Doug Nebeker <[EMAIL PROTECTED]> wrote: > I'm currently seeing something that has never happened--about 90 master > journal files (the <filename>-mjXXXXX type) that are popping up in my > database directory. I'm using 3.2.7 or 3.2.8 (have to double check) on > Windows. > > The scenario: > I have two threads that are reading and writing to the database pretty > heavily at the moment--this has worked fine for quite some time. The > 'database' consists of 9 tables, each in it's own database file. I open > a connection to one database, and then run ATTACH commands to pull the > other 8 in. The master journal file is named after one of the databases > (the one that opened the connection?). > > Each thread does all reading and writing within a transaction. > Performance is great, data looks good, everything seems to be working > well except for the extra files. > > All of the master journal files were created in the first 3 minutes of > the run--now 15 minutes later no new ones are getting created (at least > they don't hang around long enough for me to notice), but the original > 90 are still there. > > [Waiting....] > All of the read and writer threads have now finished. The database > connection was successfully closed and the application has closed down > nicely. > > I read at http://www.sqlite.org/lockingv3.html about deleting stale > master journals. Does SQLite do that automatically, or do I need to do > that myself? > > Did I have some sort of major failure that created these in the first > place? What can I do to avoid it in the future? >
The master journal files are suppose to be deleted automatically when the transaction commits. But perhaps there is a bug. I will look into it. -- D. Richard Hipp <[EMAIL PROTECTED]>