On Aug 9, 2010, at 11:30 AM, David Barrett wrote: > I'm reading up on the new write-ahead logging feature, and I'm unclear > on one point: does WAL only help alleviate multi-threaded locking (by > allowing other threads to continue reading while one is writing), or > does WAL also help between multiple processes? > > My understanding could be completely wrong, but I think the following > statements are true -- can you help fact check me? (I've read > http://www.sqlite.org/lockingv3.html but it doesn't seem to have been > updated for WAL -- assuming any updates are even needed.) > > 1) Two processes can read from the same database at the same time. > > 2) However, one process cannot read from the database while another is > writing -- WAL is irrelevant here. > > 3) One process can have two threads that read from the database from > the > same time. > > 4) And thanks to WAL, now one process can have one thread writing and > any number of threads reading at the same time. > > 5) However, WAL does not allow multiple threads to write at the same > time. > > Is this right? Or am I totally off? Are there any good pages > outlining > the intersection between multi-process database locking and WAL? > Thanks!
Unless shared-cache mode is turned on, multiple threads each using their own sqlite3* connection should behave in the same way as multiple processes do wrt to sqlite locking. If you do enable shared-cache mode, the rules here apply: http://www.sqlite.org/sharedcache.html _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
