This is a *very* old problem. The test scenario is FFS snapshot: 20190218 23:05:29 all (1/1): pageout.sh panic: handle_written_filepage: not started cpuid = 3 time = 1550527654 KDB: stack backtrace: db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe00aa9482d0 vpanic() at vpanic+0x1b4/frame 0xfffffe00aa948330 panic() at panic+0x43/frame 0xfffffe00aa948390 handle_written_filepage() at handle_written_filepage+0x3e/frame 0xfffffe00aa948400 softdep_disk_write_complete() at softdep_disk_write_complete+0xdd/frame 0xfffffe00aa9484a0 bufdone() at bufdone+0x3bb/frame 0xfffffe00aa948520 ufs_strategy() at ufs_strategy+0x83/frame 0xfffffe00aa948550 VOP_STRATEGY_APV() at VOP_STRATEGY_APV+0x90/frame 0xfffffe00aa948570 bufstrategy() at bufstrategy+0x44/frame 0xfffffe00aa9485a0 bufwrite() at bufwrite+0x249/frame 0xfffffe00aa9485f0 ffs_syncvnode() at ffs_syncvnode+0x2ae/frame 0xfffffe00aa948670 ffs_fsync() at ffs_fsync+0x22/frame 0xfffffe00aa9486b0 VOP_FSYNC_APV() at VOP_FSYNC_APV+0x88/frame 0xfffffe00aa9486d0 ufs_remove() at ufs_remove+0xdf/frame 0xfffffe00aa948730 VOP_REMOVE_APV() at VOP_REMOVE_APV+0xac/frame 0xfffffe00aa948750 kern_unlinkat() at kern_unlinkat+0x1e7/frame 0xfffffe00aa948990 amd64_syscall() at amd64_syscall+0x291/frame 0xfffffe00aa948ab0 https://people.freebsd.org/~pho/stress/log/pageout-4.txt https://people.freebsd.org/~pho/kernel+vmcore.20.r344247.mercat1.txz