(To be precise: A multipath pool created in a root shell can be imported in the GUI, after which it works fine; e.g., referencing devices dm-0...dm-15. Until you reboot. The dev references change; e.g., to sda...sdp.)
I just deployed it here on my unmothballed NAS and seems to do its job. I went for SCALE instead of CORE because it's based on the Linux kernel, in turn for built-in the NVMe-oF target support to serve ZVOLs to Windows efficiently via RoCE. Getting an eventual official support and UI for this (I'm currently using nvmetcli) would be interesting in the long shot.
What I'm less intrigued about is the Kubernetes stuff. Since this is never going to be a clustered system, I absolutely refuse to have a "control plane" eat 25-50% of a CPU plane for what's essentially no good reason in my personal scenario. So I hope that console jockeying Docker containers will survive any upgrades without need for manual intervention.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.