本文发表在 rolia.net 枫下论坛Qmail has 2 ways to log information:
1. use splogger that comes with qmail. It reads messages from stdin and feed them to syslog. You should be able to find out whether splogger is in use by running 'ps ax | grep splogger'. If it is in use, then the log for mail should be in /var/log, in the name of maillog, mail.log, mail.info, mail.err, etc.
2. use mutilog which is part of daemontools (a seperate software package also written by djb). If that's the case, qmail must be running under supervise (part of daemontools), so is multilog. You should be able to find out whether those programs are in use by running 'ps ax | grep supervise' and 'ps ax | grep multilog'.
If daemontools are in use, there must be a service directory, most likely in /service, /var/service, etc. Under that, there must be smtpd, smtpd/log, qmail, qmail/log, pop3d, pop3d/log. In each of those directories, there must be a script called run which contains commands to run those services. The run script under log directory must contain the command multilog, and the last parameter of multilog is the directory it logs information to, which is a directory called main under log directory by default. You should be able to find log files in those directories. One more thing is that the timestamp of entries in those log files are in tai64 format, you will need to run 'tai64nlocal < <logfile>' to get the human readable timestamp.
Once you can find the log file, you shoud be able to know the problem of your mail server.
Here are some pointers you might want to look into:
- Life with Qmail (http://www.lifewithqmail.org) -- highly recommended
- Qmail site (http://www.qmail.org/top.html)
- DJB's site:
-- for qmail (http://cr.yp.to/qmail.html)
-- for daemontools (http://cr.yp.to/daemontools.html)
-- for ucspi-tcp (http://cr.yp.to/ucspi-tcp.html)
Good luck!更多精彩文章及讨论，请光临枫下论坛 rolia.net
1. use splogger that comes with qmail. It reads messages from stdin and feed them to syslog. You should be able to find out whether splogger is in use by running 'ps ax | grep splogger'. If it is in use, then the log for mail should be in /var/log, in the name of maillog, mail.log, mail.info, mail.err, etc.
2. use mutilog which is part of daemontools (a seperate software package also written by djb). If that's the case, qmail must be running under supervise (part of daemontools), so is multilog. You should be able to find out whether those programs are in use by running 'ps ax | grep supervise' and 'ps ax | grep multilog'.
If daemontools are in use, there must be a service directory, most likely in /service, /var/service, etc. Under that, there must be smtpd, smtpd/log, qmail, qmail/log, pop3d, pop3d/log. In each of those directories, there must be a script called run which contains commands to run those services. The run script under log directory must contain the command multilog, and the last parameter of multilog is the directory it logs information to, which is a directory called main under log directory by default. You should be able to find log files in those directories. One more thing is that the timestamp of entries in those log files are in tai64 format, you will need to run 'tai64nlocal < <logfile>' to get the human readable timestamp.
Once you can find the log file, you shoud be able to know the problem of your mail server.
Here are some pointers you might want to look into:
- Life with Qmail (http://www.lifewithqmail.org) -- highly recommended
- Qmail site (http://www.qmail.org/top.html)
- DJB's site:
-- for qmail (http://cr.yp.to/qmail.html)
-- for daemontools (http://cr.yp.to/daemontools.html)
-- for ucspi-tcp (http://cr.yp.to/ucspi-tcp.html)
Good luck!更多精彩文章及讨论，请光临枫下论坛 rolia.net