Hi all,

 

I have pre-built IDE unraid server with 12 300GB drives. This morning, one of the drives went down. The management utility reported that the drive was disabled with errors. Before replacing the drive (a new one is on order), I wanted to try re-enabling it.

 

First, I stopped the array. A stream of commands scrolled by and the utility reported the array was stopped. At this point, the management page kept updating with the same sequence of kill commands - is this normal? In any case, I powered the system down, removed the offending drive and restarted. Upon reloading the management page, I would have expected the utility to require a manual restart of the array due to the removal of the disk. However, the array started automatically and the disk was listed as "not installed."

 

I stopped the array again, with the same results as before. I powered the system down, reinstalled the drive and restarted. Again, upon loading the management page I was expecting to be required to manually start the array. However, the array started automatically and the re-installed disk was listed as "disabled" with the same number of errors.

 

If I understand the instructions in the manual correctly, unraid should have treated the reinstalled disk as new, but this doesn't seem to have happened. Have I missed something?

 

 

 

John

Is there any data on the failed drive you wish to save, if so, now is the time to copy it from the disabled drive elsewhere.

 

To get a failed drive to be recognized again you will first need to unassign it, start the array without it, (it will show as missing ... in fact, this is how you can simulate a failed drive)

 

Then, stop the array and re-assign the drive.  If it is working, its content will be re-built based on the parity drive and remaining data drive data. If this is a big drive, this could take a few hours.  In any case, even during this rebuild, you can get to the drive's data.

 

The "failed" drive could be as simple as a bad cable, or poor connection on a cable, or a true failure.  Therefore, between the first step of un-assigning the drive, and the second of re-assigning it, power down and re-seat the cables and controller card.

 

Joe L.

Hi Joe,

 

Thanks for the swift reply!

 

When you say that I must "unassign" the failed drive, then "reassign" it, is this different form what I did? These are the steps I took:

 

1. I stopped the array, powered down and removed the failed drive. When I powered the system back up, the array started automatically and the drive was listed as "not installed." Is there something else I need to do to unassign the drive?

 

2. I stopped the array again, powered down and reinstalled the drive. Upon restart, the array again started automatically and the drive was listed as "disabled" with the same number of errors as before.

 

I appear to be missing a step, but I'm not sure where. I couldn't find anything in the documentation regarding unassigning and reassigning drives beyond the steps indicated here:

 

http://www.lime-technology.com/wordpress/?page_id=47#toc-how-do-hard-drives-become-disabled

 

 

 

John

Yup... you skipped a step.

 

After stopping the array, go to the devices page in the web-management-tool and un-assign the failed drive.

 

Then power up without it.

 

To get the array to completely forget it ever existed, after un-assigning it, you can stop the array and use the "Restore" button after checking the little checkbox below it.  If you do that a new set of parity calculations will be written to the parity drive, but without any data from the failed (and now un-assigned) drive, forgetting it, and its data had ever existed.

 

From what you have described here so far, you have a failed drive.  When you powered up the array without it, unRaid did exactly what it should... protect you from losing its data.  The disk is still part of your array, (but either missing or disabled) depending upon if errors occurred writing the disk, or if it was not detected at all.  You can still read its old contents, and you can even still write to it, even though it is physically missing.

 

You are not currently protected against a second drive failing until the first one is replaced, or until you reset the array configuration to forget it ever existed.

 

unRaid will not treat the failed drive as new unless it is not part of the current set of assigned devices.  If it can read the disk, it will not erase and re-format it unless it cannot find a reiserfs file-system on it. 

 

So... if you expect for the replacement drive to arrive in a day or two, put the new one into place then.  If you expect it to take longer, and really don't want to risk losing data if a second disk were to fail, copy all the data from the failed drive to other drives, then un-assign it, then use the "Restore" button the Set the Array configuration to the currently assigned drives.  unRaid will then have forgotten about the failed drive completely and re-built parity without it. (this process could take a few hours or more, depending on the speed of your array)

 

Joe L.

Ah . . . that explains it.

 

Unfortunately, I don't have a devices page in the management utility - just "main", "settings" and "tools." The server version appears to be 1.050930. Is there any way to do what you suggest with this version?

 

 

John