Two USB 4 disk enclosure and a panic

Ronald Klop ronald-lists at klop.ws
Fri Apr 27 08:16:15 UTC 2018


On Wed, 25 Apr 2018 14:15:03 +0200, Nenhum_de_Nos  
<matheus at eternamente.info> wrote:

> On Mon, April 23, 2018 23:18, Nenhum_de_Nos wrote:
>> Hi,
>>
>> I would like to know how to debug this. I have two 4 disk enclosures:
>>
>> Mediasonic ProBox 4 Bay 3.5' SATA HDD Enclosure – USB 3.0 & eSATA
>> (HF2-SU3S2)
>> NexStar HX4 - NST-640SU3-BK
>>
>> and both have 4 disk on them, and not all disk are equal.
>>
>> The issue comes when I plug the probox usb3 enclosure on the system. I
>> can't even read the /var/log/message, it crashes very quickly.
>>
>> I can see on the boot process up to the point where the second enclosure
>> comes to be loaded. The 4 disk are shown on the dmesg/console, and then  
>> a
>> core dump happens, the boot process goes to the debug screen and a  
>> restart
>> happens like a flash.
>>
>> The motherboard is a Intel® Desktop Board D525MW running 8GB RAM.
>> All disk use ZFS, 4 or 5 zpools, one raidz, one mirror and two or three
>> single disk pools.
>> FreeBSD xxx 11.1-RELEASE-p7 FreeBSD 11.1-RELEASE-p7 #1 r330596: Thu  
>> Mar  8
>> 06:45:59 -03 2018     root at xxx:/usr/obj/usr/src/sys/FreeBSD-11-amd64-PF
>> amd64
>>
>> The kernel is a slightly modified generic, just to have altq.
>>
>> How can I debug this? I have no idea. I have to use two machines to run
>> all those disks, and I would really like to have just one for it.
>>
>> Can it be the amount of RAM? The other box is and APU2 from PCEngines  
>> and
>> have 4GB ram. apu2 uname -a: FreeBSD yyy 11.1-RELEASE-p4 FreeBSD
>> 11.1-RELEASE-p4 #0: Tue Nov 14 06:12:40 UTC 2017
>> root at amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC  amd64
>>
>> I tried to plug the Vantec hardware on the apu2 box, but there it would
>> not panic, but won't load all vantec disks neither. I really run out of
>> ideas here :(
>>
>> thanks.
>>
>> --
>> "We will call you Cygnus,
>> the God of balance you shall be."
>
> Hi,
>
> I found some logs on the daily security output:
>
> +ZFS filesystem version: 5
> +ZFS storage pool version: features support (5000)
> +panic: Solaris(panic): blkptr at 0xfffff8000b93c848 DVA 1 has invalid  
> VDEV 1
> +cpuid = 0
> +KDB: stack backtrace:
> +#0 0xffffffff80ab65c7 at kdb_backtrace+0x67
> +#1 0xffffffff80a746a6 at vpanic+0x186
> +#2 0xffffffff80a74513 at panic+0x43
> +#3 0xffffffff82623192 at vcmn_err+0xc2
> +#4 0xffffffff824a73ba at zfs_panic_recover+0x5a
> +#5 0xffffffff824ce893 at zfs_blkptr_verify+0x2d3
> +#6 0xffffffff824ce8dc at zio_read+0x2c
> +#7 0xffffffff82445fb4 at arc_read+0x6c4
> +#8 0xffffffff824636a4 at dmu_objset_open_impl+0xd4
> +#9 0xffffffff8247eafa at dsl_pool_init+0x2a
> +#10 0xffffffff8249b093 at spa_load+0x823
> +#11 0xffffffff8249a2de at spa_load_best+0x6e
> +#12 0xffffffff82496a81 at spa_open_common+0x101
> +#13 0xffffffff824e2879 at pool_status_check+0x29
> +#14 0xffffffff824eba3d at zfsdev_ioctl+0x4ed
> +#15 0xffffffff809429f8 at devfs_ioctl_f+0x128
> +#16 0xffffffff80ad1f15 at kern_ioctl+0x255
> +CPU: Intel(R) Atom(TM) CPU D525   @ 1.80GHz (1800.11-MHz K8-class CPU)
> +avail memory = 8246845440 (7864 MB)
> +Timecounter "TSC" frequency 1800110007 Hz quality 1000
> +GEOM_PART: integrity check failed (ada0s1, BSD)
> +GEOM_PART: integrity check failed (diskid/DISK-5LZ0ZDBBs1, BSD)

This error also looks interesting.


> +ugen1.2: <Dell Dell KM632 Wireless Keyboard and Mouse> at usbus1
> +ukbd0 on uhub0
> +ukbd0: <Dell Dell KM632 Wireless Keyboard and Mouse, class 0/0, rev
> 2.00/1.30, addr
> 2> on usbus1
> +kbd2 at ukbd0
> +ZFS filesystem version: 5
> +ZFS storage pool version: features support (5000)
> +re0: link state changed to DOWN
> +uhid0 on uhub0
> +uhid0: <Dell Dell KM632 Wireless Keyboard and Mouse, class 0/0, rev
> 2.00/1.30, addr
> 2> on usbus1
> +ums0 on uhub0
> +ums0: <Dell Dell KM632 Wireless Keyboard and Mouse, class 0/0, rev
> 2.00/1.30, addr
> 2> on usbus1
> +ums0: 3 buttons and [XYZ] coordinates ID=0
> +re0: promiscuous mode enabled
> +re0: link state changed to UP
>
> For what I see can be ZFS related.
>
> If anyone have any hints, please tell :)
>
> I kinda got curious about this:
>
> +ZFS storage pool version: features support (5000)
>
> How can I figure out if my pools are from different versions and may this
> be the culprit here?

zpool upgrade [-v]

But I think you can only run this if your machine does not panic. Maybe on  
another machine.

Can you get the output of the coredump? On reboot the /etc/rc.d/savecore  
script will look for a valid core dump in the dumpdev (see /etc/rc.conf,  
/etc/defaults/rc.conf) and write it to /var/crash/. That will contain  
information which can help people debug this.

Regards,
Ronald.


>
> thanks,
>
> matheus


More information about the freebsd-stable mailing list