CURRENT: Issue with ZFS and 2TB WD HDD (WDC WD20EURS-63Z9B1 80.00A80)

[O. Hartmann]

On 11/29/10 19:32, James R. Van Artsdalen wrote:
> O. Hartmann wrote:
>> Attached to the system are three WD harddrives with ZFS as filesystem
>> on GPT partitioning scheme.
>
> Please send the output of "camcontrol devlist" and "zpool status" on
> FreeBSD-8.1.  Then export the pool in FreeBSD-8.1, boot FreeBSD-9 with
> verbose booting and send the output of "camcontrol devlist", "zpool
> import BACKUP" and the part of /var/log/messages that shows FreeeBSD
> searching for the pool during the import.

Appended, you'll find some of the requested data. There is no valuable 
dmesg/message output, due to the simple fact, that after yesterday's 
"make world" and another import try the BACKUP00 pool was imported. The 
only thing to report is, that while importing the BACKUP00 pool, which 
is on devise ada3, I got an message on console like:

Nov 29 21:13:36 <0.2> thor kernel: ZFS WARNING: Unable to attach to ada1.

This strange message always came up when importing the BACKUP00 pool 
(BACKUP is a typo, sorry for the inconvenience) and when it was about to 
fail.

>
> It might be useful to know what disk controller is being used.
>
>> =>         34  3907029101  ada3  GPT  (1.8T)
>>            34        4062        - free -  (2.0M)
>>          4096  3907025039     1  freebsd-zfs  (1.8T)
>
> It is not related to your problem, but how did that unused space wind up
> before partition 1?

Well, my fault. I initiated the partitions with the "-b 4096" option set 
for the first partition. It was so nice and at ease playing around with 
gpart ...

And here comes the data requested:

#FreeBSD 8.2-PRE/amd64
----------------------
camcontrol devlist
<WDC WD5001AALS-00L3B2 01.03B01>   at scbus4 target 0 lun 0 (pass0,ada0)
<WDC WD1001FALS-00J7B0 05.00K05>   at scbus5 target 0 lun 0 (pass1,ada1)
<WDC WD1001FALS-00J7B0 05.00K05>   at scbus6 target 0 lun 0 (pass2,ada2)
<WDC WD20EURS-63Z9B1 80.00A80>     at scbus7 target 0 lun 0 (pass3,ada3)
<HL-DT-ST DVDRAM GH22LS50 TL03>    at scbus9 target 0 lun 0 (cd0,pass4)

#FreeBSD 9.0-CURRENT/amd64
--------------------------
camcontrol devlist
<WDC WD6402AAEX-00Y9A0 01.01V01>   at scbus4 target 0 lun 0 (pass0,ada0)
<WDC WD1001FALS-00J7B0 05.00K05>   at scbus5 target 0 lun 0 (pass1,ada1)
<WDC WD1001FALS-00J7B0 05.00K05>   at scbus6 target 0 lun 0 (pass2,ada2)
<WDC WD20EURS-63Z9B1 80.00A80>     at scbus7 target 0 lun 0 (pass3,ada3)
<WDC WD5001AALS-00L3B2 01.03B01>   at scbus8 target 0 lun 0 (pass4,ada4)
<HL-DT-ST DVDRAM GH22LS50 TL03>    at scbus9 target 0 lun 0 (cd0,pass5)

(needed to swap 1st and 5th drive since FBSD 8 was installed with MBR 
and no labels and needed device special file for mounting partitions 
while on FreeBSD 9 I used the first time GPT partitioning scheme with 
labeling for mounting via /etc/fstab)

#FreeBSD 8 zpool status:
------------------------
   pool: BACKUP00
  state: ONLINE
  scrub: none requested
config:

         NAME        STATE     READ WRITE CKSUM
         BACKUP00    ONLINE       0     0     0
           ada3p1    ONLINE       0     0     0

errors: No known data errors

   pool: THOR00
  state: ONLINE
  scrub: none requested
config:

         NAME        STATE     READ WRITE CKSUM
         THOR00      ONLINE       0     0     0
           ada1p1    ONLINE       0     0     0
           ada2p1    ONLINE       0     0     0

errors: No known data errors

#FreeBSD 8 pciconf -lcv (for the requested controller)
------------------------
ahci1 at pci0:0:31:2:      class=0x010601 card=0x82d41043 chip=0x3a228086 
rev=0x00 hdr=0x00
     vendor     = 'Intel Corporation'
     device     = '6 port SATA AHCI Controller'
     class      = mass storage
     subclass   = SATA
     cap 05[80] = MSI supports 16 messages enabled with 1 message
     cap 01[70] = powerspec 3  supports D0 D3  current D0
     cap 12[a8] = SATA Index-Data Pair
     cap 09[b0] = vendor (length 6) Intel cap 2 version 0

In kernel config, I use AHCI driver modell, not the atapi one, but I 
guess this is obvious due to 'adaXX' ...

And by the way, the 2 TB drive in question is sold as a 'PVR/VCR video 
special device", I mistakenly ordered this one since it promissed low 
power, high sustain-reliability (it is used for backups), but some web 
sites claim that the controller of this specific, expensive 2 TB hdd 
deals different with errornous blocks read or written out - I can not 
verify this, since I have no technical paper nor do I understand much of 
HDD electrical engineering ...

Regards,
Oliver