Made huge cache mistake

[cambriancatalyst]

Hi everyone,

 

i just swapped cases and forgot to connect my sata cache drive (setup is a cache pool using 2 nvme and 1 sata drive all in a pool). The array started and immediately began filling the remainder of the nvme drives with, what I’m assuming, was the data on the sata through a btrfs . There is now a small amount of space left on the nvmes (approx 30 gbs) but my docker containers have all gone down... I’m assuming my image is corrupted.

 

I have now shut the array down and intend to reconnect the sata cache drive for the pool. Is this the right course of action? Is it already too late?

 

Any help is appreciated if anyone can spare the time. Thanks so much!

Edited November 29, 2020 by cambriancatalyst
Edited data to be sata and elaborated on pool config

[trurl]

If possible before rebooting and preferably with the array started
Go to Tools - Diagnostics and attach the complete Diagnostics ZIP file to your NEXT post in this thread.

[cambriancatalyst]

Hi there,

 

Thanks for the reply. I reconnected the sata drive. Prior to beginning the array I was prompted by the UI that it would be cleared if I started and the drive had a blue status symbol next to it. I unselected the drive from my pool and then reselected it and it showed a green status symbol and didn't prompt that anything would be deleted. I restarted my array and everything seems to be running correctly.

 

The only issue is that now I have lost a large amount of space on the pool (went from something like 150GB free to now only 40ish). I'm assuming there are duplicate files on the pool across drives now. Is there an easy way to rectify this?

 

I've included the diagnostic below. Thanks so much for your help! Y'all are really lifesavers!

tower-diagnostics-20201129-1853.zip

 

[cambriancatalyst]

Reading up on this a little more, would it be appropriate to run a balance/scrub on the cache to attempt to recoup the lost pool space?

[JorgeB]

You can run a balance but don't see how there could be duplicate files, pools config looks correct, note that due to a btrfs bug free space reported when using a raid1 pool with an odd number of devices will almost always be incorrect.

[cambriancatalyst]

Thanks Jorge! I ended up running the balance and reclaiming the space on the dashboard output. I'm assuming this was a visual glitch. Thanks for confirming!