[ceph-users] Why newly added OSD need to get all historical OSDMAPs in pre-boot

Hi List, We see newly added OSD take long time to become active, it is fetching old osdmaps from monitors. Curious why it need that much historical ? The cluster is stable , no osd down/out except the ongoing cap add (one osd at a time). { "cluster_fsid": "e959f744-64be-4f4e-9606-103ccaa9d2d6", "osd_fsid": "0a2d4b9d-f6e5-4e40-be7f-16daee131a20", "whoami": 38, "state": "preboot", "oldest_map": 193222, "newest_map": 298982, "num_pgs": 0 } root@host:~# ceph osd getmap | head -5 got osdmap epoch 324825 -Xiaoxi