Solved RT-AX88U Is USB NVME SSD Supported?

[Tech9]

if you connect them to a PC?

@El Mariachi ... and on the SATA interface, not with the same USB external case to USB port. There is a chance the drives are fine.

 

[El Mariachi]

After these SSDs have been "killed" what happens if you connect them to a PC? Can you see/format them there or are they completely undetected?

If I let them go on too long, when I connect them to a PC, they are detected but not readable as in Ext4 format. When I attempt to format them, the formatting job usually fails.

If I pull them from active use at first signs of I/O write errors occuring, I can typically format successfully. However, after use in the router I do see the same block ID's receiving write errors. As its the same block ID's it leads me down the thought process of permanent damage to the drives.

With everyones feedback so far, it's sounding like a voltage issue at this point possibly caused by the enclosure.

Your router may be getting low RAM condition and corrupting the drive. If you regularly do intensive transfers you better get a NAS.

I do get to high RAM but I have a 10gig swap. I also dont use the router as a media server/file server. All writes to my file server are done directly. The router is only providing internet gateway and routing services.

You are better off buying a NAS, you could use Link Aggregation with the AX88U.
View attachment 51098

I dont see how a NAS would fix my issues as I dont use the attached USB for media serving. I only use it for running merlin and script logging. I only have a Fibre 500 connection so a single gigabit link is more than sufficient for my needs.

 

[Tech9]

I do get to high RAM but I have a 10gig swap

This swap on USB won't help.

The router is only providing internet gateway and routing services.

Do you mean the high load transfers are NOT to this USB attached drive?

 

[sfx2000]

With everyones feedback so far, it's sounding like a voltage issue at this point possibly caused by the enclosure.

I would also consider the USB cable as well...

 

[El Mariachi]

This swap on USB won't help.



Do you mean the high load transfers are NOT to this USB attached drive?

Yep. The USB attached SSD is only running entware/scripts and writing to script log files on the drive. Would a series of low size high volume writes be capable of causing a too high a power draw? I feel this is pretty standard usage. I have nothing custom running. Large downloads are done via a separate device writing to its own drives. The router is simply routing web traffic to it. The router would just log what connections were made/blocked to its USB attached SSD (which is the one that is constantly failing).

@Tech9 on a basic Marlin implementation have you experienced high RAM usage? On a fresh reboot I start at 65% usage. Over the course of a few days it rises to 97%

I would also consider the USB cable as well...

USB cable is the one that came with the enclosure. I'll try a fresh enclosure/cable combination with dedicated mains power and format the WD red drive and see how I go. Have you had any experience using Low Level Disk Format Tool? Is it snake oil or a legitimate way to clear bad blocks?


EDIT: Drive is dead. Formatting fails emmediately...

 

[EmeraldDeer]

I have been using MiniTool Partition Wizard to work on my Entware external SSD from a Windows PC. The counter argument to this approach is that you should use the native tools on the router to avoid compatibility issues.

A low level format tool would likely be provided by the disk vendor. It would be unlikely that such a tool would function with a USB enclosure. You would need to put the drive in a 2280 slot.

 

[ColinTaylor]

I have been using MiniTool Partition Wizard to work on my Entware external SSD from a Windows PC. The counter argument to this approach is that you should use the native tools on the router to avoid compatibility issues.

The problem I found with MiniTool and other Windows utilities was that it was incredibly slow when creating ext* filesystem. The filesystem would also frequently be incompatible with my router. It was just a lot fast to format it on the router.

A low level format tool would likely be provided by the disk vendor. It would be unlikely that such a tool would function with a USB enclosure. You would need to put the drive in a 2280 slot.

I don't believe it's been possible to do true low level formatting since the MFM days. What the generic utilities seem to do is just overwrite the logical blocks of the raw device. You can achieve the same thing using dd (which is what amtm uses). These utilities are only really needed because of Windows' nasty habit of refusing to format drives that contain partition layouts or filesystems that it doesn't understand. This is where things like MiniTool can help.

 

[sfx2000]

If the drives are actually dead, the enclosure is my first suspect, not the router.

Everything keeps pointing towards the enclosure...

Could it be the USB-Sata bridge firmware perhaps, or the chipset itself - there were issues a long time back with certain controllers from JMicron (the infamous JM20337 controller which was very commonly used back in the day)

 

[El Mariachi]

The problem I found with MiniTool and other Windows utilities was that it was incredibly slow when creating ext* filesystem. The filesystem would also frequently be incompatible with my router. It was just a lot fast to format it on the router.


I don't believe it's been possible to do true low level formatting since the MFM days. What the generic utilities seem to do is just overwrite the logical blocks of the raw device. You can achieve the same thing using dd (which is what amtm uses). These utilities are only really needed because of Windows' nasty habit of refusing to format drives that contain partition layouts or filesystems that it doesn't understand. This is where things like MiniTool can help.

So rather than using the amtm fd command try dd directly from console and pick some custom options? Im not super familiar with unix/linux. What would be an optimal set of arguments for the dd command with an ext4 journaled drive config?

Everything keeps pointing towards the enclosure...

Could it be the USB-Sata bridge firmware perhaps, or the chipset itself - there were issues a long time back with certain controllers from JMicron (the infamous JM20337 controller which was very commonly used back in the day)

I agree @sfx2000. I'll get the drive replaced and get a new enclosure. There is a silverstone and a cooler master I am eyeing off that use a realtek chipset. If I move away from a WD RED SATA NAS m.2 drive, would you recommend a Samsung 980 Pro 500gb or something there abouts? I care about high endurance (set, forget and never die) rather than outright speed. The router wont use that level of speed anyway for its use case.

 

[ColinTaylor]

So rather than using the amtm fd command try dd directly from console and pick some custom options? Im not super familiar with unix/linux. What would be an optimal set of arguments for the dd command with an ext4 journaled drive config?

No, you're missing my point. Just use amtm fd.

 

[El Mariachi]

No, you're missing my point. Just use amtm fd.

If you are just trying to say don't use windows to format a filesystem it has issues working with, I get it. Thought there may have been a new tip/trick to learn in your suggestion. I'm always open to growing my knowledge.

I've temporarily reverted to an old external disc drive. So far no write errors or growing numbers of bad blocks in the last 24 hrs. I'll give it another 24 hrs but early indications are that it is the enclosure which seems to be the culprit. Unlucky me but glad to finally have it diagnosed.

Thank you all on this thread for your help!

Admins is there a way to update the thread title with the SOLVED prefix?

 

[ColinTaylor]

Admins is there a way to update the thread title with the SOLVED prefix?

You can edit your first post and add the SOLVED prefix to the title.