添加host:for id in {0…19};do ceph osd crush add-bucket bb_host_KaTeX parse error: Expected group after '_' at position 86: …sh move bb_host_̲id root=bb_host_group;done
将osd添加到host下面:for id in {20…39};do ceph osd crush add osd.KaTeX parse error: Expected group after '_' at position 24: …08 host=bb_host_̲id;sleep 1;done
ec pool：
创建profile:poolname=bb;ceph osd erasure-code-profile set
poolname"_host_group"
创建ruleset:poolname=bb;ceph osd crush rule create-erasure $poolname"_ruleset" $poolname"_profile"
创建资源池:poolname=bb;ceph osd pool create $poolname 128 128 erasure $poolname"_profile" $poolname"_ruleset"
设置min_size:poolname=bb;ceph osd pool set $poolname min_size 16
设置overwrite属性:poolname=bb;ceph osd pool set $poolname allow_ec_overwrites 1 --yes-i-really-mean-it
创建副本资源池,创建副本pool：
poolname=bb;ceph osd crush rule create-simple $poolname"_ruleset" $poolname"_host_group" host firstn
poolname=bb;ceph osd pool create $poolname $pg_num $poolname"_ruleset"
poolname=bb;ceph osd pool set $poolname size $type_size
poolname=bb;ceph osd pool set $poolname min_size $type_min_size
创建副本资源池,创建副本pool：
poolname=bb;ceph osd crush rule create-simple $poolname"_ruleset" $poolname"_host_group" host firstn
poolname=bb;ceph osd pool create $poolname $pg_num $poolname"_ruleset"
poolname=bb;ceph osd pool set $poolname size $type_size
poolname=bb;ceph osd pool set $poolname min_size $type_min_size
添加错了之后:
for id in {1…19};do ceph osd crush rm bb_host_$id;done
删除资源池:ceph osd pool rm aa aa --yes-i-really-really-mean-it
查看对应的rule:ceph osd crush rule ls
创建cache-tier:
ceph osd tier add storage-cold storage-hot --force-nonempty
ceph osd tier cache-mode storage-hot writeback
ceph osd tier set-overlay storage-cold storage-hot
设置tier参数:
ceph osd pool set storage-hot hit_set_type bloom
ceph osd pool set storage-hot hit_set_count 4
ceph osd pool set storage-hot target_max_bytes 1073741824 #1G //设置最大tier池的百分之八十，注意这里单位是字节
ceph osd pool set storage-hot cache_target_dirty_ratio 0.4
ceph osd pool set storage-hot cache_target_dirty_high_ratio 0.6
ceph osd pool set storage-hot cache_target_full_ratio 0.8
重新设置ruleset：重设过后会产生大量的数据重构
ceph osd pool set fs_data crush_rule fs_data1_ruleset
ceph osd pool set fs_data_meta crush_rule fs_meta1_ruleset