socket call resulted in error (98): no message

[McAllister, Andrew]

I had similar problems with 2.4.3 on CentOS 5.5 when I started shibd without the init scripts. Once started, shibd would not let go of its socket even with -f or manual kills of the processes. I tried "shibd stop", "shibd -f stop", "shibd -f start", and finally tried "kill" and eventually "kill -9". The kills terminated the process but the socket remained.

I didn't have time to chase it farther and since this was a throwaway test system, I resorted to the easiest thing that WOULD clean out the socket, a reboot. SELinux was set to permissive.

Once I started using the init.d scripts or the "service shibd start" command, the problem went away.

Andy

> -----Original Message-----
> From: users-bounces at shibboleth.net [mailto:users-
> bounces at shibboleth.net] On Behalf Of Cantor, Scott
> Sent: Monday, February 20, 2012 12:57 PM
> To: Shib Users
> Subject: RE: socket call resulted in error (98): no message
> 
> > What port should I be looking at - I believe I can use netstat to see if it is
> > active.
> 
> It's a file, it's a Unix domain socket.
> 
> See the Listener documentation in the wiki. Should be a relative file name
> that eventually ends up in /var/run/shibboleth. In a newer configuration it's
> entirely defaulted.
> 
> I checked the header files on Red Hat, 98 is "address in use". I can't think of
> any reason -f wouldn't clear the socket, so I don't know what its problem is.
> Other than clearing /var/run/shibboleth, I'd probably try an strace around
> running shibd to see what it's trying to create.
> 
> I'd still check SELinux and make sure it's off.
> 
> -- Scott
> 
> --
> To unsubscribe from this list send an email to users-
> unsubscribe at shibboleth.net