Writeback is disabled by default until a flush is encountered. Try setting "rbd_cache_writethrough_until_flush = false" in your config. -- Jason Dillaman ----- Original Message ----- > From: "Vish (Vishwanath) Maram-SSI" <vishwanath.m@xxxxxxxxxxxxxxx> > To: ceph-devel-owner@xxxxxxxxxxxxxxx, ceph-devel@xxxxxxxxxxxxxxx > Sent: Friday, February 5, 2016 11:36:29 AM > Subject: Reg:. Unable to observe the performance impact of Rbd_Cache parameter. > > Hi All, > > We are trying to experiment with Rbd_Cache parameter and expect it to improve > the performance when run on a "rbdimage" using FIO. Please find the below > details: > > 1. 1 OS - CentOS7.2 > 2. CEPH Code - Hammer Version - 0.94.5 > 3. Ceph.conf - Please find below > 4. rbd -p pool3 bench-write im3 --io-size 4096 --io-threads 256 --io-total > 10240000000 --io-pattern seq > a. With the above command we were able to observe performance improvement by > modifying Rbd_Cache (True/False) > 5. We were expecting that we could observe the same improvement if we run FIO > directly on the above pool, but we don't observe any difference with or > without Rbd_Cache (True/False) > > It would be great if someone throw some light. > > Thanks, > -Vish > > > Ceph.conf - > > > [global] > fsid = 9eda02e2-04b7-4eed-a85a-8471ea51528d > > auth_cluster_required = none > auth_service_required = none > auth_client_required = none > #auth_supported = none > cephx sign messages = false > cephx require signatures = false > > debug_lockdep = 0/0 > debug_context = 0/0 > debug_crush = 0/0 > debug_buffer = 0/0 > debug_timer = 0/0 > debug_filer = 0/0 > debug_objecter = 0/0 > debug_rados = 0/0 > debug_rbd = 0/0 > debug_ms = 0/0 > debug_monc = 0/0 > debug_tp = 0/0 > debug_auth = 0/0 > debug_finisher = 0/0 > debug_heartbeatmap = 0/0 > debug_perfcounter = 0/0 > debug_rgw = 0/0 > debug_asok = 0/0 > debug_throttle = 0/0 > > debug_journaler = 0/0 > debug_objectcatcher = 0/0 > debug_client = 0/0 > debug_osd = 0/0 > debug_optracker = 0/0 > debug_objclass = 0/0 > debug_filestore = 0/0 > debug_journal = 0/0 > debug_mon = 0/0 > debug_paxos = 0/0 > > filestore_xattr_use_omap = true > osd_pool_default_size = 1 > osd_pool_default_min_size = 1 > osd_pool_default_pg_num = 128 > osd_pool_default_pgp_num = 128 > > mon_pg_warn_max_object_skew = 10000 > mon_pg_warn_min_per_osd = 0 > mon_pg_warn_max_per_osd = 32768 > osd_pg_bits = 8 > osd_pgp_bits = 8 > > mon_compact_on_trim = false > log_to_syslog = false > log_file = /var/log/ceph/$name.log > perf = true > mutex_perf_counter = true > throttler_perf_counter = false > > [mon.a] > host = Mon1 > mon_addr = 10.10.10.150:6789 > mon_max_pool_pg_num = 166496 > mon_osd_max_split_count = 10000 > > [osd] > filestore_wbthrottle_enable = false > filestore_queue_max_bytes = 1048576000 > > filestore_queue_committing_max_bytes = 1048576000 > filestore_queue_max_ops = 5000 > filestore_queue_committing_max_ops = 5000 > filestore_max_sync_interval = 10 > filestore_fd_cache_size = 64 > filestore_fd_cache_shards = 32 > filestore_op_threads = 6 > > osd_op_threads = 32 > osd_op_num_shards = 25 > osd_op_num_threads_per_shard = 2 > osd_enable_op_tracker = false > osd_client_message_size_cap = 0 > osd_client_message_cap = 0 > objecter_inflight_ops = 102400 > objecter_inflight_op_bytes = 1048576000 > > ms_dispatch_throttle_bytes = 1048576000 > ms_nocrc = true > throttler_perf_counter = false > > [osd.0] > host = server2a > public_addr = 10.10.10.154 > > [client] > rbd_cache = true > rbd_cache_writethrough_until_flush = true > admin_socket = /var/run/ceph/$cluster-$type.$id.$pid.$cctid.asok > log_file = /var/log/ceph/ > > > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html > -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
