M.2 SSD Cache causes NAS to crash (TS-453Be & QM2-2S10G1T)

[alphaod]

I have TS-453Be and the QM2-2S10G1T PCIe adapter that gives me 2x M.2 SATA slots and 10Gb Ethernet.

If I just have 10GbE working it seems to work just fine. However if I enabled SSD caching it will initialize, but as soon as anything starts to get cached, I'll hear a "click" like a breaker jumped and the NAS will crash.

I'm using 2x 1TB Crucial MX500 M.2 SATA SSDs which is on the compatibility list. I have tried different variations of RAID 1, RAID 0, write only, read only, read/write. All end up the same: the NAS will crash if the cache gets used. Interestingly if I use the QM2 model that does not have the 10GbE or I use only the 10GbE without caching enabled, the NAS does not crash. I already exchanged the QM2-2S10G1T with a new one and the issue is still persists.

I'm wondering if I'm encountering a bug that other folks have or not?

[Don]

Have you tried with just one of the SDDs installed? Maybe one is bad.

[alphaod]

I have tested both SSDs on my desktop computer and they both passed with flying colors. I really don't think there is anything wrong with the SSDs. I mentioned that I originally had the QNAP QM2-2S-220A M.2 SATA PCIe adapter and I was able to use both SSDs without any issues. It's only on this adapter with both SSDs and 10GbE enabled that I have issues. In fact if I don't use 10GbE, it seems the NAS doesn't crash either. Of course the problem is if i don't use 10GbE the point of SSD caching becoming rather pointless for me.

The power shut off is something I am familiar with. I had several computers that would "click" then shut off; the culprit was that the system was sometimes using more current than the power supply could provide. The solution on those computer were to replace the power supplies with higher output versions. I'm wondering if this is the same issue I am encountering on my NAS. I've done some research and estimations. The NAS powered on with 4x 6TB Seagate IronWolf Pro HDDs, 2x 4GB of RAM, 2x Gigabit Ethernet bonded, 10Gb Ethernet active, and the 2x M.2 SSDs working, the NAS is drawing about peak 66W of power. I don't know how much power the motherboard draws. Although the power supply is rated for 90W, we know that there is always some efficiency losses. I am using the NAS on a 1500W Pure Sine output UPS which is at only 20% load.

So with everything powered on and active at the same time, is my system drawing more current than the power supply can provide? Or perhaps is there a bigger power supply I can use?

[mlai]

I have the same problem on my TS453B but with a 2P10G1T. And the NVMe SSDs are tested fine on computers but the TS453B will crash with them. Now even after the SSDs are removed, I am left with a corrupted file system on the NAS and the NAS frequently unclean reboots and comes back with check file system flagged.

[P3R]

Now even after the SSDs are removed, I am left with a corrupted file system on the NAS and the NAS frequently unclean reboots and comes back with check file system flagged.

Doesn't checking the file system help?

To both you and the OP, please contact Qnap support. This isn't likely something that the user community can solve.

[mlai]

Now even after the SSDs are removed, I am left with a corrupted file system on the NAS and the NAS frequently unclean reboots and comes back with check file system flagged.

Doesn't checking the file system help?

To both you and the OP, please contact Qnap support. This isn't likely something that the user community can solve.

Check file system did not help. And already contacted QNAP support. They remoted in and did a clean file system. And it did not help either.

[alphaod]

I contacted QNAP support in January. They told me that because I migrated from one QNAP to another, that my network adapter was the culprit and there was a bug in the [at the time] current release of QTS and the only resolution was to wait it out for a fix.

The problem is in the recently two releases they both state there are fixes to the network adapter issues from migrating disks from one NAS to another, but I still have problems with my NAS if I enabled M.2 SSD caching. It doesn't matter if I choose read, write, read/write, RAID 0, or RAID 1 combos. They all cause the NAS to crash.

And I still get the error message "Network & Virtual Switch fixed errors in your original system configuration during the migration to QTS 4.3.5. See the details below." "DNS settings are available on each adapter. If you previously selected "Fixed DNS" before migration, the first two active settings on each adapter are now selected."

The whole ordeal is quite a bit frustrating especially since there still seems to be no fix in sight after two months.

[alphaod]

I figure I'd update this thread.

Seems after a while QNAP fixed the 10GbE virtual switch problems which is good.

The SSD cache crashing issue I think might be because of the crappy Crucial MX500 M.2 SSDs. I warrantied both SSDs and they still causes a lot of problems. Seems writing more than 50GB or so in short period of time would cause the Crucial SSDs to crash and basically disconnect themselves; this would cause the NAS to lock up because the M.2 SSD are not hot pluggable. Fortunately I tested it by using as a read-only cache so I didn't have any file corrupt unlike the other poster.

I think QNAP needs to remove these drives from their compatibility list because they are junk drives and will cause data corruption issues if I wasn't testing these in a read-only cache environment.

The result is I ended up buying two 1TB Samsung 860 EVO M.2 SSDs; same exact setup and it has been working flawlessly for a week now.