Coral Servers Problem

Bill Murray bmurray at snf.stanford.edu
Tue Dec 3 08:29:04 PST 2002


John,

I should finish up all the loose ends this week which leaves two weeks to
get rosen updated.  I should be able to handle that by myself in two weeks.
Would you rather that we try to get atu ready instead?  The upside of that 
would be that we could include all the runtime additions and avoid another
migration later.  Perhaps, the most prudent approach would be to update
rosen, then, if time permits, move to atu.  We don't lose much time that
way.

The two unknowns on atu are system/network configuration issues which prevent
remote clients from running and our inability to get FormulaOne running.

Bill

On Tue, 3 Dec 2002, John Shott wrote:

> Bill:
> 
> Well, the "auto-detect and restart" script did it's job.  We had a WRITE lock
> condition develop at 11:36 last night and the servers re-started themselves at
> 11:41 ... apparently before we received any complaints.
> 
> Nonetheless, it would be good to update rosen to include the new and improved
> server changes, the multiple reservation stuff, and the new maintenance client
> ... and, I think, we could probably migrate to BouncyCastle ... everything but
> the runtime stuff.
> 
> Can we come up with a plan to be able to do that before the shutdown? That, of
> course, fits with other developments and constraints?
> 
> Thanks,
> 
> John
> 




More information about the coral mailing list