[TriLUG] 512/4096B cluster size incompatibilities

Tue Jan 1 16:58:36 EST 2019

On Tue, 1 Jan 2019, David Burton wrote:

> That's very odd, Joe.
>
> This one is *wrong:*
> Serial Number:    Z4Z408QQ
> Firmware Version: 0958
> Sector Size:      512 bytes logical/physical
>
> This one is* right:*
> Serial Number:    Z4Z5LZJG
> Firmware Version: CC26
> Sector Sizes:     512 bytes logical, 4096 bytes physical
>
> (Note: you swapped the two firmware versions in your email; according to
> the smartctl output you pasted, it's the 0958 firmware which is reporting
> the wrong physical sector size, not the CC26 firmware.)

Hi David,

Thanks for popping up again.

Sorry about assigning the drives incorrectly and thanks for figuring it out.

> It appears that the 0958 version firmware is reporting incorrect information. 
> Maybe that's causing fdisk to do the wrong thing when partitioning the drive, 
> though that is surprising to me. I think modern fdisks should handle "512 
> bytes logical/physical" exactly the same as "512 bytes logical, 4096 bytes 
> physical".
>
> Steven is correct, the Seagate web site says there's no firmware update 
> available for this model 
> <https://apps1.seagate.com/downloads/request.html?userPreferredLocaleCookie=en_US_>. 
> (See pasted web page, below.) Since you have two *different* firmware versions 
> for the exact same model drive, it seems pretty clear that the Seagate web 
> site is lying.

:-(

well I'll download their software and run it under Windows. Last time I tried a 
disk manufacturer's test software (for a return), all it did was run what 
appeared to be a readonly test on the surface, so I decided never to bother with 
manufacturer's testing software again.

> So you could contact Seagate tech support, and try to get a firmware update to 
> version CC26 for the Z4Z408QQ drive.

OK.

Well 0958 is the one that doesn't work with smartctl -A. Are we still having a 
problem with the firmware versions swapped?

> Ah... did you use fdisk? Or something else, to partion? Maybe the partitioning 
> tool that you used is not doing the right thing. You could try a different 
> partitioning tool.

I haven't partitioned either disks. I'd freshly badblocked them from their 
previous job, and had put both into a dual drive usb external drive case (for 
use with ZFS, which I just got running yesterday - yay! - thanks to the talk we 
had in the fall) and noticed that one drive wouldn't give the smartctl -A 
output, while the other did. ZFS is happy with both drives.

On looking at how ZFS has partitioned the drives, I find that both are 
partitioned with 512B clusters.

I just find it a pain to not be able to run smartctl on a drive.

Both drives give the expected smartctl output on a regular internal SATA port, 
but not in the same external case.

> I would suggest zeroing the drive before rebooting and then partitioning it:
> ddrescue -F /dev/zero /dev/sdb
> (you can abort it after it's written a few GB)
>
> You could even try zeroing the drive, and then moving it to a Windoze
> machine, and partitioning it there. That should create a "very standard"
> partition table.
>
> Or you could partition & format the Z4Z5LZJG (CC26 firmware) drive (which
> *should* be correct!), and then use ddrescue to image it onto the Z4Z408QQ
> (0958 firmware) drive.

at the moment, I don't care what the partition table looks like anymore (I've 
given up).

I also don't trust ddrescue between devices with different logical block sizes.

What I would like is to get the smartctl output.

> This is going to sound stupid, but...

We're already in stupidland. We're just trying to figure out where the stupidity 
is. Bring on the stupidity.

> If I were you, I'd shutdown, then swap the SATA cables between the two
> drives (i.e., disconnect them at the drive, not at the MB).

I'd thought of that, but by the time I did, the disks were doing things and I 
didn't want to un-ZFS them (since I'm new to ZFS). I then promptly forgot about 
it. But glad you brought it up. It has to be done to see whether the problem is 
with the disk or the disk controller.

Here we go

--

before swapping the drives in the external enclosure.

/dev/sdb (firmware 0958, 512/512) doesn't give the smartctl -A output

homer: 01 Jan 
2019:/src/distros/slackware/slackware-14.2/slackbuilds-14.2-20181229.1# smartctl -d sat -a /dev/sdb
smartctl 6.5 2016-05-07 r4318 [x86_64-linux-4.4.14] (local build)
Copyright (C) 2002-16, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Family:     Seagate Barracuda 7200.14 (AF)
Device Model:     ST2000DM001-1ER164
Serial Number:    Z4Z408QQ
Firmware Version: 0958
User Capacity:    2,000,398,934,016 bytes [2.00 TB]
Sector Size:      512 bytes logical/physical

/dev/sdc (firmware CC26, 512/4096) does give the smartctl -A output

homer: 01 Jan 
2019:/src/distros/slackware/slackware-14.2/slackbuilds-14.2-20181229.1# smartctl -d sat -a /dev/sdc
smartctl 6.5 2016-05-07 r4318 [x86_64-linux-4.4.14] (local build)
Copyright (C) 2002-16, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Family:     Seagate Barracuda 7200.14 (AF)
Device Model:     ST2000DM001-1ER164
Serial Number:    Z4Z5LZJG
LU WWN Device Id: 5 000c50 0917d2705
Firmware Version: CC26
User Capacity:    2,000,398,934,016 bytes [2.00 TB]
Sector Sizes:     512 bytes logical, 4096 bytes physical

--

after swapping slots in the external enclosure

/dev/sdb (firmware 0958, 512/4096) doesn't give the full output with smartctl

homer: 01 Jan 
2019:/src/distros/slackware/slackware-14.2/slackbuilds-14.2-20181229.1# smartctl -d sat -a /dev/sdb
smartctl 6.5 2016-05-07 r4318 [x86_64-linux-4.4.14] (local build)
Copyright (C) 2002-16, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Family:     Seagate Barracuda 7200.14 (AF)
Device Model:     ST2000DM001-1ER164
Serial Number:    Z4Z5LZJG
Firmware Version: 0958
User Capacity:    2,000,398,934,016 bytes [2.00 TB]
Sector Size:      512 bytes logical/physical

.
.
.

SMART Error Log not supported

SMART Self-test Log not supported

Selective Self-tests/Logging not supported

/dev/sdc (firmware CC26, 512/4096) is just fine

homer: 01 Jan 
2019:/src/distros/slackware/slackware-14.2/slackbuilds-14.2-20181229.1# smartctl -d sat -a /dev/sdc | more
smartctl 6.5 2016-05-07 r4318 [x86_64-linux-4.4.14] (local build)
Copyright (C) 2002-16, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Family:     Seagate Barracuda 7200.14 (AF)
Device Model:     ST2000DM001-1ER164
Serial Number:    Z4Z408QQ
LU WWN Device Id: 5 000c50 08759a553
Firmware Version: CC26
User Capacity:    2,000,398,934,016 bytes [2.00 TB]
Sector Sizes:     512 bytes logical, 4096 bytes physical

--

summary

firmware 0958, 512/512 is /dev/sdb in both slots (I guess it answers first) and 
does not output the smartctl -A results.

The external drive has an eSATA port. I got the same results with it.

I assume the problem is with the drive and not the slot in the external drive 
case.

--

I wasn't expecting much help from Seagate and I expect I'll have to live with 
it. I was just hoping that someone on the list had something to say about the 
problem. Maybe like "Seagate does this all the time. I buy from ... "

This seemed unlikely too. If Seagate was shipping drives randomly assigned to 
512B or 4096B logical clusters, I would have expected people who use disks in 
sets to complain bitterly.

> Download Finder

thanks for the info from Seagate

Joe

--
Joseph Mack NA3T EME(B,D), FM05lw North Carolina
jmack (at) wm7d (dot) net - azimuthal equidistant 
map generator at http://www.wm7d.net/azproj.shtml 
Homepage http://www.austintek.com/ It's GNU/Linux!