--- In ib-support@y..., Louis van Alphen <lja@b...> wrote:
Hi, Louis. I never used IB5x, jumped from IB4 to Firebird, so my
observations may be wrong for you. But I don't see differences between
IB4 and Firebird in this behavior.
> - Getting the 'Object in use' message. Then everybody has to log off
> have to do a backup & restore cycle to be able to use the
> index/table/SP/whatever again
Yes with Drop Table, Drop Index, Drop Constraint. If object is
cashed by one of connections, even your own. No need b/r, just
disconnect all including youself. Not pleasant but not deadly.
> - Triggers not firing. Goodness knows why not.
Never encountered. Maybe part of recompiling problem below - fires
but as previous version.
> - SP not running correctly. We have had the following problem that
> boggles the mind. We have also repeated this to make sure we are not
> stupid mistakes. We compile a SP sucessfully. If we run it, the old
> still runs.
It is described in Data Definition Guide "Altering and dropping
procedures in use". It is the only safe way to allow change SPs with
another connections established and be shure all use the same version.
Once more - no need b/r, just all disconnect. Some people reported
that if all chain of SPs/triggers calling each other will be
recompiled from bottom to called by client, clients will work with new
version without disconnect/connect, but I did'nt tried this.
> - Triggers same as SPs.
> - Abnormal server terminations
Some bugs are in any software. Check the buglist having in mind your
programming technology. But more often it is hardware/OS problems.
Pay attention to Doug Chamberlin's recommendation. I'm not so
severe, all 3 of us change metadata but:
1. At least 80% of data structures was burn from the primary script,
constructed as a result of looong "paper" period.
2. Each of us is responsible for own limited part and changes are
discussed in the sense "don't break neighbour's functionality".
Changes on developer's database are made in single-user mode after we
reach agreement. And on production one - in maintenance period.
I think any RDBMS can't stand against chaotic development.