TerraMaster Forum

Raid 5 problems… I Will try to be brief… NEW Terra F4 424Pro..

August 9 2024 - Fresh install of up to TOS 6.0.436 Beta… NAS Operating OK…

NVME 1 - 1.02TB WDC PC SN530 SDBTPZ-100, serial xxxad461120
NVME 2 - 512GB KBG50ZNT512G LS KIOXIA, serial xxx1EWVU (used for cache)

And 4 x 18TB HDD Seagate IronWolf Pro
Slot 1 - Disk 1 – serial xxxPCQ
Slot 2 - Disk 2 – serial xxx3VB
Slot 3 - Disk 3 – serial xxxZMW
Slot 4 - Disk 4 – serial xxxZQS

Performed quick and long SMART tests on ALL drives.. all tested OK… Performed
IronWolf Health Management – ALL disks Healthy

August 26, 9am NAS reported Slot 3 - Disk 3 – Serial xxxZMW to be failed or degraded and detached. I turned off the NAS, took the drive off, put it back in, NAS continued to work OK

August 29, 1:11am NAS reported BOTH Slot 3 - Disk 3 – serial xxxZMW and Slot 4 - Disk 4 – serial xxxZQS as failed or detached AT THE SAME TIME (very very strange). I turned of the NAS as before, reattached the disks, NAS wouldn’t boot.

I took all the 4 disks out, installed Disk Genius as suggested by Terramaster in one of the support forums… Ran a complete Sector Scan of all the disks. All 4 disks tested OK, no bad blocks. This took a couple of days.

On September 2nd, thinking I lost all the data (I have backups), I re-inserted the disks in this order (swapped slots 1-2, 3-4 to 3-4, 1-2):

Slot 1 - Disk 3 – serial xxxZMW
Slot 2 - Disk 4 – serial xxxZQS
Slot 3 - Disk 1 – serial xxxPCQ
Slot 4 - Disk 2 – serial xxx3VB

After doing this, the NAS booted again, the Raid 5, previously inaccessible, reappeared with ALL the data. I then proceeded to update to TOS 6.0.461 and everything was working.

On September 4th (today), the NAS again reported at 2:01am that BOTH Slot 3 - Disk 1 – serial xxxPCQ
And Slot 4 - Disk 2 – serial xxx3VB as failed or detached AT THE SAME TIME, very strange AGAIN…

So this is where I am at today. I don’t think is the disks, as all 4 are new, tested all SMART tests OK, performed ALL blocks scan and reported no errors… I swapped the 4 disks and the problem persisted on Slots 3 and 4 with different disks. How can I proceed? Right now the NAS is turned on and responding. I don’t want to turn it off unless you need any additional info from it.

This is frustrating… Please help TERRAMASTER….

Based on your troubleshooting and detailed description, it is suspected that there may be a problem with the slots or the power supply.
Are you using the original power adapter? Is your power environment stable? Does the voltage meet the requirements of the power adapter?
Or you can try another power adapter with the same output voltage and interface.
When you are sure that it is not a problem with the power adapter, it should be a problem with the slots, please contact the after-sales service.

A plausible explanation at face value. However, the history of this type of issue (and there have been several) is that it only happens with IronWolf drives and TOS6. Factor that in. IMO there is an issue with how that particular combination interprets SMART values - particularly 'Command time out 188'.
{It may happen in later versions of TOS5 as well, but I have not followed that and have turned off 'Health Notifications' for my device using Tos5 and IronWolf drives.}

Just FYI - an interesting discussion on Seagate drives, SMART 188 Command Time Out and OS monitoring. It is a fairly long read, especially if you follow the links but still interesting. In the case of TNAS, for any references to "bad data cables, etc" think Backplane connectors (as earlier) of which there will generally be 2 (models vary); BP to motherboard and HDD to backplane.

https://forums.unraid.net/topic/43310-s ... unchanged/

TMzethar wrote: ↑05 Sep 2024, 11:05 Based on your troubleshooting and detailed description, it is suspected that there may be a problem with the slots or the power supply.
Are you using the original power adapter? Is your power environment stable? Does the voltage meet the requirements of the power adapter?
Or you can try another power adapter with the same output voltage and interface.
When you are sure that it is not a problem with the power adapter, it should be a problem with the slots, please contact the after-sales service.

Hi. I submitted earlier today the log report and everything else to the support email, getting the log directly with the tool provided by the NAS itself. The NAS is "brand new", the power supply is brand new, the UPS is maybe 6 months old and the only thing connected to it is the NAS. What do you suggest now?

Also, reading the post from @gremlin above... and working on that line, isn't there something you should fix/tweak in the next TOS beta release? perhaps a quick fix? Right now I am sitting on $650 USD worth of NEW TerraNAS, $1400 USD worth of NEW 4x18TB Seagate Ironwolf HDDs, and I can't do anything with them, as I had originally planned.

Thanks @TMzethar for your post, but I need to get a little more reassuring answer. It's very unlikely that two slots would come defective in a brand new piece of hardware... Please look at my logs I submitted to [email protected] through my email carlosdesedas AT gmail DOT com.

Best regards,

Carlos

Gremlin wrote: ↑05 Sep 2024, 16:24 A plausible explanation at face value. However, the history of this type of issue (and there have been several) is that it only happens with IronWolf drives and TOS6. Factor that in. IMO there is an issue with how that particular combination interprets SMART values - particularly 'Command time out 188'.
{It may happen in later versions of TOS5 as well, but I have not followed that and have turned off 'Health Notifications' for my device using Tos5 and IronWolf drives.}

Thank you Gremlin.. but it's strange that it only happens on slots 3-4. I swapped the drives and the problem persisted on those 2 slots. All 4 drives are 18TB Seagate Ironwolf drives.... As a matter of fact.. since that happened I have left the NAS on, I can access it.. it shows no volumes and no storage pools (obviously, as it's missing slots 3-4 to complete the riad 4), but the drives on the first two slots are still recognized and mark OK/Green/Healthy. One would think that eventually those two drives should be dropped too, but they haven't been.

What is the smart data telling you?

Just as an example, I have 4 Iron wolf x 4TB drives in Traid under Tos5.1. For the Smart 188 code these show the RAW values of

HDD1 - Slot 1 - 7
HDD2 - Slot 2 - 5
HDD3 - Slot 3 - 65537
respectively

None of these drives are reported as anything but healthy (including under the IRONWOLF health Management routine). None of my disk checking software on other machines have ever reported them as anything but healthy.
{Now, interestingly - for those who are suspicious and don't believe in coincidence - that last number is a Prime and also the fifth Fermat prime (F4)}.
Make of it what you will.

As it happens I have just completed a long SMART test {> 8 hours} on HDD3 and the numbers did not change.

Sorry, that should be 3 drives in the array. (The fourth is an unused spare)

Gremlin wrote: ↑06 Sep 2024, 05:12 What is the smart data telling you?

OK.. right now I have only slots 1-2 populated. slots 3-4 I have the drives physically detached...

The two connected show Connect Timeout value of 100