On 21 Feb 2014, at 12:45am, Richard Hipp <d...@sqlite.org> wrote: > Long-running or overlapping readers can prevent the checkpoint from > occurring. The checkpoint will be retried again and again, but if there is > always a read transaction open on a transaction other than the most recent > transaction, the the checkpoint will never have an opportunity to run to > completion and reset the WAL file.
Suppose this happens and another process tried to write. Will the write be held up until the checkpoint can complete, or will the expected WAL size be exceeded ? Reading blocks checkpointing blocks writing ? Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
