Hello, everyone!  I've just built myself an unRAID server and am still learning a lot.  Have begun transferring my data into the array, but stopped when I came across this error.  So far, all my content has been written to Disk 1 due to the allocation method (high-water) and split level configuration (auto split top 3) I'm using.

 

Yesterday, I noticed the dashboard is showing a SMART error for Disk 2 in the array.  I've run a SMART short self-test that passed and am now running an extended self-test which has been hanging at 30% for a good portion of the day.  SMART log included below.  Apologies if the formatting is off a bit.

 

In the Attributes tab of Disk 2, I also have a UDMA CRC error count with a value and worst that equals 100 and Raw Value of 1.  Should I be concerned about any of this?

 

ATA Error Count: 1

• CR = Command Register [HEX]
• FR = Features Register [HEX]
• SC = Sector Count Register [HEX]
• SN = Sector Number Register [HEX]
• CL = Cylinder Low Register [HEX]
• CH = Cylinder High Register [HEX]
• DH = Device/Head Register [HEX]
• DC = Device Command Register [HEX]
• ER = Error register [HEX]
• ST = Status register [HEX]

Powered_Up_Time is measured from power on, and printed as DDd+hh:mm:SS.sss where DD=days, hh=hours, mm=minutes, SS=sec, and sss=millisec. It "wraps" after 49.710 days.

 

Error 1 occurred at disk power-on lifetime: 30 hours (1 days + 6 hours)

When the command that caused the error occurred, the device was active or idle.

 

After command completion occurred, registers were:

ER ST SC SN CL CH DH

-- -- -- -- -- -- --

84 43 00 00 00 00 00 Error: ICRC, ABRT at LBA = 0x00000000 = 0

 

Commands leading to the command that caused the error were:

CR FR SC SN CL CH DH DC Powered_Up_Time Command/Feature_Name

-- -- -- -- -- -- -- --                      ----------------              --------------------

60 40 50 d0 4b da 40 08    20:32:46.579        READ FPDMA QUEUED

60 40 60 90 51 da 40 08    20:32:46.573        READ FPDMA QUEUED

60 80 58 10 51 da 40 08     20:32:46.564       READ FPDMA QUEUED

60 40 48 90 46 da 40 08    20:32:46.564       READ FPDMA QUEUED

60 40 40 50 41 da 40 08    20:32:46.564        READ FPDMA QUEUED

Attach Diagnostics to your NEXT post in this thread. 

Diagnostics attached.  SMART test has progressed to 40% so I'm just going to leave it be.

sol-diagnostics-20250127-1631.zip

Edited January 27, 20251 yr by mergahtrern
Updating SMART scan status.

From others on the forums UDMA CRC errors can be caused by a loose SATA cable.  You can try re-seating the cable at both ends, if that does not make a difference try a different cable.  Did you run a preclear on the 3 WD22 Reds before adding them to the array?  If you didn't I would try a preclear (clear and erase disk) after your long SMART test completes.  You will need to take drive 2 out of your array to run the preclear on that drive.  Since you just started and if you don't mind copying your data again I would run a preclear on all your drives.  It takes time but it is worth it to stress test the drives before adding them to the array.  I had one of my drives show a single smart error once.  The drive corrected and reset itself after I ran a preclear on it, not sure if you will be that lucky.  In my case I think it was caused by a network issue during a file copy.  I see timeout errors in your log for your network share, that might have been caused by the network or the drive, no way to know for sure.  I hope this helps.

Edited January 28, 20251 yr by SShadow

For the timeouts, I had an issue with my current NAS which was resolved by rebooting it.  The transfer I had going in unRAID was stuck because of this, but automatically continued once the NAS booted back up and became available on the network again.

 

Only 4 TB of 24 TB copied so far so I don't mind copying data again.  Once the SMART test is done (50% now), I'll first reseat the SATA cable on both ends, then remove the drives from the array and preclear them.  Can they all be precleared at once?  I do have some extra SATA cables lying around so swapping out to another won't be a problem if issues persist.

Under the Tools tab there is a Preclear Disk option.  You can click on the blue play box icon (the first option in the row of icons on the upper right side) to start multiple sessions at the same time.  I tried to run multiple preclears on some old drives last year and one of them froze during the preclear.  I also ran my server normally as these were old drives not part of the system, which might have had something to do with it.  I ended up just doing them one at a time without any issues.  I have read about others running multiple preclears on drives without a problem.  I would give it a try running all three at the same time since the system won't be doing anything else. 

So I ran preclear on all three disks and it completed successfully.  Also swapped out two of the three SATA cables that each threw 1 CRC error with a pair that were being used in another system.

 

This past weekend, I began a 20 TB transfer from my NAS via SMB which started out strong, but speeds have slowly decreased to 9 MBps.  At this rate, the transfer won't complete for another two weeks or more. 🙃 Progress is at 24% so I wouldn't be opposed to cancelling the transfer, figuring out why this is happening, then kick it off again as doing all that would probably let me get this migration of data complete in less than 2 weeks.

 

Another diagnostics is attached for review.  I am noticing in the Main tab that the disk currently being written to (Disk 1) has equal read and write rates with the parity disk showing the same rate of writes and Disk 2 showing equal rates of read.  Parity sync has been paused since the array came online this weekend until a couple minutes ago after I just cancelled it.  This doesn't appear to have had an impact on the read/write stats I shared above.  My plan was to wait until everything was moved over to the array, then run a parity sync.

 

FWIW, all other transfers I ran into the array before this one ran anywhere between 50MBps - 90MBps+, but the largest one was just 4 TB and the others were far smaller.

sol-diagnostics-20250203-1839.zip

The slowest thing disks do is seeking, moving the read/write heads to position them for reading/writing.

 

Building parity while writing to the array is going to involve a lot of seeking.

 

Best plan would be to unassign parity until initial data load is complete, then build parity after.

That would require I stop the array which would consequently stop/cancel the current transfer, correct?

yes

Cool.  Cruising at 95-101 MBps now.  Will check it out in the AM.

 

Thanks for the quick replies, @trurl.

I'm back down to 20 - 25 MBps this afternoon.  Should I copy stuff over in smaller batches to avoid this?

sol-diagnostics-20250204-1609.zip

Many things affect the transfer speed.  Small files really bog things down for one example.  I also seem to remember that 22TB WD Reds are SMR disks rather than the faster CMR 20TB drives. 

Edited February 4, 20251 yr by Frank1940
Bad memory...

It's probably best for me to exercise some patience and just let it roll.  Speeds bumped back up to 36-45 MBps.  Another TB to go before unRAID should start writing content to the second disk in the array.  Will be interesting to see what happens then.

To put a bow on this, I ended up copying everything over in batches.  I noticed transfer speeds suffered after X number of items were being copied at once.  Everything is now in my array and parity sync has completed.

@mergahtrern good to hear.  Hope you are enjoying Unraid.  There are some SMB network tweaks I had to do on my system to get better network performance.  In the SMB Extras section of the SMB Settings page and after some trial and error I added:

 

[global]
server min protocol = SMB3_11  
client ipc min protocol = SMB3_11
aio read size = 1
aio write size = 1

 

The aio read/write settings gave me similar performance to a Windows PC to Windows PC transfer.  I also enabled SMB Multi-Channel but the aio read/write I think made the biggest difference.  I run my server with bonded NICs running LACP.  If you are running anything MAC on your network that needs to be SMB shared to the Unraid server don't add the server and client min protocol lines.  Those lines just force the connection protocol to the highest version.  Not sure if it will help you with your system and hardware setup but might be worth some testing.  Good luck.  

Edited February 11, 20251 yr by SShadow

It is awesome so far, @SShadow.  Thinking of what else to do with it besides Plex and Minecraft, haha!

 

Am even considering a second Unraid server build for off site backup/storage so I can re-use these WD Reds from my now-retired Drobo 5N and some other hardware I have.  That or throw them in the current server as a pool and backup internally.

@mergahtrern it does not take long to want a 2nd server 😁.  It took me two years of running one server to build a smaller 2nd server.  The small server is mainly for running pfSense in a VM with a pass-through Ethernet card, 2 Ubuntu server VMs, a 2nd Pi-Hole instance on each of my VLANs (I have the main Pi-Hole instance for each VLAN on my main server), and a few other docker containers.  On my main server I made a second pfSense VM as a backup so if I need to work on the small server I would only take the internet down for 30 seconds to swap my WAN cable over to the other server.  Makes server maintenance much easier and it is nice to have redundancy.  Not exactly high availability for pfSense but close enough for me for my setup.  I did exactly what you are thinking, I reused some drives I had pulled out of an old HTPC and I internally backup everything from the original array drives to those drives.  It did come in handy a few months ago to have the 2nd backup.  I must have made a mistake testing one of my rsync scripts and had some missing files on the main drives of the array.  I was able to easily recover the missing files from the 2nd backup drives.  Sounds overkill but you can never have too many backups.