tomspark wrote:
That's an LXC /proc/meminfo quirk, not a SparkBox bug — the dashboard reads memory from /proc/meminfo, and inside an LXC that often reports the host's view or a tiny cgroup number unless lxcfs is wired in. Since everything actually works, it's purely cosmetic — the container really does have your 16 GB. If you want the dashboard to show the right figure, enabling lxcfs on the Proxmox host (and making sure the CT uses it) fixes /proc/meminfo for every tool inside the container, not just us. I'll note it so we can add an LXC-aware memory read down the line. (v1.6.115)