You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@spamassassin.apache.org by bu...@bugzilla.spamassassin.org on 2004/02/01 16:12:18 UTC

[Bug 2991] New: spamd processes get stuck, eating all processor time

http://bugzilla.spamassassin.org/show_bug.cgi?id=2991

           Summary: spamd processes get stuck, eating all processor time
           Product: Spamassassin
           Version: 2.63
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: major
          Priority: P3
         Component: spamc/spamd
        AssignedTo: spamassassin-dev@incubator.apache.org
        ReportedBy: gnn@freemail.hu


Every now and then some spamd processes do not finish processing.  
  
This is a laptop, sometimes dialup Internet, sometimes LAN. I am using  
fetchmail to get my POP3 mails (couple hundreds at one time), gives them to  
exim4, then to procmail, and procmail calls spamc. spamd is running all the  
time.  
  
Usually during the first burst of messages I get 2-5 spamd processes stuck.  
I have no clue as to why they are stuck. All I see is that they don't eat up  
too much memory (3%), but they are in R status, and try to get all processor  
time available.  
Only way to get rid of them is kill -9  
I tried running spamd with debug switch once, but did not see anything  
helpful.  
  
I tried to wait to see if they finish processing sooner or later, but no, they  
don't. So it's not about a big message processed for a long time.  
  
Also it is not about heavy load. Yesterday I was killing the stuck processes  
whenever I saw them. I hardly got 1-2 mails in 15 minutes (when fetchmail  
runs), and still, some processes got stuck again.  
  
I used Debian woody with another laptop I had the same setup, but I did not  
have this problem. (Should have been version 2.20).  
Then I got a new laptop, installed Debian sarge (spamassassin 2.61, and I  
had this problem. I have upgraded to sid's 2.63, the problem remains)  
  
I tried to figure out what went wrong, but I don't have the skills it seems.  
Since I can reproduce this problem (not intentionally though), if you ask me  
to perform some tests, that's OK.  
  
Meanwhile some output:  
ps -ef f  
F S UID        PID  PPID  C PRI  NI ADDR SZ WCHAN  STIME TTY        TIME CMD 
1 S root       385     1  0  73   5 -  6396 select 14:00 ?        
0:01 /usr/sbin/spamd -c -m 10 -a -H -d --pidfile=/var/run/spamd.pid 
5 R gee        555   385 39  78  10 -  6633 -      14:02 ?         23:01  
\_ /usr/sbin/spamd -c -m 10 -a -H -d --pidfile=/var/run/spamd.pid 
5 R gee       1746   385 26  78  10 -  6699 -      14:16 ?         11:47  
\_ /usr/sbin/spamd -c -m 10 -a -H -d --pidfile=/var/run/spamd.pid 
 
netstat -tp 
Active Internet connections (w/o servers) 
Proto Recv-Q Send-Q Local Address           Foreign Address         State       
PID/Program name 
tcp        0      0 localhost:spamd         localhost:32774         CLOSE_WAIT 
555/spamd.pid 
tcp        0      0 localhost:spamd         localhost:33908         CLOSE_WAIT 
1746/spamd.pid 
 
I don't know what process they were communicating with, but there are no spamc 
processes running. I think there are nothing on the other end of these 
connections. 
 
swordfish:~# top -b -n1 
top - 15:11:55 up  1:12,  0 users,  load average: 3.30, 3.32, 3.20 
Tasks:  74 total,   5 running,  69 sleeping,   0 stopped,   0 zombie 
Cpu(s):  24.7% user,  19.7% system,  50.4% nice,   5.2% idle 
Mem:    256524k total,   247368k used,     9156k free,    26628k buffers 
Swap:   498952k total,    30660k used,   468292k free,    68660k cached 
 
  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND 
  532 root      14   0  1052 1052  860 R 28.1  0.4  
13:49.81 /usr/bin/fetchmail --daemon 300 --syslog -f /etc/fetchmailrc 
-i /var/mail/.fetchmail-UIDL-cache 
  555 gee       17  10 23848 7324 7216 R 16.8  2.9  24:50.38 /usr/sbin/spamd 
-c -m 10 -a -H -d --pidfile=/var/run/spamd.pid 
 1746 gee       17  10 24156  18m 7420 R 16.8  7.2  13:36.72 /usr/sbin/spamd 
-c -m 10 -a -H -d --pidfile=/var/run/spamd.pid 
 1486 root       5 -10 67920  33m 1892 S  1.9 13.4   0:39.09 /usr/bin/X11/X 
-dpi 100 -nolisten tcp 
 9295 root      11   0   952  952  756 R  1.9  0.4   0:00.04 top -b -n1 
 9357 root      15   0   712  708  580 R  1.9  0.3   0:00.01 sh -c uname -r 
... 
  385 root      13   5 22864 9728 6700 S  0.0  3.8   0:01.34 /usr/sbin/spamd 
-c -m 10 -a -H -d --pidfile=/var/run/spamd.pid 
 
So I think this is a serious issue. If I am not there to make spamd processes 
nice, or kill them, there will be many, and my computer can't do it's useful 
tasks. 
 
Best regards, 
Gabor Nagy



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.