I'm also using RPI 4B 8GB + SSD, plainly running bitcoind and Fulcrum, the --fast-sync value of 5000 and 400 made the app OOM[1]. But with the value of 512, it works just fine. And the process took less than 48H.
...
[1] https://github.com/cculianu/Fulcrum/issues/96#issuecomment-1099346565
My RaspiBlitz node had running at least: Bitcoin Core, LND, Bitcoinexplorer, mempool.space, Fulcrum (so there were some memory eaters nibbling on the 8GB RAM)
Yes, I had --fast-sync in place, first with 2048 and likely changed this to 1024 as memory consumption seemed too high with the larger value for my RPi, though it didn't ever run OOM.
I used some of these settings in my fulcrum.conf:
worker_threads = 1
db_mem=1024
# for 8GB RAM
#db_max_open_files=500
db_max_open_files=300
#fast-sync = 2048
fast-sync = 1024
...
# tweaks for huge address histories
max_history=500000
max_buffer=25000000
subnets_to_exclude_from_per_ip_limits = 127.0.0.1/32, ::1/128, 192.168.178.
txhash_cache = 256
#workqueue = 15000 (default)