newbie, afs file path question

[bil]

On Friday, September 7, 2001, at 05:53 PM, Henry B. Hotz wrote:

> At 12:03 PM -0700 9/7/01, Troy Goodson wrote:
>> I use two different machines, one Mac OS X and the other HP-UX. Both 
>> have some form of AFS installed.  On the HP-UX machine, the following 
>> path is valid and gives me access to my file:
>>
>> /afs/jpl/user/t/tgoodson/www/index.html
>>
>> On my Mac OS X machine, with Arla, the above path does not work, and I 
>> have to use this path:
>>
>> /afs/jpl.nasa.gov/user/t/tgoodson/www/index.html
>>
>> What's going on?
>
> Probably arlad is being started with the -D option, which is needed for 
> the GUI configuration tool to work completely.  This means that the 
> contents of /afs/* are determined by that tool and there is no 
> provision for aliases to the "standard" of using the DNS name of the 
> cell.  Unless the developers have added some alias capability recently 
> I don't have a really good recommendation for you.
>
> On your HP you are probably seeing a lot of things like /afs/jpl2 and 
> such which are defined by our AFS administrators for convenience.
>
> If you take the -D option off the arlad command in 
> /usr/arla/sbin/startarla (I think) then you will see the same stuff in 
> /afs/* as on the HP.  The big downside is that the finder wants to 
> cache a lot of info about all the links there and it will take longer 
> for the /afs volume to mount on your desktop than it does to start 
> classic emulation.  This is what I do, but it's frustrating.

FWIW, if that's the case, you might try trimming out the entries in the 
CellServDB file for cells you don't use--I'm still running an older copy 
of arla on one machine, with only 2 cells listed, and it mounts those 
cells in a reasonable time, and the alias we have set for our cell works 
ok. (cs.unc.edu is aliased to unc in our system).

> Give me a call if you need some more explanation.
> -- __________________________________________________________
> The opinions expressed in this message are mine,
> not those of Caltech, JPL, NASA, or the US Government.
> Henry.B.Hotz at jpl.nasa.gov, or hbhotz at oxy.edu