In the case of nested transactions, if the transaction is a parent
transaction with unresolved (neither committed or aborted) child
transactions, the child transactions are aborted and the commit of the
parent will succeed.
In the case of nested transactions, if the transaction is a child
transaction, its locks are not released, but are acquired by its parent.
While the commit of the child transaction will succeed, the actual
resolution of the child transaction is postponed until the parent
transaction is committed or aborted, i.e., if its parent transaction
commits, it will be committed, and if its parent transaction aborts, it
will be aborted.
If the -nosync option is not specified, a commit log record is
written and flushed to disk, as are all previously written log records.
The options are as follows:
-nosync
Do not synchronously flush the log. This means the transaction will
exhibit the ACI (atomicity, consistency and isolation) properties, but
not D (durability), i.e., database integrity will be maintained but it is
possible that this transaction may be undone during recovery instead of
being redone.
This behavior may be set for an entire Berkeley DB environment as part of
the berkdb env interface.
-sync
Synchronously flush the log. This means the transaction will exhibit all of
the ACID (atomicity, consistency and isolation and durability) properties.
This behavior is the default for Berkeley DB environments unless the
-nosync option was specified to the berkdb env or
envtxn interfaces.
Once txncommit has been called, regardless of its return, the
txn handle may not be accessed again. If txncommit
encounters an error, then this transaction and all child transactions
of this transaction are aborted.
The txncommit command returns 0 on success, and in the case of error, a Tcl error
is thrown.