Random SSD disconnects

Minky

Cadet
Joined
Sep 18, 2023
Messages
1
Good morning everyone,

maybe I'm just missing something, maybe i'm doing something wrong.
I read a lot of information about flash only Vdevs here but didn't find the real answer to my problem.

I'm running a small TRUENAS SCALE - "Homelab" NAS.
The NAS is build with an Intel Core i5-6500, FUJITSU Motherboard, LSI 9300-16I HBA (IT-Mode).

I'm running a HDD-Array and wanted to do a second SSD-only array for my proxmox VMs and as a gaming partition for my pc.
The SSDs are four cheap INTENSO TOP 1TB SSDs all buyed them to different times - so serial numbers are not in line. ;-)
I know not the best idea to use customer SSDs but for me on this point was the usecase - there are a lot of reads and writes (VMs and Gameupdates)
So i was thinking - build an 4 ssd array and change them everytime they wear out - until now I'm not sure how long they will run.

Everything is working fine with also the performance is good, but in the SSD array I'm randomly losing ssds.
For me at the moment I didn't see any indicator what's happening that the SSD will turn off.
They are shown as removed then.

At the moment I'm thinking, that it's an firmwarebug in the INTENSO SSD because:

In the beginning I just have everything pluged to my HBA card.
When the errors start I was wondering what could happen there and find some information, that on some SSD you need to set "-C 0" on smart arguments - tried it - doesn't work.
Then I found a thread where someone has written, that this could be a trim error - so i diabled auto trim - doesn't fix my problem.
But the strange thing here is:
When the ssd is shown as removed they won't come back in truenas at all - tried other HBA ports and I tried to connect to the motherboard itself without luck a hdd or another ssd are recognized directly.
So for the beginning it looks like the SSD is completely gone, but when I take them out of the system and pack it with a cheap usb adapter to my pc it firstly doesn't show up - so i reconnected it once more and the ssd is shown on my pc. I then checked the smart sector and it looks perfect.

When I put it back to my truenas again, the ssd is there directly and is working fine for sometimes a few hours, sometime a fews days, sometimes weeks.
And this could happen on all four ssds in the array.

Until last week it was running on latest TRUENAS SCALE 22 update.
Since the data is backed up always I also tried to update to truenas 23.10 BETA1 but there's no difference at all.

At the moment I'm running out of ideas.
Maybe anyone has the seen something similar and could give me a hint here.
 

Attachments

  • Screenshot 2023-09-19 075917.png
    Screenshot 2023-09-19 075917.png
    33.2 KB · Views: 63
Top