Drive errors. SMART extended self-test passed on all drives

[Feducu]

Hi guys,

I'm getting this errors on my unRAID 6.5.0 servers. I made a SMART extended self-test on ALL drives and all of them passed (The 12tb one last for 17hs...) But, almost one hour of running I start to get this errors:

Mar 21 10:03:20 unRAID kernel: XFS (md6): metadata I/O error: block 0x58 ("xfs_trans_read_buf_map") error 5 numblks 8
Mar 21 10:03:21 unRAID kernel: XFS (md3): metadata I/O error: block 0x3e5b10 ("xfs_trans_read_buf_map") error 5 numblks 8

On random drives, sometimes is drive 7 or like on this time are drive 3 and 6. The thing that is really odd is that, if I reboot the system (Force reboot) then, all the drives start working again. I tought this was a temp issue, but is not the drives are between 32°C-39°C

 

I don't know what to do

 

Thanks, 

 

Feducu

 

PS: Sorry bad english, not native languague

[trurl]

https://lime-technology.com/wiki/Check_Disk_Filesystems

[Feducu]

1 hour ago, trurl said:

https://lime-technology.com/wiki/Check_Disk_Filesystems

 

I did a repair of the 12TB drive (the others drives seemed OK), let's see if this solves the issue! Thanks for answering!

[Feducu]

It happened again :/. I did the repair and now its working all right. Will I have to do this every 2 days (give or take). Why is my filesystem getting corrupt?

[trurl]

19 minutes ago, Feducu said:

Why is my filesystem getting corrupt?

 

You should always use the webUI to shutdown if possible instead of just using the power switch. And you should not allow any drive to get full. Those are just general suggestions to prevent filesystem corruption.

 

Other than that, you haven't really provided any of the information we normally ask for when trying to help. In particular, go to Tools - Diagnostics, preferably after the problem occurs but before rebooting, and post the complete diagnostics zip.

 

 

[Feducu]

Yeah, I read that post and did the extract, I'll attach the diagnostics of the first error.

 

I did not shutdown nor reboot the server since the last repair, and the drive is not even a 1/4 of the capacity.

 

Thanks!

unraid-diagnostics-20180321-1011.zip

 

EDIT: Now I'm doing a read-check test

[trurl]

No SMART report for many of your disks, and no parity disk assigned. Check connections, both SATA and power, reseat controller. You need to get your hardware problems squared away or you are going to keep having problems. Post another diagnostic after you fix things.

 

Why don't you have a parity disk? I hope you don't have any important data on your server.

[JorgeB]

The Marvell controller on those Asrock boards is a known problem, as it keeps dropping disks, never use the first 4 white SATA ports.

[Feducu]

16 minutes ago, trurl said:

No SMART report for many of your disks, and no parity disk assigned. Check connections, both SATA and power, reseat controller. You need to get your hardware problems squared away or you are going to keep having problems. Post another diagnostic after you fix things.

 

Why don't you have a parity disk? I hope you don't have any important data on your server.

 That's the odd thing, I did run a extensive self-test on all disk...  I've already cheacked all connections

[JorgeB]

14 minutes ago, Feducu said:

I've already cheacked all connections

In case you missed my posted when you were replying, this is your problem:

 

19 minutes ago, johnnie.black said:

The Marvell controller on those Asrock boards is a known problem, as it keeps dropping disks, never use the first 4 white SATA ports.

 

[Feducu]

14 minutes ago, johnnie.black said:

In case you missed my posted when you were replying, this is your problem:

 

 

Yeah, I saw that, I'll try that now