[OpenAFS] Changes for Mosaic's AFS cell...

Rodney M Dyer rmdyer@uncc.edu
Thu, 06 Apr 2006 21:10:11 -0400

Ok great response.  I believe I now have a "clue" on where OpenAFS is 
relating to all my questions.

1.  Given the responses I believe we will continue to wait patiently for 
1.4.1.  I understand that no version will ever be completely bug free, but 
we really would like the server side stuff to be as solid as 
possible.  Using "betas" and "release candidates" on the clients is a less 
problematic issue.

I think one issue here is that the OpenAFS.org client "product" doesn't 
work like other software "products" in the industry.  Usually a software 
package is released, then if it has bugs, the customer receives "patches" 
that solve a problem.  A patch could even be an entirely new client, 
however it is still called a patch with some version number (Transarc used 
to do this).  In OpenAFS.org's case, all we get is a "production release", 
then everything after that is just "...on to the next version with unstable 
betas and rc's".  I understand this has more to do with the nature of open 
source in that the "customer" can download a "source patch" and recompile 
to fix a particular problem, however it is still kind of irritating that no 
official "patches" are issued.  Eg, there's no place on the web site under 
"downloads" where the patches to fix particular bugs are listed.

2.  Regarding the xlock/xscreensaver/KRBCCNAME issue, we are running RHEL3, 
so yes, we are using older stuff.  We are using the K5 PAM that came from 
RedHat.  In light of what was stated, a simple upgrade to RHEL4 should fix 
this issue.  I believe we tried RHEL4 just after it became available and 
had some issues with the OpenAFS compile (or applications) so it was 
decided to wait for a while until we could examine the problems further.