Hi all.

Hoping you can help me out here.

I have a machine running unraid and it works without a flaw.

Exept for one disk that every once in a while reports as bad.

First time this happened it came up with errors and disk fail so i returned the disk to Seagate and they sent me a new one.

The new disk was installed in the machine and everything worked fine for a while.

Then suddenly the new disk fails with read and write errors so i think that i am just extreamly unlucky and i send the new disk to Seagate and the send me another disk.

I install the new disk #2 and the same thing happens, the disk is fine for some weeks and then it fails.

Now i start to think something is right so i put the disk in another computer, delete the partition, try making ntfs partition and then some other partitions and then i do a read and write test on the disk and everything is fine.

So after testing i delete the last partition that i made and install the disk in my unraid machine but this time i change the power cable, sata port and cable to the disk.

Now everything has been working fine for some weeks and this morning i have the same error and the disk is not in the game anymore.

Any idea what could be causing this?

Best regards

dalek-diagnostics-20190414-1104 (1).zip

Disk dropped offline, so there's no SMART report, but assuming it looks OK my main suspect would be the controller, it's a Marvell chipset similar to the one on the SAT2-MV8 and at least with that controller there are known issues with more recent devices, disks or SSDs, would suggest replacing it with an LSI HBA.

Forgot to mention, you need to run a filesystem check on disk1, it was already corrupt before dropping out.

Hi Johnnie.

Thanks for the quick reply.

When this happened before the disk was connected to the onboard sata controller on the motherboard.

Can the Marvel controller have impact on the disk if it is not connectet to the controller?

I mean, can this controller be affecting,  lets say, all disks in the machine if they are larger than X Gb ?

This filesystem was only about 3-4 weeks old and i can not understand why only this disk (or same type of disks and same size because this is disk number 3 that acts like this) is having problems.

18 hours ago, Aparass said:

Can the Marvel controller have impact on the disk if it is not connectet to the controller?

No, if it also failed in a different controller it can't be that, if SMART looks fine then maybe a power issue.

Now this has happened again.

Last time this happened i took the disk out. Put it in another computer, deleted the partition, set up NTFS partition, Checked the disk with the Seagate tool (took over a day), installed it again in my unraid machine, changed the power supply, new sata cable, did the "clean new disk" or whatever it is called and had it do a read and write scan and that took almost two days and everything turned out fine and no errors, put the disk in the raid again and everything worked fine for a while.

This time i lost a lot of data so this is getting very frustrating.

Smart is fine, the disk always is fine, the only thing is that once in a while the disk registers as Bad in the raid and can not be used until the partition has been deleted and unraid gets to make a new one.

Has anyone here seen anything like this or does anyone have any ideas what i could check out next ?

Disk connect under Marvel controller ？

Hi guys.

No. The disk is not connected to the marvel  controller.

Connected to the onboard controller.

Any other suggestions?

If the same disks keeps failing with different controllers (and assuming also different cables) it's likely a disk problem.