Hi all,
Actually there are 2 optimizations:
- Sequoia JDBC driver (from client app to Sequoia): transparently caches
connections so if you close/re-open fast enough, you will get the same
connection (this can be disabled by an option in the JDBC URL)
- Database JDBC driver: Sequoia uses a connection pool per backend, only
the SimpleConnectionManager open/close connection every time, all the
others keep open connections in the pool.
Hope this helps,
Emmanuel
Stefan Lischke wrote:
Hi alex,
AFAIK Sequoia is using a connection pool, so there are always a special
number of connections open to the DB and if you connect via the sequoia
jdbc driver one of these connections is reused.
Another cause for speed maybe the cache that sequoia uses.
hth
stefan
Alex wrote:
I take it that Sequoia controllers usually keep a connection to the
backend database open? (Anyone know of docs on this issue - I've not
found a good reference on this yet.)
I just ran some JDBC stress tests (using The Grinder) - first, against
a naked MySQL database, and then against a Sequoia virtual DB with a
single MySQL database behind it containing the same data. The Sequoia
test ran much faster - presumably because Sequoia is holding the
connection to the actual database open... (My test closes the
connections and reopens for each query.)
Alex
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
