Errors during badblocks test on two brand new drives in same/similar spots

[imaack]

I'll preface by saying that I'm a total FreeNAS newbie and a moderate linux user (more ubuntu than freebsd)

I am building a new FreeNAS machine and have installed FreeNAS on it. During the installation process my two storage hard drives (each 4TB WD Red) were configured in a Mirror layout I believe (I just followed the wizard).

I'm following this guide on how to burn-in my drives:
https://forums.freenas.org/index.php?threads/how-to-hard-drive-burn-in-testing.21451/

I have executed the earlier steps and I'm now executing this command (on both drives simultaneously) using tmux :

Code:

badblocks -b 4096 -ws /dev/adaX

On what I think was the third pass I start to get errors on both drives and currently the status says something like this:

Drive 1:
(0/0/215464)

Drive 2:
(0/0/212512)

The output also lists blocks/sectors (whatever the numbers reference) that match exactly between the two drives, e.g:
976754656
976754657
976754658
976754659
976754660
976754661

I find it unlikely that I received two drives that have bad blocks in exactly the same places and my hunch is that there is something writing to the volume (and FreeNAS is mirroring things).

Another potentially pertinent tidbit is I am using a KVM to switch between my workstation and the FreeNAS box and every time I switch I get a log output like this:

Code:

ums0: at uhub-, port 9, addr 13 (disconnected)class 0/0, rev 2.00/1.12, addr 13> on usbus0

I switched a few times early on but had stopped before the third pass (that generated the errors) began... so I am confused.

My question is this: Is there something on FreeNAS that is writing to the volumes that I should disable? Is it the KVM switch? Is the fact that the totals for both drives are not the same cause for concern (0/0/215464) vs (0/0/212512)?

Appreciate the help!

 

[Stux]

Running badblocks in destructive mode across disks in an active pool would be bad.

You should detach the pool (see manual) and you might as well destroy the data too (it's a checkbox).

Rerun the badblocks test.

Then recreate your pool.

 

[imaack]

I don't have any actual data on there so no damage done but it is an active pool. I will follow those steps. Thanks for your suggestions.

 

[SweetAndLow]

Running badblocks on drives that are part of a pool will give lots of work because zfs will try to fix things and cause strange badblocks results.

Sent from my Nexus 5X using Tapatalk

 

[joeschmuck]

I don't have any actual data on there so no damage done but it is an active pool. I will follow those steps. Thanks for your suggestions.

You don't understand, you should not run badblocks on an active pool. Getting errors doesn't surprise me.

Let me explain... You are running badblocks which writes data across the drive and verifies the data it wrote. Now your active pool starts to write something in the same area that badblocks is trying to work and badblocks reads back data it didn't expect thus a failure.

I'm not saying this is what is happening however both drives failing in the same spot has got to be very rare.

Stop badblocks, destroy your pool, start up badblocks again.