Hi All, 

 

Been using unraid for awhile now, since 6.0b13 if memory serves and the system has been decent until today. I've had 6.3.3 installed since the day it was made available and upgraded from 6.3.2 with no issues. 

 

Just today though all of a sudden my shares are completely blank, all drives were still online and when browsing them through the main -> array devices I could see all the folder structure of the original shares, although the shares were completely blank.

 

I've pulled diagnostics and attached for pre and post reboot to see if there's anything outstanding. I've removed the name of the files that were moved from the syslog but that's it. 

 

One thing I noticed is that the shares themselves are gone from within the shares folder from the diagnostics yet after the reboot they're all there.....

 

Any assistance will be greatly appreciated.

 

 

 

 

Edited May 9, 20179 yr by jaybird2203

Haven't looked at your diagnostics, but the symptoms sound like a corrupt cache drive. Try starting the array with the cache drive unassigned.

should I back up my dockers from the cache only share first?

19 minutes ago, jaybird2203 said:

should I back up my dockers from the cache only share first?

Won't hurt.

 

You've got corruption on your cache drive.  You need to check disk filesystems on it. https://lime-technology.com/wiki/index.php/Check_Disk_Filesystems  (BTW, XFS is the better file system to use on the cache drive if you're not using a pool as it is more mature (ie: stable) and less sensitive to things like unclean shutdowns, etc.

 

Also, intermittently your CPU is overheating, and to prevent it melting down it is throttling itself down.

ok cool. 

 

I'll look into the overheating on the CPU shortly, I've recently moved and the CPU/MOBO combo has been installed for ages so wouldn't be surprised if it shifted slightly recently.

Ok so ran the check disk filesystems using the scrub option given the btrfs and cache pool and it couldn't repair the error, specifically "error details: csum=1"

 

Once that was completed I stopped array, unassigned the cache drives and restarted the array. 

 

Immediately got a message from the "fix common errors" about detecting hardware errors, so installed mcelog plugin via nerdpack then restarted complete tower.

 

Didn't see the same error but pulled diagnostics and attached. 

 

Since then I've backup the contents of the cache drive and I'm open to completely destroying the pool and starting from scratch although I can't seem to find the location where I can reformat from the gui. I have the preclear plugin installed but didn't think that'd be the proper way to format the drives and start from scratch.

 

Edited May 9, 20179 yr by jaybird2203

1 minute ago, jaybird2203 said:

Since then I've backup the contents of the cache drive and I'm open to completely destroying the pool and starting from scratch although I can't seem to find the location where I can reformat from the gui.

Easiest way from the gui would be to set the number of cache slots to 1 then assign one of the SSD's and change the desired format to XFS. After you start the array you will be given the option to format all unformatted volumes, be SURE the only volume it wants to format is the single SSD in question. After it's formatted, stop the array, change to the other SSD and repeat. After both are XFS, you should be able to change the number of cache slots to account for all the drives you wish to assign to the pool, assign them all, and start the array again. It will automatically change to BTRFS when you change to multiple cache slots, and ask to format the pool.

 

After that's done, you should probably post diagnostics and someone can verify things are progressing as desired.

On 5/7/2017 at 11:06 PM, jonathanm said:

Easiest way from the gui would be to set the number of cache slots to 1 then assign one of the SSD's and change the desired format to XFS. After you start the array you will be given the option to format all unformatted volumes, be SURE the only volume it wants to format is the single SSD in question. After it's formatted, stop the array, change to the other SSD and repeat. After both are XFS, you should be able to change the number of cache slots to account for all the drives you wish to assign to the pool, assign them all, and start the array again. It will automatically change to BTRFS when you change to multiple cache slots, and ask to format the pool.

 

After that's done, you should probably post diagnostics and someone can verify things are progressing as desired.

Sweet, just did the steps and looks like the cache drives are back to the pool. Just copying the files back. 

 

Here's the diagnostics for ref.

 

Edited May 9, 20179 yr by jaybird2203