The mailing list for listmasters using Sympa

[Nick Holden <address@concealed>]

Dear all,

Having done some more research I've cleared those two messages (yes,
there were only two) from out of the bounce spool, and the bounced.pl
error messages have gone away.

But sympa itself is still not working properly, and taking up vast
amounts of system resources:

top - 12:26:12 up  2:21,  1 user,  load average: 2.80, 2.60, 2.20
Tasks: 127 total,   4 running, 123 sleeping,   0 stopped,   0 zombie
Cpu(s): 90.9% us,  8.6% sy,  0.0% ni,  0.0% id,  0.3% wa,  0.0% hi,
0.2% si
Mem:   1035848k total,   936064k used,    99784k free,    50240k buffers
Swap:  4008120k total,      848k used,  4007272k free,   145768k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+
COMMAND                                                                       
                                                      
 1925 sympa     25   0  143m 136m  10m R   98 13.5 120:06.03
sympa.pl                                                                      
                                                      
 1931 sympa     16   0 90644  83m 7984 R   26  8.3   4:05.40
task_manager.pl                                                               
                                                      
 3985 qscand    17   0  7772 5380 3868 S    5  0.5   0:00.16
perl5.8.8                                                                     
                                                      
 1827 snort     16   0 46900  42m 2392 S    0  4.2   0:13.84
snort                                                                         
                                                      
 3840 root      16   0  2288 1172 1880 R    0  0.1   0:00.10 top     


Mail is passing through the qmail service on the server without
problems, but mail to sympa and its aliases is not. Lists are not being
served and there are now several hundred messages, going back over two
days, in /home/sympa/spool/msg - how can I get them moving again? There
also seem to be tasks backing up in /home/sympa/spool/task

Any ideas?

Thanks,

Nick



On Wed, 2009-07-15 at 09:32 +0100, Nick Holden wrote:
> Hi all,
> 
> In recent days the amount of spam coming to our sympa server seems to
> have increased significantly, partly because some spammer has decided to
> spoof our sympa address in their spam, so the server is getting
> non-delivery reports.
> 
> One of the results of this is that bounced.pl is repeatedly trying to
> process bounce messages, including some which have no valid email
> address in them. The logs show lines like:
> 
> Jul 15 09:14:45 theweaa1 bounced[11105]: error: no address found in
> message from  for list ius 
> Jul 15 09:14:45 theweaa1 bounced[11105]: Cannot
> create /address@concealed/OTHER/ 
> Jul 15 09:14:49 theweaa1 bounced[11105]: error: no address found in
> message from  for list iufit 
> Jul 15 09:14:50 theweaa1 bounced[11105]: Cannot
> create /address@concealed/OTHER/ 
> Jul 15 09:14:50 theweaa1 bounced[11105]: error: no address found in
> message from  for list ius 
> Jul 15 09:14:50 theweaa1 bounced[11105]: Cannot
> create /address@concealed/OTHER/ 
> Jul 15 09:14:54 theweaa1 bounced[11105]: error: no address found in
> message from  for list iufit 
> Jul 15 09:14:55 theweaa1 bounced[11105]: Cannot
> create /address@concealed/OTHER/ 
> 
> It seems to be just those same two lists receiving either the same
> message over and over, or many messages. As it is just those two lists,
> I fear that there is some kind of loop of the same messages going over
> and over. 
> 
> Whatever it is starting to have an impact on server resources. Sympa,
> the task-manager and bounced.pl are starting to be the major drains on
> the system, which they haven't been up until now...
> 
> Tasks: 131 total,   4 running, 126 sleeping,   0 stopped,   1 zombie
> Cpu(s): 91.5% us,  6.6% sy,  0.0% ni,  0.0% id,  1.2% wa,  0.2% hi,
> 0.5% si
> Mem:   1035848k total,   980228k used,    55620k free,    20324k buffers
> Swap:  4008120k total,    98736k used,  3909384k free,   346192k cached
> 
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+
> COMMAND                                                                     
>                                                             
> 11094 sympa     25   0 77136  68m 9.9m R   52  6.8   6:51.31
> sympa.pl                                                                    
>                                                             
> 11275 www-data  15   0 41860  18m  27m S   33  1.8   9:09.31
> apache                                                                      
>                                                             
> 11109 sympa     15   0 84592  75m 9.8m S   33  7.5   1:16.51
> task_manager.pl                                                             
>                                                             
> 12991 sympa     22   0 43280  37m 8028 R   32  3.7   0:35.49
> perl5.8.8                                                                   
>                                                             
> 13011 www-data  15   0 31904 9.9m  25m S   11  1.0   0:01.91
> apache                                                                      
>                                                             
> 11105 sympa     16   0 29392  21m 9.8m S    0  2.2   0:03.14
> bounced.pl                                                                  
>                                                             
> 13210 root      16   0  2288 1172 1880 R    0  0.1   0:00.03
> top            
> 
> Can anyone suggest (a) what is happening here, and (b) what I should do
> to stop it? If it is the same two messages going round and round, can I
> get them out of the 'bounced.pl' queue somehow?
> 
> This is a sympa installation which has been working excellently for a
> long time up until this point.
> 
> Thanks,
> 
> Nick
> 
>