SUMMARY: RPC: Timed out with Calendar Manager

From: Don Baune 581-6088 MIRL (
Date: Thu Aug 15 1991 - 18:31:04 CDT


Thanks to all who took the time to answer this one. The list came through
with an answer that I can live with AGAIN.

It appears that sun didn't do a very clean job on this one and the way to
get it to work is a kluge at best, but after 3 days of testing I haven't
had one complaint of lost data or any more timeouts so it appears to work.
The answer came from, thanks.

It turns out that if I create a shell script named cm_remote (sorry about
the name change Anthony I prefer to keep the utilities as close to what sun
put out and then call them with something diffrent if need be)


#! /bin/sh
/usr/openwin/bin/xhost + doug > /dev/null
rsh doug \
        "setenv LD_LIBRARY_PATH /usr/openwin/lib:/usr/lib;\
        /usr/openwin/bin/xview/cm $* -display `hostname`:0"

doug being our main server machine. This allows someone to login from any
machine and then run the cm on the main machine. This way all are using
the same deamons and cache to keep things current.

Thanks to all who replyed as each answer gave insites as to why things
were acting strange. Sometimes just a explaination of the real problem
can lead to the solution. (Meg Grice)
Bryan McDonald <> Milt Ratcliff
Mike Raffety <> (Anthony Worrall) (MaryBeth Gallacchi)

Selected responses below better explain the problem and the work around.
Your not doing anything wrong its Sun's problem. I wanted to do
exactly what you want to do. I got it to work under 4.1 , but
could not get it to work under 4.1.1. I called Sun they said
it will not work with one directory from the server mounted
on all the workstations -- that /var/spool/calendar had to
be local on each machine. This is needless to say the pits.
Fixing calendar so that it can be an nfs mount point is supposedly
on of the things that Sun is working on.

The syntax error is because the file must contain certain
information if it is to exist, if not then the server complains
as you see below.

Be very careful with global mounting all the callog's from a central
location. If more the one person tries to use any one
calendar, or the same person uses it from another machine,
then a race condition occurs and you loose some data.

You probably don't want to NFS mount /var/spool/calendar ... that's how
you lose appointments (because of race conditions between multiple processes
trying to access the same file).

Instead, just have the user specify "<user>@<server>", like "miker@trinity",
so they access the rpc.cmsd on the server, instead of on individual

It seems that some people in SUN have not seen the company's trade mark
"The Network is the Computer" and still think that a user only uses one
machine. Perhaps this due to them having a machine on their desk.

Any way my soloution was to move cm to cm_local and replace cm with
a script which rsh cm_local on our server. This is not ideal because the
user must be able to rsh tghe server and the server must be added to xhosts.
This seems to work but we are not using cm to a large degree.

#! /bin/sh
/usr/openwin/bin/xhost + server > /dev/null
rsh server \
        "setenv LD_LIBRARY_PATH /usr/openwin/lib:/usr/lib;\
        /usr/openwin/bin/xview/cm_local $* -display `hostname`:0"

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:20 CDT