Re: panic in zfs(?) on stable/12

My choice would be to test the memory and try 13.
BTW: What checksum algorithm are you using? Something standard or a more obscure one?


Regards,
Ronald.


Van: Juraj Lutter <juraj@lutter.sk>
Datum: 13 juli 2022 10:08
Aan: stable@freebsd.org
Onderwerp: panic in zfs(?) on stable/12

> 
> 
> Hi.
> 
> every now and then I’m getting:
> 
> panic: vm_fault: fault on nofault entry, addr: 0xfffffe01d9620000
> cpuid = 5
> time = 1657698436
> KDB: stack backtrace:
> #0 0xffffffff80c2e775 at kdb_backtrace+0x65
> #1 0xffffffff80be4ba8 at vpanic+0x178
> #2 0xffffffff80be4a23 at panic+0x43
> #3 0xffffffff80f63b1b at vm_fault+0x20fb
> #4 0xffffffff80f61920 at vm_fault_trap+0x60
> #5 0xffffffff81100270 at trap_pfault+0x1e0
> #6 0xffffffff810d7c28 at calltrap+0x8
> #7 0xffffffff824bf80d at abd_iterate_func+0x12d
> #8 0xffffffff8257ced6 at zio_checksum_generate+0x66
> #9 0xffffffff8257b26d at zio_execute+0xad
> #10 0xffffffff80c412b4 at taskqueue_run_locked+0x144
> #11 0xffffffff80c426a2 at taskqueue_thread_loop+0xd2
> #12 0xffffffff80ba592e at fork_exit+0x7e
> #13 0xffffffff810d8c5e at fork_trampoline+0xe
> Uptime: 3d11h13m45s
> 
> 
> Machine is stable/12 (very recent, built from clean sources), zpool upgraded to version 28.
> 
> I’m now on a crossroad: Is it better to thoroughly test the memory, upgrade to stable/13, switch to out-of-tree OpenZFS or diagnose the ZFS further?
> 
> Thanks!
> otis
> 
> —
> Juraj Lutter
> otis@FreeBSD.org
> 
> 
> 
> 
> 
>