On Mon, May 27, 2013 at 04:31:25PM +0100, Simon Slavin wrote: > > On 27 May 2013, at 4:22pm, Woody Wu <[email protected]> wrote: > > > If Yaffs2 is the cause, how can I write an effective test to exposure it? > > Do you have an opportunity to format the same drive in a different format ? > I'm not telling you to change your long-term practises, just to try a > different format for testing. Reformat it, run your identical testing > program, and see if it fails in the same way. > > Also I liked Richard's point: are you testing on a cheap Flash drive ? > Perhaps one intended only for use with FAT and therefore only tested for use > with FAT ?
Can you experts explains why a cheap Flash drive can harm an application such as sqlite? Does you mean these cheap drive was cheating with ECC? Otherwise, filesystem should be able to capture IO error (fix it or forward), right? > > > [snip] > > > > Another thing is very suspicious: after a while of running the test > > program, I can see from my syslog that there are a log of warning messages, > > something like: Yaffs2 trunk nnnn was not erased. > > Well, that sounds like it might be relevant. See if you can find why yaffs2 > should write a report like that. What does it indicate about what happened ? > > > On one hand, I am not > > sure if this is caused by sqlite or Yaffs2 itself, on the other hand, I > > also cannot prove this really means bad things since the program at that > > moment was still running fine. > > The chances that this is a sign of a fault in SQLite3 are extremely low. > There are literally millions of SQLite3 installations on all sorts of > hardware. It's part of the OSes that run all the popular designs of > smartphone (iPhone, Android, Windows Phone) all of which, of course, use > Flash Memory for storage. It's possible that the VFS you're using is not > compatible with yaffs2, and the way to test that is what we've discussed > above. > > Simon. > _______________________________________________ > sqlite-users mailing list > [email protected] > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- woody I can't go back to yesterday - because I was a different person then. _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
