Am Montag, 30. Juli 2007 10:09 schrieb Robert Felber: > On Sat, Jul 28, 2007 at 09:59:02AM +0200, Thomas Krieger wrote: > > Hello, > > > > last night I encountered a serious problem with policyd-weight. It > > crahsed and therefore mails bounced with 450 Server configuration > > problem. > > Did the master became unresponsive as well? If so, any messages > from the master process? > > You may want to try out following: > > send a PIPE sig via kill to a child process > after that, try to telnet to 12525 if you get a connection > then things "should" be ok. > > If not, then I have to rethink about signal trapping and signal > logging. It works here, on a FBSD, gentoo (kern 2.6) and redhat 7.1 (kern > 2.4) > > Which wouldn't explain why the children got a SIGPIPE, though. > > > I use the latest devel version of policyd-weight. No changes made to the > > program. > > > > # DATE: Thu Jul 19 08:43:25 CEST 2007 > > # VERSION: 0.1.14 beta-6 > > > > Here's what I found in the mail server's logfile: > > > > Jul 28 05:36:00 mail postfix/policyd-weight[13047]: weighted check: > > IN_DYN_PBL_SPAMHAUS=3.25 NOT_IN_SBL_XBL_SPAMHAUS=-1.5 NOT_IN_SPAMCOP=-1.5 > > NOT_IN_BL_NJABL=-1.5 CL_IP_NE_HELO=4.75 REV_IP_EQ_HELO=-1.25 > > NOK_HELO_SEEMS_DIALUP=5 (check from: .ballco. - > > helo: .dsl-189-155-176-14.prod-infinitum. - helo-domain: > > .prod-infinitum.) FROM/MX_MATCHES_NOT_UNVR_HELO(DOMAIN)=4.85 > > CLIENT_NOT_MX/A_FROM_DOMAIN=4.75 CLIENT/24_NOT_MX/A_FROM_DOMAIN=4.75; > > <client=189.155.176.14> > > <helo=dsl-189-155-176-14.prod-infinitum.com.mx> <[EMAIL PROTECTED] > > o.de> <[EMAIL PROTECTED]>; rate: 21.6 > > Jul 28 05:36:00 mail postfix/policyd-weight[13047]: decided action=550 > > Mail appeared to be SPAM or forged. Ask your Mail/DNS-Administrator to > > correct HELO and DNS MX settings or to get removed from DNSBLs; please > > relay via your ISP (ballco.de); Please use DynDNS; > > <client=189.155.176.14> > > <helo=dsl-189-155-176-14.prod-infinitum.com.mx> <[EMAIL PROTECTED] > > lco.de> <[EMAIL PROTECTED]>; delay: 7s > > Jul 28 05:36:00 mail postfix/policyd-weight[13047]: warning: Got SIGPIPE. > > Child 13047 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[13047]: warning: child: err: > > Died at /usr/local/sbin/policyd-weight-devel line 1264, <GEN54868> line > > 15. Jul 28 05:36:00 mail postfix/policyd-weight[13047]: warning: Got > > SIGPIPE. Child 13047 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[13043]: warning: Got SIGPIPE. > > Child 13043 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[13043]: warning: child: err: > > Died at /usr/local/sbin/policyd-weight-devel line 1264. > > Jul 28 05:36:00 mail postfix/smtpd[318]: warning: premature end-of-input > > on 127.0.0.1:12525 while reading input attribute name > > Jul 28 05:36:00 mail postfix/policyd-weight[13043]: warning: Got SIGPIPE. > > Child 13043 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[13035]: warning: Got SIGPIPE. > > Child 13035 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[13035]: warning: child: err: > > Died at /usr/local/sbin/policyd-weight-devel line 1264. > > Jul 28 05:36:00 mail postfix/smtpd[333]: warning: premature end-of-input > > on 127.0.0.1:12525 while reading input attribute name > > Jul 28 05:36:00 mail postfix/policyd-weight[13035]: warning: Got SIGPIPE. > > Child 13035 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[10656]: warning: Got SIGPIPE. > > Child 10656 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[10656]: warning: child: err: > > Died at /usr/local/sbin/policyd-weight-devel line 1264, <GEN85426> line > > 15. Jul 28 05:36:00 mail postfix/policyd-weight[10656]: warning: Got > > SIGPIPE. Child 10656 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[10545]: warning: Got SIGPIPE. > > Child 10545 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[10545]: warning: child: err: > > Died at /usr/local/sbin/policyd-weight-devel line 1264. > > Jul 28 05:36:00 mail postfix/policyd-weight[10545]: warning: Got SIGPIPE. > > Child 10545 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[10470]: warning: Got SIGPIPE. > > Child 10470 terminated. > > Jul 28 05:36:00 mail postfix/policyd-weight[10470]: warning: child: err: > > Died at /usr/local/sbin/policyd-weight-devel line 1264. > > Jul 28 05:36:00 mail postfix/policyd-weight[10470]: warning: Got SIGPIPE. > > Child 10470 terminated. > > > > I hope this will help to find the cause of the problem. > > Unfortunately not. It doesn't tell why the children received > SIGPIPE. It doesn't appear to be reproduceable either.
Ok, I searched the logfiles for policyd warnings. I did on 27th and 28th, the day of the crash. Please find attached the output as gz file. But I do not find any master messages in the file. But there are a view messages concerning syslog and the crash happend during log rotation on the productive server. Testing on my test server killing a child with SIGPIPE does not kill the master. Everything goes on fine. Strange problem I think. I will investigate this further on. Kind regards Thomas -- In fiction the recourse of the powerless is murder; in life the recourse of the powerless is petty theft.
policyd-warn.log.gz
Description: GNU Zip compressed data