PM863 FW File

[starshipeleven]

Hello I have some Samsung PM863A with model (shown by the samsung tool) MZ7LM960HMJP-00003.

I have tried upgrading the firmware in all the ways mentioned in this thread. With hdparm the foodesha method and firmware it completes correctly but firmware version does not change later so I don't know if this did anything.
With the three Samsung tools from @DarkServant it either says the drive is unsupported like with @bleomycin above or it fails to flash. All three detect the drive and see the same firmware version.

the drives have firmware 51F3Q as mentioned by @pimposh here https://forums.servethehome.com/index.php?threads/pm863-fw-file.18084/post-355889 so I suspect they come from a Netapp storage appliance.

Sadly even if I use my workplace account for NetApp I can't get to the firmware packages, it complains that I don't have the permission to do that even if I'm logged in.

The only thing I've been able to see are the two articles I attach here (with account name redacted), that talk about what drives were affected and what firmware versions are bad. Strangely, the firmware 51F3Q is not mentioned in these articles even if it's more recent than some of the bugged ones.

some excerpts from the articles:
-------
Samsung reported a firmware bug on PM863a drive firmware versions GXT5104Q, GXT5204Q, and GXT5003Q (non self-encrypted version) that can cause premature drive failures leading to data unavailability and potential data loss. This bug can result in concurrent drive failures in NetApp HCI and SolidFire AFA storage nodes.

-----
NetApp HCI or SolidFire AFA clusters with Samsung drives model PM863a with firmware version GXT5104Q, GXT5204Q, or GXT5003Q (the non self-encrypted or “non-SED” version) are exposed to two, potentially impactful bugs. The Samsung bugs can result in drive(s) prematurely failing out of the cluster and in rare instances lead to multiple drives failing in quick succession causing double-faults.

Impacted Models:

• PM863A: 480GB (non SED) MZ7LM480HMHQ-00003
• PM863A: 960GB (non SED) MZ7LM960HMJP-00003
• PM863A: 1.92TB (non SED) MZ7LM1T9HMJP-00003
• PM863A: 3.84TB (non SED) MZ7LM3T8HMLP-00003
• PM863A: 480GB (SED) MZ7LM480HMHQ-00005
• PM863A: 960GB (SED) MZ7LM960HMJP-00005
• PM863A: 1.92TB (SED) MZ7LM1T9HMJP-00005
• PM863A: 3.84TB (SED) MZ7LM3T8HMLP-00005

-------------------------------

1. Serial model number “MZ7LM” (example: scsi-SATA_SAMSUNG_MZ7LM48S2UJNX0) and Firmware Version “GXT5104Q” or “GXT5204Q”

-------------

• Error "The drive in slot X should be using firmware version GXT5404Q, but is using unsupported version GXT5104Q."

 

[crlt]

Hello I have some Samsung PM863A with model (shown by the samsung tool) MZ7LM960HMJP-00003.

I have tried upgrading the firmware in all the ways mentioned in this thread. With hdparm the foodesha method and firmware it completes correctly but firmware version does not change later so I don't know if this did anything.
With the three Samsung tools from @DarkServant it either says the drive is unsupported like with @bleomycin above or it fails to flash. All three detect the drive and see the same firmware version.

the drives have firmware 51F3Q as mentioned by @pimposh here https://forums.servethehome.com/index.php?threads/pm863-fw-file.18084/post-355889 so I suspect they come from a Netapp storage appliance.

Sadly even if I use my workplace account for NetApp I can't get to the firmware packages, it complains that I don't have the permission to do that even if I'm logged in.

The only thing I've been able to see are the two articles I attach here (with account name redacted), that talk about what drives were affected and what firmware versions are bad. Strangely, the firmware 51F3Q is not mentioned in these articles even if it's more recent than some of the bugged ones.

some excerpts from the articles:
-------
Samsung reported a firmware bug on PM863a drive firmware versions GXT5104Q, GXT5204Q, and GXT5003Q (non self-encrypted version) that can cause premature drive failures leading to data unavailability and potential data loss. This bug can result in concurrent drive failures in NetApp HCI and SolidFire AFA storage nodes.

-----
NetApp HCI or SolidFire AFA clusters with Samsung drives model PM863a with firmware version GXT5104Q, GXT5204Q, or GXT5003Q (the non self-encrypted or “non-SED” version) are exposed to two, potentially impactful bugs. The Samsung bugs can result in drive(s) prematurely failing out of the cluster and in rare instances lead to multiple drives failing in quick succession causing double-faults.

Impacted Models:

• PM863A: 480GB (non SED) MZ7LM480HMHQ-00003
• PM863A: 960GB (non SED) MZ7LM960HMJP-00003
• PM863A: 1.92TB (non SED) MZ7LM1T9HMJP-00003
• PM863A: 3.84TB (non SED) MZ7LM3T8HMLP-00003
• PM863A: 480GB (SED) MZ7LM480HMHQ-00005
• PM863A: 960GB (SED) MZ7LM960HMJP-00005
• PM863A: 1.92TB (SED) MZ7LM1T9HMJP-00005
• PM863A: 3.84TB (SED) MZ7LM3T8HMLP-00005

-------------------------------

1. Serial model number “MZ7LM” (example: scsi-SATA_SAMSUNG_MZ7LM48S2UJNX0) and Firmware Version “GXT5104Q” or “GXT5204Q”

-------------

• Error "The drive in slot X should be using firmware version GXT5404Q, but is using unsupported version GXT5104Q."

The -00003 is also used by Cisco and those drives that I had did not accept the original Samsung firmware. I have version MZ7LM1T9HMJP00003 and flashed from firmware GXT50F3Q to GXT51F3Q. I searched Cisco and google and could not find any later firmware, even in the later dated UCS packages it still ships with the 51F3.

They should work for all capacities. It seems to be important that the first four digits (hardware-revision/generation) and the sixth&seventh match (e.g. GXT5*04Q, HXM7*04Q, EDA5*02Q --> *=firmware-revision, 0 to F).

Based on this schema, GXT51F3Q is at the same update as GXT5103Q. So there is no worry and our drives are at the latest release.

 

[luckylinux]

Stock for PM863a attached. Flashed, tested on many drives. Works okay. Data is intact after upgrade, but always be on sunny side with backup.

That is only for the non-SED according to the Filename it would seem (-00003 Models).

What about those of us that got the SED Variant (-00005 Models) such as the MZ7LM1T9HMJP-00005 ?

I just won an Auction and the Item has FW GXT5204Q written on the Package. Maybe it has been updated since then (I will check via smartctl), but otherwise I would need an updated Firmware to flash on it.

Looking at https://forums.servethehome.com/index.php?threads/pm863-fw-file.18084/page-4 it's quite bad the BUG affecting ALL PM863A pretty much.

 

[rockwr123]

i have a two PM863 one has got GXT34Z3Q and the other one has GXT1003Q. All Samsung branded.

The GXT1003Q had ERRORMOD and i was able to recover it with a secure erase. but i cant seem to be able to upload any other firmware to any of the drives. DC tools doesn’t detect it at all.

hdparm flashes the file but doesn’t change anything.

i did the " Or it's still the "Get Dell Firmware & Trim a few Bytes then flash using hdparm" Procedure that needs to be carried out ? " – still no change in the firmware.

I used the GA3B.fwh, then trimmed it down. anyone else had any luck?

 

[luckylinux]

Thanks, I can confirm, that it works on PM863a 1.92tb with firmware GXT5204Q and reports correctly as GXT5404Q after fw upgrade.

Did you flash a -00003 (non-SED) or a -00005 (SED) Version with that File though ? I got a MZ7LM1T9HMJP-00005 (SED) and the Firmware seems to be only for non--SED according on the Filename.

 

[rockwr123]

Which firmware do you have currently?

Guess you can leave it as it is and wait till I can update mine haha maybe see if dctool can even see your drive, then there might be a chance.

I even tried the PM863a posted earlier on this thread, still no chnage to the firmware.

Tbh i cant even figure out which is the latest firmware for the PM863.

I wouldn't use it as a main drive without a backup or raid setup, mine just died without notice, luckily i had mine a in a raid 5.

At least they can be reverted from the errormod status.

 

[luckylinux]

Hi,
i've uploaded some new firmwares for the pm893, pm897, pm1653. Updated the ssdmanager tool too (DFDB_Basic.enc).
I think now the samsung V-NAND v6 enterprise SATA drives are probably quite stable, or use the Intel/Solidigm alternatives D3-S4520/S4620 if you can get them. Samsung eSSD Firmware Repository

Hope it helps

Any hope of having also a PM863 (non-A) Version ?

Or it's still the "Get Dell Firmware & Trim a few Bytes then flash using hdparm" Procedure that needs to be carried out ?

That way seems a bit hit or miss, as Smartctl might now show Firmware Change, Drive gets "renamed" or some Statistics get out of whack (cannot remember what/where I saw that, just went through the entire Thread for the 3rd Time).

 

[pimposh]

Which firmware do you have currently?

Guess you can leave it as it is and wait till I can update mine haha maybe see if dctool can even see your drive, then there might be a chance.

I even tried the PM863a posted earlier on this thread, still no chnage to the firmware.

Tbh i cant even figure out which is the latest firmware for the PM863.

I wouldn't use it as a main drive without a backup or raid setup, mine just died without notice, luckily i had mine a in a raid 5.

At least they can be reverted from the errormod status.

Still have one PM863 1.92T (non-A) running in some system with GXT3003Q so at least that one exists.

 

[rockwr123]

Still have one PM863 1.92T (non-A) running in some system with GXT3003Q so at least that one exists.

Did that firmware come with the drive or did you update it? Im trying to update mine to GXT3003Q

 

[luckylinux]

i have a two PM863 one has got GXT34Z3Q and the other one has GXT1003Q. All Samsung branded.

The GXT1003Q had ERRORMOD and i was able to recover it with a secure erase. but i cant seem to be able to upload any other firmware to any of the drives. DC tools doesn’t detect it at all.

hdparm flashes the file but doesn’t change anything.

i did the " Or it's still the "Get Dell Firmware & Trim a few Bytes then flash using hdparm" Procedure that needs to be carried out ? " – still no change in the firmware.

I used the GA3B.fwh, then trimmed it down. anyone else had any luck?

I was waiting for some Feedback after my Post, and I didn't try anything in the Meantime. Kinda feeling stuck between risking breaking if attempting upgrade or risking some BUG resulting in Drive dying early ...

 

[pimposh]

Was shipped with it. Never seen anything later in vanilla Samsung flavour for non-A.

 

[DarkServant]

I never got smart out of the F/W numbering of the pm863... some have an GXT1##3Q while others the GXT3##3Q, these should indicate different hardware revisions. I have to give up on these firmware's i think.

The only one i now of, is that there is an GXM1203Q for the sm863 somewhere out in the net.
I still have an sm863 480gb, and done nothing, it sits since years unpowered on a dusty shelf, still GXM1003Q.
The older ones are pretty hard to get, while with the newer ones it's way easier to get them somewhere.


The flash of these V-NAND drives is very good and durable, especially the v1 24-layer (40K P/E-cycles?) and v2 32-layer ones.
IF there would be an genius open-source PCB with a low-power 8+ channel multicore RISC-V MCU/Controller, a top-notch PMIC, E-Fuse, discrete firmware NOR-flash, a configurable PLP-circuit and an open-source Firmware for it, where everybody can solder their own NAND-Packages down.

 

[luckylinux]

Which firmware do you have currently?

Guess you can leave it as it is and wait till I can update mine haha maybe see if dctool can even see your drive, then there might be a chance.

I even tried the PM863a posted earlier on this thread, still no chnage to the firmware.

Tbh i cant even figure out which is the latest firmware for the PM863.

I wouldn't use it as a main drive without a backup or raid setup, mine just died without notice, luckily i had mine a in a raid 5.

At least they can be reverted from the errormod status.

I didn't check it yet, I had installed Proxmox on top of a Mirror (RAID-1) between this Samsung PM863 and a Micron ECO 5200.

 

[DarkServant]

Seems like Samsung Enterprise Drives are a real Hell with lack of Firmware Support to end Customers .

I bought:

a) Samsung PM863 1.92TB - MZ7LM1T9HCJM-00003 - Sticker on the Drive says Firmware GXT3003Q

b) Samsung PM863a 1.92TB - MZ7LM1T9HMJP-00005 - Sticker on the Drive says Firmware GXT5204Q

No I don't know the Version ACTUALLY on the Drives right now as I'm still setting up the System they were intended to.

Yepp, not even the minimum of "support"... Firmware updates! F...planetfirst Samsung!

I got pissed off and created the Repository of Firmwares i and others collected over the years.

Via smartctl in Linux or CrystalDiskInfo or GSmartControl in Windows, you get some essential information about drives (power-on hours, wear leveling count, and more severe SMART values like uncorrectable error count, erase fail count, and so on. With GSmartControl you can even start an drive-self-test on SATA drives.
For a) you have to get one from Dell, trim some bytes off and stuff it on the drive via hdparm...
For b) there is one actual (GXT5404Q)

 

[rockwr123]

My drives are PM863
MZ7LM1T9HCJM-00005 - GXT1003Q
MZ7LM1T9HCJM-000FB - GXT34Z3Q

I have another 4 GXT1003Q but i dont want to use them till i get a firmware without the errormod issue.

Does anyone know if this error was ever fixed in any of the firmwares.

 

[rockwr123]

To be honest I don't even know the Firmware Release Chain/Order. The Naming makes not much sense to met at least.

Like ... what is the Ancient, Oldest, Older, New, Newer, Newest, etc ?

It does NOT seem like it's a Case of "Bigger is better" ...

And I could NOT find ANYTHING about my GXT33P3Q Firmware at all. Like, it doesn't appear on Google even !

100% it doent make sense.

The label dates on the drive

2016:
MZ7LM1T9HCJM-00005 - GXT1003Q

2015:
MZ7LM1T9HCJM-000FB - GXT34Z3Q

 

[acc]

Hi alll, very interesting thread, my searches on trying to fix my drives keep coming back here but so far i am unable to get out of errormod.

I have a lot of the above disks from a HPE Nimble storage box and slowly but surely they are all dying, the Samsung SSD doctor reports them in errormod and attempts the update the firmware with the files from Cloud GMX does not work, also i am unable to run the erase command from the same program.

Can anyone tell me the firmware i need, and the tool needed please? I worry more disks will fail and I dont have money to purchase replacements yet.
Thank you

 

[blunden]

View attachment 44210

Hi alll, very interesting thread, my searches on trying to fix my drives keep coming back here but so far i am unable to get out of errormod.

I have a lot of the above disks from a HPE Nimble storage box and slowly but surely they are all dying, the Samsung SSD doctor reports them in errormod and attempts the update the firmware with the files from Cloud GMX does not work, also i am unable to run the erase command from the same program.

Can anyone tell me the firmware i need, and the tool needed please? I worry more disks will fail and I dont have money to purchase replacements yet.
Thank you

I think I used the GXT5404Q file from the Cloud GMX link. My situation was a little different though since I intentionally glitched the hardware (connecting the enable pin on one of the chips to ground using a wire while it was running) to force the drive into ERRORMOD in order to bypass the unknown ATA maximum security password set.

I then used hdparm in Linux to secure erase and update the firmware.

Secure erase:

sudo hdparm --security-erase NULL /dev/[sdX]

Update firmware:

sudo hdparm --fwdownload [firmware.bin] --yes-i-know-what-i-am-doing --please-destroy-my-drive /dev/[sdX]

EDIT: Updated with the secure erase command as well.

 

[luckylinux]

I never got smart out of the F/W numbering of the pm863... some have an GXT1##3Q while others the GXT3##3Q, these should indicate different hardware revisions. I have to give up on these firmware's i think.

The only one i now of, is that there is an GXM1203Q for the sm863 somewhere out in the net.
I still have an sm863 480gb, and done nothing, it sits since years unpowered on a dusty shelf, still GXM1003Q.
The older ones are pretty hard to get, while with the newer ones it's way easier to get them somewhere.


The flash of these V-NAND drives is very good and durable, especially the v1 24-layer (40K P/E-cycles?) and v2 32-layer ones.
IF there would be an genius open-source PCB with a low-power 8+ channel multicore RISC-V MCU/Controller, a top-notch PMIC, E-Fuse, discrete firmware NOR-flash, a configurable PLP-circuit and an open-source Firmware for it, where everybody can solder their own NAND-Packages down.

Seems like Samsung Enterprise Drives are a real Hell with lack of Firmware Support to end Customers .

I bought:

a) Samsung PM863 1.92TB - MZ7LM1T9HCJM-00003 - Sticker on the Drive says Firmware GXT3003Q

b) Samsung PM863a 1.92TB - MZ7LM1T9HMJP-00005 - Sticker on the Drive says Firmware GXT5204Q

No I don't know the Version ACTUALLY on the Drives right now as I'm still setting up the System they were intended to.

 

[acc]

I think I used the GXT5404Q file from the Cloud GMX link. My situation was a little different though since I intentionally glitched the hardware (connecting the enable pin on one of the chips to ground using a wire while it was running) to force the drive into ERRORMOD in order to bypass the unknown ATA maximum security password set.

I then used hdparm in Linux to secure erase and update the firmware.

sudo hdparm --fwdownload firmware.bin --yes-i-know-what-i-am-doing --please-destroy-my-drive /dev/[sdX]

Great, I am on it now, hopefully I'll have more success in Linux (hopefully a live version too, as my computer is a NUC with just the single SATA)

 

[blunden]

Great, I am on it now, hopefully I'll have more success in Linux (hopefully a live version too, as my computer is a NUC with just the single SATA)

Yes, live booting a Linux distro of your choice works fine for this purpose.

EDIT: I double checked the checksum of that file and can confirm that the file on the GMX download site is the same as the one I used.

Below is the thread about my little adventure, specifically the successful process I used:

https://forum.hddguru.com/viewtopic.php?p=305555#p305555

 

[luckylinux]

I got pissed off and created the Repository of Firmwares i and others collected over the years.

I know, that's why I asked you if you had the PM863 (non-A) ... well also the PM863a now that I bought those as well - Firmware Versions, since you had collected most of what you could find (but maybe you weren't using the "old" PM863 Drives, I don't know).

 

[acc]

Yes, live booting a Linux distro of your choice works fine for this purpose.

EDIT: I double checked the checksum of that file and can confirm that the file on the GMX download site is the same as the one I used.

Below is the thread about my little adventure, specifically the successful process I used:

https://forum.hddguru.com/viewtopic.php?p=305555#p305555

Wow, it worked! looks like you helped me bring 2 SSD's back to life!! I owe you (serious) let me know if I can send you a couple ££'s!

I found it didnt take right away,originally it resulted in an error SG_IO bad/missing sense data, but reading the thread you linked mentioned the --security-erase NULL and after running that the disks popped up at full capacity on the fly. I did still run the f/w update though as something really screwy for 2 out of my 5 disks to go down like that. Now I wonder if I should do the others which are running fine atm, I think so after doing a little data backup.

 

[luckylinux]

@DarkServant: I just checked, two Samsung PM863 (MZ7LM1T9HCJM-00003) are running Firmware Version GXT33P3Q

I cannot find any Information on it online though

 

[blunden]

Wow, it worked! looks like you helped me bring 2 SSD's back to life!! I owe you (serious) let me know if I can send you a couple ££'s!

I found it didnt take right away,originally it resulted in an error SG_IO bad/missing sense data, but reading the thread you linked mentioned the --security-erase NULL and after running that the disks popped up at full capacity on the fly. I did still run the f/w update though as something really screwy for 2 out of my 5 disks to go down like that. Now I wonder if I should do the others which are running fine atm, I think so after doing a little data backup.

Awesome! The true heroes are the ones in that thread.

Yeah, I got several of those "SG_IO bad/missing sense data" errors too, at least when I used a USB to SATA adapter.

I decided to update the firmware on my second PM863a too, just in case. Btw. it's worth knowing that some of the SMART data will have been wiped from the drive once the drive entered the ERRORMOD state. Other than that, I haven't noticed any negative effects.

 

[luckylinux]

Would Samsung GA3B for model number(s) MZ7LM120HCFD00D3, MZ7LM480HCHP00D3, MZ7LM960HCHP00D3, MZ7LM1T9HCJM00D3, MZ7LM3T8HCJM00D3, MZ8LM120HCFD00D3, MZ8LM480HCHP00D3 and MZ8LM960HCJH00D3. | Driver Details | Dell Singapore be the latest Dell Driver for Samsung PM863 MZ7LM1T9HCJM-00003 ? The Firmware Page mentions Dell Part Number MZ7LM1T9HCJM00D3, NOT MZ7LM1T9HCJM-00003 ...

 

[pimposh]

SM=!PM.
Why of GC5B while if GC55 and they both are different models?

 

[pimposh]

What are you saying exactly ????

For the moment i thought that these cmdlines are yours and wondered why things are so mixed up.

Dell trimmed versions indeed do flash but do not change existing version visible via SMART.
Then How one knows if it works at all (flashing fw) ?

Assumption is it does since there are numerous reports of getting drives out of errormod (yet version does not change).

 

[luckylinux]

My drives are PM863
MZ7LM1T9HCJM-00005 - GXT1003Q
MZ7LM1T9HCJM-000FB - GXT34Z3Q

I have another 4 GXT1003Q but i dont want to use them till i get a firmware without the errormod issue.

Does anyone know if this error was ever fixed in any of the firmwares.

To be honest I don't even know the Firmware Release Chain/Order. The Naming makes not much sense to met at least.

Like ... what is the Ancient, Oldest, Older, New, Newer, Newest, etc ?

It does NOT seem like it's a Case of "Bigger is better" ...

And I could NOT find ANYTHING about my GXT33P3Q Firmware at all. Like, it doesn't appear on Google even !

 

[blunden]

As I said, I also tried the GMX File for PM863a, but it did NOT work for me unfortunately . Probably somebody mixed up PM863 and PM863a in that Folder, I don't know.

I successfully flashed two PM863a drives with the unbranded GXT5404Q firmware and their firmware numbers updated correctly (after a power cycle I believe). I mentioned that earlier in this thread.