We've been struggling with this issue for well over two years, so I am, once again, hoping that someone has some solid information and/or workarounds regarding the NVFS issues apparent on the E2 handheld.
What we are seeing are seemingly random "dmErrCorruptDatabase" errors when attempting to open and/or close database(s) on the Tungsten E2. No problems on the Tungsten E or the newer Tungsten TX! Not surprisingly, this seems to almost exclusively affect the most dynamic databases in our application suite which are opened, accessed, modified (DmResizeRecord) or added to (DmNewRecord), and closed often, but maintaining and reusing the DmOpenRef has not seemed to help. These databases can become quite large, but not large enough, even with several other helper databases that could be open, to consume the 10MB of dbcache. Also, I have spent days redesigning the affected database in order to save results per user rather than per group of users in order to minimize the database sizes to, at the most, 300-400K, but this did not address the issue. Rather than a single dmErrCorruptDatabase error, I received several - one for each of several users - although not always on the same database. I've pushed this vast quantity of data through the emulator with no issues. Out of many thousands of deployed Palm handhelds (Tungsten E/E2/TX), *ALL* of the handhelds known to have this issue are Tungsten E2 handhelds, so it's difficult not to suspect this is an early implementation of NVFS issue. Can anything be done to address or work around the E2 NVFS problems - or is there something obvious I'm missing here!? Thanks for any help, Russell -- For information on using the ACCESS Developer Forums, or to unsubscribe, please see http://www.access-company.com/developers/forums/
