Hi

I am having problems with my cash drives, I sometimes get errors where Docker and VMs will not load in the web UI, When I look in my log I get

BTRFS error (device loop2) : Bad tree Block start.

This requires a complete reboot of my server. The problem then goes away for about 2 days and then comes back again.

Is this an SSD failing? 

As you can see I have 3 SSDs which are configured in software raid 5 for speed for VMs and so that each SSDs would share the loads. Do I need hardware raid which i found SilverStone SST-ECS01 https://www.amazon.co.uk/dp/B00NNPKOLE/?coliid=I17W5DQF5EQCG9&colid=13QNL4KHHVIHI&psc=1

 

I have run a smart test on all 3 SSDs and the array and i get message Completed with out an error.

Can any one help?

Thanks

Go to Tools - Diagnostics and attach the complete diagnostics zip file to your next post.

morpheus-diagnostics-20190117-2113.zip

Edited January 17, 20197 yr by Sharpey212

Unfortunately diags are after rebooting, but looks like a cache device is dropping offline, though not seeing any btrfs stat errors on mount, and they usually appear when there are some, still check with:

 

btrfs dev stats /mnt/cache

 

If all zeros grab new diags before rebooting if/when it happens again.

Also, the loop devices are vdisks, such as your docker image. Though these are contained on a disk, it is possible to corrupt them without actually having a disk problem.

 

Any particular reason you have allocated 40G to your docker image? You should never need more than 20.

I stated with 20GB and Fix Common Problems kept message says docker images getting full so kept adding 10GB to it. I need shrinking now because I have deleted a few dockers.

 

Edited January 17, 20197 yr by Sharpey212

3 minutes ago, Sharpey212 said:

I stated with 20GB and Fix Common Problems kept message says docker images getting full so kept adding 10GB to it. I need shrinking now because I have deleted a few dockers.

 

If you were filling 20GB then you definitely had something misconfigured, regardless of how many dockers you were using. Did you figure out what the real problem was? Increasing the size doesn't do anything except make it take longer to fill up.

 

See the "Docker Image Filling Up" section in the Docker FAQ:

 

https://forums.unraid.net/topic/57181-real-docker-faq/

 

I have uploaded a new file with the error morpheus-diagnostics-20190125-1851.zip

Edited January 25, 20197 yr by Sharpey212

Cache1 dropped offline:

Jan 25 18:39:23 Morpheus kernel: ata8: hard resetting link
Jan 25 18:39:28 Morpheus kernel: ata8: softreset failed (1st FIS failed)
Jan 25 18:39:28 Morpheus kernel: ata8: reset failed, giving up
Jan 25 18:39:28 Morpheus kernel: ata8.00: disabled

 

At the same time cache2 dropped offline:

Jan 25 18:39:28 Morpheus kernel: ata7: softreset failed (1st FIS failed)
Jan 25 18:39:28 Morpheus kernel: ata7: reset failed, giving up
Jan 25 18:39:28 Morpheus kernel: ata7.00: disabled
Jan 25 18:39:28 Morpheus kernel: ata7: EH complete

 

Cache pool can't continue to work with two devices missing, check cables on both devices, there could also be a problem with the Asmedia controller.

 

 

morpheus-diagnostics-20190204-2038.zip

Same problem again:

 

Feb  4 17:46:06 Morpheus kernel: ata7.00: disabled
Feb  4 17:46:06 Morpheus kernel: ata7: EH complete
Feb  4 17:46:06 Morpheus kernel: ata8: softreset failed (1st FIS failed)
Feb  4 17:46:06 Morpheus kernel: ata8: reset failed, giving up
Feb  4 17:46:06 Morpheus kernel: ata8.00: disabled

 

 

I would suggest testing with a different controller, both devices going down at the same time suggest a controller problem, or anything else the devices share, like a SATA power splitter, etc.

 

 

Edited February 4, 20197 yr by johnnie.black

Hi thanks for taking a look at my diagnostics my SSDs and the ray both connected directly to my motherboards SATA ports so if it is the controller than i think might need a new motherboard. 

Those two SSDs are connected to an Asmedia controller, it might be onboard but it's not the main 6 port Intel SATA controller, you can always buy a similar ad-don two port controller and test.