All nodes crash at the same time

You may find more information here,

my tip is higher timeouts and defragmentation of the mft.
temporarily(?) loglevel to fatal

oh, and maybe update bios, network and mainboard driver.