Another follow-up in the hope someone can shed light on the situation or ask for further specific testing.
I am told the only thing we change from the default settings is that synchronous=0. Apparently full mode was far too slow. Reading the documentation it clearly states that there are limited circumstances in which corruption can occur but that they involve O/S crash or power loss between the time the application writes data and the point that the hardware actually commits it to physical media. An engineer that went in to "rescue" one of our sites got out app up and running on a fresh db, then managed to get corruption again inside a few minutes - he confirms there was absolutely no crash of any kind. This is most worrying... Filip Navara from this list kindly asked to inspect a sample database and has told me that the journal appears partially written. Our support department reports that the frequency of sites becoming corrupt is in the region of 1-2 per week and it's been happening for several weeks (since we picked up on the pattern). We're looking at roughly 30 instances so far. Thoughts and recommendations welcome. James _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
