For anyone who doesn't care about dataloss and loves to live on the edge.
In the example below the pool is called "danger"
You will get a health warn state in the cluster due to having a pool with no replicas but it will work, well until you lose an OSD or a node reboots.etc
I did it very specifically on my homelab for a secondary disk in VM that's used as the data disk by lancache (lancache.net), this is literally just cached copies of steam/game downloads so if/when this explodes in my face there's no great loss.
In the example below the pool is called "danger"
Code:
ceph config set global mon_allow_pool_size_one true
ceph osd pool set data_pool min_size 1
ceph osd pool set danger min_size 1
ceph osd pool set danger size 1 --yes-i-really-mean-itYou will get a health warn state in the cluster due to having a pool with no replicas but it will work, well until you lose an OSD or a node reboots.etc
I did it very specifically on my homelab for a secondary disk in VM that's used as the data disk by lancache (lancache.net), this is literally just cached copies of steam/game downloads so if/when this explodes in my face there's no great loss.