Is there any difference between the terms checkpoint and savepoint?

Thanks,
Tanu

tmalik (AT) cs (DOT) jhu.edu wrote:

Yes. They're spelled differently, for starters.

Savepoints usually apply to an individual transaction; it marks a
point to which the transaction can be rolled back, so subsequent
changes can be undone if necessary.

Checkpoints usually apply to whole database systems - the terminology
gets confusing between manufacturers. Anyway, it marks a point at
which the DBMS knows that what is on disk is consistent with what's in
memory. Of course, fuzzy checkpoints make things more complex, but
the idea remains valid. It gives a point in time at which recovery
can start if the DBMS crashes - a point where the system knows where
all the data is (which may be in a variety of places on disk other
than just the obvious place - think of logical logs, physical logs,
rollback segments and what have you).

--
Jonathan Leffler #include <disclaimer.h>
Email: jleffler (AT) earthlink (DOT) net, jleffler (AT) us (DOT) ibm.com
Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/

Yes. There is a difference.

Checkpoints are a feature where buffers are flushed and updates are
guaranteed to be written to disk, and for some implementations, logged.
Checkpointing is usually intricately involved with recovery features of a
DBMS.

A savepoint is a nested transaction concept which allows developers to have
control on how much of a transaction is rolled back on error or failure. It
gives more flexibility than an all-or-nothing transaction.

Dan

<tmalik (AT) cs (DOT) jhu.edu> wrote