CephL版Config

本文档提供 Ceph Luminous (L版) 版本的完整配置文件示例，包含性能优化和常用配置项。

Ceph L 版配置文件

ceph.conf 配置文件示例：

## ceph.conf [global] fsid = 00bf66fc-dec3-4f55-8e90-eb2fd3c4725e mon_initial_members = sze0-opsinit-test01, sze0-opsinit-test02, sze0-opsinit-test03 mon_host = 10.225.97.5,10.225.97.6,10.225.97.7 auth_cluster_required = cephx auth_service_required = cephx auth_client_required = cephx public_network = 10.225.97.0/24 cluster_network = 10.225.97.0/24 mon_max_pg_per_osd = 2000 rbd_default_features = 3 rbd_cache = false rbd_cache_writethrough_until_flush = false rbd_default_format = 2 # cephx require signatures = False # cephx sign messages = False ms_crc_header = False ms_crc_data = False ms type = async perf = True rocksdb_perf = True [osd] objecter inflight ops = 0 osd objectstore = bluestore osd_min_pg_log_entries = 10 osd_max_pg_log_entries = 10 osd_pg_log_dups_tracked = 10 osd_pg_log_trim_min = 10 osd_pool_default_size = 3 osd_pool_default_min_size = 1 osd_pool_default_pg_num = 1024 osd_pool_default_pgp_num = 1024 osd memory target = 9415205847 osd_max_backfills = 1 osd_recovery_op_priority = 10 osd_max_write_size = 512 osd_client_message_size_cap = 2147483648 osd_deep_scrub_stride = 131072 osd_heartbeat_interval = 6 osd_map_cache_size = 1024 osd_client_op_priority = 63 osd_recovery_max_active = 1 osd_recovery_max_chunk = 131072 osd_recovery_threads = 2 osd_scrub_chunk_min = 5 osd_scrub_chunk_max = 10 osd_scrub_sleep = 0.2 osd_deep_scrub_interval = 1209600.000000 osd_scrub_begin_hour = 0 osd_scrub_end_hour = 24 osd_recovery_sleep = 0.5 osd_enable_op_tracker = true osd_num_op_tracker_shard = 32 osd_op_threads = 8 osd_disk_thread_ = 1 osd_op_num_shards = 24 osd_op_num_threads_per_shard = 2 osd_op_thread_timeout = 100 osd_op_thread_suicide_timeout = 300 osd_recovery_thread_timeout = 100 osd_recovery_thread_suicide_timeout = 300 bluestore_cache_size = 0 bluestore_cache_autotune = 0 bluestore_csum_type = none bluestore_cache_kv_max = 200G bluestore_cache_kv_ratio = 0.2 bluestore_cache_meta_ratio = 0.8 bluestore_cache_size_ssd = 20G bluestore_cache_size_hdd = 10G bluestore_extent_map_shard_min_size = 50 bluestore_extent_map_shard_max_size = 200 bluestore_extent_map_shard_target_size = 100 bluestore_rocksdb_options = compression=kNoCompression,max_write_buffer_number=64,\ min_write_buffer_number_to_merge=32,recycle_log_file_num=64,\ compaction_style=kCompactionStyleLevel,write_buffer_size=4MB,\ target_file_size_base=4MB,max_background_compactions=64,\ level0_file_num_compaction_trigger=64,level0_slowdown_writes_trigger=128,\ level0_stop_writes_trigger=256,max_bytes_for_level_base=6GB,\ compaction_threads=32,flusher_threads=8,compaction_readahead_size=2MB [mon] mon_allow_pool_delete = true # mon pg warn max per osd = 800 mon_clock_drift_allowed = 2 mon_osd_allow_primary_affinity = true mon_osd_down_out_interval = 3600 mon_max_pool_pg_num = 166496 mon_osd_max_split_count = 10000 [client] rbd_cache = false rbd_op_threads = 8 # rbd_cache_size = 268435456 rbd_cache_writethrough_until_flush = false # rbd cache max dirty object = 8