Hi all,

 

In the last week I managed to successfully swap out my old cache disk with a new Kingston SA400S37 (960GB) drive connected to a 6 port ASM1166 PCIe card (port 1).

 

Today I plugged in an old 2TB disk in preparation to pre-clear/check it. In doing so the SSD needed to be unplugged (both power and data) and then re-plugged in. After booting Unraid, I was alerted to SMART errors on the SSD, and sure enough the drive was showing 100,000+ 199 SATA CRC error count (which is referred to 199 UDMA CRC error count for all other drives in the system).

 

Understanding that this SMART metric tends to point to phsycial cabling issues, I shut the array down and checked/re-seated the cable at the SDD and controller end and powered back on. After a few minutes I was seeing the CRC values steadily increase. At this point I decided to swap out the SATA cable for a known good one, however the problem still persisted. Next I tried another port on the ASM1166 card (port 2) with another new cable, same issue.

 

I have now run through the cache replacement procedure and put the original cache disk back in using the original port 1 on the ASM1166 card and the original SATA cable and re-tested... the old cache drive has no UDMA CRC error counts!

 

So my question - is this new SSD bad, or is there some compatibility issues between the SSD and ASM1166 chipset?

 

I've attached diagnostics, however the 'faulty' SSD is no longer in the array.

 

Sample output from checking the SMART values when copying data to the Kingston SSD:

 

Linux 6.1.79-Unraid.
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
262395
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196859
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196859
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196859
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
131322
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
131322
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
131320
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
131318
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
131318
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
131318
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
131318
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196858
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196858
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196858
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196858
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196858
root@Pyxis:~#  smartctl -a /dev/sdg | grep "199 SATA" | awk '{print $10}'
196859

pyxis-diagnostics-20241202-1407.zip

you may need to run a syslinux grub boot comnad due to some nvme bug

unraid web UI > Main ? flash > syslinux options:

 

nvme_core.default_ps_max_latency_us=5500 default_hugepagesz=1G hugepagesz=1G transparent_hugepage=always pci=noaer pcie_aspm=off

*Add the above to your syslinx boot at the end of bzroot
 

Thank you bmartino1 for the info, my cache disk is a standard 2.5" SATA3 SSD, rather than NVMe, do you still think the above would be relevant?

20 hours ago, teakmeister said:

or is there some compatibility issues between the SSD and ASM1166 chipset?

Could be, can you connect it to a different controller and retest?

 

1 hour ago, teakmeister said:

do you still think the above would be relevant?

It's not.

1 hour ago, JorgeB said:

Could be, can you connect it to a different controller and retest?

 

It's not.

 

Thank you JorgeB. Good point, the one thing I didn't try yesterday as the only free ports are SATA-2 on the motherboard, I'll shuffle a few drives round and give it a re-test. Since yesterday I've had the SSD in a USB3 caddy and the CRC errors have been jumping around there too... not convienced of the drive, but will re-test in the unraid server.

 

Values whilst in the caddy (RAW and decimal):

 

40104 - 262,404 @ 16:59
30104 - 196,868 @ 18:50
20104 - 131,332 @ 22:17
40104 - 262,404 @ 10:00
50104 - 327,940 @ 11:20

I've popped the Kingston SSD into another unraid device (as the original one is busy pre-clearing a bunch of disks) using the onboard SATA ports and seeing the same oddities with the SMART count. I've copied some files to it and polled the SMART info during the copies:

 

root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131325
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131318
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131318
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131316
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131316
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131315
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131315
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
196852
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
196852
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
196852
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
196852
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131316
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131316
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131316
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131315
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131315
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131315
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131310
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131310
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131310
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131309
root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
131309

 

Now correct me if I'm wrong, I didn't think that the CRC errors logged in SMART actually went down, unlike some of the other values which can if the issue is corrected.

 

Value 2 mins after the last copy (my mind boggles...):

 

root@Tucana:~# smartctl -a /dev/sdf | grep "199 SATA" | awk '{print $10}'
196845

Edited December 3, 20241 yr by teakmeister

12 minutes ago, teakmeister said:

I didn't think that the CRC errors logged in SMART actually went down

Nope, they should only go up, possibly a device issue.

5 minutes ago, JorgeB said:

Nope, they should only go up, possibly a device issue.

 

This was my thought too, will send the device back, thank you for your input.