"Eric Thomas (Heidelberg Sc. Center)" <ERIC@DHDIBM1>
Fri, 24 Apr 1987 08:13 ESZ
|
I think you have missed the point. I find it unacceptable that maybe four
sites in the network are preventing 15 sites from placing a new software into
production. Period. If there is a way technically to avoid a crash in other
people's servers, I will do it, but if there is no way, tant pis.
It is not possible to send 'customized' PEERS NAMES to the various servers.
This would require me to prepare one PEERS NAMES per site and send it via
SENDFILE, ie around 70 files of 1,000 records sent from FRECP11 via SENDFILE
and more than 1 hour of work.
You say understaffed. FRECP11 has a staff of ONE person to maintain FOUR
machines on TWO different buildings with no remote login facility. And the
machine which causes the less troubles is the 4381, so guess which one gets
the smallest amount of 'staff time'? I'm not saying that you should moan on
our status, just that I am fully aware of the problems of understaffing.
If I send mail to a user and tell him "be careful, in 1 month we will have
to remove FORTVS release 456.781 from the system, be sure to have migrated to
FORTVS release 457.0 by then", and one month later he is still using release
456.781, then well, what can I do for him? The other 50 fortran users are
pressing me to install release 457.0 because of the new facilities, and for
some technical reason we can't keep the two, so we have to move forward.
Please note that the 1.4 and 1.5b servers are not affected by this since the
PUT command is unknown to them. The 1.5d to 1.5h owners can still disable this
PUT PEERS NAMES to avoid the problem, and I think that the right way is that
THEY should spend 2 minutes of their time instead of me spending 2 hours to
find a solution.
Enough blabbered for today.
Eric
|
|
|