[Bug 241980] panic: I/O to pool appears to be hung on vdev

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=241980

--- Comment #20 from Eugene Grosbein <eugen at freebsd.org> ---
This is dedicated server in Hetzner. Well, it has another SATA controller seen
as:

ahci0 at pci0:0:17:5:      class=0x010601 card=0x07161028 chip=0xa1d28086 rev=0x09
hdr=0x00
    vendor     = 'Intel Corporation'
    device     = 'C620 Series Chipset Family SSATA Controller [AHCI mode]'
    class      = mass storage
    subclass   = SATA

But first I'd like to disable panicing by zfs_deadman and see if distinct ZFS
boot pool is alive in such moment. I already tried booting patched zfs.ko by
means of nextboot.conf and
module_path="/boot/nextboot;/boot/kernel;/boot/modules". Additionally, I've
enabled hardware watchdog in BIOS setup (the server has distinct public IP for
console via virtual KVM) and enabled watchdogd.

Sadly, I made a misprint and it booted with default unpatched zfs.ko and I
missed that, so another same panic that occured today gave me little but
interesting information: the watchdog works and dmesg buffer is not cleared
between reboots, so I have KDB trace in /var/log/kernel.log after reboot
(kern.* messages logged there). I've fixed the misprint, will reboot again and
wait.

Thank you for your remark about early warning, now I see you are right. I'll
fix it.

-- 
You are receiving this mail because:
You are the assignee for the bug.