No subject
Adrian Voinea
adrian at gds.ro
Sun Feb 8 00:31:49 GMT 2004
Hello, here's my problem:
Sometimes, mailscanner uses a lot of cpu, for a long amount of time.
The mail server's load average rises to ~13 and stays that way for a while.
You can see the cpu hogs' activity in the strace output.
(I searched for "shmat and umoven: Input/output error" on google and this
error is obviously related to
the high cpu usage:
http://www.google.com/search?sourceid=navclient&ie=UTF-8&oe=UTF-8&q=shmat+umoven+cpu)
What do you think is wrong? Did anyone else encounter this error?
System configuration:
Celeron 2.4
512 ram
Kernel 2.4.23
perl 5.6.1
sendmail 8.12.11
mailscanner 4.26.7, sa+bayes enabled
spamassassin 2.63
#ps auxf output(some of the processes that consume lots of cpu time):
root 7823 0.8 21.7 117172 112232 ? S 07:48 1:42 \_
/usr/bin/perl -I/opt/MailScanner/lib /opt/MailScanner/bin/MailScanner
/opt/MailScanner/etc/MailScanner.conf
root 8149 1.0 19.7 117284 102152 ? S 08:06 1:47 \_
/usr/bin/perl -I/opt/MailScanner/lib /opt/MailScanner/bin/MailScanner
/opt/MailScanner/etc/MailScanner.conf
#strace -p 7823 output:
setup() = 0
time([1076057709]) = 1076057709
chdir("/var/spool/mqueue.in") = 0
open(".", O_RDONLY|O_NONBLOCK|0x18000) = 7
fstat64(0x7, 0xbffffb3c) = 0
shmat(7, 0x89eab58, 0x2ptrace: umoven: Input/output error
) = ?
ipc_subcall(0x7, 0x96db440, 0x1000, 0xb) = 368
ipc_subcall(0x7, 0x96db440, 0x1000, 0xb) = 0
close(7) = 0
umask(0177) = 077
umask(077) = 0177
time([1076057709]) = 1076057709
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
rt_sigaction(SIGCHLD, NULL, {SIG_DFL}, 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
nanosleep({30, 0}, {30, 0}) = 0
time([1076057739]) = 1076057739
chdir("/var/spool/mqueue.in") = 0
open(".", O_RDONLY|O_NONBLOCK|0x18000) = 7
fstat64(0x7, 0xbffffb3c) = 0
shmat(7, 0x89eab58, 0x2ptrace: umoven: Input/output error
) = ?
ipc_subcall(0x7, 0x96db440, 0x1000, 0xb) = 568
stat64(0x86254e0, 0x80f51e0) = 0
stat64(0x86254e0, 0x80f51e0) = 0
ipc_subcall(0x7, 0x96db440, 0x1000, 0xb) = 0
close(7) = 0
umask(0177) = 077
open("/var/spool/mqueue.in/qfi168t9Rc012372", O_RDWR|0x8000) = 7
fstat64(0x7, 0x80f5380) = 0
shmat(7, 0x96c5298, 0x2ptrace: umoven: Input/output error
) = ?
flock(7, LOCK_EX|LOCK_NB) = -1 EAGAIN (Resource temporarily
unavailable)
close(7) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
open("/var/spool/mqueue.in/qfi168tRg2012377", O_RDWR|0x8000) = 7
fstat64(0x7, 0x80f5380) = 0
shmat(7, 0x97440b8, 0x2ptrace: umoven: Input/output error
) = ?
flock(7, LOCK_EX|LOCK_NB) = 0
open("/var/spool/mqueue.in/dfi168tRg2012377", O_RDWR|0x8000) = 8
fstat64(0x8, 0x80f5380) = 0
shmat(8, 0x96cd8a0, 0x2ptrace: umoven: Input/output error
) = ?
flock(8, LOCK_EX|LOCK_NB) = 0
stat64(0x9648758, 0x80f51e0) = 0
stat64(0x9648758, 0x80f51e0) = 0
stat64(0x963a680, 0x80f51e0) = 0
stat64(0x963a680, 0x80f51e0) = 0
fstat64(0x7, 0xbffff95c) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
0) = 0x403aa000
_llseek(7, 0, [0], SEEK_SET) = 0
read(7, "V6\nT1076057730\nK0\nN0\nP32901\nF8bs"..., 4096) = 928
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
open("/opt/MailScanner/incoming/7823/i168tRg2012377.header",
O_WRONLY|O_CREAT|O_TRUNC|0x8000, 0666) = 9
fstat64(0x9, 0x80f5380) = 0
shmat(9, 0xb95fd18, 0x2ptrace: umoven: Input/output error
) = ?
The only error I get in the logs is:
Feb 6 11:01:11 kiki MailScanner[8096]: SpamAssassin timed out and was
killed, consecutive failure 1 of 50
Thanks,
Adrian
More information about the MailScanner
mailing list