Errors: Unable to write to cache, unable to write to docker image

[NotSoAlien]

Hey everyone,

 

I am running Version: 6.8.0-rc9:

 

I have had this error three times  now so it must be something I'm doing wrong. When trying to manage my dockers, I am unable to delete them or make any changes to them. It just throws a system error. If I reboot my entire server everything works fine for a while but then if I starting installing or removing dockers I get the errors again.  I looked in the "fix common problems" plugin and it shows the following messages:

 

Here are my drives, they are not full or even close to full. Does anyone have any suggestions?

 

 

Thanks for your time.

[trurl]

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

[NotSoAlien]

5 minutes ago, trurl said:

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

 

Thanks, I have attached what you requested. tower-diagnostics-20191211-1952.zip

[JorgeB]

There's filesystem corruption on the cache pool, best way forward is to backup, re-format and restore data.

[NotSoAlien]

Thank you, how do I reformat it? I have googled and found a few articles saying to change the file system, save, start/stop, then change it back. I don't have any other file systems to chose from (there is just an AUTO and encrypted). Is there a guide on how I should wipe them? I looked in the FAQ and it just talks about new drives, unless I'm missing something.

[JonathanM]

3 hours ago, NotSoAlien said:

how do I reformat it?

After you are positive there are no files on the cache pool that you wish to keep or recover, you can erase them by issuing a blkdiscard /dev/sdX at the console.

At the time of the screenshot you posted, they were /dev/sdd and dev/sdf, but you need to verify those designations at the time you are ready to go with erasing them.

 

This is a very final command, don't do it until you are absolutely sure there is nothing of value on the cache pool.

 

Once again, don't do this until you are ready for the cache pool and all its contents to be gone forever. Verify your backups of that data multiple times.

[NotSoAlien]

Thanks Jonathanm,

 

I was trying stuff before you responded. It was not wise but a good learning experience.

 

What I did. *before your response*.

1.) I changed the file system in the /config/disk.cfg to XFS. on one of the drives, since I had already removed the other.

2.) I tried reformatting it with the status saying to format it, it wouldn't do it because it said it was in a cache pool.

3.) Only way I knew how to get it out of the pool was to go to Tools->New config. so I made a new config. same parity disk and same data drives but only one SSD for cache.

4.) saved everything and then it kicked off a parity refresh which I'm still waiting on. It takes my parity writes/updates about 15 hours to complete. After it finishes in a few hours I will try to move my back-up data back onto the cache drive.

5.) if that doesnt work, ill just have to re-download and re-configure all my dockers.

 

If this happens again I will def try your method as it seems 100% faster/easier/cleaner.

 

Thanks for all the help.

[trurl]

8 minutes ago, NotSoAlien said:

t kicked off a parity refresh

Since your parity was presumably already valid you could have just checked a box during New Config and it wouldn't have rebuilt parity.

 

Please don't adlib with your disk configurations. Many people have lost data that way. In your case since there was nothing to lose in your cache pool it was OK, but if you ever find yourself unsure how to proceed with your parity array, wait for advice.