A/UX concerns

[Ken Lindahl 642-0866]

In article <49580 at apple.Apple.COM> ksand at Apple.COM (Kent Sandvik) writes:
>In article <1991Feb22.175718.6395 at agate.berkeley.edu> lindahl at violet.berkeley.edu (Ken Lindahl   642-0866) writes:
>>In article <1991Feb21.202509.11608 at ni.umd.edu> steveg at ni.umd.edu (Steve Green) writes:
>>    I work with several other programmers on an X windows application that
>>    currently runs on Macintoshes w/ A/UX, DECstations, VAXstations, Sun
>>    3/xx's, Sun SPARCstations, RISC/6000s, and PS/2s w/ AIX. ...

Oops, that's a little confusing. I (Ken Lindahl) wrote that paragraph, not
Steve Green. I take full responsibility and/or blame, if you can catch me.

>>                                                         ... The only way
>>    to maintain this application is to use the same source files for all
>>    of these different platforms, and to NFS-mount the source directory
>>    wherever it is needed. Currently, the source resides on a DECstation
>>    hard disk. Below the source directory is a subdirectory for each
>>    platform; compiler output (i.e. `.o' files) and the linked application
>>    go into the appropriate subdirectory. This strategy works admirably
>>    for every machine but (you guessed it) the Macintosh. In order to get
>>    the Macintosh version to work, I must compile and link on a non-NFS
>>    disk. (Actually, I've never tried compiling on NFS disks and then
>>    linking on local disk, too much work.)
>>
>>My conclusion is that A/UX 2.0's NFS is in some way incompatible with NFS on
>>the DECstation, but I don't know enough about NFS to be more precise. I've
>>had several people offer impromptu explanations/rationalizations, but
>>nothing I felt was truly definitive.
>
>I would like to personally know if this only happens with the object files
>residing on the DECStation. Is it OK with the Sparcstation used as 
>the source server? Nota Bene the NFS code is pure Sun NFS code, so the
>implementation should be the same as the Sun specs.

An interesting question -- we've never tried using the SPARCstation as the
source server because there isn't enough disk space. However, I can dummy
up a test and let you know later. Followups by mail, unless enough people
tell me they want to see it here.

>Most of the problems I've seen with NFS are Yellow Pages configuration
>problems concerning access. If in this case you have access to the binaries/
>object code, and something does not work, is there maybe some dependencies
>of files that have to be accessed from other filesystems? As you see
>it's quite hard to define the problem without having a little bit more
>evidence and configuration information.

For what it's worth, we're not using Yellow Pages, excuse me, NIS on any of
the machines in question. I'm can't quite understand what you're suggesting.
Since we're talking about compiling and linking here, access problems should
be pretty obvious, no? If you want more specifics about the configuration,
I can certainly provide them. I never intended my posting to be problem
report -- I was merely offering some corroboration to Steve Green's posting
describing problems with NFS.

>Well, back to work.

You mean this isn't work? Damn, I thought it was.

>Kent
>
>-- 
>Kent Sandvik, Apple Computer Inc, Developer Technical Support
>NET:ksand at apple.com, AppleLink: KSAND  DISCLAIMER: Private mumbo-jumbo
>Zippy++ says: "C++ was given to mankind, so that we might learn patience"


Ken Lindahl				lindahl at violet.berkeley.edu
Advanced Technology Planning,
Information Systems and Technology
University of California at Berkeley