Hi, I'm seeing hot journal frequently in our SQLite installation. We use VFS which does some checksum and other operations. There are no abnormal scenarios such as power off, process crash, abandoned transactions... DB is opened by 10-20 processes each of which can have multiple threads.
Old version of SQLite: 3.7.13 Process A is in write mode. Process B which is reading a value. Process B used to sees hot-journal. Now, we upgraded to 3.9.2 Process A is in write mode. Process B is trying to read a value. B sees a hot-journal and recovers it. Process A will get error introduced in new new package: "The xDelete method in the built-in VFSes now return SQLITE_IOERR_DELETE_NOENT if the file to be deleted does not exist" Few things stand out: Hot journal was reported in 3.7.13. There are no logs to indicate it's recovery. Hot journal is seen in 3.9.2. There are logs to indicate it's rollback. Committer Process(Process A) gets error. I know some optimizations in HardDisk or OS is creating the problem. What other layers could contribute to this anomaly? How could installation endup with Hot journals without system experiencing crash or power loss. -- Sent from: http://sqlite.1065341.n5.nabble.com/ _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
