SSD failure or Debian 10 Incompatible ?

Moderator: Andrey.Mazlin

XDm9wf8aS2aZsKh
Posts: 16
Joined: Sat Jul 13, 2019 2:05 pm

SSD failure or Debian 10 Incompatible ?

Post by XDm9wf8aS2aZsKh »

Hi,

I have a Mintbox Mini Pro 2 which I really love.

But since the beginning, I found the SSD a little bit slow, but it was ok for me.
I use Rambox and parameters are often lost. Maybe not saved correctly ?
SSD device is NT-128.

I updated from Debian 9 to Debian 10. But now I have strange behaviour.
Randomly system puts himself to read only mode. I am forced to hard reboot.

Yesterday, NT-128 completely disappears from BIOS, and from gparted live.
Today NT-128 is back.

I tried to investigate.
gsmartcontrol from gparted live says : short test + long test : all is OK.
gsmartcontrol from debian says : sometimes all is OK and sometimes gsmartcontrol crash and can't continue.

But I found that, when I do : smartctl -t long /dev/sda
System always crash and becomes read only after some seconds.
I attached screenshots after crash.

I try to add these grub parameters : irqpool all_generic_ide noacpi
System still crash after smartctl.

Can you help me ?
Problem with SSD ?

Thank you.
Attachments
IMG_20190713_154945.jpg
IMG_20190713_154945.jpg (1.67 MiB) Viewed 37295 times
IMG_20190713_152351.jpg
IMG_20190713_152351.jpg (1.41 MiB) Viewed 37295 times
Last edited by XDm9wf8aS2aZsKh on Thu Jul 18, 2019 7:59 am, edited 1 time in total.

tamir
Site Admin
Posts: 419
Joined: Mon Dec 25, 2017 4:21 pm

Re: SSD failure ?

Post by tamir »

The failure you observe from the smartctl output does not necessarily indicate an issue with the SSD.

The technical explanation is that S.M.A.R.T. is originally designed for mechanical HDD where monitoring the aging is critical to generate a warning before a total failure.
SSD aging follows a completely different pattern and there is no way to test the cell for aging. In general, S.M.A.R.T. self test on SSD cannot be trusted.
Also, S.M.A.R.T. structure for SSD is slightly different from HDD.

Although the above, disappearance of the storage device at the BIOS menu and other performance issues you might encounter might do indicate issues with the SSD.
The most efficient test is to eliminate this HW issue using an alternative SSD device if available.

Unfortunately, if that's the case, you may RMA the machine for repairs. The procedure is documented at the following link:
https://www.fit-pc.com/web/support/rma/

XDm9wf8aS2aZsKh
Posts: 16
Joined: Sat Jul 13, 2019 2:05 pm

Re: SSD failure ?

Post by XDm9wf8aS2aZsKh »

Hi,

Thanks for informations about SMART and SSD.
I very hope this is not a SSD failure.

I added this parameter to grub : libata.force=noncq

Now I have this error :

Code: Select all

[mer. juil. 17 08:25:21 2019] ata1.00: exception Emask 0x10 SAct 0x0 SErr 0x4050000 action 0xe frozen
[mer. juil. 17 08:25:21 2019] ata1.00: irq_stat 0x00000040, connection status changed
[mer. juil. 17 08:25:21 2019] ata1: SError: { PHYRdyChg CommWake DevExch }
[mer. juil. 17 08:25:21 2019] ata1.00: failed command: WRITE DMA
[mer. juil. 17 08:25:21 2019] ata1.00: cmd ca/00:08:a0:09:d0/00:00:00:00:00/e4 tag 4 dma 4096 out
                                       res 50/00:00:88:09:d0/00:00:00:00:00/e4 Emask 0x10 (ATA bus error)
[mer. juil. 17 08:25:21 2019] ata1.00: status: { DRDY }
[mer. juil. 17 08:25:21 2019] ata1: hard resetting link
[mer. juil. 17 08:25:22 2019] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[mer. juil. 17 08:25:22 2019] ata1.00: configured for UDMA/133
[mer. juil. 17 08:25:22 2019] ahci 0000:00:12.0: port does not support device sleep
[mer. juil. 17 08:25:22 2019] ata1: EH complete
[mer. juil. 17 08:25:22 2019] ata1.00: exception Emask 0x10 SAct 0x0 SErr 0x4050000 action 0xe frozen
[mer. juil. 17 08:25:22 2019] ata1.00: irq_stat 0x00000040, connection status changed
[mer. juil. 17 08:25:22 2019] ata1: SError: { PHYRdyChg CommWake DevExch }
[mer. juil. 17 08:25:22 2019] ata1.00: failed command: WRITE DMA
[mer. juil. 17 08:25:22 2019] ata1.00: cmd ca/00:08:98:2b:14/00:00:00:00:00/ee tag 28 dma 4096 out
                                       res 50/00:00:20:2a:14/00:00:00:00:00/ee Emask 0x10 (ATA bus error)
[mer. juil. 17 08:25:22 2019] ata1.00: status: { DRDY }
[mer. juil. 17 08:25:22 2019] ata1: hard resetting link
[mer. juil. 17 08:25:23 2019] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[mer. juil. 17 08:25:23 2019] ata1.00: configured for UDMA/133
[mer. juil. 17 08:25:23 2019] ahci 0000:00:12.0: port does not support device sleep
[mer. juil. 17 08:25:23 2019] ata1: EH complete
[mer. juil. 17 08:25:24 2019] EXT4-fs (sda2): mounted filesystem with ordered data mode. Opts: (null)
On the internet, they often talk about SATA cable problems. But there is no cable with M2 SSD.

Any other solutions ?
It is very strange that it happened just after upgrading to Debian 10 !

I use this bios version : FLT2.0.40.01.00 (05/14/2018). Upgrade can help ?

tamir
Site Admin
Posts: 419
Joined: Mon Dec 25, 2017 4:21 pm

Re: SSD failure ?

Post by tamir »

We have additional reports about issues with Debian 10 not working well on the fitlet2.
Debian 9 seems to work fine on the fitlet2, but best way to eliminate any HW issue is to install Linux mint which is fully supported on the fitlet2 platform (being its official platform).
You may find an installation guide documented for the fitlet2:
http://fit-pc.com/wiki/index.php?title= ... _Mint_19.1

fred44nl
Posts: 29
Joined: Tue Jul 24, 2018 8:37 am

Re: SSD failure ?

Post by fred44nl »

XDm9wf8aS2aZsKh wrote:Any other solutions ?
It is very strange that it happened just after upgrading to Debian 10 !
from which logfile did you get the above ??

I also upgraded from Debian 9 to Debian 10
and it is not going well, meaning unexpected lock-ups

fred44nl
Posts: 29
Joined: Tue Jul 24, 2018 8:37 am

Re: SSD failure ?

Post by fred44nl »

tamir wrote:We have additional reports about issues with Debian 10 not working well on the fitlet2.
Debian 9 seems to work fine on the fitlet2
I understand yr reaction, but if one is familiar with Debian, it is not that simple to change to Mint.

XDm9wf8aS2aZsKh
Posts: 16
Joined: Sat Jul 13, 2019 2:05 pm

Re: SSD failure ?

Post by XDm9wf8aS2aZsKh »

Linux Mint was very slow on the fitlet2.
When I put Debian, it was blazzing fast !
I use this mintbox mini pro 2 as my main computer.

I can get log from dmesg -T or when it fully crashes, I go to « Ctrl + ALT + F1 ».

I actually found a temporary fix. It lets me work without crash.
Add this grub parameters : libata.force=1.5G

SSD should be slower, but I don't have any errors since 24 hours. Before this, I had errors after few minutes or few hours.

But even with this parameter, if I launch smartctl => crash

If Linux Mint works, Debian should work too ! Linux Mint is based on Debian ...
Is it a kernel problem, or a simple config to edit ?

fred44nl
Posts: 29
Joined: Tue Jul 24, 2018 8:37 am

Re: SSD failure or Debian 10 Incompatible ?

Post by fred44nl »

I use my Mint Mini Box 2 as a media-server, completely headless.
removed as many packages as possible.
I am now down to 958 packages.
so, I don't have smartctl.

mine seems to crash when I use apt.
that was so from the upgrade to Buster.
it does not crash when I use Synaptic; this goes over rdp, as I don't have a keyboard, nor a mouse, nor a screen.

now I have removed apparmor.
may be that helps :)

I will try your grup parameter later on.

XDm9wf8aS2aZsKh
Posts: 16
Joined: Sat Jul 13, 2019 2:05 pm

Re: SSD failure or Debian 10 Incompatible ?

Post by XDm9wf8aS2aZsKh »

mine seems to crash when I use apt.
that was so from the upgrade to Buster.
I also had problems when upgrading, and when using apt. In fact, I have problems when there are many writing to disk.

I don't understand why. But « libata.force=1.5G » seems to fix it.
Normal speed should be 6G.

fred44nl
Posts: 29
Joined: Tue Jul 24, 2018 8:37 am

Re: SSD failure or Debian 10 Incompatible ?

Post by fred44nl »

can you help me here ??
fred44nl@Debian:~$ sudo nano /etc/default/grub
added libata.force=1.5G at the bottom
fred44nl@Debian:~$ sudo update-grub
/usr/sbin/grub-mkconfig: 34: /etc/default/grub: libata.force=1.5G: not found
I also did

Code: Select all

dd if=/dev/zero of=test bs=1048576 count=2048
and

Code: Select all

dd if=test of=/dev/null bs=1048576
went well :)

edit:
look back at syslog from this morning and found several occurencies like this:
Jul 18 07:20:58 Debian kernel: [55099.498604] INFO: task deluged:431 blocked for more than 120 seconds.

Post Reply

Return to “fitlet2 hardware”