Split Two Disk Cache Pool - Data Still Mirrored?

[Technab]

Hello, everyone!

 

I originally had a two disk cache pool named "Cache" which I then split into "Cache-apps" and "Cache-ingest". I've been running into an issue where the "fix common problems" plug-in says that the "appdata" share lives on both the "cache-apps" and "cache-ingest" pools.

 

I shut down the docker service, and used rsync to sync the "appdata" folder from "cache-ingest" to "cache-apps" and then I deleted the folder from "cache-ingest". For some reason that deleted the "appdata" folder from BOTH cache drives at the same time.

 

Looking at the cache drives now, I see all of my docker containers have their folders mirrored on both drives.

 

 

I've attached my diagnostics dump. Am I doing something wrong? Did I mess something up by having the drives combined and then splitting them out? Any help would be appreciated!

tower-diagnostics-20220212-0859.zip

[JorgeB]

Both pools are currently the same filesystem, using both devices, this means the device wasn't correctly removed from the pool before assign it to the new pool, to fix this you can try this:

 

Stop the array, disable docker/VMs services, unassign the devices from both pools, start array, stop array, assign both devices to the same pool (there can't be a "all data on this device will be deleted" red warning for any pool device) , start array, stop array, unassign one the pool devices, start array so it can be removed from the pool, stop array, you can now assign the removed device to a new pool, start array, format new pool, all should be good now, if it isn't post new diags, you can also re-enable docker/VMs services.

[Technab]

7 hours ago, JorgeB said:

Both pools are currently the same filesystem, using both devices, this means the device wasn't correctly removed from the pool before assign it to the new pool, to fix this you can try this:

 

Stop the array, disable docker/VMs services, unassign the devices from both pools, start array, stop array, assign both devices to the same pool (there can't be a "all data on this device will be deleted" red warning for any pool device) , start array, stop array, unassign one the pool devices, start array so it can be removed from the pool, stop array, you can now assign the removed device to a new pool, start array, format new pool, all should be good now, if it isn't post new diags, you can also re-enable docker/VMs services.

 

Process looked daunting on the surface, but wasn't that bad! "cache-ingest" is now empty while "cache-apps" only has "appdata" which is what I expect.

 

Thank you!