20000131: testing mcidas7.604 (cont.)

>From: David Fitzgerald <address@hidden>
>Organization: Millersville University of Pennsylvania
>Keywords: 200001262144.OAA16099 McIDAS-X 7.6

Dave (hi Rich),

>The only thing I can add to what you found when you were checking things out
>here, is that I compiled mcidas on twister, a solaris 7 machine, which ALL
>of our clients are, however snowball is a solaris 2.6 machine, so thats
>probably why you got the missing library error when you were on snowball as
>mcidas.

Yes, that explains the message about libgen.so being missing.  I should
have remembered that from the same sort of situation here.

>Snowball IS the fileserver that holds the mcidas account but that
>shouldn't be the problem should it?

No, not at all.

>Why should snowball care if it is
>serving a program compiled for solaris 7 or not, since the clients are all
>solaris 7 and supposedly solaris 7 is binary compatable with solaris 2.6.

It shouldn't.

>Snowball's filesystem doesn't really care what the binary is since it is the
>client machine doing the "real" work right?

You are absolutely correct.

>Or am I wrong?

Nope.

>I've done things
>this way in the past, ie garp, and others, and have not had any trouble.

It should work fine.  In fact, our main file server is still running 5.6
and a number of clients are running 5.7.  BTW, I just saw that Sun is
supposed to be releasing Solaris 8 (aka 5.8) this spring.

>I've been holding off upgrading snowball since we won some grant money to
>replace snowball with an Ultra-10 server.

Good move.

>I'll reboot snowball and twister later tonight after people go home.

OK, I'll try and jump on later to see if the connectivity is better.

Later...

Tom