Hi,

 

I've been lurking the forums and i saw there was a big issue with 6.8.x causing btrfs cache volumes to "constantly loop writes" and eventually killing the ssd, is this still a thing or was it ever solved in 6.8x?

 

My unraid install has 2x btrfs mirrored, will this be an issue? Is it recommended to convert this to xfs encrypted rather then running btrfs in 6.8.3?

 

Back when 6.7.2 was the latest version the recommended cache setting was btfs thats why i have it, i have no real idea what the upside/downside of having it really is. The rest of my array is xfs encrypted.

 

Thanks.

Edited July 29, 20214 yr by je82

v6.8.x should be the same as v6.7.x for that, v6.9.x has has some changes to decrease those writes, one of them requires the pool to be reformatted.

30 minutes ago, JorgeB said:

v6.8.x should be the same as v6.7.x for that, v6.9.x has has some changes to decrease those writes, one of them requires the pool to be reformatted.

oh really? oh well i've been running my unraid for over a year straight so i can only imagine how teared down my ssds are, i might as wel run them into the dirt would the writes decrease if i remade the cache volume to use xfs encrypted rather then btrfs? what would be the downside of not having btrfs?

20 minutes ago, je82 said:

what would be the downside of not having btrfs?

You can't have multiple disks in the pool unless using btrfs. 6.9+ allows multiple pools though. I have a "fast" pool as xfs for dockers and a "cache" pool as 2x btrfs raid1 for redundancy.

10 hours ago, trurl said:

You can't have multiple disks in the pool unless using btrfs. 6.9+ allows multiple pools though. I have a "fast" pool as xfs for dockers and a "cache" pool as 2x btrfs raid1 for redundancy.

The multi pool support is cool for sure, separate the cache away from the dockers is a good idea, we all know how bad it is if you accidentally fill the cache on unraid thanks for the tip!