA8) = 0
time() = 1102092631
getuid() = 0 [2053]
write(25, " [", 1) = 1
write(25, " 2 0 0 4 / 1 2 / 0 3 1".., 19) = 19
write(25, " , 0", 3) = 3
write(25, " ] ", 2) = 2
write(25, " s m b d / o p e n . c", 11) = 11
write(25, " :", 1) = 1
write(25, " o p e n _ m o d e _ c h".., 15) = 15
write(25, " ( 5 5 5", 4) = 4
write(25, " )\n", 2) = 2
getuid() = 0 [2053]
write(25, " o p e n _ m o d e _".., 164) = 164
kill(21819, SIG#0) = 0
time() = 1102092631
getuid() = 0 [2053]
write(25, " [", 1) = 1
write(25, " 2 0 0 4 / 1 2 / 0 3 1".., 19) = 19
write(25, " , 0", 3) = 3
write(25, " ] ", 2) = 2
write(25, " l i b / u t i l . c", 10) = 10
write(25, " :", 1) = 1
write(25, " s m b _ p a n i c", 9) = 9
write(25, " (", 1) = 1
write(25, " 1 0 5 5", 4) = 4
write(25, " )\n", 2) = 2
getuid() = 0 [2053]
write(25, " P A N I C : o p e".., 69) = 69
getuid() = 0 [2053]
write(25, " \n", 3) = 3
getuid() = 0 [2053]
sigaction(SIGABRT, 0x00000000, 0xFFBEEFB0) = 0
llseek(0, 0, SEEK_CUR) = 0
sigaction(SIGABRT, 0xFFBEEE80, 0xFFBEEF00) = 0
sigprocmask(SIG_UNBLOCK, 0xFFBEEF30, 0x00000000) = 0
getpid() = 22241 [21295]
kill(22241, SIGABRT) = 0
Received signal #6, SIGABRT [default]
siginfo: SIGABRT pid=22241 uid=0
*** process killed ***
In samba I have all op locks options turned on. Is the NFS directory( on a
snap server) not responding correctly to the op locks, or is there another
problem?
Thanks in advance.
David
On Fri, Dec 03, 2004 at 05:47:08PM -0000, ALLEN, David wrote:> Hi, > > Has anyone any ideas what is causing my smbd processes to crash? The files > being accessed via samba are on an NFS mounted directory. In the samba log > file I am getting the following errors: > > [2004/12/03 16:31:33, 0] smbd/oplock.c:request_oplock_break(1026) > request_oplock_break: no response received to oplock break request to pid > 2181 > 9 on port 53336 for dev = 2c00002, inode = 297939333 > for dev = 2c00002, inode = 297939333, tv_sec = 41b09217, tv_usec = 2be1b > [2004/12/03 16:31:33, 0] smbd/open.c:open_mode_check(555) > open_mode_check: exlusive oplock left by process 21819 after break ! For > file > 5000_photos/Thumbs.db, dev = 2c00002, inode = 297939333. Deleting it to > continue > ... > [2004/12/03 16:31:33, 0] lib/util.c:smb_panic(1055) > PANIC: open_mode_check: Existant process 21819 left active oplock.I'm currently looking into this on another bug report. It may be that one of my paranoia checks is a little too strict - the person looking into this has just commented the check out and it seems to be ok - but I'm going to look at this closely. I have another problem to fix on my plate first though. Jeremy.
Hi, We may have solved this problem by switching the NFS mount from tcp to udp (specify proto=udp in vfstab). Has anyone any suggestions why this should solve the problem, and are there any disadvantages? Thanks in advance. David> -----Original Message----- > From: ALLEN, David > Sent: 03 December 2004 17:47 > To: 'samba@lists.samba.org' > Subject: NFS mounted directory > > Hi, > > Has anyone any ideas what is causing my smbd processes to crash? The files > being accessed via samba are on an NFS mounted directory. In the samba log > file I am getting the following errors: > > [2004/12/03 16:31:33, 0] smbd/oplock.c:request_oplock_break(1026) > request_oplock_break: no response received to oplock break request to > pid 2181 > 9 on port 53336 for dev = 2c00002, inode = 297939333 > for dev = 2c00002, inode = 297939333, tv_sec = 41b09217, tv_usec = 2be1b > [2004/12/03 16:31:33, 0] smbd/open.c:open_mode_check(555) > open_mode_check: exlusive oplock left by process 21819 after break ! For > file > 5000_photos/Thumbs.db, dev = 2c00002, inode = 297939333. Deleting it to > continue > ... > [2004/12/03 16:31:33, 0] lib/util.c:smb_panic(1055) > PANIC: open_mode_check: Existant process 21819 left active oplock. > > and a truss -d for the process which has crashed gives the following: > > *** SUID: ruid/euid/suid = 0 / 2053 / 2053 *** > *** SGID: rgid/egid/sgid = 0 / 500 / 500 *** > poll(0xFFBEED08, 2, 32000) (sleeping...) > poll(0xFFBEED08, 2, 32000) = 0 > time() = 1102092631 > getuid() = 0 [2053] > write(25, " [", 1) = 1 > write(25, " 2 0 0 4 / 1 2 / 0 3 1".., 19) = 19 > write(25, " , 0", 3) = 3 > write(25, " ] ", 2) = 2 > write(25, " s m b d / o p l o c k .".., 13) = 13 > write(25, " :", 1) = 1 > write(25, " r e q u e s t _ o p l o".., 20) = 20 > write(25, " (", 1) = 1 > write(25, " 1 0 2 6", 4) = 4 > write(25, " )\n", 2) = 2 > getuid() = 0 [2053] > write(25, " r e q u e s t _ o p".., 133) = 133 > getuid() = 0 [2053] > write(25, " f o r d e v = ".., 75) = 75 > fcntl(14, F_SETLKW64, 0xFFBEEEA8) = 0 > time() = 1102092631 > getuid() = 0 [2053] > write(25, " [", 1) = 1 > write(25, " 2 0 0 4 / 1 2 / 0 3 1".., 19) = 19 > write(25, " , 0", 3) = 3 > write(25, " ] ", 2) = 2 > write(25, " s m b d / o p e n . c", 11) = 11 > write(25, " :", 1) = 1 > write(25, " o p e n _ m o d e _ c h".., 15) = 15 > write(25, " ( 5 5 5", 4) = 4 > write(25, " )\n", 2) = 2 > getuid() = 0 [2053] > write(25, " o p e n _ m o d e _".., 164) = 164 > kill(21819, SIG#0) = 0 > time() = 1102092631 > getuid() = 0 [2053] > write(25, " [", 1) = 1 > write(25, " 2 0 0 4 / 1 2 / 0 3 1".., 19) = 19 > write(25, " , 0", 3) = 3 > write(25, " ] ", 2) = 2 > write(25, " l i b / u t i l . c", 10) = 10 > write(25, " :", 1) = 1 > write(25, " s m b _ p a n i c", 9) = 9 > write(25, " (", 1) = 1 > write(25, " 1 0 5 5", 4) = 4 > write(25, " )\n", 2) = 2 > getuid() = 0 [2053] > write(25, " P A N I C : o p e".., 69) = 69 > getuid() = 0 [2053] > write(25, " \n", 3) = 3 > getuid() = 0 [2053] > sigaction(SIGABRT, 0x00000000, 0xFFBEEFB0) = 0 > llseek(0, 0, SEEK_CUR) = 0 > sigaction(SIGABRT, 0xFFBEEE80, 0xFFBEEF00) = 0 > sigprocmask(SIG_UNBLOCK, 0xFFBEEF30, 0x00000000) = 0 > getpid() = 22241 [21295] > kill(22241, SIGABRT) = 0 > Received signal #6, SIGABRT [default] > siginfo: SIGABRT pid=22241 uid=0 > *** process killed *** > > In samba I have all op locks options turned on. Is the NFS directory( on a > snap server) not responding correctly to the op locks, or is there another > problem? > > Thanks in advance. > > David