RE: CA broadcast problems on Solaris 2.6?

[wright@odin.atdiv.lanl.gov (Rozelle Wright)]

edd and cau which were built on Solaris 2.5.1 with Epics libraries built on
2.5.1 work fine from the Solaris 2.6 machine.  I will try rebuilding the
Channel access libraries on that machine and see what happens.
Rozelle
> From johill@lanl.gov Wed Mar 18 10:54 MST 1998
> To: "'Dave Barker'" <barker@acrux.ligo.caltech.edu>,
>         "tech-talk@aps.anl.gov"
> 	 <tech-talk@aps.anl.gov>
> Subject: RE: CA broadcast problems on Solaris 2.6?
> Date: Wed, 18 Mar 1998 10:35:11 -0700
> Organization: LANL
> X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211
> Mime-Version: 1.0
> Content-Transfer-Encoding: 7bit
> X-Lines: 66
> 
> 
> Dave,
> 
> On Tuesday, March 17, 1998 6:03 PM, Dave Barker [SMTP:barker@acrux.ligo.caltech.edu] wrote:
> > I'm getting the following problems when running MEDM on my
> > solaris 2.6 machines:
> > 
> > CA.Client.Diagnostic..............................................
> >     Message: "No internet interface with broadcast available"
> >     Severity: "Warning" Context: "caAdd : ca_build_and_connect failed
> > "
> >     Source File: ../medmCA.c Line Number: 680
> > ..................................................................
> > 
> > but when I ifconfig -a, I get:
> > 
> > lo0: flags=849<UP,LOOPBACK,RUNNING,MULTICAST> mtu 8232
> >         inet 127.0.0.1 netmask ff000000
> > hme0: flags=863<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST> mtu 1500
> >         inet 206.61.208.109 netmask ffffff00 broadcast 206.61.208.255
> > 
> 
> Quick workaround:
> Set EPICS_CA_AUTO_ADDR_LIST to NO and specify an explicit broadcast
> address in EPICS_CA_ADDR_LIST (see the CA reference manual).
> 
> Long term:
> The code involved is in base/src/ca/if_depn.c
> in routine caDiscoverInterfaces(). One possibility is that the code has to
> be recompiled on Solaris 2.6 in order for the interface flags or ioctl codes
> to be defined correctly. You might look in the release notes for Solaris 2.6. 
> You could also build EPICS base with debug (ie HOST_OPT = NO) and then
> step through this routine in the debugger to get a closer look at the 
> situation.
> 
> I just spoke with Rozelle here, and it turns out that we recently installed a 
> Solaris 2.6 machine here. Rozelle said that she will try to reproduce 
> the problem here.
> 
> Jeff
> 
> 
> 
> > the broadcast address is set correctly for a Class C network. I have
> > seen this problem
> > on all types of Solaris 2.6 machines (E2 and Ultra30).
> > 
> > Can anyone help with this?
> > 
> > thanks,
> > 
> > Dave Barker.
> > 
> > --
> > 
> > ---------------------------------------------------------------
> > Dave Barker                | Tel    Office: 509 373 6203 (msg service)
> > LIGO Hanford Observatory   |        Lab 0:  509 373 3580
> > P.O. Box 1970; MS S9-02    |        Lab 1:  509 372 3976
> > Richland                   | Fax    509 372 2178
> > WA 99352                   | e-mail barker@ligo.caltech.edu
> > United States of America   | S/W Engineer, UNIX Sys Admin.
> > ---------------------------------------------------------------
> > 
> > 
> > 
>