I recently downgraded my server to save power and make a little money so I went from 256GB of ram to 96GB. I noticed that the system is always using 35+GB of memory even on boot. I thought it wasn't an issue but when I transcode to ram, it never frees up any of that system space and my transcode will fail. I've tried everything I can think of to find out what is using the system ram but have not been able to do so.

I don't have Unraid Connect installed, but I still see lots of instances of /usr/local/unraid-api/dist/main.js for unknown reasons...

harley-diagnostics-20260930-0646.zip

Your boot configuration contains:

hugepagesz=2M hugepages=16384

This reserves 32 GiB of RAM for applications configured to use huge pages. Your diagnostics show 16,373 of those 16,384 pages still free within that reserved pool. Ordinary applications cannot use that memory, even under memory pressure.

Your saved 16 GiB VM configuration also requests huge pages. Do you still need that setup? If you do, we should review the reservation size. If you no longer need it, remove the VM’s huge-page requirement while it is stopped, remove those two parameters from the active boot entry, and reboot.

The Unraid API is expected even without Connect installed: it has been built into Unraid since 7.2. Your startup log reports one API instance; repeated entries alone do not establish a memory leak.

This explains much of the memory use at boot, but the diagnostics do not show the transcode failure. Which application are you using, what host path is mapped for RAM transcoding, and what exact error does it report?

You are spot on. I had set it up this way with the previous Xeon setup I had and totally spaced it out.