Hi all,

 

One of my 12TB drive(disk 7) failed , I am trying to replace it with a smaller one(11.8 TB, diskX).

 

I have done the following：

1. Set disk 7 to "no device"

2. Mount disk 7 and the new diskX via ssh

3. Copied all folders from disk 7 to DiskX using "cp -vR"

   Not sure if it is a better way to do this, like copy files from "emulated array drive" instead of the failed physical hard drive

 

My plan is to performe a "new config" procedure as blow:

1. Run array new config, keep everything else except disk 7 assginment.

2. Set the new DiskX as disk 7.

3. Rebuild the parity.

 

Is it ok to do so? Thank you!

 

 

 

Edited July 24, 20241 yr by huminsk

In what way did the disk fail? Bad connections are more common than bad disks. Can you get a SMART report from the disk?

22 minutes ago, trurl said:

In what way did the disk fail? Bad connections are more common than bad disks. Can you get a SMART report from the disk?

 

trurl, thanks for replying. S.M.A.R.T report and syslog uploaded. 

 

The symptom is UNRAID reported thousands of Read Errors on disk 7 and then disabled it. 

It makes funny noise during unmount phase of stopping array(Maybe it still could be a cable issue related to power connector?). I will try the failed drive when I get the new cable.

 

35000cca2594f2e2c-2024-02-15 disk7 (sdm) - DISK_DSBL.txt syslog.txt

The fact you indicated 'funny noises' suggests to me either the drive really is failing, or alternatively it is getting insufficient power.

 

Running the Extended SMART test is a good indication of a drive's health.  If it passes then the drive is probably OK, and a failure normally means it should be replaced.

12 hours ago, huminsk said:

copy files from "emulated array drive" instead of the failed physical hard drive

If the disk is really failing copying from the emulated disk might be better.

On 7/22/2024 at 6:35 PM, itimpi said:

The fact you indicated 'funny noises' suggests to me either the drive really is failing, or alternatively it is getting insufficient power.

 

Running the Extended SMART test is a good indication of a drive's health.  If it passes then the drive is probably OK, and a failure normally means it should be replaced.

 

On 7/22/2024 at 11:10 PM, trurl said:

If the disk is really failing copying from the emulated disk might be better.

 

Two more hard drives reported Read Error when rebuilding data. All these 3 drives were connected to a same Molex MiniSAS-SAS cable.

I have changed the cable and rebuilt the data of the "failed" disk, so far so good.  So it was a cable issue as you guys suspected.

Thank you for your help.  @trurl @itimpi