Physically locating a failed drive

[unknownerrors]

I have a 14 drive array running on a RAID controller that does not support pass though, so on the controller each drive is co figured in a single disk RAID 0 because of this none of the drives smart data or serial numbers are accessible to the OS. The system is reporting one drive as removed (I'm banking its failed)
I need to figure out which of the 14 is in trouble and at this point the only thing I can think of is the bring the volumes on it offline to ensure there is no activity and start yanking drives until I find the right one. My question is, if I do that when I reconnect the working drives will the system automatically detect and reactivate them or will it cause more problems?

Thank you so much for your time and consideration.

~UE

 

[blanchet]

If a disk is missing, normally you can find it with the embedded software of the RAID controller.
On DELL server, you connect to iDRAC, then go to the Storage | Physical Disk

 

[Etorix]

You should NOT use a RAID controller with ZFS, not even passing individual drives as "single drive RAID 0". That's just a proven recipe to losing your data.

 

[unknownerrors]

Thank you for your assistance

If a disk is missing, normally you can find it with the embedded software of the RAID controller.
On DELL server, you connect to iDRAC, then go to the Storage | Physical Disk

Sadly the server is a home brew built on a tyan, so iDrac is not an option.

 

[unknownerrors]

You should NOT use a RAID controller with ZFS, not even passing individual drives as "single drive RAID 0". That's just a proven recipe to losing your data.

Good to know, however if I am going to be able to backup my some 20Tb of data I'd like to do it at the highest speed possible.
Back to my original question. Can I safely unslot drives in a process of elimination and not expect issues when I slide the working drives back in?

 

[Etorix]

If it were pure ZFS, yes. With the (unnamed) RAID controller in the middle, make sure there's no issue on this side.
This RAID controller should have some UI, at least from the BIOS, where it should be possible to identify the serial number of the failing drive.

The priority is to make sure your data is safe. Once it's backed up elsewhere, replace the RAID controller by a proper HBA (or flash it to IT mode), reformat drives and rebuild the pool. Note that it is recommended to keep to a maximum of about 10-12 drives per vdev; 14-wide is too wide.