Subject Re[4]: [ib-support] Unsuccessful metatdata update: Index is in use
Author Dieter Tremel
Hello Claudio.,

am Montag, 25. März 2002 um 10:00 schrieben Sie:

CVC> Hence, the reloading
CVC> process picks the second trigger (the trigger that implements the
CVC> update-cascade whose associated FK we don't drop yet) and compiles it. Since
CVC> that trigger uses detail.b, the underlying index in detail.b is taken as a
CVC> needed resource.

CVC> When you try to drop the second constraint, the engine again has to modify
CVC> both relations because there's a cascade operation defined. When it goes to
CVC> modify detail to drop the index in b, that index can't be dropped because
CVC> the trigger in master has it taken (the trigger has not been unloaded).

Thank You for your detailed explanation, helped a lot, I think I
caught the point.
Couldn't the server 'unload' the index after the compilation of the
master index? After the first DDL is finished the trigger 1 has gone
and trigger 2 is recompiled and active the state could be the same
freshly opened without ddl.

Dieter Tremel