Re: SATA SSD keeps disconnecting

From: David Christensen <dpchrist_at_holgerdanske.com>
Date: Sat, 26 Feb 2022 20:09:02 UTC 
On 2/26/22 09:34, Andrea Venturoli wrote:
> 
> Hello.
> 
> I've got a desktop system with two SSD drives (gmirror swap + ZFS 
> mirrored pool) and one of them keeps disconnecting.
> 
> Smart says the drive is good and I've already tried replacing the cables.


I collected many red SATA I, II, and III cables over the years.  Only a 
few were marked as to their speed.  I ran into a lot of problems with 
these cables; and may have incorrectly blamed and replaced a few drive 
racks.  Finally, I replaced all of my cables with new Cable Matters 
locking black marked 6 Gbps cables, and the problems went away.


> Relevant dmesg:
>> FreeBSD 12.3-RELEASE-p1 releng/12.3-n234208-c1494385203 AAAAA amd64


Good.


>> FreeBSD clang version 10.0.1 (git@github.com:llvm/llvm-project.git 
>> llvmorg-10.0.1-0-gef32c611aa2)
>> VT(vga): resolution 640x480
>> CPU: Intel(R) Core(TM) i5-3470 CPU @ 3.20GHz (3200.09-MHz K8-class CPU)
>>   Origin="GenuineIntel"  Id=0x306a9  Family=0x6  Model=0x3a  Stepping=9


Good.


>> ahci0: <Intel Panther Point AHCI SATA controller> port 


Good.


>> ada0 at ahcich0 bus 0 scbus0 target 0 lun 0
>> ada0: <PNY CS900 120GB SSD CS900613> ACS-4 ATA SATA 3.x device

>> ada1: <PNY CS900 120GB SSD CS900613> ACS-4 ATA SATA 3.x device


I had a PNY video card with a "lifetime" warranty back in the day.  It 
died after ~2 years.  I contacted PNY.  They define "lifetime" as 
"lifetime of the product", which means "until they stop making it".  The 
card was out of production and had warranty.  I have not bought PNY 
products since.


I prefer Intel enterprise desktop drives:

ada0: <INTEL SSDSC2CW060A3 400i> ACS-2 ATA SATA 3.x device


> And this is what I get in the logs:
>> Feb 26 12:01:01  kernel: ada0 at ahcich0 bus 0 scbus0 target 0 lun 0
>> Feb 26 12:01:01  kernel: ada0: <PNY CS900 120GB SSD CS900613> s/n 
>> PNY14200238640200E14 detached
>> Feb 26 12:01:01  kernel: GEOM_MIRROR: Device swap: provider ada0p2 
>> disconnected.
>> Feb 26 12:01:01  kernel: (ada0:ahcich0:0:0:0): Periph destroyed
>> Feb 26 12:01:01  kernel: ada0 at ahcich0 bus 0 scbus0 target 0 lun 0
>> Feb 26 12:01:01  kernel: ada0: <PNY CS900 120GB SSD CS900613> ACS-4 
>> ATA SATA 3.x device
>> Feb 26 12:01:01  kernel: ada0: Serial Number PNY14200238640200E14
>> Feb 26 12:01:01  kernel: ada0: 600.000MB/s transfers (SATA 3.x, UDMA6, 
>> PIO 8192bytes)
>> Feb 26 12:01:01  kernel: ada0: Command Queueing enabled
>> Feb 26 12:01:01  kernel: ada0: 114473MB (234441648 512 byte sectors)
>> Feb 26 12:40:56  kernel: ada0 at ahcich0 bus 0 scbus0 target 0 lun 0
>> Feb 26 12:40:56  kernel: ada0: <PNY CS900 120GB SSD CS900613> s/n 
>> PNY14200238640200E14 detached


Yup.


> Anything I can try on the software side?
> Some more logging?
> Some sysctl?
> Other tweaks?


If you are confident in the cables (and/or racks), replace the drive(s).


David