[Zope] Zope won't start: "Protocol wrong type for socket"

Dieter Maurer dieter@handshake.de
Mon, 24 Sep 2001 21:22:30 +0200 (CEST)


David writes:
 > I use redhat 7.0 but will a number of packages from rawhide to bring me a bit 
 > more up to date.
 > 
 > Kernel 2.4.3
 > Zope 2.4.1
 > python 2.1.1
 > 
 > I have tried both zserver and pcgi using the packages from 
 > http://www.taupro.com/Downloads/Zope/
 > 
 > Zope fails to start with the following message
 > 
 > Sep 21 01:08:50 CO3011387-A z2[4335]: Startup exception : error: (91, 
 > 'Protocol wrong type for socket')
 > 
 > So I did a little investigation and found that the following command
 > 
 > /usr/bin/env python2.1 z2.py  
 > -u  zope  
 > -z      /usr/share/zope                 
 > -Z    /var/run/zwatchdog.pid  
 > -w      8080    
 > -f      8021
 > -m      ''              
 > -D                
 > -l      /var/log/zope    >>      /var/log/zope 2>&1     &
 > 
 > 
 > 
 > returns
 > ------
 > 2001-09-24T03:44:38 PANIC(300) z2 Startup exception
 > Traceback (innermost last):
 >   File z2.py, line 606, in ?
 >   File /usr/share/zope/ZServer/medusa/logger.py, line 163, in __init__
 >   File /usr/share/zope/ZServer/medusa/m_syslog.py, line 144, in __init__
 > error: (91, 'Protocol wrong type for socket')
There seems to be a "ZSYSLOG*" environment variable that triggers
"syslog" logging (rather than file logging as you call for with
the "-l" above).
The variable either has a wrong value or it triggers a bug
in Zope code.

I would locate the environment variable and eliminate it....


 > Yet when I include the -r flag for read only ...
 > everything okay
In this case, logging goes to stdout (though I do not know why!).
The wrongly configured syslog logging is not used....


Dieter