[SOLVED] Upgrade 4.7 to 5.0 RC6-test - Wrong disc error

[Glimmerman911]

I just finished a parity sync on 4.7 successfully, and upon first boot after upgrade 4.7 to 5.0 RC6-test I have a problem.

 

The array is stopped, and it says "upgrading disc".  My 6th disc reporting as "Wrong" and re-balled.  The 6th disc is detected as the exact same disc as it used to be, why is it detecting it as a new hard drive?

 

How do I fix this?  I need to upgrade to 5.0 so I can use 3gb drives.

[Glimmerman911]

Even worse, I am trying to copy my previous 4.7 back onto the flash. 

When I try to open my flash drive, I can get to the server in windows, and see the flash sahre, but windows is saying it is referring to a location that is not available!

 

I can still get to the web gui just fine, what is going on and how do I fix this!?

 

*Update*

I powered down the server, pulled the USB and put it into a PC and copies back bzroot & bzimage.

Booted the server back up, now I have nothing, no webUI, no telnet connection.

 

[Helmonder]

The errors in windows are because of changed permission structure in unraid. Have you run the new permissions script ?

 

Also you need to remount al your shares in windows using new credentials (nobody, not root).

 

The red ball however has nothing to do with this, if you have been tinkering on the inside of the system then recheck cabling, if you haven't done anything like that then the most probable cause is that you are experiencing a drive failure, bad luck, possibly triggered by the reboot..

[Glimmerman911]

No I have not ran any script, where do I find that, was it in the release notes?

 

I just ran a parity check that finished today, and right when I reboot after an upgrade it red-balls?

 

And what is the "Wrong" error with red-ball, it is asking me to re-build onto the same disc, if the drive failed would it be asking me this!?

 

*Update*

I did a full copy back of my 4.7 version of flash onto my USB, and Unraid booted up just fine, no red-balled drive. 

 

How do I get 5.0 fixed so I can upgrade and use 3tb drives?

[Helmonder]

Yes. when you go from 4.7 to 5.0 you need to:

 

- run the permissions script

- remount all your shares (if you mounted them as root until now, most people did)

 

I cannot comment on the wrong / red ball... It sounds like your array is thinking that your drive is not the correct one.. In that case rebuilding on the same drive would probably solve this (parity will rebuild on the "new" old drive), however you are in unprotected mode at the moment. Hold out for an expert on the forum to confirm and or deny my statement, there might be a better way to do this.

[Glimmerman911]

And the red-balled drive is back to green when I restored 4.7.

 

I really want to upgrade to 5.0 for 3tb drives.

[Helmonder]

You are working ok on 4.7 now without any issues ?

 

Do you have a syslog of the "wrong disk" situation ?

[Glimmerman911]

No, but I can try putting 5.0 RC6 back on see if the disc red-balls again or not, then running the syslog if it does. 

 

*Update*

 

Re-installed 5.0 RC6, red-balled drive issue "Wrong disc" came back. 

What reports do you want me to pull, and how do I create them?

 

 

[Helmonder]

The syslog .. It's in /var/log

 

If you copy it to the /boot from console you can copy it from the flashdrive share... If you are using unmenu there is a button in there to get it also ..

[Glimmerman911]

Syslog attached

SMART report for the red-balled disc also attached.

 

*Update*

I have re-installed 4.7 and once again the drive is back to green.

I am running a parity sync right now, takes about 11 hours.

I will update whether there are any parity issues when I know.

 

syslog.txt

smart.txt

[JonathanM]

My best guess is that this

Aug 25 17:44:14 Tower kernel: ata6.00: HPA detected: current 3907027055, native 3907029168
Aug 25 17:44:14 Tower kernel: ata6.00: ATA-8: ST32000542AS, CC34, max UDMA/133
Aug 25 17:44:14 Tower kernel: ata6.00: 3907027055 sectors, multi 16: LBA48 NCQ (depth 31/32)

is your problem.

 

I would wait for confirmation from someone more experienced with the 5. series though, I'm still happily running 4.7

[Helmonder]

SMART results are ok... I cannot see anything wrong that I identify... Pending an unraid genius's idea this is what I would do:

 

1) move your data off the drive that will redball if you upgrade;

2) upgrade

3) "restore" the wrong drive onto itself

4) parity check

 

But as said... wait for an expert, there might be another way to solve this.

[Glimmerman911]

Thank you for your help, for now I am going to buy a 2tb drive in thethe morning to get me by until this is resolved. 

[dgaschk]

My best guess is that this

Aug 25 17:44:14 Tower kernel: ata6.00: HPA detected: current 3907027055, native 3907029168
Aug 25 17:44:14 Tower kernel: ata6.00: ATA-8: ST32000542AS, CC34, max UDMA/133
Aug 25 17:44:14 Tower kernel: ata6.00: 3907027055 sectors, multi 16: LBA48 NCQ (depth 31/32)

is your problem.

 

I would wait for confirmation from someone more experienced with the 5. series though, I'm still happily running 4.7

 

See here: http://lime-technology.com/forum/index.php?topic=10866.msg106957#msg106957

[Glimmerman911]

OK so I appadantly i have hpa on,  why would that cause an issue with 5.0 but not 4.7?

 

I just switched to a brand new mobo last week, I wonder if it was the old board or my new board that has hpa.

[Joe L.]

It s being detected as a different size. It may have an HPA which must be removed before upgrading, (I seem to remember this being mentioned in the upgrade instructions in the release notes)

[Glimmerman911]

Ok, so I followed the instructions from your other post joe, at

http://lime-technology.com/forum/index.php?topic=5072.msg46903#msg46903

 

 

Now when I boot unraid, it shows the disk that I removed the HPA from as blue-ball, and unraid wants to re-build it.

 

Should I now do this?  I should I upgrade to 5.0 and then I won't have to rebuild?

Is this normal?

[Glimmerman911]

I started the drive re-build after verifying that the syslog was not reporting any HPA

 

Thank you for your help, once the re-build and parity sync complete successfully I will report back before upgrading to 5.0.

[Glimmerman911]

Parity sync completed successfully, no errors found.  I thought there would have been a few errors near the end of the disc from the HPA, but I guess not.

Proceeding with 5.0 upgrade.

[Glimmerman911]

I upgrade to 5.0 RC6-test, and everything went well, the removal of HPA worked.

 

I started the array, then started the permissions process as per the release notes.

 

It says it is a background process, but I cannot access any of my drives right now, or the Web GUI.  I can telnet into the server.

 

Is this normal?  I didn't expect to lose access to my files, how long will the process take, I have a 15gb server that is full.

 

*Update*

I now have Web GUI back, though very slow, and I can see the reads and writes increasing on disc1, so I assume the utility is still running.

[Glimmerman911]

*update*

 

I re-started the process and just left the webGUI on the same screen, and can see the script progressing through disc1 now.

[JonathanM]

The permissions script seems to be complete, though I see no way of confirming this.

 

The read/writes have stopped increasing on disc 1, but nothing has occurred on the other discs.

 

When I try to access any of my files, it give me an error that I don't have permission.

 

How do I fix this so I can access my files?

Run the permissions script again, but this time don't close the web page. Apparently there is a bug in the current new permissions script that causes it to exit if you close the page, but there is no progress indicator, or other way of telling you to not close the page. It could take many hours to fully process a server with thousands of files.

[Glimmerman911]

Thank you, that is what I have re-done, and I will leave the webGUI up while it completes.

 

*Update*

 

permissions script has completed, everything appears to be working, I will be adding my 3TB drive next.

[Reginald]

Sorry to float this but I wanted to thank Joe and Glim for working through this, it helped me save my array.

 

Edit: just wanted to say, I finished the update and I am now running the New Perms script and it is working like a charm. Thanks for the help!