SuperMicro X7DAL-E with Samsung EVO 870 SSD

G

Geekazoid

Good Day! I am hoping that there is a sysctl to adjust. We are receiving an error since upgrading to FreeBSD 13.0. The issue did not appear when running 12.x.

We have been getting an error:
Code: 
ada0 at ahcich0 bus 0 scbus1 target 0 lun 0
ada0: <Samsung SSD 870 EVO 500GB SVT01B6Q> s/n ... detached
g_vfs_done():ada0s1d[READ(offset=186027343872, length=32768)]error = 6
(ada0:ahcich0:0:0:0):Periph destroyed
After that, everything that tries to write to the device fails with lots of errors. Essentially, the system is unusable until a hard reboot.

We are running the GENERIC kernel. The machine will run for several hours and then repeat this error.

Any advice is GREATLY appreciated! Also, if this is the wrong forum, my apologies. Please move the message. Otherwise, Thank You In Advance!!!

glenn
 
Did you check for the obvious culprits, like cables, power and such?
 
I would start with three things: check physical hard drive connection to avoid loose cables - check both signal and power, check disk with smartctl (from smartmontools) , in your case it would be smartctl -a /dev/ada0 , check if there are errors. Then boot into single mode and check disk with fsck.
 
Even if there are no errors in the smartctl output, check a few of the parameters for plausibility. For example: Is the temperature within a reasonable range (say 30 to 40)? How about power-on hours and total data written? Error rates?

If the disk is reachable via smartctl, then 3301's suggestion of fsck is good. Before that, you could attempt reading the disk, with dd if=/dev/adaX, of=/dev/null, bs=1048576, and see what you get. The read should proceed at about 1/2 GByte/second, so reading the whole disk (assuming it is 1TByte in size) should take about 1/2 hour, give or take a factor of two. This would be a good stress test to verify that the disk is healthy. Obviously, fsck is an even better test, but at a higher level, and it seems that you have a low-level hardware problem here.

And obvious, second the idea of checking cables and power first.
 
That's a very old board, the PCIe and SATA must very old gen, and the BIOS will be years out of maintenance. Possibly you're running into low level compatibility issues when trying to use it with newer peripherals ?
 
Thank You all for your responses! I greatly appreciate it!
SirDice - Cables have been checked (more than once actually.) :)
3301 - I have just run the fsck on each of the filesystems in single user mode. Each returned clean. When I cat /etc/fstab, the error popped up on the screen and the periph was destroyed.
ralphbsz - Temperatures are fantastic. I have air conditioning blowing on the open case. The drive is a 500GB Samsung. This is the 2nd SSD I have tried with the same results. Both were brand new, from different batches, when they were installed.
ajmcc - I figured this might actually be the problem. Alas, is there a way to confirm? Did the AHCI drivers change that much from 12.x to 13.x?

Any other thoughts? :-/

Again, Thank You All for your assistance!!! This one has been puzzling me deeply!!!

glenn
 
Geekazoid said:
Any other thoughts? :-/
Click to expand...
Install newest firmware for this SSD.
Check disk on a different computer or check on the same machine but with different SATA controller - I also have an old board (Tyan S7010) which has built-in SATA 2 but somewhere on this forum found advice on buying faster controller (I bought LSI SAS 9212-4i4e for decent money) and now it works like a charm.
 
Just to close things out on this issue. I do NOT believe that FreeBSD 13.x is compatible with the SATA controller on the SuperMicro X7DAL-E. The board is just too old!

3301, I did try 2 PCI ATA controllers I had laying around. Same results. :-/ Thank You for the idea though!!!

End result was to grab a different motherboard I had around. It is a SuperMicro X7DB3. This board is also old but newer than the X7DAL-E. Suffice to say, it has been up for 17 hours so far. :)

To that end, Thank You All for sharing ideas!!! I believe we are functional again around here! Have a Great Day!!!

glenn
 
Glad you've got a fix.
If you keep an eye on Ebay you should be to pick up a cheap X9 gen which gets you PCI & SATA 3.0.
 
You must log in or register to reply here.
Back
Top