| Request | MaildropHost -- bug report -- by Johnny Souza |
| Posted on | Oct 18, 2007 11:45 am |
| Subscribe |
| Resolve by Jens Vagelpohl on Oct 18, 2007 2:08 pm | |
|
Thanks for the patch, it's been committed: http://svn.dataflake.org/?view=rev&revision=1434 |
|
|
|
| Initial Request by Johnny Souza on Oct 18, 2007 11:45 am | |
|
Hi, When the daemon (maildrop.py) receive a SIGTERM signal or occour an error the pidfile is not removed. It make a error in the next maildropctl start. If uses cron to start daemon again, the daemon is not started because the pidfile exists. ERROR MESSAGE: Error: maildrop daemon already started. SOLUTION: Register a function that remove the pidfile in atexit and as handler os SIGTERM. Diff bellow. Index: maildrop.py =================================================================== --- maildrop.py (revisao 1433) +++ maildrop.py (copia de trabalho) @@ -17,6 +17,8 @@ import time import rfc822 import imp +import atexit +import signal FATAL_ERROR_CODES = ('500', '501', '502', '503', '504', '550', '551', '553') MaildropError = 'Maildrop Error' @@ -214,6 +216,17 @@ pid_file.close() + +def exit_function(pidfile_path): + # Remove the daemon pid file + try: os.unlink(pidfile_path) + except: pass + + +def handle_sigterm(signum, frame): + sys.exit(0) + + if __name__ == "__main__": try: if len(sys.argv) < 2: @@ -340,6 +353,8 @@ print >>sys.stderr, "fork #2 failed: %d (%s)" % ( e.errno, e.strerror) sys.exit(1) + atexit.register(exit_function,MAILDROP_PID_FILE) + signal.signal(signal.SIGTERM,handle_sigterm) else: print '***** Starting in DEBUG mode *****' print '***** All log messages are shown on the console *****' |
