[Help] F6-424 Max Rebooting Hardware? Software?

[TheHawk]

My F6-424 Max has been running fine on TOS 6.x, I upgraded to 7.x and all seemed to be well too.

Intel i5-1235U, 4 x 16TB Drive in RAID 5 + 1TB NVME.
It runs file sharing + Plex and hasn't missed a beat until now.

Last week I did two things
1. I installed immich and shared my 200k photo library (read only) and it was chugging away on that from Friday.
2. On Saturday I picked up some more RAM, a single stick of Lexar LD5DS016G 16GB to go alongside the stock 8GB module until I get some more money and/or RAM prices come down.

It rebooted fine and kept working without issue, happily running Plex the whole time.

On Monday it appears have rebooted itself with a message saying "Device Chesterv has encountered an abnormal shutdown, possibly caused by a power outage or system error. It is recommended to check the device and system status as soon as possible.". It sits behind a dedicated UPS which hasn't reported an outage... so it's at least not mains.

On boot it decided that the Volume was corrupt and seems to have started to automatically rebuild itself. That seemed to (eventually) come good this morning so I figured I'd check things out give it the latest software update it had been asking for, so I downloaded the latest 7.0.1169 did that and it decided it was corrupt again and started Syncronising again, although this time the shared folders started being accessible after only a few %.

I had done a pretty quick and dirty RAM test with stressapptest, this didn't immediately fail, but I figured

None of the services seem to have restarted themselves automatically, but manually enabling Plex seems to get it running again. I haven't started Docker for immich again yet, I figure I'll let the system finish synchronising and see what happens first.

Next steps?

Once it's done synchronising, I figure I remove the extra RAM stick and see if it stays stable. If it reboots itself again before that happens, I will pull the RAM earlier, not sure I want to interrupt the current process otherwise.

Test 1: Longer RAM Test? Maybe memtest instead of the inline stressapptest.
Test 2: Stock 8GB
Test 3: 1 x 16GB Lexar module?
Test 4: Back to this configuration?

Anyone seen anything like this before?

[rfbjr]

I don't have the MAX just the F6-424 model. Yes in April I bought OWC 16gig "compatible" with this model. Yes it would reboot at various times usually at night or early morning. I gave up with the 16gig and stuck with the 8gig it came with. Since in my case that device is a a target of various backups, I really didn't need the extra RAM.

[TheHawk]

Immich likes the RAM, but otherwise I don’t need it. I could move that to my promox machine, but that is much slower and would be reading over the network, albeit at 2.5G.

Did you ever do any testing of the RAM itself, or just took it out and the problems went away? I wonder if a stress test can highlight issues or if it’s an obscure and random thing you can’t really test for?

I’d consider the very expensive Terramaster RAM buts it’s not available at the moment.

[rfbjr]

Yes I remember running some tests on the ram but nothing came of it. My guess and it’s only that is the systems are designed for TM RAM only.
Also stopped for other non-technical reasons but a few of the crashes did trigger RAID resync and there went a day or two.

[CursaYang]

We recommend using MemTest to test the memory.

In addition, to further confirm the issue, please enable debug mode. When the issue occurs again and the device restarts, download the debug logs and send them to us.[support(at)terra-master.com]

If possible, please note the time when the issue occurred — this will be very helpful for our troubleshooting.

The debug logs consist of three files: kern.log, syslog, and atop_log. When downloading, please check whether they are being blocked by your browser.

[TheHawk]

We recommend using MemTest to test the memory.

In addition, to further confirm the issue, please enable debug mode. When the issue occurs again and the device restarts, download the debug logs and send them to us.[support(at)terra-master.com]

If possible, please note the time when the issue occurred — this will be very helpful for our troubleshooting.

The debug logs consist of three files: kern.log, syslog, and atop_log. When downloading, please check whether they are being blocked by your browser.

Thanks for that, I'll look into doing all of that over the coming days. For now I'm remotely monitoring the Sync which is going to take a whole bunch more hours. I'm guessing it's best not to interrupt that process if possible.

[TheHawk]

EDIT: I have to give some credit, the Support Team have been very responsive. They say the logs point to the RAM despite it passing testing. I've got the last round of logs with them again now.

*********************
Where I am right now:

Memtest passes with both modules.
Memtest passes with only the new 16GB module installed.

The synchronisation task takes about 24 hours after a crash which means it takes time before I can test things again. But following the last synchronisation, I removed the stock 8GB module leaving only the new 16GB, as above it passed memtest. But after a booting back into TOS it crashed a couple of times overnight! Seems this NAS really doesn't like that Lexar module, even if it tests OK?

I don't have another DDR5 laptop with removable RAM to test it in... so is it actually faulty or does the Terramaster just not like it for some reason???

In any case, currently I have only the stock 8GB module installed and it's synchronising again. Tomorrow I will be able to start up the services again and see if it survives running immich in docker (which would really like more RAM).

I will also find out later if the local computer store will accept a return/replacement on the Lexar module. If they want to test it and it passes.... well, then I'm stuck with a module that doesn't work for reasons I don't understand.

But assuming they do replace it, my choices are Crucial, Kingston or Team Group Elite (or ADATA if I just got a second 8GB module). Has anyone used anything successfully that I might have better luck with??

Sadly Terramaster RAM just isn't available to buy right now otherwise I might just pay the premium and call it a day.

[CursaYang]

Memory test errors may be related to the SA GV feature enabled in the BIOS. This feature can affect the compatibility of third-party memory under certain conditions, causing errors during testing. Disabling the SA GV feature in the BIOS settings can improve compatibility with third-party memory.
viewtopic.php?t=9413

[TMroy]

TerraMaster has not implemented any measures to restrict memory configuration. However, the potential for incompatibility between memory modules and the motherboard is an indisputable fact.

[rfbjr]

In all fairness and I just checked the site again, memory is out of stock. Yes I looked at the compatibility site and the only memory listed is from TerraMaster. Granted the device I was going to increase the memory on is basically the target of backups, so a bump up of memory wasn’t going to do much.