Hi,

 

After a power loss, I lost my vms & docker containers. My Dashboard tab says No apps available to show even with the show all apps ticked on.

 

When I start my array I have the following error:

 

btrfs error (device loop0) in cleanup_transaction:1850: errno=-5 IO failure (Error while writing out transaction)

 

The data and shares looks OK but when I try to create a new VM I got: 

 

operation failed: unable to find any master var store for loader: /usr/share/qemu/ovmf-x64/OVMF_CODE-pure-efi.fd

 

I have already tried to reset the files on my usb stick with clean ones without effects.

 

Thanks.

 

EDIT: I've attached the diagnostics as suggested

tower-diagnostics-20170912-0807.zip

Edited September 12, 20178 yr by mathieuNls

You should post your diagnostics.  

 

loop0 is corrupted (will be either the docker.img or the libvirt.img), but the underlying problem may be on the cache drive itself being corrupted.

 

And get a UPS...

Edited September 12, 20178 yr by Squid

7 minutes ago, Squid said:

You should post your diagnostics.  

 

 

Done.

 

7 minutes ago, Squid said:

loop0 is corrupted (will be either the docker.img or the libvirt.img), but the underlying problem may be on the cache drive itself being corrupted.

1

 

How can I know which one is corrupted? 

 

7 minutes ago, Squid said:

And get a UPS...

 

Definitely.....

The underlying problem is that your cache pool is showing as being 100% full, and BTRFS does not respond well to that situation.

 

@johnnie.black would be able to tell if there's corruption on the cache pool.  I don't think so to my less well-trained eyes.  But you're probably going to have to free up some space on the cache drive, and you will more than likely have to delete the docker.img file and recreate it.

6 minutes ago, Squid said:

The underlying problem is that your cache pool is showing as being 100% full, and BTRFS does not respond well to that situation.

 

 

 

I've added a second drive to my cache pool a few days before the power loss. It shows that the total size is 360 GB even though the disks sum up to 740 GB. It also says that there is 120 GB free :x.

 

Are all the disks in cache pool supposed to be the same ?

 

6 minutes ago, Squid said:

But you're probably going to have to free up some space on the cache drive, and you will more than likely have to delete the docker.img file and recreate it.

 

 

Well, that's acceptable. The newly downloaded docker containers should be able to use the data from the old ones right ?

 

Also, do you think the cache pool problem is also behind the vms problems?

 

operation failed: unable to find any master var store for loader: /usr/share/qemu/ovmf-x64/OVMF_CODE-pure-efi.fd

 

5 minutes ago, mathieuNls said:

I've added a second drive to my cache pool a few days before the power loss. It shows that the total size is 360 GB even though the disks sum up to 740 GB. It also says that there is 120 GB free :x.

 

Available space is incorrectly reported when using 2 different size devices in raid1, real available space is the size of the smaller device, 240GB, hence the out of space errors.

1 hour ago, johnnie.black said:

 

Available space is incorrectly reported when using 2 different size devices in raid1, real available space is the size of the smaller device, 240GB, hence the out of space errors.

 

For whatever reason, I thought cache pool was a raid0 config. It makes sense now. The biggest drive is marked unmountable when used alone in the cache pool. 

 

I'll save the cache data on one of the HDD, format the 480GB Drive and put the cache data back on it. Then, only use that one in the cache pool.

 

Hopefully, it'll do the trick and I'll be able to see my dockers & vms ...

 

 

32 minutes ago, mathieuNls said:

I'll save the cache data on one of the HDD, format the 480GB Drive and put the cache data back on it. Then, only use that one in the cache pool.

 

It's easier to use the remove cache device procedure:

 

https://forums.lime-technology.com/topic/46802-faq-for-unraid-v6/?do=findComment&comment=480418

 

33 minutes ago, mathieuNls said:

Hopefully, it'll do the trick and I'll be able to see my dockers & vms ...

 

You'll may still need to redo the docker and vm images.