Reducing Size of Ceph fs

[djfreak]

Greetings,

I currently have a 1.12 TB CEPH fs which has 60GB of data on it that I don’t want to lose. However 1.11TB is much too large. I need to reduce it to 512GB or even 256GB.

I have set the target size to 512GB however the available size of the fs refuses to move from 1.11TB. Why? What else do I need to do?

Thanks in advance.

 

[Azunai333]

I think you misunderstood the "Targe Size".
With "Target Size" you configure the PG-Autoscaler [1]. This is your expected pool target size when it is completely filled. You either configure "Target Size" OR "Target Ratio" for all your pools. The Autoscaler will then calculate the optimal amount of PGs for your pool to optimize performance.

A pool in ceph will share its free space with all other pools (if they use the same storage devices and have the same amount of replicas).

[1] https://pve.proxmox.com/wiki/Deploy_Hyper-Converged_Ceph_Cluster#pve_ceph_pools

 

[djfreak]

Ok after several days of reading I think I understand. I have 16TB of physical hard drives. All of my pools are set to size 3 with a minimum of 2. Now that I understand what that means, now I understand why 16TB of hard drives gives me approx 5.3TB of actual storage. That’s almost exactly what I have. About 3.6TB of ceph storage and 1.1TB of fs storage. Not sure where the .6 are going but perhaps to my other pools. I understand that pools will draw from one another if they need to however I’ll obviously upgrade drives before then so that will probably never happen.

Here is my new question. When I buy more hard drives to add more CEPH storage, will that fs continue to grow past 1.1TB? I *really* need it to stop growing. Yes I know my ceph pool will draw from fs but I’m never going to get there so it’s pointless. Is there a way to set that fs size as a static value?