LITHOSP stores temporary versions of files in TESTFILES to indicate they have a lock on them. Often times you won't actually see these files because they don't stay for long before getting deleted, but if you do an "ls" in that directory repeatedly you'll see them come and go. If one instance of LITHOSP crashes then all of the others instances will halt until that file is deleted.
To get all instances of LITHOSP to resume, sometimes all you'll need to do is delete the files in TESTFILES. However, the problem landmark still hasn't iterated. In that case you'll need to run that landmark again, so do a find_nofitP and look for the landmark that has a "!" next to it, and run that landmark again (i.e. LITHOSP < XXX.INN).
Other times you'll have to poke more to get it to resume. Perhaps there's a file missing or you've run out of disk space or something like that. *Add more solutions as you find them.*