Ok, this isn't working.  I followed the swap-disable directions to the letter, but right now I seem to be in the same place as these guys. 

 

I figured out which drive was bad by turning the system on and off just once, not even getting past POST, and listening to see which drive made the horrible clunking noise.  I removed the drive, then booted the system up fully so I could get the serial number of the parity drive from the web interface. When I got into the interface, it showed what I expected - a missing disk, the array was stopped, and parity data was valid.  I got the serial number of the parity disk & did a power down via the web interface.

 

With the system shut down, I found the parity drive, moved it to the SATA port previously occupied by the dead drive, then took my new 2TB drive and put it on the same SATA port that the parity drive previously inhabited.  Booted up the system, log into the web interface, and I have red "lights" next to my parity disk and disk 1 (previously the dead disk, now my old parity drive).  The parity slot shows that it is configured properly to be the new 2TB drive, and the disk1 assignment shows that it is my previous parity drive, so both are where they're supposed to be.  But I cannot start the array, and next to the start button is the error message "Too many wrong and/or missing disks!"

 

I cannot, at the moment, put things back as they were and copy the data from the failed drive off onto another disk, then rebuild the array as was suggested in those two older threads I linked.  I don't have the disk space, and I'm really not confident in a couple of my remaining hard drives lasting through more than one array rebuild.  Is there anything you guys can suggest, or should I just email Tom directly?

Ok, this isn't working.  I followed the swap-disable directions to the letter, but right now I seem to be in the same place as these guys. 

 

I figured out which drive was bad by turning the system on and off just once, not even getting past POST, and listening to see which drive made the horrible clunking noise.  I removed the drive, then booted the system up fully so I could get the serial number of the parity drive from the web interface. When I got into the interface, it showed what I expected - a missing disk, the array was stopped, and parity data was valid.  I got the serial number of the parity disk & did a power down via the web interface.

 

With the system shut down, I found the parity drive, moved it to the SATA port previously occupied by the dead drive, then took my new 2TB drive and put it on the same SATA port that the parity drive previously inhabited.  Booted up the system, log into the web interface, and I have red "lights" next to my parity disk and disk 1 (previously the dead disk, now my old parity drive).  The parity slot shows that it is configured properly to be the new 2TB drive, and the disk1 assignment shows that it is my previous parity drive, so both are where they're supposed to be.  But I cannot start the array, and next to the start button is the error message "Too many wrong and/or missing disks!"

 

I cannot, at the moment, put things back as they were and copy the data from the failed drive off onto another disk, then rebuild the array as was suggested in those two older threads I linked.  I don't have the disk space, and I'm really not confident in a couple of my remaining hard drives lasting through more than one array rebuild.  Is there anything you guys can suggest, or should I just email Tom directly?

Power down once more...

Double check the cables to the new parity disk.  Odds are either the power or data cable is loose, or (low possibility) the replacement parity drive is defective.  Usually it is just the cables.

 

Whatever you do, do not press "restore", or set a new initial config, or press "format" as they have nothing to do with what you are trying to do and would make the swap impossible.  The only thing that looks odd in your swap is the new "parity" drive is not working.

I don't think that's the problem, since on the main home page of the unraid web interface, it shows both the new and old parity disks' specifications correctly - the model number & the serial number for both.  I also just checked both cables a couple of times, booted it back up, and here's a screenshot of the interface as it currently stands.

 

Did you go to the devices tab and make sure the drop downs matched the drives? I think that's a long shot since it's probably already set but if you have not checked that then at least take a look.

 

Peter

I don't see anything you are doing wrong either...  I know there was an issue with the swap process in version 4.4.1 of unRAID, but that was fixed.

 

Attach a syslog to this thread, also, type the following and provide the output.  There may be a clue in its output.

/root/mdcmd status

Here ya go, both are attached to this post.

syslog-11.19.2010.txt

mdcmd-status.txt

...anyone?  Bueller?

send an e-mail to [email protected]

 

I see nothing unexpected in the syslog.  The parity swap should be detected, but obviously it is not.  Tom may have an idea.

 

Point him to this thread.  Send him a PM too.

 

Joe L.

Looking at it...

From the screenshot you posted it looks like the bad disk (disk1 correct?) was never 'disabled' to begin with.  What you should do is put things back the way they were.  Then boot server & see if it comes up (it should), and then post back here & I'll tell you what to do next.

 

From reading your first post, you say there was a "bad" drive (clunking noise), but did it still work?

From reading your first post, you say there was a "bad" drive (clunking noise), but did it still work?

He did say:

I figured out which drive was bad by turning the system on and off just once, not even getting past POST[/color], and listening to see which drive made the horrible clunking noise.  I removed the drive, then booted the system up fully so I could get the serial number of the parity drive from the web interface. When I got into the interface, it showed what I expected - a missing disk, the array was stopped, and parity data was valid. 

Yes but if he didn't Start the array in the state where a drive is missing, it would not have disabled the drive.

Yes but if he didn't Start the array in the state where a drive is missing, it would not have disabled the drive.

Oh.... I see...

So if he puts everything back, and the array starts ok (though with a noisy disk1 apparently), I would  have him:

1. Shutdown

2. Remove noisy drive.

3. Boot, observe webGui says disk is missing, but Start array anyway.

3. Array should start, will use parity reconstruct to satisfy I/O requests to now disabled disk.

4. Shutdown

5. Move old parity to disk1 slot, install new (bigger) disk in parity slot.

6. Boot, observe webGui shows 'parity-swap' situation present.

7. Start array - wait for copy to finish (probably several hours) - when done, parity rebuild should be taking place

 

So if he puts everything back, and the array starts ok (though with a noisy disk1 apparently), I would  have him:

1. Shutdown

2. Remove noisy drive.

3. Boot, observe webGui says disk is missing, but Start array anyway.

3. Array should start, will use parity reconstruct to satisfy I/O requests to now disabled disk.

4. Shutdown

5. Move old parity to disk1 slot, install new (bigger) disk in parity slot.

6. Boot, observe webGui shows 'parity-swap' situation present.

7. Start array - wait for copy to finish (probably several hours) - when done, parity rebuild should be taking place

 

So what you are saying is the "swap-disable" procedure is not valid unless the original disk to be replaced is "disabled" first.

 

What about a person who is simply upgrading a working data disk.  Today it makes little sense to purchase a 500Gig PATA drive when a 2TB SATA drive is often less expensive.  If the ports were available and the 2TB was larger than parity, your statement implies they must perform the upgrade in two steps.

 

1.  Replace the existing smaller parity drive with the new 2TB drive.  Build parity on it.

2.  Replace the smaller data drive with the old parity drive, reusing it as a data drive.

 

They cannot do it in one step using swap-disable?  True?

 

Joe L.

I'll see if I can do that, but we're talking clunking so bad I couldn't even get the system to fully power down via a soft-off... it got most of the way off, but then tried to (I'm just saying what it sounded like here) spin up the drive just so it could spin it back down, it wouldn't spin up, just make that loud clunking noise, then it would reset the process and try to spin it up again, then that continued for like 3-4 minutes until I got fed up and just held down the power button to force it off.  Regardless, I'll restore the original configuration and see if I can get unraid booted all the way so it'll mark the drive as disabled.

 

Just in the case that I can't actually get things booted correctly, is there any way to manually flag the dead drive as disabled, then do the swap-disable procedure?

Why attempt to start with the failed drive installed. What's wrong with just putting only the parity drive back and starting the array?

 

Just going to the devices tab and taking the new Hitachi off the list and then moving the WD back to the parity slot should allow the array to at least be started.

 

Peter

 

Why attempt to start with the failed drive installed. What's wrong with just putting only the parity drive back and starting the array?

 

Just going to the devices tab and taking the new Hitachi off the list and then moving the WD back to the parity slot should allow the array to at least be started.

 

Peter

 

Exactly.  About all you should need to do, if I understand correctly is leave the bad disk disconnected.  Assign the old parity drive as parity once more.  Un-assign the new drive, for the moment.

 

Start the array.  It should recognize the missing drive and disable it.  Then,

you can stop the array, do the "swap-disable" (assigning the new 2TB drive to the parity slot, and the old parity to the disabled data disk slot) and then start the array once more.

Well, I actually tried connecting the dead drive in addition to removing the new parity drive and putting the old parity drive back in place, but what wound up happening was exactly the situation you just described.  The dead drive was so toasted it wasn't even detected by the SATA controller, but once I started the array with the single "missing" drive, stopped the array, shut down, and then put things back into the new configuration, the swap-disable config was detected fine, and it's happily copying the parity info to the new parity drive now.  Kinda wish I hadn't bothered to drive back to my office to retrieve the dead drive, but hell, I'm just happy things are working.

 

Thanks Tom and Joe for all your help!