Hello Mark,
having seen that you recently made version 1.7.0 available, I checked it out and found that still my replication setups were causing core dumps at the receiving side.
I am using asynchronous replication with regularly scheduled shutdowns and application of the log files using replogtool and in one case using watchdir.
Using the cores I could determine a number of problems, many of these are related to the shutdown processing, with some threads still being active while the DB environment seems to be half-shut down already. Also some crashes occured in libtokyocabinet when being called with an empty tree, there are now extra checks to avoid that.
With these fixes applied, I do not have core dumps any longer.
The fixes are available when defining BESI_FIX in the Makefile.
Also, I added configuration options for the logfiles, so that I can have split log files when running more than one instance of lessfs on a server (which I do).
This is available when defining BESI_CONF in the Makefile.
I hope you will find this helpful and suggest to check and possibly add this to the common code.
Best Regards,
Hajo