> (summary to date) fb 1.5.1, fedora core
>
> I'm trying to explain how/why this is happening. It has to be in firebird
> itself rather than the database as rebooting the machine will fix it and
> a backup/restore has no affect.
>
> I have a situation where, sometimes, a database that copied to another
> identical machine will only allow 1 connection, any additional attempt
> to connect just waits.
>
> exit the first connection and a waiting connection completes the connect.
> I've done this with isql so as to eliminate client/network issues.
>
> Now granted that this is a database that is copied rather than backup,
> copy backup file, restore but I've had it happen with a database where
> I've done backup, restore on same machine to new name, copy to new
> machine. Nobody has connected to the restored database on the original
> machine and it's still had this problem.
>
> I can do the exactly same copy operations and one copy will allow multiple
> connections and the other will only allow one.
>
> fb_lock_manager has permissions -r-sr-s--- and is running as firebird.
>
> All clients were the same group, the same id in fact (multiple shells
> from same login) and were attempting to attach as sysdba

fascinating....
so you are running classic everywhere? I had assumed super server
Do these databases, by any chance, have a lot of garbage?
After this first connection which is slow (and may be doing garnage
collection) finishes what it is doing, you can then connect OK on all
subsequent connection eh.
Try doing a backup (without the -g switch if you are currently using it in
your backups) of the source database before you copy it.
Is there any chance that you are doing a mass delete on the source before
copying? Or some other operation which would create large volumes of
garbage?
Alan