[SOLVED] Disk read errors During Parity check, and after

[arombo]

Hello All,

 

So during a parity check yesterday my Disk2 started reporting back read errors, over 20k. This system has never had any read errors before,and has been nice and rock solid for months.

Poked around on the forums for some troubleshooting ideas and went into maintenance mode for the array, and ran the file system check on all of the drives, nothing came back stating there were errors, or that I needed to take additional steps. Restarted the array normally, no issues, but today again I'm starting to get read errors.

 

The only changes made to the system before were the addition of the Plug-ins: CA Auto Turbo Write Mode, Dynamix Cache Directories, and Tips and Tweaks. 

Please find below my diagnostics. 

 

Right now I'm just looking for guidance on either repairing the drive with errors, or if it's too late to salvage and I should go out and get a replacement drive.

I appreciate any and all help that can be provided. 

 

monolith-iii-diagnostics-20170711-0850.zip

[JorgeB]

Don't use the Marvell ports (1st 4 white ports) on your board, they are a known problem, there's a firmware update on Asrock'ss site that helps but appears to not completely fix the issue.

 

Intel ports are more the enough for your current number of disks, if/when you need more buy an LSI controller.

[arombo]

Thank you Johnnie.black,

 

I thought I had all the drives on the Intel controlled SATA ports. I will move some cables this evening and then restart the server to see if the issue resolves. 

 

Unrelated question, the ASRock EP2C602 motherboard I'm using has a group of 4 SAS ports on it. Are they not worthwhile over an LSI add in card?

[JorgeB]

2 minutes ago, arombo said:

ASRock EP2C602 motherboard I'm using has a group of 4 SAS ports on it

 

Those are also Intel ports, though not AHCI, they use a different driver, but you can use them without a problem, only after the 10 Intel ports total are occupied you need an aditional controller.

[arombo]

6 minutes ago, johnnie.black said:

 

Those are also Intel ports, though not AHCI, they use a different driver, but you can use them without a problem, only after the 10 Intel ports total are occupied you need an aditional controller.

 

Thank you again Johnnie.black! 

So even though the driver is different, the ports on the board that are labeled SAS will function like regular SATA ports when a SATA drive is attached to them?

There is still room in the case I'm using for 2 more 3.5" hard drives, and I'm just trying to be certain that I can still do that in the the future.

 

[JorgeB]

2 minutes ago, arombo said:

So even though the driver is different, the ports on the board that are labeled SAS will function like regular SATA ports when a SATA drive is attached to them?

 

Yep.

[arombo]

Just now, johnnie.black said:

 

Yep.

 

Awesomesauce!

 

Thanks again Johnnie.black!!!

 

[arombo]

After safely shutting down the server last night and moving the SATA cables over to the Intel ports, everything appeared to be back up and functioning with no issues.

However, after checking it this morning I'm still getting disk read errors on my disk2.

I've attached the diagnostics file freshly pulled from this morning. 

monolith-iii-diagnostics-20170712-0555.zip

[JorgeB]

SMART still looks good, if the cables are the same replace them, if they are new it's probably a bad disk.

[arombo]

The Cables definitely are a little suspect to me, I noticed they weren't aligning 100% square to the SATA port when I moved them last night because of the tight squeeze from the drive cage to the motherboard. The cables are only nine months old though. 

I was already googling last night for some thinner\more flexible SATA cables to make things fit better and without putting stress on the motherboard ports. 

 

[arombo]

So far so good on new SATA cables. Still showing no errors, and everything has been running just fine. (knocks on all the wood)

 

Here's the latest Log just for records sake. I feel pretty confident that this issue can now be marked solved.

monolith-iii-diagnostics-20170726-1131.zip