Getting ATA bus errors while initializing RAID 5

Triple · 14. Dezember 2020

Hi,

I'm new to the openmediavault forum and

I just setup a new NAS and wanted to create a RAID 5 with 4*14TB harddisks.

Everything went fine, but now while initializing the RAID 5, There are some errors:

ata4.00: exception EMask 0x10 SAct 0x0 SErr 0x8000000 action 0x6 frozen

ata4.00: irq_stat 0x08000000, interface fatal error

ata4.00: SError: [ LinkSeq ]

ata4.00: failed command: READ_DMA EXT

ata4.00: cmd ... tag 10 dma 688128 in

res ... Emask 0x10 (ATA bus error)

ata4.00: status [ DRDY ]

I already updated the grub file, but still get sometimes that error.

The initializiation broke two times before, currently its running (20% finished) but the error occurred 8 times.

I changed the SATA 3 connectors to SATA2, but with no effect.

Does anyone know this erro or where it comes from?

Triple · 15. Dezember 2020

So I deactivated two harddisks and created just a stripe RAID (RAID 0). And I still got the same issue.

It was not possible to create the filesystem.

After that I used some new SATA3 cables. And now it worked for the two HDDs as RAID0.

I just ordered new SATA3 cabel for the remaining two HDDs and will try it.

So it seems that the issue is realted to the SATA cabels. Don't save money on that

cabrio_leo · 15. Dezember 2020

Zitat von Triple

I just setup a new NAS and wanted to create a RAID 5 with 4*14TB harddisks.

You should really think about if a RAID5 array created with 14 TB disks is still an appropriate way nowadays. If you had to replace a disk in the future the resilvering process takes forever and entails the risk that a second disk fails. Which then means a complete loss of the array.

Zitat von Triple

So I deactivated two harddisks and created just a stripe RAID (RAID 0).

You are aware, that RAID0 has no data redundancy? If one disk fails you lost the array. A RAID 0 is normally not useful in a NAS, especially with a GB Lan connection.

How do you want to organize your data backup system?

Triple · 15. Dezember 2020

The RAID0 I created is just vor testing, if the error still occurs and if the HDDs are working and I had only two new SATA3 cables.

And yes, RAID5 with 4*14TB will take a loooong time to be restored if I have to change a harddisk, But this is just an archive data, so I can wait a few days if the cases really happens, that one disc is failing (hopefully only one ;)). For the data of the daily use, I have a backup concept, which I can access atonce.

Thanks for your feedback !

cabrio_leo · 15. Dezember 2020

Zitat von Triple

But this is just an archive data, so I can wait a few days if the cases really happens, that one disc is failing (hopefully only one ;))

If this NAS is only for archived data you really should have a look at SnapRaid / mergerfs. Especially snapraid is perfect for data which is not changing permanently and has the ability to check for silent data corruption which RAID did not. Today I would no longer use a solution that did not guarantee data integrity.

One of many available SnapRaid tutorials: DIY NAS with OMV, SnapRAID, MergerFS, and Disk Encryption

Triple · 16. Dezember 2020

Zitat von cabrio_leo

If this NAS is only for archived data you really should have a look at SnapRaid / mergerfs. Especially snapraid is perfect for data which is not changing permanently and has the ability to check for silent data corruption which RAID did not. Today I would no longer use a solution that did not guarantee data integrity.

I like this and will try this out. Unless I don't have data on my NAS, I can try out some things. Thanks a lot.

Btw. now with all new SATA3 cabels, all discs are running without errors. If anybody would have told me to change the SATA cables I would not believe him.

Jetzt mitmachen!