failed to read SMART Attribute Data

[kenster]

I'm getting intermittent SMART errors (below) in FreeNAS installed on a Supermicro 36 bay chassis. How do I begin troubleshooting this? I'm not convinced it's a SMART drive error.
I have 12 6TB drives in a ZFS pool with 2 vdevs and WD-WD60EFAX-68SHWN0 drives. The drives are running about 31-34C

New alerts:
* Device: /dev/da8 [SAT], Read SMART Error Log Failed
* Device: /dev/da8 [SAT], failed to read SMART Attribute Data
* Device: /dev/da8 [SAT], not capable of SMART self-check
* Device: /dev/da8 [SAT], Read SMART Self-Test Log Failed

 

[Chris Moore]

Here is a link to a very well written troubleshooting guide by joeschmuck :

Hard Drive Troubleshooting Guide (All Versions of FreeNAS)

UPDATE: 22 September 2018 - Added Drive Data Refreshing UPDATE: 2 April 2017 - Added support for FreeNAS Corral (FreeNAS 10 and beyond) UPDATE: 1 November 2020 - Added ID 1 and 7 description for Seagate drives at bottom of Appendix B This guide...

www.truenas.com

Here is a whole long list of useful links:

Links to useful threads

I created this resource so I could share this list of links with the forum. I accumulated these links over the last few years to help me remember where to find some important information. I hope that others will find this helpful. Most are...

www.truenas.com

 

[Chris Moore]

PS. If you have an empty bay, you can swap the drive to another bay as a test to see if there is some problem with the connection.

You might also share the output of the smartctl –a so we can take a look at the details for you. Someone on here has probably seen a similar error before.

 

[Alecmascot]

May be a timeout because the SMR drive is busy with housekeeping.

 

[Chris Moore]

WD-WD60EFAX-68SHWN0

May be a timeout because the SMR drive is busy with housekeeping.

I didn't look the model number up. If all your drives are that model, it is bad news for your pool. The SMR drives are terrible with ZFS. So bad that WD came out with a new model of drive recently.

WD Red DM-SMR Update 3 Vendors Bail and WD Knew of ZFS Issues

We discuss other WD Red DM-SMR performance experiences, share that WD-HGST knew using DM-SMR was not good for ZFS, and how this could have passed testing

www.servethehome.com

 

[Chris Moore]

PS. I didn't realize, when I linked to that article, that they have a photo of me in the article... Honest.

 

[joeschmuck]

PS. I didn't realize, when I linked to that article, that they have a photo of me in the article... Honest.

You just can't stop being the center of attention. Well I guess another star is born.

 

[kenster]

I didn't look the model number up. If all your drives are that model, it is bad news for your pool. The SMR drives are terrible with ZFS. So bad that WD came out with a new model of drive recently.

WD Red DM-SMR Update 3 Vendors Bail and WD Knew of ZFS Issues

We discuss other WD Red DM-SMR performance experiences, share that WD-HGST knew using DM-SMR was not good for ZFS, and how this could have passed testing

www.servethehome.com

this issue re-occured, so I checked. They are all SMR WD drives. Looks like I need to replace them one by one as they smart test fail. the weird thing is that they show SMART test passed ok when I run smartctl -a

 

[NugentS]

Its not a smart issue (per se). ZFS and SMR simply do not mix well. You need to replace all of the disks, not just one at a time.
Everytime you replace a disk the pool re-silvers - this will take forever and trash the crap out of the existing HDD's. Then you do it again, and again, and again, and again, and again (rinse and repeat). This process could take months during which your data will be at risk

Backup the pool, remove and replace all disks then restore the pool

 

[Ericloewe]

If you can attach the new disks in parallel, you can actually resilver as many as will fit in one go.

 

[kenster]

Its not a smart issue (per se). ZFS and SMR simply do not mix well. You need to replace all of the disks, not just one at a time.
Everytime you replace a disk the pool re-silvers - this will take forever and trash the crap out of the existing HDD's. Then you do it again, and again, and again, and again, and again (rinse and repeat). This process could take months during which your data will be at risk

Backup the pool, remove and replace all disks then restore the pool

this is crazy. you're saying the SMR drives won't survive the re-silvering time it will take to replace them? I have 12 drives 2 6 drive vdevs. The pool is in a 36 bay super micro so I could build a new pool and copy the data from the one pool to the other easy enough. But you're saying these SMR drives are so crap I need to replace them now?

If you can attach the new disks in parallel, you can actually resilver as many as will fit in one go.

I have 2 vdevs with 6 6TB drives in each. You're saying I could replace one drive in each vdev simultaneously.

 

[Ericloewe]

Without loss of redundancy, you can replace as many disks as you can attach at once. If you can't attach additional disks and have two RAIDZ2 vdevs or something similar, which you imply you do, you could also conceivably remove one drive from each vdev and thus replace two disks at a time.

 

[NugentS]

As you haven't posted most of your hardware spec including specifics of your pool makeup I simply gave you the worst case

Testing WD Red 4TB SMR vs CMR HDDs in a ZFS NAS Array - Skip SMR

Check out Will's main site article here: https://www.servethehome.com/wd-red-smr-vs-cmr-tested-avoid-red-smr/We tested the WD Red WD40EFAX drive that uses DM...

www.youtube.com

 

[kenster]

As you haven't posted most of your hardware spec including specifics of your pool makeup I simply gave you the worst case

Testing WD Red 4TB SMR vs CMR HDDs in a ZFS NAS Array - Skip SMR

Check out Will's main site article here: https://www.servethehome.com/wd-red-smr-vs-cmr-tested-avoid-red-smr/We tested the WD Red WD40EFAX drive that uses DM...

www.youtube.com

FreeNAS-11.2-U8
128 GB
Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz

2 vdevs
6 6TB WD Red drives in each
SuperMicro 36 Bay Chassis
LSI SAS9300-8i

 

[NugentS]

and how are the vdevs configured?

 

[kenster]

and how are the vdevs configured?

I have 2 vdevs with 6 6TB drives in each.

 

[Ericloewe]

RAIDZ1/2/3 or 6-wide mirrors? I'm going to guess it's not the latter.

 

[NugentS]

Well clearly you appear to have no idea how the vdevs are configured
So - the only options I can suggest are:
1. Build an entirely new Pool with new CMR disks and copy the data over. Then send the SMR disks back to WD for replacement as not fit for purpose. You might get lucky. This is by far the better solution - much much safer
2. Replace a single disk in each vdev and resilver, repeat five more times. Make sure you have a good backup cos the disks may not survive (and you may not survive the wait)

Read the link to STH that I posted.

Also read the forum rules - please post your hardware specs. For examples, look at my signature, Look at @Ericloewe's signature.

Mod note:
Edited for tone.

 

[kenster]

RAIDZ1/2/3 or 6-wide mirrors? I'm going to guess it's not the latter.

it's RAIDZ2

 

[Ericloewe]

So yeah, if you can't attach additional disks, at least you can replace two at a time fairly safely.