RT-86U - USB 2.0 port not mounting at boot.

[miazza]

https://pastebin.com/40BPA5Q0

I don't know if this is useful. I have taken it now after abot 20h from the issue.

Now PEN DRIVE is dead and cannot be anymore formatted by any OS.

 

[ColinTaylor]

Unfortunately the log doesn't go back far enough to show us some of the more useful information.

You have two devices plugged in. Can you describe each one (pen drive, ssd, hdd) and how it is formatted?

 

[miazza]

I was afraid the log was not usefull ...

I have a 16Gb pen drive attached in the USB 3 port that shoud be NTFS (I'm not sure otherwise it's FAT32 and not linux).
I use it for SMB share of temporary documents and video.

In the USB 2 port I have installed three different pen drive (all killed); they all were generic pen drive 4Gb linux formatted in single partition ext4.

The feeling is that I have started to have this killing effect after having installed the Diversion Status script plug-in.
Before that my pen drives were experiencing a very safe good and long life

Now I have no pen drive in USB 2 port and the rooter is quite calm, almost no CPU usage and about 100 Mb free RAM.
With entware installed and Diversion + Div-status the CPU usage was very variable and often high with only 30 Mb free RAM (even with swap file).

 

[ColinTaylor]

The device in USB 3 was FAT32.

Can you trying plugging the faulty device back into the USB 2 port. If it shows as mounted in the GUI then unmount it first. Then try the following commands (I don't know whether they exist on you router) and post the output.

umount /dev/sdb1

hdparm -r /dev/sdb
hdparm -r /dev/sdb1

hdparm -r0 /dev/sdb
hdparm -r0 /dev/sdb1

e2fsck -p /dev/sdb1

 

[miazza]

The faulty key never mounts.

ASUSWRT-Merlin RT-AC86U 384.17_0 Sun Apr 26 02:28:53 UTC 2020
miazza@RT-AC86U-3150:/tmp/home/root# umount /dev/sdb1
umount: can't unmount /dev/sdb1: Invalid argument
miazza@RT-AC86U-3150:/tmp/home/root# hdparm -r /dev/sdb
-sh: hdparm: not found
miazza@RT-AC86U-3150:/tmp/home/root# hdparm -r /dev/sdb1
-sh: hdparm: not found
miazza@RT-AC86U-3150:/tmp/home/root# hdparm -r0 /dev/sdb
-sh: hdparm: not found
miazza@RT-AC86U-3150:/tmp/home/root# hdparm -r0 /dev/sdb1
-sh: hdparm: not found
miazza@RT-AC86U-3150:/tmp/home/root# e2fsck -p /dev/sdb1
Entware: Superblock has an invalid journal (inode 8).
CLEARED.
*** ext3 journal has been deleted - filesystem is now ext2 only ***

Entware: Resize inode not valid.

Entware: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
        (i.e., without -a or -p options)
miazza@RT-AC86U-3150:/tmp/home/root#

 

[ColinTaylor]

Run

e2fsck -f /dev/sdb1

 

[john9527]

Also try as a double check

fdisk -l /dev/sdb

 

[miazza]

ASUSWRT-Merlin RT-AC86U 384.17_0 Sun Apr 26 02:28:53 UTC 2020
miazza@RT-AC86U-3150:/tmp/home/root# e2fsck -f /dev/sdb1
e2fsck 1.42.13 (17-May-2015)
Superblock has an invalid journal (inode 8).
Clear<y>?

 

[ColinTaylor]

Press "y".

 

[miazza]

https://pastebin.com/9LrHi2hH

 

[ColinTaylor]

Now do this:

e2fsck -fy /dev/sdb1

 

[miazza]

https://pastebin.com/qy4fNw5x

 

[ColinTaylor]

Well I'm at a loss really.

We can see from the fsck outputs that even though it looks like it's repairing the filesystem it isn't really because the device is still read-only. If you look in the router's syslog I'd expect to see lots of write errors.

Is there anything common about the USB flash drives that have this problem? Are they all from the same manufacturer? I know that SanDisk devices have this "feature" where they go into read-only mode if it thinks there's a problem with the drive.

 

[miazza]

Haha I arrived at the same loss after having tried any recovery tool .
I even tried to delete the partition manually but every tool used says it's ro.

I really don't know if there is something common about all the failed USB drive. They look different but for the router they are all Generic Drive.
All drive are 4Gb and all are presenet from suppliers

The other drives have different behaviour: I can still format them and use them in FAT32 or NTFS. May be not reliable but they can be used for some file exchange. If I try to use them in the router they fail after very short time.

Am I the only one having this issue ? Indeed all have failed shortly after having installed Diversion Status script.

 

[miazza]

After having invested some time on the broken device, i found the drive has a lot of bad damaged block. Such a quantity of bad block has triggered a Read Only protection that cannot be reverted.
As a matter of facts I cannot even delete the partition and I cannot do anything.

On the net I found people who experienced this kind of issues due to the high numbers of R/W cycles and this I believe it's the case also considering the SWAP file.

Now I have to say that the EXT 4 choice for a normal pend drive might not be a good choice: every time you write to a EXT 4 drive, the journal must be updated which causes more writing and life reduction.
This is not critical for SSD or for M2.

I wonder if the formatting script shall really recommend the Ext4 partion for whatever drive we use or if the Ext2 is more appropriate for the Entware and SWAP file use.

 

[ColinTaylor]

After having invested some time on the broken device, i found the drive has a lot of bad damaged block. Such a quantity of bad block has triggered a Read Only protection that cannot be reverted.

This was what I was refering to in my previous post. Although I've only ever seen it in branded flash drives and not cheap generic ones. The generic ones just keep on going regardless of the bad blocks in my experiance.

Now I have to say that the EXT 4 choice for a normal pend drive might not be a good choice: every time you write to a EXT 4 drive, the journal must be updated which causes more writing and life reduction.

This should not be the case. Since kernel 2.6 writes to the swapfile bypasses the filesystem and journal and writes directly to disk. Link

I wonder if the formatting script shall really recommend the Ext4 partition for whatever drive we use or if the Ext2 is more appropriate for the Entware and SWAP file use.

I would still definitely recommend a journaled filesystem even when using a swap file. Having the occasional flash drive failure is par for the course, but the rate you're getting through them makes me think this is a problem unique to your setup. Otherwise why isn't this forum flooded with "me too" posts? Rather than being a software problem perhaps there is an electrical issue with your router's USB socket, or it's getting too hot....

If the problem is caused by the volume of write operations then I'd say that either you're using exceptionally poor quality flash drives (my router's flash drive has been in use for 6 years) or your router is thrashing the flash drive. If it's thrashing then you need to find out why.

 

[miazza]

I would still definitely recommend a journaled filesystem even when using a swap file. Having the occasional flash drive failure is par for the course, but the rate you're getting through them makes me think this is a problem unique to your setup. Otherwise why isn't this forum flooded with "me too" posts? Rather than being a software problem perhaps there is an electrical issue with your router's USB socket, or it's getting too hot....
If the problem is caused by the volume of write operations then I'd say that either you're using exceptionally poor quality flash drives (my router's flash drive has been in use for 6 years) or your router is thrashing the flash drive. If it's thrashing then you need to find out why.

You are for sure right ... the fact I'm alone in this suggest an issue with the USB socket (hope not because I don't know how to debug it). The feeling is that it is not too hot ...

Poor quality drives ...

The point is that it worked well for almost 1 year and now this rate of failure is suggesting something has changed in the router functionality more than a drive issue.
In any case I will try with a better quality one ....

Your router is thrashing the flash drive.

Most likely you are right but again I really do not know how to debug...

In any case thank you for your kind assistance.

 

[ColinTaylor]

Most likely you are right but again I really do not know how to debug...

You could start by running top and looking at the io usage (which should be 0 almost all the time). But unless it's happening 24/7 you'll have to monitor it over time to see anything. That would require you to install software like sar or vmstat.

 

[miazza]

I guess I need to install sysstat. Is there an easy isntall process or tutorial ?
I found this https://github.com/RMerl/asuswrt-merlin.ng/tree/master/release/src/router/sysstat-10.0.3
but for my skill it is not enought.

May be there is an apt-get install ?

 

[ColinTaylor]

I'm surprised but maybe it's installed already with current Merlin. I had to install it from Entware using John's fork.

Check it with with command:

sar 1 5