smartd: Currently unreadable (pending) sectors

[basnek]

Hello everyone,
A warning has started to pop up on the console
"...Dec 8 10:15:24 truenas 1 2023-12-08T10:15:24:24.771405+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, 1 Currently unreadable (pending) sectors..."

Every 30" I have a warning about it
Consulting the logs I have done a grep in the “messages” logs and I see that everything seems to start on December 7, 2023

Show : Grep messages

Dec 7 22:06:03 truenas (ada0:ata2:0:0:0): READ_DMA. ACB: c8 00 18 7a 58 48 00 00 00 00 80 00
Dec 7 22:06:03 truenas (ada0:ata2:0:0:0): CAM status: ATA Status Error
Dec 7 22:06:03 truenas (ada0:ata2:0:0:0): ATA status: 51 (DRDY SERV ERR), error: 40 (UNC )
Dec 7 22:06:03 truenas (ada0:ata2:0:0:0): RES: 51 40 20 7a 58 08 08 00 00 80 00
Dec 7 22:06:03 truenas (ada0:ata2:0:0:0): Retrying command, 3 more tries remain
Dec 7 22:06:07 truenas (ada0:ata2:0:0:0): READ_DMA. ACB: c8 00 18 7a 58 48 00 00 00 00 80 00
Dec 7 22:06:07 truenas (ada0:ata2:0:0:0): CAM status: ATA Status Error
Dec 7 22:06:07 truenas (ada0:ata2:0:0:0): ATA status: 51 (DRDY SERV ERR), error: 40 (UNC )
Dec 7 22:06:07 truenas (ada0:ata2:0:0:0): RES: 51 40 20 7a 58 08 08 00 00 80 00
Dec 7 22:06:07 truenas (ada0:ata2:0:0:0): Retrying command, 2 more tries remain
Dec 7 22:15:24 truenas 1 2023-12-07T22:15:24.853741+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, 1 Currently unreadable (pending) sectors
Dec 7 22:15:25 truenas 1 2023-12-07T22:15:25.343202+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, ATA error count increased from 0 to 2

I have also done the grep in the daemon.log where you can see the temperature ones and another one about "Seek_Error_Rate"

Show : Grep daemon.log

Nov 19 11:39:25 truenas 1 2023-11-19T11:39:25.847838-08:00 truenas.local smartd 1729 - - Device: /dev/ada0, SMART Usage Attribute: 7 Seek_Error_Rate changed from 100 to 200
Nov 28 18:06:37 truenas 1 2023-11-28T18:06:37.908024+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, SMART Usage Attribute: 194 Temperature_Celsius changed from 126 to 127
Dec 7 09:45:24 truenas 1 2023-12-07T09:45:24.682225+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, SMART Usage Attribute: 7 Seek_Error_Rate changed from 200 to 100
Dec 7 22:15:24 truenas 1 2023-12-07T22:15:24.853741+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, 1 Currently unreadable (pending) sectors
Dec 7 22:15:25 truenas 1 2023-12-07T22:15:25.343202+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, ATA error count increased from 0 to 2
Dec 8 15:45:24 truenas 1 2023-12-08T15:45:24.350671+01:00 truenas.local smartd 1699 - - Device: /dev/ada0, 1 Currently unreadable (pending) sectors

After that message, the one who told you that it repeats itself every 30 “continually”

Does this mean I have to change that disk?
I hope that with what I tell you and the attachments you can help me with what I should do. Thank you

Show : See My TrueNAS system

TrueNAS-13.0-U5.3
SuperMicro X8DTL-3
Processor (CPU) - 2 x Intel Xeon CPU E5620 @2.40GHz
Threads: 16
24 GB RAM
4 HD x 4TB

 

[NugentS]

Can you expand you hardware list to say what HDD's you are using
Also the results of the following in code blocks please

"smartctl -a /dev/ada0"
"zpool status -v"

 

[basnek]

Fist of All Thx for you soon anser
we attach the outputs of every command
I hope I did well
Best Regards

 

[NugentS]

First: You aren't running long smart tests - which you should be doing.

Next:
That drive is 6 years powered on - so its doing well

Whats happened is that there is a possible faulty sector - this in itself is not an issue - unless that number starts to expand. However you don't actually know the state of the drive as short tests only test the electronics, and not the drive surface which means you MIGHT have an issue, but don't know about it. A long test actually attempts to read the drives surface - so may turn up a bunch of errors that you don't yet know about.

So - do you have a backup?

You need to run a long smart test on all the drives, and schedule one every month at least. Then see what happens

 

[basnek]

It's good to read your comments

I'll put you in context a month or so ago I was forced to change the server and now I have a supermicro to which I put two 4tb disks that came from my previous server (a poweredge) and I bought two other 4t disks and intercalated them but Apparently I did it by putting from left to right new - old - new - old but from what I see the reading of the ada was the other way around and I have ahrao old - new - old - new
When I saw what the command you recommended showed me, I realized this detail and of course this is one of the two old disks from when I had the poweredge there in 2017

Regarding your comments about the long smart tests. Could you recommend how to do this? Excuse my newness.

Luckily I have a backup made after the new installation of my new server.

I would appreciate information regarding what you are saying about the long smart tests or at least where I can find out so I don't do nonsense.

Thanks a lot!

 

[NugentS]

You can schedule short or long tests using the gui. I can't remember exactly where on Core. Also don't pay any attention to ada0/1/2/3 etc. Every time TN boots these can change

or "smartctl -t long /dev/ada0" at the command line. Do this for all 4 drives then wait the required time+ and see what happens. Its better to schedule these tests say 1/month (long) and say 1/week (short) for best effect. SMART isn't perfect - but it does tend to warn you of upcoming failure. All the long tests can run at the same time - its the drive testing itself. You see the results with "smartctl -a /dev/ada0"

Given that you haven't been running tests - you MIGHT have a nasty shock coming your way - I hope not

 

[Redcoat]

I can't remember exactly where on Core.

Filling in for @NugentS here:

Go to Tasks>SMART Tests;

Click ADD and configure to your taste per the docs at https://www.truenas.com/docs/core/13.0/coretutorials/tasks/runningsmarttests/#automatic-smart-tests:

 

[Etorix]

Irrespective of the result of long SMART tests, "pending sectors" means that the drive is growing bad sectors (which will end up in "reallocated"). Which, after six years of use, is fair. One sector is not much, but one may argue that it's one too many, and things can only get worse from here. So, plan to replace the drive. Maybe not right now, but in the near future.

 

[basnek]

Irrespective of the result of long SMART tests, "pending sectors" means that the drive is growing bad sectors (which will end up in "reallocated"). Which, after six years of use, is fair. One sector is not much, but one may argue that it's one too many, and things can only get worse from here. So, plan to replace the drive. Maybe not right now, but in the near future.

totally agree

 

[basnek]

Filling in for @NugentS here:

Go to Tasks>SMART Tests;

Click ADD and configure to your taste per the docs at https://www.truenas.com/docs/core/13.0/coretutorials/tasks/runningsmarttests/#automatic-smart-tests:

View attachment 73286

Thx for the info!

 

[basnek]

You can schedule short or long tests using the gui. I can't remember exactly where on Core. Also don't pay any attention to ada0/1/2/3 etc. Every time TN boots these can change

or "smartctl -t long /dev/ada0" at the command line. Do this for all 4 drives then wait the required time+ and see what happens. Its better to schedule these tests say 1/month (long) and say 1/week (short) for best effect. SMART isn't perfect - but it does tend to warn you of upcoming failure. All the long tests can run at the same time - its the drive testing itself. You see the results with "smartctl -a /dev/ada0"

Given that you haven't been running tests - you MIGHT have a nasty shock coming your way - I hope not

I also hope not, but it's good that all the bad things are that. For now I will be asking for another album for which we know is already showing signs of fatigue :) and what you recommend is totally valid so it will be what it has to be. But hey, at least I want to have a disk on hand for whatever it takes. And I will cross my fingers for the new albums it has since as I told you I have 2 old ones and 2 new ones so that the new ones are not bad I cross my fingers hahaha thank you for all of your help