Hello everyone and Merry Christmas!

 

I'm having an issue that is infrequent but quite annoying. I have my docker image (10gb) on my cache drive which is an Intel 530 120gb. I don't actually use my cache drive as a cache it is actually my application drive really.

 

Now every now and then my docker image becomes corrupted. I know when it happens because either my dockers start acting up or they just won't start with no apparent errors in the logs.

 

If I stop docker and then run a BTRFS scrub on the image I will find errors that it says it is unable to repair. Then I have to go through the process to delete the image and re-install the dockers. OK this takes like 5 minutes but I'd like to get to the bottom of how this is occurring. Does the BTRFS scrub work if you only have one drive? Do you require two drives for fault tolerance?

BTRFS requires 2 drives for data fault tolerance. Since it's an image file on the cache drive I don't believe having mirrored cache drives would help unless it was an issue with the underlying cache drive filesystem causing it.

Thanks Lionelhutz, does the BTRFS scrub work with a single drive / image though? Can it repair errors in the file system?

 

I appreciate you don't have fault tolerance with one drive but surely it can repair errors in the file system?

Scrubbing likely will not be needed in the future if you actually get to the bottom of the problem. I suspect you are filling up your docker.img and that is how it gets corrupted. How many dockers are you running? 10G may be large enough, or you might need to make it a little larger. I have run probably a dozen or so in 20G.

 

Have you read the Docker FAQ?

No, scrubbing won't fix data corruption errors with a single filesystem. It can fix the filesystem metadata and tell you there is a data issue.