Whenever I run a lot of xmlrpc requests, after a while the xmlrpc process
becomes unresponsive. It is still there, it doesn't crash, the rest of
openser works as expected, but the xmlrpc process simply doesn't answer
any request anymore until restarted. The number of requests I have to
issue to make this happen is variable. Sometimes is takes only 100
requests, other times is need up to 20000, but the end result is always
the same.
The xmlrpc request I ran is refreshWatchers(account, 'presence', 0) called
repeatedly until it no longer receives a reply. The account is a
subscriber that is present in the system. But I also tried pwd and ps and
got the same result.
Below is the strace right before and while it got stuck. As it can be seen
it got stuck in futex() which seems to indicate some sort of locking
mechanism that has entered a deadlock. The futex is not called before
that point in the whole strace output, and when it is called it locks.
--- SIGCHLD (Child exited) @ 0 (0) ---
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 4353
close(9) = 0
waitpid(-1, 0xbfdcca6c, WNOHANG) = -1 ECHILD (No child processes)
sigreturn() = ? (mask now [])
getpeername(8, {sa_family=AF_INET, sin_port=htons(38926),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|
SIGCHLD, child_tidptr=0xb7c69bc8) = 4354
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
accept(6, {sa_family=AF_INET, sin_port=htons(38928),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 9
--- SIGCHLD (Child exited) @ 0 (0) ---
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 4354
close(8) = 0
waitpid(-1, 0xbfdcca6c, WNOHANG) = -1 ECHILD (No child processes)
sigreturn() = ? (mask now [])
getpeername(9, {sa_family=AF_INET, sin_port=htons(38928),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|
SIGCHLD, child_tidptr=0xb7c69bc8) = 4355
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
accept(6, {sa_family=AF_INET, sin_port=htons(38930),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 8
--- SIGCHLD (Child exited) @ 0 (0) ---
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 4355
close(9) = 0
waitpid(-1, 0xbfdcca6c, WNOHANG) = -1 ECHILD (No child processes)
sigreturn() = ? (mask now [])
getpeername(8, {sa_family=AF_INET, sin_port=htons(38930),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|
SIGCHLD, child_tidptr=0xb7c69bc8) = 4356
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
accept(6, {sa_family=AF_INET, sin_port=htons(38932),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 9
--- SIGCHLD (Child exited) @ 0 (0) ---
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 4356
close(8) = 0
waitpid(-1, 0xbfdcca6c, WNOHANG) = -1 ECHILD (No child processes)
sigreturn() = ? (mask now [])
getpeername(9, {sa_family=AF_INET, sin_port=htons(38932),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|
SIGCHLD, child_tidptr=0xb7c69bc8) = 4357
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
accept(6, {sa_family=AF_INET, sin_port=htons(38934),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 8
--- SIGCHLD (Child exited) @ 0 (0) ---
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 4357
close(9) = 0
waitpid(-1, 0xbfdcca6c, WNOHANG) = -1 ECHILD (No child processes)
sigreturn() = ? (mask now [])
getpeername(8, {sa_family=AF_INET, sin_port=htons(38934),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|
SIGCHLD, child_tidptr=0xb7c69bc8) = 4358
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
accept(6, {sa_family=AF_INET, sin_port=htons(38936),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 9
--- SIGCHLD (Child exited) @ 0 (0) ---
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 4358
close(8) = 0
waitpid(-1, 0xbfdcca6c, WNOHANG) = -1 ECHILD (No child processes)
sigreturn() = ? (mask now [])
getpeername(9, {sa_family=AF_INET, sin_port=htons(38936),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|
SIGCHLD, child_tidptr=0xb7c69bc8) = 4359
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
accept(6, {sa_family=AF_INET, sin_port=htons(38938),
sin_addr=inet_addr("10.0.0.146")}, [16]) = 8
kill(4359, SIG_0) = 0
--- SIGCHLD (Child exited) @ 0 (0) ---
waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 4359
close(9) = 0
futex(0xb7dc9140, FUTEX_WAIT, 2, NULL
--
Dan
_______________________________________________
Devel mailing list
Devel@lists.openser.org
http://lists.openser.org/cgi-bin/mailman/listinfo/devel
