A process runs on a split database environment - connecting locally to
the database db1 and remotely to the database db2 (via database
links). It deletes a bunch of rows in a table A in database db1 for a
specific account and updates another table B in the database db2 for
the same account. The access to table B is via a view (with a select
clause to read all the rows of the table B via the database link).

When we have 8 processes running at the same time, all of them die with
the error "ORA-02049: timeout: distributed transaction waiting for
lock" one after another after processing for some time.

I know that An ORA 2049 is signalled if

o you are blocked
o you are in a distributed transaction (you are using a DB link, even
just to
read)
o you wait for longer than distributed lock timeout

In this case, the processes update distinct rows on the table B in the
remote database db2. Also, even if only one process is run, it dies
eventually with the above error after running for a long time. It seems
to die immediately after the update on the table B in the remote
database.

I read that this could be due to ITL shortage deadlock. But, why would
the error come even when only one process is running ?

I found that the PCTFREE was 20 and INITTRANS was 1 on the table B in
the remote database db2. Will increasing these values have any impact ?

Any help would be really appreciated.