Ugraded to V6.0.1 from V5.0.5 and Array will not start - Help!

[mifronte]

HELP!

 

So I plugged in the spare unRAID V6.0.1 flash drive (configured with my drive config files to upgrade instead of clean install) and booted up. In the process of unRAID loading, I accidentally press the IPMI Ctrl+Alt+Delete button.  So I let unRAID boots again and it came up with the array stopped with a message "Unclean shutdown detected." next to the Stopped status.

 

I went ahead and remove my cache drive and added my two new SSD drives to create a drive pool.  Then I press start and the message goes:

Spinning up all drives...Start Array...Refresh SMB

Then the webgui flashes and the array is still showing as stopped.

 

All my drives are green balls except the two new cache SSD drives.  They are showing a blue square.

 

Attached is my log file.

 

Edit:

The two SSDs are not actually new, they were a from a Windows machine that had them as a striped RAID 0 system disk.  I figured I would used them as my cache pool and let unRAID format them.  Is this a correct assumption?

beanstalk-syslog-20150824-0019.zip

[RobJ]

Just a quick comment, you're braver than most of us!  You do something major like upgrading to v6, which includes almost completely overhauling your flash drive and starting a new 64 bit OS, and before it's ready, you build a cache pool, when you don't yet know if you even have a running operational system!  That's confidence!

[mifronte]

Yeah, I may have too much confidence in unRAID since it has been simple, trouble-free, and rock solid 

 

The good thing is that I have two licenses and so I always use the spare for upgrades or testing and never have to jeopardize my current production flash drive.  All is well just as long as I don't muck with any array drives.

 

unRAID V6 true technical workings are scattered all over the forum, especially in the beta threads, and so I never know if I am reading the definitive specs.  The Wiki is too general.  So I am left with just jumping in with two feet and hope the new webgui is very intuitive.

 

I don't know why the array will not start and there is absolutely no logging activities from the moment I press Start.

[mifronte]

Would the fact that the SSDs do not have a valid partition table be the problem?  The two for my cache pool had invalid partition tables since they were in a RAID 0  set for a Windows system.  The third SSD is brand new.

 

I just went into fdisk to properly create a primary Linux partion for each of the SSDs.

 

I am contemplating just going with a clean install and reassigning all 20+ drives to keep it simple.

[mifronte]

So I decided to do a clean install and reassigns all my drives.  Can someone confirm the following before I start the array?

 

[*]All my drives are showing up as a new device (blue square).

[*]The FS column for all my data drives are showing Auto.

[*]Parity drive FS column is blank.

 

So can I safely start the array?  I am afarid that unRAID might start to format all my drives.

[RobJ]

It should not format them.  Make sure you check the box that says something like "Parity is already valid", and I believe it will start a parity check, to ensure parity is valid and correct.  Parity has no file system, so it should be blank.  Blue balls (well, blue squares now) mean they are new to this array, which is correct when starting clean.

 

Just make sure they are assigned correctly, especially the parity drive, and Start the array.

[mifronte]

Still no go    The array just refuse to start.  The status is Stopped Initial Configuration.

 

I will now have to try a pristine clean install without any of my shares cfg files. I even checked the disk.cfg file for the "startState=" as specified in the upgrade guide troubleshooting.  This is not funny any more.

[jonp]

Remove the flash device, plug it into a Windows PC and perform a check disk.  I'm wondering if because you did a hard-reboot in the middle of the intial boot of 6.0.1 you caused not only an unclean shutdown, but made it so the flash is booting in read-only mode, which can cause numerous problems.

[mifronte]

The second go around, I had reformatted the flash and just did a clean install where I only migrated my files in /config/shares.  So I recreated all my users and reassigned all drives to their proper slots.  There must be something in one of the share cfg files.

 

I just reformatted the flash again for a third go around.  I even ran a check disk.  I will now unzip unRAID 6.0.1 to the flash drive and see if a pristine install will get the array running.  This means I will recreate all shares, settings, users, and disk assignments.  BTW, the downloaded unRAID 6.0.1 zip file MD5 checksum matches with what is on the download page.

 

I will report back to see if a pristine install will get the array up and running.  BTW, I am using the latest Chrome browser on Windows 7 Ultimate if that matters.

[mifronte]

Third time is NOT a charm 

 

This third time, I prep the flash as a brand new install.  So the only thing on the flash are what is in the unRAID download.  I only edited the ident.cfg to change Tower to my preferred server name.

 

After booting, I went ahead and configured all my settings as follows:

[*]Added users

[*]Configured Date & Time

[*]Edit Disk Settings to enable auto start, spin group, and modify Tunable to 600

[*]Edit Identification to add model information

[*]Edit Global Shares settings enable user shares and cach disk

[*]Assigned all devices and edit properties (spin group, comments)

[*]Assigned cache device (cache drive and not pool)

[*]Checked parity is valid and then start

 

So server went Sping Up Disks...Start Array...Refresh SMB and then the webgui flashes and the array is still stopped.

 

So I decided to Reboot and see what happens.  The server came back up with all settings except for the array device assignments!  It remembered the assigned cache device and that's it.  All other settings like users, time & date, global shares settings were also persisted.

 

I give up unless someone has a good idea what could be going wrong.

 

Reverting back to 5.0.5. for now  :'(

 

Edit:

The log file is after the reboot and I reassigned the devices again.  This time I tried to configure a cache pool instead of a cache drive.  Same result, the array refuses to start.

beanstalk-syslog-20150824-2326.zip

[bonienl]

Can you try your steps again but without spinup groups (just leave the defaults) ?

 

[mifronte]

I had power down the server and then powered it back up and so all the array devices, with the exception of the cache devices, were unassigned again.

 

I then proceeded to reassign the array devices again, but this time, I edited the disk properties to remove the spin group, comment, and reset the spin down to the default.  I also disabled spin group on the Disk Settings.

 

After all that, the array will still not start.  Attached is the latest log.

 

Edit:

I will try one last clean install tomorrow before throwing in the towel.  This time I am just going to assign the devices and see if the array starts before changing any settings.

beanstalk-syslog-20150825-0013.zip

[mifronte]

For my final attempt I kept everything as simple as possible and still the array will not start.  Here are my steps (no other actions beyond what I listed) and attached is the latest log file.

[*]Reformat flash drive.

[*]Extract unRAID zip to flash.

[*]Ran make_bootable.bat

[*]edit ident.cfg changing Tower and WORKGROUP to my preferred names.

[*]Copied Pro license key to config folder.

[*]Verified all files and folders on flash is not marked as read only.

[*]Safely disconnect USB flash from Windows.

[*]Boot up unRAID 6.0.1 with newly prepared flash.

[*]Assigned array devices and cache drive (not pool) to match my current V5.0.5 assignments.

[*]Checked "Parity is already valid." and clicked start.

[*]Status at bottom of page reads: Spinning up all drives...Start Arrary...Restart SMB...

I don't know how much simpler I can make it and still the array will not start.  I have 27 total devices connected as follows:

• 1 parity
  
• 1 cache drive
  
• 20 assigned to array
  
• 5 unassigned outside of array
  

Another note: When I reboot, all my assigned devices with the exception of the cache drive is gone.  Guess I am stuck on unRAID V5.0.5 for now  :'(

beanstalk-syslog-20150825-0932.zip

[bonienl]

For my final attempt I kept everything as simple as possible and still the array will not start.  Here are my steps (no other actions beyond what I listed) and attached is the latest log file.

[*]Reformat flash drive.

[*]Extract unRAID zip to flash.

[*]Ran make_bootable.bat

[*]edit ident.cfg changing Tower and WORKGROUP to my preferred names.

[*]Copied Pro license key to config folder.

[*]Verified all files and folders on flash is not marked as read only.

[*]Safely disconnect USB flash from Windows.

[*]Boot up unRAID 6.0.1 with newly prepared flash.

[*]Assigned array devices and cache drive (not pool) to match my current V5.0.5 assignments.

[*]Checked "Parity is already valid." and clicked start.

[*]Status at bottom of page reads: Spinning up all drives...Start Arrary...Restart SMB...

I don't know how much simpler I can make it and still the array will not start.  I have 27 total devices connected as follows:

• 1 parity
  
• 1 cache drive
  
• 20 assigned to array
  
• 5 unassigned outside of array
  

Another note: When I reboot, all my assigned devices with the exception of the cache drive is gone.  Guess I am stuck on unRAID V5.0.5 for now  :'(

 

Likely your problem is the total number of devices present in your system.

 

In version 6.0.x the Pro license doesn't allow you to start when you have more than 26 devices. You can try version 6.1 (latest is 6.1rc6) which has a correction for this.

 

[RobJ]

He got it.  I'd physically remove power to the 2 OCZ SSDs, for now.  I would also not try to start clean, but copy your good super.dat, it has all of the drive assignments except the Cache drive.  I wanted to help more but another afib episode is making it really hard to think right now.  When you can, provide a full Diagnostics zip (Tools -> Diagnostics).

[mifronte]

Reducing the number of connected devices down to 25 did the trick!  Looks like my array is back online under V6.0.1 with the cache pool working.

 

Thanks for all the help.  Now you will probably see a few threads with some minor issues to iron out.