Intel 540s periodic refresh blocks broken?

[andrey_v1]

ssd have old partitions and old files around 4-5 years old on them. These partitions have enough free space (spare blocks, never touched). I tried ddrescue on these partitions&files, the reading's speed was very slow compared to another fresh partitions, 33mb/sec average with often drops to 0. Some files had badblocks around some tens.

Smart values:

E9 - 100%

F9 - 3.5Tb on 1tb ssd

Spare blocks - 100-300Gb were never used.

My questions:

1) Why didn't the firmware refresh degraded blocks on such huge period?

2) What does mean huge smart BB counter with thousands value after reading old partition? I have badblocks, but not so much. Are the data still consistent with exception known badblocks?

3) Why didn't the firmware refresh degraded blocks after 5+times readings in row?

4) Is it warranty case?

[JosafathB_Intel]

Hello, andrey_v1.

Thank you for contacting Intel® Memory and Storage support.

As we understand, you need assistance with your Intel® SSD 540s Series. If we inferred correctly, we would appreciate it if you provide us with the following information:

1- We would appreciate it if you can provide us with the system details by extracting the SSU logs.

A- Go to https://downloadcenter.intel.com/download/25293/ (for Windows*) and to https://downloadcenter.intel.com/download/26735/ (for Linux) download the software.

B- When you finish downloading it, open it.

C- Attach the file generated by the tool in your reply.

2- We would like to check the SMART attributes of your Intel® SSDs and review them to see if they show any error, using this system or another system that recognize the drives:

*Please extract the following logs by using the Intel® Memory and Storage Tool:

https://downloadcenter.intel.com/download/29337/Intel-Memory-and-Storage-Tool-CLI-Command-Line-Inter...

- SMART attributes: intelmas show -smart -intelssd “SSD_index”.

-The SSD information using show -all command from Intel MAS CLI: intelmas show -all -intelssd “SSD_index”.

A complete User Guide is available below. Replace the “SSD_index” with the actual index number of your SSD.

User Guide:

https://downloadmirror.intel.com/30059/eng/Intel_Memory_And_Storage_Tool_CLI_User%20Guide-Public-342...

Video on Intel® SSD Firmware Update and Health Monitoring Using Intel® Memory and Storage Tool (Intel® MAS)

https://www.intel.com/content/www/us/en/support/articles/000057393/memory-and-storage.html

Based on the Intel® Memory and Storage Tool CLI Release Notes, page 14, in the Intel® SSD 540s Revision History section:

https://downloadmirror.intel.com/30556/eng/Intel-MAS-CLI-1.9-Release-Notes-342335-011US.pdf

Intel® SSD 540s Revision History

Date Revision Description July 2020 043C/017C

This firmware has the below fixes:

• Optimizations for NAND data refresh algorithms

This firmware update improved the data refresh algorithms.

We will be looking forward to your reply, including the logs requested.

We will follow up with you again on Friday, the 16th of July 2021, if we do not receive an answer from you. Please, let us know if you would like to schedule a different date for the follow-up.

Best regards.

Josh B.

Intel® Customer Support Technician.

[andrey_v1]

Spoiler

INTEL SSDSC2KW010X6 

ID Attribute Description Threshold Value Worst Data Status
05 Reallocated Sector Count 0 100 100 839 OK: Always passes
09 Power-On Hours Count 0 100 100 501 OK: Always passes
0C Power Cycle Count 0 100 100 4326 OK: Always passes
AA Available Reserved Space 10 98 98 0 OK: Value is normal
AB Program Fail Count 10 100 100 0 OK: Value is normal
AC Erase Fail Count 10 100 100 0 OK: Value is normal
AE Unexpected Power Loss 0 100 100 246 OK: Always passes
B7 SATA Downshift Count 0 100 100 1 OK: Always passes
B8 End-to-End Error Detection Count 90 100 100 0 OK: Value is normal
BB Uncorrectable Error Count 0 100 100 20310 OK: Always passes
BE Airflow Temperature 0 36 52 28, 52, 36 OK: Always passes
C0 Unsafe Shutdown Count 0 100 100 246 OK: Always passes
C7 CRC Error Count 0 100 100 59 OK: Always passes
E1 Host Writes 0 100 100 5.45 TB OK: Always passes
E2 Timed Workload Media Wear 0 100 100 0 OK: Always passes
E3 Timed Workload Host Read/Write Ratio 0 100 100 0 OK: Always passes
E4 Timed Workload Timer 0 100 100 0 OK: Always passes
E8 Available Reserved Space 10 94 94 0 OK: Value is normal
E9 Media Wearout Indicator 0 99 99 0 OK: Always passes
F1 Total LBAs Written 0 100 100 5.45 TB OK: Always passes
F2 Total LBAs Read 0 100 100 23.29 TB OK: Always passes
F9 Total NAND Writes 0 100 100 3645 GB OK: Always passes
FC <vendor-specific> 0 100 100 6 OK: Always passes

The statement about firmware is a lie. It contains OPTIMIZATION, not BUG FIXED.

My questions still the same:

Spoiler

1) Why didn't the firmware refresh degraded blocks on such huge period?

2) What does mean huge smart BB counter with thousands value after reading old partition? I have badblocks, but not so much. Are the data still consistent with exception known badblocks?

3) Why didn't the firmware refresh degraded blocks after 5+times readings in row?

4) Is it warranty case?

But actually they are the statements. My position:

You have an old device model(s) with undiscovered Refresh Period bug due to firmware update descriptions. My host was powered everyday, with some breaks, to do enough reprogramming stuff in background. Firmware algo even didn't try to refresh blocks after slow read.

Also, i  noticed, that 09 counter contains 500hrs, but should contain  10000-16000hrs. Your firmware even have overflow bug on simple counter. I guess on 16384 value.

I am asking you:

0) To invite proper engineer or team

2) To investigate the problem with refresh period and poweron counter.

( It's interesting case, this device (and all others loyd stars) is under expiring warranty period and has such a longplay bug. And the old information on it is close to death. And i don't understand how to use it in future without autorefreshing blocks (or trust into working autorefreshing ) Probably, it's not warranty case. But i've  bought it  because of 5years warranty period and now wanna to get a good customer service.  )

3) to extend warranty period for a problem device with low usage PE blocks (or better solution).

[andrey_v1]

1

[andrey_v1]

Why was my reply deleted?