sorry about that.
here's a copy of his reply. If you need further clarification, let me know and I can probably fill that in
---------------------------
Stratis is a volume-managed filesystem and hides things in the background. If you already went through the RH134 course (System Admin 2 Course), you would have learned that stratis extrapolates and hides all the operations for you and that it manages the "PVs and LVs" physical volumes/logical style volumes on the back-end.
Stratis also manages the filesystem, so at this time, that is done using XFS. What actually happens is that stratis will create your "thin-provisioned" storage pools and filesystem and manage all of this for you and the OS and users when looking at a stratis volume/filesystem will see like 1TB available.
Unfortunately, the actual physical disks making up the stratis volumes only equate to 500GB for all stratis volumes across all storage pools. Even if there was a single pool available, 500GB < 1 TB. Once you hit the physical limit of storage space available, there is no longer room to write data to a disk, meaning those bits of data that "should" be written to a disk are discarded.
So, it is very important to monitor stratis pool sizes and how much data is actually being used vs. how much data you physically have left, so when you are approaching the limit of storage space available provided by the drives you have time to add another physical disk and make it available to the stratis storage pool.
One other thing to point out, in case it wasn't made clear in the course, Stratis was in Tech Preview for RHEL 8.2 and it remains in Tech Preview as of RHEL 8.4
https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html-single/8.4_release_not...
Hope this helps,
Travis
--------------------
in a separate replay he also included this:
...
The real issue here is that normal users and system administrators are not aware Stratis is being used on the back end, so they think they have unlimited space. It is only when you are using Stratis to manage the filesystem that you know you're out of space.
This has happened quite often to me when I was a system integrator and worked as a consultant and architect. Enterprise SAN administrators would provision LUNs to be used for a group, but these too would be thin-provisioned from a shared storage pool on the SAN. Only the SAN administrator fully knew how much physical storage was backing up the pool and it was their responsibility to add additional (physical) storage to the pool to ensure that the already provisioned LUNs had the physical space they needed.
It is a similar principle when dealing with Stratis because it is a volume-managed filesystem, you want to use the Stratis tools and not the OS-level tools to provide accurate information on how much storage space is really available.
