> > I have a process with several threads working in the same time on a
> > database file. I have 0 or 1 writers and 0 or N readers at a moment. All > > write accesses are isolated in transactions, I am using WAL and shared > > cache, but sometimes during the DELETE statemens I get "database table is > > locked" errors, either in the write or read thread. > Have you set a timeout value for all your connections ? If you haven't, SQLite never does any waiting for locks at all, it just returns an error. You were right the busy timeout was set only for one of the connections. But, anyway the table lock continued to be triggered immediately not waiting to pass 2500 ms. I thought that WAL should've solved this aspect about the concurrency, but it seems that for DELETE statements it still needs to lock the table somehow. Regards Sandu _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users