< img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=177525256911397&ev=PageView&noscript=1" />

Expansion SSD causes instability

KevinG

New Member
Not sure if it's a bad cable or something, but attaching a separate 2.5" SSD via the provided cable results in a guaranteed system freeze anywhere from 2 to 8 hours after boot.

This has happened with CentOS 7, CentOS Stream (9), Ubuntu 20.04 Server, and Ubuntu 20.04 Desktop. I tried two different known good SSDs (1TB and 4TB) formatted to ext4. The drives work fine while the system is up, but eventually it locks up in less than one day.

I removed the 2.5" SSD and the system remained functional for over seven days before I powered down and reattached the SSD. The system froze after about seven hours from there.

Any suggestions? Is the provided SSD connector something I can find a replacement for or is it a custom cable unique to these products? Willing to try anything.

Thanks
 

Carbon fib'er

Active Member
Does your box have 1 or 2 bays? My hx90 has 2, and I would wonder what happens if the drive were in the other bay, the with the other cable (the cables are made for the bay). I assume yours has one bay, or you would have thought of that too.

All your tests were with Linux. I wonder what a Windows install would do? Sometimes linux can be finicky. I assume you have tried these drives in another machine with the same Linux?

I know "trimming" can be problematic. It's never clear to me whether to use fstab's "discard" option for realtime trimming, or the fstrim command periodly (or sysctrl scheduled fstrim?). I've read of people having lockups and needing to use discard or the command, not the scheduled trim. Something like that, I don't remember exactly. I've read of discard being a problem. Or, doing no trimming leads to degregation after some time (depending on how heavy disk activity is).

If it's something about your box, I'd think it's the chips/driver/bios, not the cable. I can't think of how a cable would slowly (repeatedly) fail that way. It sound much more like that trimming matter.
 

KevinG

New Member
I ran Windows briefly just to see it run, but this was intended to be a mini home Linux server so I installed over it almost immediately.

I have the JB95, so just one extra SSD bay. Both drives have been used in other systems (before and after these failures) without any issue. Additionally, I use these exact drives for work, partitioned with a script that uses parted to set everything up, and have never seen a failure. I doubt it's the fault of the OS, since it has the same issue from CentOS 7 through 9, and Ubuntu 20.04. I haven't found any reports of a similar issue, so maybe I just have a bad board. Should have kept the packaging.

Is there somewhere I can look for an updated BIOS to try? The only post/comment that mentioned a BIOS update has a dead link.

Edit: The attached SSD has basically no activity, simply mounted as extra storage (and empty).
 

KevinG

New Member
Small update...I used a SATA->USB adapter with the same drive, and the OS locked up much quicker (tens of minutes).

Outside of a BIOS update I don't expect any resolution...I'll have to look for another box to fit my needs. :confused:
 

Carbon fib'er

Active Member
Is there somewhere I can look for an updated BIOS to try? The only post/comment that mentioned a BIOS update has a dead link.
You mentioned that you formatted the windows partition. That could prevent you from upgrading the bios if there were a new version available. I always replace the device containing the OS with a new drive, and use that. Then I can resinstall the windows drive, update the bios, remove the drive again. (Seems wasteful, but until bios-update sofrware is more basic and free to use, it's practically required to keep the drive as an important part of the sytem. I think minisforum's updater is a windows program. I don't know for sure. But, I think I saw that.).

I think you have to email support@minisforum.com if you feel you need support. It doesn't seem to happen here (I'm curious if the email interactions are productive. I don't think I've seen complaints about it.).
 

KevinG

New Member
Ah I see what you mean, with other systems I've worked with I put the BIOS update files into a FreeDOS bootable USB and run it from there, but you're saying it's a Windows application. I've seen that before too.

I emailed the support email on March 31st with no response.

I've found an alternative solution to fit my needs, not a mini box like one of these but maybe I can repurpose this for another task. Really needed the extra storage space, so I'm just going to call this a bad learning experience.

Thanks for chiming in.
 

Carbon fib'er

Active Member
You don't think there's something in the bios that might influence this? (I have the HX90 with a high-end cpu, and I don't know what I'm even looking at in the bios. To the extent you have config options, maybe it's something there?). Could it be temperature? I've worried adding two drives to my mini might increase the temperature too much.

A common recommendation is to press the pinhole reset button 30 seconds after disconnecting the power (and any powered devices connected to it). You might also try going into the bios and exit with saving (even if you don't change anything). More than once that's fixed a problem after adding memory, or flashing the bios. Another idea: run memtest86. You might think "the memory is fine, I didn't touch it." But, I've seen memory not exhibit a problem until a certain condition exists. For example, maybe the the presence of the sata drive caused drivers to be loaded, and now a bad mem location is being touched? (I had a computer that would lockup with Linux, but run Win 98 fine. I thought it was something incompatible about the hardware. This was back when you had to be more careful about the hardware you bought. Incompatibility was worse then. But, for some reason I ran that memtest, and discovered I had bad memory. Replaced the stick, and then Linux ran fine. I've seen the same kind of wierdness when I didn't verify the checksum of a downloaded iso. Something was corrupt, and it caused a strange problem like this (instability). But, you said this happens with different distros. So, it's not likely they were all corrupted the same way.

For the speed/performance of that machine's cpu, I've found good used laptops on cralgslist for half the price. I've gotten some good stuff that way, just to leave running as servers for 2-3 years.
 
Top Bottom