Did you check the crash time from some other logs? My point is that crashes might happen at a later time too, and the file is still 0 MB.

For me crashing with can log vfs happens quite randomly. Every time a crash happens the capture file size is zero, even if the capture was running for a long time already.. My workaround for this has been stop&start longer captures periodically (joining multiple crtd is a trivial task: just remove the header from follow ups and join).

Jarkko