Critical uptime question (Was "Looking for some upgrade advice")
John Esak
john at valar.com
Tue May 30 00:34:08 PDT 2006
>
> >
> John,
>
> I realize that the amount of work necessary is related to how large the
> database is. Given what you're telling me it does seem that while it
> may be possible, the amount of work necessary would almost be like
> rewriting every operation from scratch.
>
> Boaz
Yeah, had this database... let's call it a huge disjointed data set... not
grown like topsy over the past 20 years, maybe things would be different.
Certainly, if I had the ability to start from scratch now... I would write
absoltuely everything with transational considerations in mind. But, still,
there is that "change-voer", "fail-over" point which is still impossible to
do "atomically". Meaning, filePro would still not be the application to
expect perfect results in this situation. We are very nearly "tied" to
strange hardware... soon, perhaps, we will be *directly* tied to some more
hardware in the way of electronic scales. A button is pushed, the weight is
thrown into a filePro field, and a bar code is rinted. Can't be done on
indivicual machines, has to be handled by a "sever" whihc is intimately
twined with the database at large. Imagine 40 machines or so, each putting
out a new "item" every couple minutes or so... randomly staggered amongst
the 40 machines... and the 40 machines are at differnet locations. Items are
coming off the production lines almost every few seconds... very little time
to bring down a server, or transfer all these filePro users to another
server.
it's funny, I have built in a nd very cool method for updating the software
being used on the fly...no problems... it is just transferring the
users/processes to another system that is difficult to impossible.
John
More information about the Filepro-list
mailing list