On 1/18/24 23:23, gene heskett wrote:
On 1/19/24 00:55, David Christensen wrote:
I am unclear if those errors are inside the SSD or if they are the
SATA communications link between the SSD and the motherbaord or HBA
port and/or main memory (?). Does dmesg(1) show anything?
I'm not sure what I should be looking for, and I don't see anything that
is looping to correct an error. Suggested grep targets?
Here is a dmesg(1) excerpt from 2014 -- Debian 7, good SSD, bad SATA cable:
[ 2.086360] ata3.00: ATA-9: INTEL SSDSC2CW060A3, 400i, max UDMA/133
[ 2.086365] ata3.00: 117231408 sectors, multi 16: LBA48 NCQ (depth
31/32), AA
[ 2.096265] ata3.00: configured for UDMA/133
[ 14.718054] EXT4-fs (dm-0): mounted filesystem with ordered data
mode. Opts: (null)
[ 18.449227] EXT4-fs (sda1): mounted filesystem with ordered data
mode. Opts: (null)
[ 20.157693] ata3.00: exception Emask 0x10 SAct 0x400000 SErr 0xc00001
action 0x6 frozen
[ 20.157699] ata3.00: irq_stat 0x08000000, interface fatal error
[ 20.157703] ata3: SError: { RecovData Handshk LinkSeq }
[ 20.157709] ata3.00: failed command: WRITE FPDMA QUEUED
[ 20.157716] ata3.00: cmd 61/08:b0:a0:e0:61/00:00:00:00:00/40 tag 22
ncq 4096 out
[ 20.157721] ata3.00: status: { DRDY }
[ 20.157727] ata3: hard resetting link
[ 20.473489] ata3: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 20.484835] ata3.00: configured for UDMA/133
[ 20.484847] ata3: EH complete
[ 21.059825] ata3.00: exception Emask 0x10 SAct 0x4000 SErr 0x400100
action 0x6 frozen
[ 21.059831] ata3.00: irq_stat 0x08000000, interface fatal error
[ 21.059835] ata3: SError: { UnrecovData Handshk }
[ 21.059840] ata3.00: failed command: WRITE FPDMA QUEUED
[ 21.059848] ata3.00: cmd 61/08:70:50:e2:61/00:00:00:00:00/40 tag 14
ncq 4096 out
[ 21.059853] ata3.00: status: { DRDY }
[ 21.059859] ata3: hard resetting link
[ 21.376135] ata3: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 21.397234] ata3.00: configured for UDMA/133
[ 21.397246] ata3: EH complete
[ 22.590805] ata3.00: exception Emask 0x10 SAct 0x600 SErr 0x400100
action 0x6 frozen
[ 22.590811] ata3.00: irq_stat 0x08000000, interface fatal error
[ 22.590815] ata3: SError: { UnrecovData Handshk }
[ 22.590819] ata3.00: failed command: WRITE FPDMA QUEUED
[ 22.590826] ata3.00: cmd 61/08:48:f0:ee:1d/00:00:00:00:00/40 tag 9
ncq 4096 out
[ 22.590831] ata3.00: status: { DRDY }
[ 22.590834] ata3.00: failed command: WRITE FPDMA QUEUED
[ 22.590840] ata3.00: cmd 61/08:50:70:ef:1d/00:00:00:00:00/40 tag 10
ncq 4096 out
[ 22.590844] ata3.00: status: { DRDY }
[ 22.590851] ata3: hard resetting link
[ 22.909955] ata3: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 22.921525] ata3.00: configured for UDMA/133
[ 22.937878] ata3: EH complete
[ 22.938635] ata3: limiting SATA link speed to 3.0 Gbps
[ 22.938638] ata3.00: exception Emask 0x10 SAct 0x400000 SErr 0x400100
action 0x6 frozen
[ 22.938640] ata3.00: irq_stat 0x08000000, interface fatal error
[ 22.938642] ata3: SError: { UnrecovData Handshk }
[ 22.938645] ata3.00: failed command: WRITE FPDMA QUEUED
[ 22.938648] ata3.00: cmd 61/60:b0:20:28:66/00:00:00:00:00/40 tag 22
ncq 49152 out
[ 22.938650] ata3.00: status: { DRDY }
[ 22.938652] ata3: hard resetting link
[ 23.257418] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
[ 23.269251] ata3.00: configured for UDMA/133
[ 23.285387] ata3: EH complete
In any case, make sure that you are using SATA III 6 Gbps cables with
locking connectors for your drives and that all the connections are good.
That's hard to verify once the cables are removed from the packing. all
are black, with locking clips There is a cable maker under every tree
in china so I'n not swearing any are up to specs, I've had cable problem
in the past but usually a magenta colored on that is over 2 years old,
If you have a known good src on straight on cables, please share. You
would be doing everyone a favor.
https://www.cablematters.com/pc-187-156-3-pack-straight-60-gbps-sata-iii-cable.aspx
https://www.cablematters.com/pc-188-156-cable-matters-3-pack-90-degree-right-angle-60-gbps-sata-iii-cable-18-inches.aspx
Test what you have by taking a wooden stick and moving each one a
centimeter or so, if the log blows up with sata resets, bingo, bad
cable. replace it asap.
I call that the "wiggle" test.
David