Hi, We have performance issues with SQLite running on our file system (remote FS using Atomic Write Sector).
In fact we are limited with the number of write by second allowed by our file system, let’s say we have 10 write/s After a series of benchmarks the log file (journal) seems to be the bottleneck, SQLite writes to the log file many times for a simple transaction (insert), and it consumes most of the write/s allowed. One solution consists to use the journal only in memory (PRAGMA journal_mode = MEMORY). We can continue to use the rollback systems in case of nominal mode, but on application crash or power failure the SQLite recovery can’t be used, so we plan to perform our own recovery system based on SQL queries. And the question is : On Application crash or power failure with an Atomic write sector FS does anybody knows if the SQLite database file system will be corrupted and will need some specific SQLite recovery ? Kind Regards. -- View this message in context: http://www.nabble.com/Corrupted-database-with-Atomic-write-sector---tp23757941p23757941.html Sent from the SQLite mailing list archive at Nabble.com. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
