I think I have resolved my issue with vnstat stopping unexpectedly. It coincides with a scheduled reboot, and at that time dn-vnstat starts but vnstat itself doesn't. I couldn't find out where in the starting sequence precisely it failed, but the S33 starting code is not necessary for me after the initial installation, I think, so I hard-coded the choice of jffs or usb storage location and removed the S32 tests. Now it starts on a reboot.
I seem to recall other instances where a lot of Sxx code created a race condition or some other starting issue. Not sure why, as other add-ons seems to run ok even with a lot; Scribe has a lot offloaded to another file that gets called, for example.
I seem to recall other instances where a lot of Sxx code created a race condition or some other starting issue. Not sure why, as other add-ons seems to run ok even with a lot; Scribe has a lot offloaded to another file that gets called, for example.