The stress test scenario is: "umount -f" of an active FS: 20190211 12:26:07 all (1/1): pfl3.sh panic: handle_written_inodeblock: live inodedep 0xfffff80261dbc400 cpuid = 45 time = 1549884371 KDB: stack backtrace: db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe010ac22000 vpanic() at vpanic+0x1b4/frame 0xfffffe010ac22060 panic() at panic+0x43/frame 0xfffffe010ac220c0 handle_written_inodeblock() at handle_written_inodeblock+0x669/frame 0xfffffe010ac22110 softdep_disk_write_complete() at softdep_disk_write_complete+0x462/frame 0xfffffe010ac22170 bufdone() at bufdone+0x3bb/frame 0xfffffe010ac221f0 g_io_deliver() at g_io_deliver+0x298/frame 0xfffffe010ac22240 g_io_deliver() at g_io_deliver+0x298/frame 0xfffffe010ac22290 md_kthread() at md_kthread+0x258/frame 0xfffffe010ac222f0 fork_exit() at fork_exit+0x84/frame 0xfffffe010ac22330 fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe010ac22330 --- trap 0, rip = 0, rsp = 0, rbp = 0 --- KDB: enter: panic [ thread pid 35083 tid 101903 ] Stopped at kdb_enter+0x3b: movq $0,kdb_why db> x/s version version: FreeBSD 13.0-CURRENT #0 r343968: Sun Feb 10 20:35:52 CET 2019\012 pho@flix1a.netperf.freebsd.org:/usr/src/sys/amd64/compile/PHO\012 db> https://people.freebsd.org/~pho/stress/log/pfl3.txt
Created attachment 204257 [details] panic bt screenshot Just hit this after an unclear shutdown. Boot into single user and two fsck -y runs marked the FS clean. 13.0-CURRENT r340668 (amd64, GENERIC-NODEBUG)
The problem is not seen in: FreeBSD 14.0-CURRENT #0 main-n244671-b3c6fe663bb
Hi Team, does anyone one knows which commit fixed this issue in 14.0. We are facing similar problem on FBSD 13.0 and want to backport the changes from 14.0 if possible.
(In reply to Rupesh Pilania from comment #3) I do not know if this helps, but I just ran the same test on stable/13-f22bd73428dd, without any problems seen.