Samsung SSD PM1643 3.8TB SAS 12Gb/s MZ-ILT3T8A 2.5" @ $289.98

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
A

atreyu79

New Member
Feb 17, 2023
5
2
3
www.workdangit.com
Smart says:

Revision: ESFA

Two of the drive labels say:

F/W VFT50Z2Q

So I'm not really sure, TBH.

Update to my earlier post... after formatting to 512, I created a new GPT table and a new partition and now I'm dd'ing the drives to wipe them. I also did a fsck for good measure... which didn't report any problems. Watching dmesg, I haven't seen the errors on the drives for about 3 hours now.
 
Last edited:
  • Like
Reactions: gb00s and Fritz
Y

yukas

Member
Jun 3, 2022
34
2
8
Hi, i have several PM1643 and PM1633a fails drives with Hardware Error
FW EQHD
ESV8
ESFA

all drives in

for any command in SG -reply this:
mode sense(10):
Fixed format, current; Sense key: Hardware Error
Additional sense: Logical unit failure
Raw sense data (in hex), sb_len=64, embedded_len=18
70 00 04 00 00 00 00 0a 00 00 00 00 3e 01 00 00
00 00
MODE SENSE (10) command: Medium or hardware error, type: sense key (plus blank check for tape)

are not flashed with either array, IT HBA mode or SSD Manager - no way.
How to remove the lock, is there a factory key or a universal algorithm?
 
G

geonap

Member
Mar 2, 2021
76
75
18
so i'm rather confused.

i have about 15 MZ-ILS3T80 (PM 1633 1632 REV 0) ish drives.

Model: P1633N38 EMC3840
Revision: EQPC

unfortunately i can't read this one in my possession but i also bought some other drives that are showing up as:
(1633a)
Product: PA33N3T8 EMC3840
Revision: EQL8
@ 33834 hours currently

how does one actually know if the ssd's are affected?

thanks
 
Last edited:
T

TonyArrr

Active Member
Sep 22, 2021
141
75
28
Straylia
geonap said:
so i'm rather confused.

i have about 15 MZ-ILS3T80 (PM 1633 1632 REV 0) ish drives.

Model: P1633N38 EMC3840
Revision: EQPC

unfortunately i can't read this one in my possession but i also bought some other drives that are showing up as:
(1633a)
Product: PA33N3T8 EMC3840
Revision: EQL8
@ 33834 hours currently

how does one actually know if the ssd's are affected?

thanks
Click to expand...
did you ever find an answer to this? A lot of MZ-ILS3T80 drives are popping up at good prices around the net and I’m wondering if they are a safe buy or if I’d be poisoning my arrays…
 
Y

yukas

Member
Jun 3, 2022
34
2
8
I've repaired a few dozen of these drives - they have a lot of problems not only with the smart time counter, but also with the "crooked" handling of SAS 6GBit controllers.
In general you can only protect yourself with newer firmware versions.
 
M

mrpasc

Well-Known Member
Jan 8, 2022
579
320
63
Munich, Germany
Do you have by any chance some never firmware for those EMC ones (Model: P1633N38 EMC3840
Revision: EQPC) for share?
Would be very appreciated.
 
K

keekles

New Member
May 5, 2023
6
0
1
I have 4 of these EMC branded disks. I'm looking at how to confirm the firmware is up to date and set the sector size to 4096. I've tried the sg n
format command under linux with failure. I have the setblocksize program running on it now and will let it format until tomorrow and see if that work. The SSDManager program for linux segfaults as well.

Code: 
=== START OF INFORMATION SECTION ===
Vendor:               SAMSUNG
Product:              P1633N38 EMC3840
Revision:             EQPC
Compliance:           SPC-4
User Capacity:        3,840,774,504,448 bytes [3.84 TB]
Logical block size:   512 bytes
Physical block size:  4096 bytes
LU is resource provisioned, LBPRZ=1
Rotation Rate:        Solid State Device
Form Factor:          2.5 inches
Logical Unit id:      0x5002538a7640a170
Serial number:        NZNCAH400934
Device type:          disk
Transport protocol:   SAS (SPL-3)
Local Time is:        Sat May  6 05:15:29 2023 UTC
SMART support is:     Available - device has SMART capability.
SMART support is:     Enabled
Temperature Warning:  Disabled or Not Supported

=== START OF READ SMART DATA SECTION ===
SMART Health Status: OK

Percentage used endurance indicator: 3%
Current Drive Temperature:     37 C
Drive Trip Temperature:        60 C

Manufactured in week 13 of year 2016
Accumulated start-stop cycles:  98
Specified load-unload count over device lifetime:  0
Accumulated load-unload cycles:  0
Elements in grown defect list: 0

Error counter log:
           Errors Corrected by           Total   Correction     Gigabytes    Total
               ECC          rereads/    errors   algorithm      processed    uncorrected
           fast | delayed   rewrites  corrected  invocations   [10^9 bytes]  errors
read:          0        0         0         0          0    2149689.551           0
write:         0        0         0         0          0     949348.927           0
verify:        0        0         0         0          0     264089.006           0

Non-medium error count:       39

SMART Self-test log
Num  Test              Status                 segment  LifeTime  LBA_first_err [SK ASC ASQ]
     Description                              number   (hours)
# 1  Background short  Completed                   -   51664                 - [-   -    -]
# 2  Reserved(3)       Completed                   -   51357                 - [-   -    -]
# 3  Reserved(3)       Completed                   -   51357                 - [-   -    -]
# 4  Reserved(3)       Completed                   -   51357                 - [-   -    -]
# 5  Reserved(3)       Completed                   -   51357                 - [-   -    -]
# 6  Reserved(3)       Completed                   -   51356                 - [-   -    -]
# 7  Reserved(3)       Completed                   -   51352                 - [-   -    -]
# 8  Reserved(3)       Completed                   -   51350                 - [-   -    -]
# 9  Reserved(3)       Completed                   -   51347                 - [-   -    -]
#10  Reserved(3)       Completed                   -   51345                 - [-   -    -]
#11  Reserved(3)       Completed                   -   51343                 - [-   -    -]
#12  Reserved(3)       Completed                   -   51341                 - [-   -    -]
#13  Reserved(3)       Completed                   -   51339                 - [-   -    -]
#14  Reserved(3)       Completed                   -   51337                 - [-   -    -]
#15  Reserved(3)       Completed                   -   51335                 - [-   -    -]
#16  Reserved(3)       Completed                   -   51333                 - [-   -    -]
#17  Reserved(3)       Completed                   -   51330                 - [-   -    -]
#18  Reserved(3)       Completed                   -   51328                 - [-   -    -]
#19  Reserved(3)       Completed                   -   51326                 - [-   -    -]
#20  Reserved(3)       Completed                   -   51324                 - [-   -    -]

Long (extended) Self Test duration: 90 seconds [1.5 minutes]
 
A

atreyu79

New Member
Feb 17, 2023
5
2
3
www.workdangit.com
Not the exact same drive, but I found the most stability with mine by leaving it at 512 logical/4096 physical. It's a home server... suits my use case just fine. So if you're having issues converting it, I'd say just leave it. For some reason, when I converted mine to 4096 all hell broke loose. Switched it back, been solid since. I think others have had this experience as well.

=== START OF INFORMATION SECTION ===
Vendor: SAMSUNG
Product: P043S3T8 EMC3840
Revision: ESFA
Compliance: SPC-5
User Capacity: 3,840,774,504,448 bytes [3.84 TB]
Logical block size: 512 bytes
Physical block size: 4096 bytes
LU is resource provisioned, LBPRZ=1
Rotation Rate: Solid State Device
Form Factor: 2.5 inches
Logical Unit id: 0x5002538b08b52d10
Serial number: ZWNE0KB02378
Device type: disk
Transport protocol: SAS (SPL-3)
Local Time is: Sat May 6 11:14:28 2023 MST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled
Temperature Warning: Enabled
 
K

keekles

New Member
May 5, 2023
6
0
1
atreyu79 said:
For some reason, when I converted mine to 4096 all hell broke loose.
Click to expand...
Two questions to this:

1. What did you use to convert it to 4096?
2. What do you mean by all hell broke loose?

My plan is to use these as ZFS special devices in a 3x mirror for my 261tb spinning rust pool.

In testing on my staging computer, I'm seeing some poor throughput on the SSD, like 100 mbytes/s of write, well below the spinning disks, but this could be the setup. I'm going to my datacenter now to do testing on one my "production" boxes.
 
K

keekles

New Member
May 5, 2023
6
0
1
atreyu79 said:
It's all written out a couple pages back. See here: https://forums.servethehome.com/ind...2gb-s-mz-ilt3t8a-2-5-289-98.32725/post-368042
Click to expand...
Yes, that's a slightly different part number/firmware so didn't think it was applicable.

I did some testing with mine, a bit of a let down as the direct write performance was only about 40% better than the spinning disks and 40% less than the older 1.6 TB SSD's I had as spares. I'm wondering if this is a firmware issue. I was surprised to see the older SSD's perform so much better in raw reading.

Did you do any performance testing?

Code: 
#testing on thunderpussy.keekles.org: 6g sas :-(
#Broadcom / LSI SAS2308 PCI-Express Fusion-MPT SAS-2

#Testing on the Samsung P1633N38 EMC3840 FW EQPC 3.84 tb with 512b sectors:

# read
# dd of=/dev/null if=/dev/sdi bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 116.356 s, 461 MB/s

root@ThunderPussy:~# dd of=/dev/null if=/dev/sdi bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 116.227 s, 462 MB/s

# write
# dd if=/dev/zero of=/dev/sdi bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 186.933 s, 287 MB/s

# dd if=/dev/zero of=/dev/sdi bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 184.313 s, 291 MB/s

#Testing on the Samsung P1633N38 EMC3840 FW EQPC 3.84 tb with 4k sectors:

#Read
# dd of=/dev/null if=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 118.784 s, 452 MB/s

# dd of=/dev/null if=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 118.342 s, 454 MB/s

#write
# dd if=/dev/zero of=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 183.659 s, 292 MB/s

# dd if=/dev/zero of=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 184.987 s, 290 MB/s

#Testing on the HGST 1.6tb disks HUSMM1616ASS204   FW C350 512b sectors
Read:
# dd of=/dev/null if=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 100.917 s, 532 MB/s

# dd of=/dev/null if=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 100.95 s, 532 MB/s

Write:
# dd if=/dev/zero of=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 112.101 s, 479 MB/s

# dd if=/dev/zero of=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 113.285 s, 474 MB/s

#Testing on the  SEAGATE  ST1200MM0088     N004 10.5k RPM disks
Read:
# dd of=/dev/null if=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 251.903 s, 213 MB/s

# dd of=/dev/null if=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 252.365 s, 213 MB/s

Write:
# dd if=/dev/zero of=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 262.665 s, 204 MB/s

# dd if=/dev/zero of=/dev/sdh bs=1M count=51200
53687091200 bytes (54 GB, 50 GiB) copied, 263.432 s, 204 MB/s
 
W

Whaaat

Active Member
Jan 31, 2020
358
187
43
keekles said:
I did some testing with mine, a bit of a let down as the direct write performance was only about 40% better than the spinning disks
Click to expand...
Did you enable write caching on this ssd? They usually arrive from ebay sellers with write cache disabled because of controller policy those ssds were used with
 
K

keekles

New Member
May 5, 2023
6
0
1
Whaaat said:
Did you enable write caching on this ssd? They usually arrive from ebay sellers with write cache disabled because of controller policy those ssds were used with
Click to expand...
I didn't even think of checking that. How do I check that?
 
K

keekles

New Member
May 5, 2023
6
0
1
Whaaat said:
sdparm --get=WCE /dev/sdi
Click to expand...
It looks like the Samsung have the write cache enabled. This would explain the speed I see at first (1000 mbyte/s+) and then drops off over the 50g file.

None of the other disks have it enabled, assuming 1 here is enabled.

Code: 
# sdparm --get=WCE /dev/sdl
    /dev/sdl: SAMSUNG   P1633N38 EMC3840  EQPC
WCE           1  [cha: y, def:  0, sav:  1]

sdparm --get=WCE /dev/sdd
    /dev/sdd: HGST      HUSMM141CLAR1600  C350
WCE           0  [cha: y, def:  0, sav:  0]

sdparm --get=WCE /dev/sdh
    /dev/sdh: SEAGATE   ST1200MM0088      N003
WCE           0  [cha: y, def:  1, sav:  0]
So it would appear these are just much slower, no?
 
K

keekles

New Member
May 5, 2023
6
0
1
mochouinard said:
Anyone figure out a good firmware for it ?
The firmware EQPC should be in this file... BUT if gpg encrypted !!! (Thanks Dell!!!)
Click to expand...
Lovely. At least I can confirm the disks I have have the newest firmware on them. (EQPC)

I'm at a loss as to why they would give firmware out encrypted. To decrypt it, their updater software must have the private key in it. It makes no sense.

Are you able to find the updater application that process the gpg file?
 
W

Whaaat

Active Member
Jan 31, 2020
358
187
43
keekles said:
I'm at a loss as to why they would give firmware out encrypted. To decrypt it, their updater software must have the private key in it. It makes no sense.
Click to expand...
HP and Cisco also do it. For instance, you can download a whole bunch of firmware updates for HP drives in one file, but surprise, they are encrypted, and there is also tiny 10kb executable file in the folder, which will encrypt all the .fw files in its location over again once being started.
 
N

nepeat

New Member
Jul 9, 2020
2
1
3
keekles said:
Lovely. At least I can confirm the disks I have have the newest firmware on them. (EQPC)

I'm at a loss as to why they would give firmware out encrypted. To decrypt it, their updater software must have the private key in it. It makes no sense.

Are you able to find the updater application that process the gpg file?
Click to expand...
I bought 8x of these and found out they each all have 51.6K hours on them. That thankfully confirms they are patched for the 32k bug. GBs written to my batch are not bad with them ranging from 979106 - 1177170.920 for the drives I have, with a single outlier of 23644.558 GB.


Code: 
root@sea420-hv-2:~# grep Revis smart
Revision:             EQPC
Revision:             EQPC
Revision:             EQPC
Revision:             EQPC
Revision:             EQPC
Revision:             EQPC
Revision:             EQPC
Revision:             EQPC
root@sea420-hv-2:~# grep hours: smart
Accumulated power on time, hours:minutes 51673:54
Accumulated power on time, hours:minutes 51688:02
Accumulated power on time, hours:minutes 51682:13
Accumulated power on time, hours:minutes 51619:21
Accumulated power on time, hours:minutes 51633:19
Accumulated power on time, hours:minutes 51596:45
Accumulated power on time, hours:minutes 51687:52
Accumulated power on time, hours:minutes 51713:10
root@sea420-hv-2:~# grep write: smart
write:         0        0         0         0          0    1055649.361           0
write:         0        0         0         0          0     974407.712           0
write:         0        0         0         0          0    1063623.852           0
write:         0        0         0         0          0     979106.454           0
write:         0        0         0         0          0      23644.558           0
write:         0        0         0         0          0    1177170.920           0
write:         0        0         0         0          0    1035296.232           0
write:         0        0         0         0          0    1009782.962           0
 
You must log in or register to reply here.
Top Bottom