Posted this message on borland.public.interbase.general and it was suggested to try here so:

We had a situation on a customer site that created a limbo transaction. A
backup and restore on one of the databases fixed the problem.

Looking at our error logging - the following error occurred just before the
re-curring limbo transaction exception:

Exception 'EIB_ISCError' in module IBO40CRT_D5.bpl at 00077A09
ISC ERROR CODE:335544349

ISC ERROR MESSAGE:
attempt to store duplicate value (visible to active transactions) in unique
index "RDB$INDEX_32"

Source file: IB_Session.pas, Line 1146
...(call stack information follows)...

The call stack revealed that the exception was raised as a two phase commit
transaction was being commited (ie. txnMyTransaction.commit;)

RDB$INDEX_32 is a unique index on RDB$TRANSACTIONS.RDB$TRANSACTION_ID. Both
of the databases have RDB$INDEX_32 as the same thing (system index on this
field).

Whenever this particular routine was called after this exception, the
following exception is raised:

Exception 'EIB_ISCError' in module IBO40CRT_D5.bpl at 00077A09
ISC ERROR CODE:335544459

ISC ERROR MESSAGE:
record from transaction 62067 is stuck in limbo

Source file: IB_Session.pas, Line 1146


Any thoughts as to why interbase would be attempting to store a duplicate
value in a system table?

Cheerio



Josh

IB 6.0.1.6
IBOBJECTS 4 something (latest I think)



[Non-text portions of this message have been removed]