FreeNAS on ESXi with passthrough SATA

Donixon

Cadet
Joined
Apr 4, 2019
Messages
5
Hi all,

Before we start, hereby the system I am running:
- Intel I7
- 4x 4GB RAM
- Asrock-z77-extreme4 motherboard
- 240GB SSD (NEW)
- 4x 2TB HDD (NEW)

ESXi 6.7 is running on a 32GB USB-stick and FreeNAS 11.2-stable is installed on the 240GB SSD. I got the 4x 2TB HDD connected to FreeNAS with a pass-through workaround: https://gist.github.com/Hengjie/1520114890bebe8f805d337af4b3a064 . This does work all fine and I see the HDD's in FreeNAS.
The 4 HDD's are in a ZFS-pool RAIDZ.

I have a problem with FreeNAS running on ESXi. FreeNAS keeps giving the message: "Device: /dev/da1, 8 currently unreadable (pending) sectors". It started with the number 8 and is now at 56 over a couple of days. Today it also stared giving the message for /dev/da1 (8).

My first reaction was to check the HDD's for badsectors. I did this on FreeNAS layer with the commands "badblocks". Did this for all 4 HDD's. On one it gave back that there were 16 badblocks, i manually checked them and i could read them out correctly, so it looks like this was a false-positive.

Conclusion; it looks like that the HDD's are all fine, but FreeNAS keeps giving the message that there are unreadable sectors...

Can you guys help me to troubleshoot this and hopefully fix the problem!
 

Donixon

Cadet
Joined
Apr 4, 2019
Messages
5
Also did a smartctl check:
"smartctl --health /dev/da1"


=== START OF READ SMART DATA SECTION ===
SMART Status command failed: scsi error unsupported field in scsi command
SMART overall-health self-assessment test result: PASSED
Warning: This result is based on an Attribute check.
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,681
This isn't a problem with running on ESXi. It's just bad blocks. I would return the drives and get some other kind.

Drives will reallocate bad blocks and replace them with spares. It is unlikely this was a "false positive".

The other deficiencies in your system, I am not going to address, since it has nothing to do with your question.
 

Donixon

Cadet
Joined
Apr 4, 2019
Messages
5
Hi @jgreco , thanks for your reaction.

Like I said, I ran a badblocks scan for some hours on all 4 HDD's. There was nothing found. The HDD's are bought in different batches.
My first thoughts also where "a broken disk". But now that there are TWO disks that give this error, I don't trust the message from FreeNAS.
All the scans i run give feedback that everything is healthy...

EDIT;
Just looked again at FreeNAS and at the errors /dev/da1 was out 56 badblocks, and now at 32? How come this fluctuate that much?
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,681
Running a badblocks scan isn't going to find already-remapped sectors. Hard drive manufacturers know that when you have billions of blocks out there, statistically some of them will suffer from defects. The drives have a mechanism to cope with that. It is a limited mechanism, though.

It is the drive itself that is alerting the "unreadable (pending)" sector alert. FreeNAS has absolutely nothing to do with that, other than it watches for it and presents that info.

What are the results of your SMART long tests looking like?
 

Donixon

Cadet
Joined
Apr 4, 2019
Messages
5
Output from "smartctl -t long /dev/da1"

root@freenas0[~]# smartctl -t long /dev/da1
smartctl 6.6 2017-11-05 r4594 [FreeBSD 11.2-STABLE amd64] (local build)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF OFFLINE IMMEDIATE AND SELF-TEST SECTION ===
Sending command: "Execute SMART Extended self-test routine immediately in off-line mode".
Command "Execute SMART Extended self-test routine immediately in off-line mode" failed: scsi error unsupported field in scsi command
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,681
Well that just raises more questions, like what kind of drives are these.
 
Top