mooncake_master: Warning: SetUsageMessage() never called Flags from /__w/Mooncake/Mooncake/mooncake-store/src/master.cpp: -allocation_strategy (Allocation strategy for segments, random | free_ratio_first | cxl | ssd_free_ratio_first | local_first) type: string default: "random" -allow_evict_soft_pinned_objects (Whether to allow eviction of soft pinned objects during eviction) type: bool default: true -batch_oplog_retry_timeout_sec (Maximum time to retry transient batch OpLog standby errors.) type: uint32 default: 180 -client_ttl (Seconds a client stays considered alive after the last heartbeat. If this TTL elapses without a refresh, the master treats the client as disconnected and may unmount its segments) type: int64 default: 10 -cluster_id (Cluster ID for the master service, used for kvcache persistence in HA mode) type: string default: "mooncake_cluster" -config_path (master service config file path) type: string default: "" -cxl_path (DAX device path for CXL memory) type: string default: "/dev/dax0.0" -cxl_size (CXL memory size in bytes) type: uint64 default: 8589934592 -default_kv_lease_ttl (Default lease time for kv objects. Supports raw milliseconds or duration strings with ms, s, m, or h suffixes) type: string default: "10000" -default_kv_soft_pin_ttl (Default soft pin TTL for kv objects. Supports raw milliseconds or duration strings with ms, s, m, or h suffixes) type: string default: "1800000" -dynamic_replication_admission_qps_threshold (Per-key access frequency threshold for dynamic MEMORY replication) type: double default: 0.80000000000000004 -dynamic_replication_heat_window_seconds (Heat observation window for dynamic MEMORY replication) type: uint32 default: 10 -dynamic_replication_max_memory_replicas (Maximum MEMORY replicas allowed for one dynamically replicated key) type: uint64 default: 2 -dynamic_replication_mode (Dynamic MEMORY replica fanout mode: off, observe, or enforce) type: string default: "off" -enable_cxl (Whether to enable CXL memory support) type: bool default: false -enable_disk_eviction (Enable disk eviction feature for storage backend (default: true)) type: bool default: true -enable_ha (Enable high availability using the configured HA backend) type: bool default: false -enable_http_metadata_server (Enable HTTP metadata server instead of etcd) type: bool default: false -enable_kv_events (Enable RFC #1527 KV cache event publisher over ZMQ) type: bool default: false -enable_metadata_cleanup_on_timeout (Enable cleanup of HTTP metadata (mooncake/ram/*, mooncake/rpc_meta/*) when client heartbeat times out. Works in two modes: (1) co-located (enable_http_metadata_server=true) via in-process removal, or (2) separately-deployed metadata server via async HTTP DELETE.) type: bool default: false -enable_metric_reporting (Enable periodic metric reporting) type: bool default: true -enable_multi_tenants (Enable strict multi-tenant namespace and quota admission) type: bool default: false -enable_offload (Enable offload availability) type: bool default: false -enable_oplog (Enable HA metadata replication through batch-record OpLog) type: bool default: false -enable_snapshot (Enable periodic snapshot of master data) type: bool default: false -enable_snapshot_restore (enable restore from snapshot) type: bool default: false -etcd_endpoints (Endpoints of ETCD server, separated by semicolon, required in HA mode) type: string default: "" -eviction_high_watermark_ratio (Ratio of high watermark trigger eviction in Memory) type: double default: 0.90000000000000002 -eviction_ratio (Ratio of objects to evict when Memory space is full) type: double default: 0.050000000000000003 -global_file_segment_size (Size of global NFS/3FS segment in bytes) type: int64 default: 9223372036854775807 -ha_backend_connstring (HA backend connection string. If unset, only backend_type=etcd falls back to etcd_endpoints for backward compatibility) type: string default: "" -ha_backend_type (HA backend type, e.g. etcd | redis | k8s) type: string default: "etcd" -http_metadata_server_host (Host for HTTP metadata server to bind to) type: string default: "0.0.0.0" -http_metadata_server_port (Port for HTTP metadata server to listen on) type: int32 default: 8080 -kv_events_additional_salt (Deprecated: not emitted on events; use indexer POST /register) type: string default: "" -kv_events_backend_id (backend_id for published KV events (cache owner identity)) type: string default: "" -kv_events_bind_endpoint (ZMQ PUB bind endpoint for KV events, e.g. tcp://0.0.0.0:5557) type: string default: "" -kv_events_block_size (Deprecated: not emitted on events; use indexer POST /register) type: uint32 default: 0 -kv_events_dp_rank (Deprecated: not emitted on events; use indexer POST /register) type: uint32 default: 0 -kv_events_emit_legacy_compat (Include vLLM/SGLang-compatible type/block_hashes fields) type: bool default: true -kv_events_emit_object_key (Include Mooncake object_key in published KV events) type: bool default: true -kv_events_lora_name (Deprecated: not emitted on events (no LoRA context in master)) type: string default: "" -kv_events_model_name (Deprecated: not emitted on events; use indexer POST /register) type: string default: "" -kv_events_queue_capacity (Maximum pending KV events; oldest events are dropped when the queue is full (0 = unbounded)) type: uint32 default: 65536 -kv_events_tenant_id (Deprecated: tenant_id comes from each object on events) type: string default: "default" -max_kv_soft_pin_ttl (Maximum request-level soft pin TTL for kv objects. Supports raw milliseconds or duration strings with ms, s, m, or h suffixes) type: string default: "86400000" -max_retry_attempts (Maximum number of retry attempts for failed tasks) type: uint32 default: 10 -max_threads (Maximum number of threads to use (deprecated, use rpc_thread_num)) type: int32 default: 16 -max_total_finished_tasks (Maximum number of finished tasks to keep in memory) type: uint32 default: 10000 -max_total_pending_tasks (Maximum number of pending tasks to keep in memory) type: uint32 default: 10000 -max_total_processing_tasks (Maximum number of processing tasks to keep in memory) type: uint32 default: 10000 -memory_allocator (Memory allocator for global segments, cachelib | offset) type: string default: "offset" -metrics_host (Address for the HTTP metrics/admin server to listen on. Use "::" to listen on IPv6 (and IPv4 on dual-stack hosts)) type: string default: "0.0.0.0" -metrics_port (Port for HTTP metrics server to listen on) type: int32 default: 9003 -nof_eviction_high_watermark_ratio (Ratio of high watermark trigger eviction in NoF SSD) type: double default: 0.90000000000000002 -nof_eviction_ratio (Ratio of objects to evict when NoF SSD space is full) type: double default: 0.050000000000000003 -nof_heartbeat_failures_threshold (Consecutive NoF heartbeat failures required before unmounting a NoF segment) type: uint32 default: 3 -nof_heartbeat_interval_sec (How often master probes each mounted NoF segment) type: int64 default: 10 -nof_heartbeat_probe_timeout_ms (Timeout in milliseconds for a single NoF heartbeat probe) type: uint32 default: 1000 -offload_cap_ratio (Per-cycle offload cap as a fraction of offloading_queue_limit. Controls how many objects can be queued for offload in a single eviction cycle before falling back to force-evict) type: double default: 0.5 -offload_force_evict (Force-evict objects exceeding offload cap without disk offload) type: bool default: false -offload_on_evict (Defer LOCAL_DISK offload to eviction time instead of PutEnd) type: bool default: false -offloading_queue_limit (Maximum number of objects allowed in the offloading queue per local disk segment. Increase to allow more objects to be offloaded to SSD before force-eviction kicks in) type: uint64 default: 50000 -oplog_batch_max_entries (Maximum number of committed/reserved entries in the open batch-record OpLog waiting batch.) type: uint32 default: 1024 -oplog_poll_interval_ms (Batch-record standby poll interval.) type: int32 default: 1000 -pending_task_timeout_sec (Timeout in seconds for pending tasks (0 = no timeout)) type: uint64 default: 300 -pod_name (Pod name for K8s label-based routing (default: $POD_NAME)) type: string default: "" -pod_namespace (Pod namespace for K8s label-based routing (default: $POD_NAMESPACE)) type: string default: "" -port (Port for master service to listen on (deprecated, use rpc_port)) type: int32 default: 50051 -processing_task_timeout_sec (Timeout in seconds for processing tasks (0 = no timeout)) type: uint64 default: 300 -promotion_admission_threshold (Min CountMinSketch count for a key before promotion fires (set 1 to disable second-touch gating)) type: uint32 default: 2 -promotion_max_per_heartbeat (Max promotion tasks returned to a single client per PromotionObjectHeartbeat call. Each task is a synchronous SSD-read + RDMA-write on the client; serializing them avoids blocking past the client-liveness window. Default 1 is conservative.) type: uint32 default: 1 -promotion_on_hit (Promote LOCAL_DISK-only keys to MEMORY on read access (mirror of offload_on_evict)) type: bool default: false -promotion_queue_limit (Max in-flight promotion tasks across all shards) type: uint32 default: 50000 -put_start_discard_timeout_sec (Timeout for discarding uncompleted PutStart operations) type: uint64 default: 30 -put_start_release_timeout_sec (Timeout for releasing space allocated in uncompleted PutStart operations) type: uint64 default: 600 -quota_bytes (Quota for storage backend in bytes (0 = use default 90% of capacity)) type: uint64 default: 0 -root_fs_dir (Root directory for storage backend, used in HA mode) type: string default: "" -rpc_address (Address for RPC server to bind to, required in HA mode) type: string default: "0.0.0.0" -rpc_conn_timeout_seconds (Connection timeout in seconds (0 = no timeout)) type: int32 default: 0 -rpc_enable_tcp_no_delay (Enable TCP_NODELAY for RPC connections) type: bool default: true -rpc_interface (Network interface name for RPC server address resolution. When set, its IPv4 address overrides rpc_address) type: string default: "" -rpc_port (Port for RPC server to listen on (0 = use port, preferred over port)) type: int32 default: 0 -rpc_thread_num (Number of threads for RPC server (0 = use max_threads, preferred over max_threads)) type: int32 default: 0 -snapshot_backup_dir (Optional local directory for snapshot and restore backup. If empty, local backup is disabled) type: string default: "" -snapshot_catalog_backend_connstring (Deprecated alias of --snapshot_catalog_store_connstring) type: string default: "" -snapshot_catalog_backend_type (Deprecated alias of --snapshot_catalog_store_type) type: string default: "" -snapshot_catalog_store_connstring (Optional connection string for snapshot catalog store) type: string default: "" -snapshot_catalog_store_type (Snapshot catalog store type: ''/'embedded' or 'redis' ('payload' is kept as a deprecated alias)) type: string default: "" -snapshot_child_timeout_seconds (Timeout for snapshot child process in seconds) type: uint64 default: 300 -snapshot_interval_seconds (Interval in second between periodic snapshots of master data) type: uint64 default: 600 -snapshot_object_store_type (Snapshot object store type: 'local' for local filesystem, 's3' for S3 storage) type: string default: "" -snapshot_payload_backend_type (Deprecated alias of --snapshot_object_store_type) type: string default: "" -snapshot_payload_store_type (Deprecated alias of --snapshot_object_store_type) type: string default: "" -snapshot_retention_count (Number of recent snapshots to keep (older snapshots will be automatically deleted)) type: uint32 default: 2 -tenant_quota_connector_type (Tenant quota policy connector type) type: string default: "file" -tenant_quota_connector_uri (Tenant quota policy connector URI) type: string default: "" Flags from /__w/Mooncake/Mooncake/mooncake-store/src/utils.cpp: -mmap_arena_pool_size (Arena allocator pool size in bytes) type: uint64 default: 8589934592 -use_mmap_arena_allocator (Enable the lock-free mmap arena allocator for mmap buffers) type: bool default: false Flags from /builddir/build/BUILD/gflags-2.1.2/src/gflags.cc: -flagfile (load flags from file) type: string default: "" -fromenv (set flags from the environment [use 'export FLAGS_flag1=value']) type: string default: "" -tryfromenv (set flags from the environment if present) type: string default: "" -undefok (comma-separated list of flag names that it is okay to specify on the command line even if the program does not define a flag with that name. IMPORTANT: flags in this list that have arguments MUST use the flag=value format) type: string default: "" Flags from /builddir/build/BUILD/gflags-2.1.2/src/gflags_completions.cc: -tab_completion_columns (Number of columns to use in output for tab completion) type: int32 default: 80 -tab_completion_word (If non-empty, HandleCommandLineCompletions() will hijack the process and attempt to do bash-style command line flag completion on this value.) type: string default: "" Flags from /builddir/build/BUILD/gflags-2.1.2/src/gflags_reporting.cc: -help (show help on all flags [tip: all flags can have two dashes]) type: bool default: false currently: true -helpfull (show help on all flags -- same as -help) type: bool default: false -helpmatch (show help on modules whose name contains the specified substr) type: string default: "" -helpon (show help on the modules named by this flag value) type: string default: "" -helppackage (show help on all modules in the main package) type: bool default: false -helpshort (show help on only the main module for this program) type: bool default: false -helpxml (produce an xml version of help) type: bool default: false -version (show version and build info and exit) type: bool default: false Flags from src/logging.cc: -alsologtoemail (log messages go to these email addresses in addition to logfiles) type: string default: "" -alsologtostderr (log messages go to stderr in addition to logfiles) type: bool default: false -colorlogtostderr (color messages logged to stderr (if supported by terminal)) type: bool default: false -drop_log_memory (Drop in-memory buffers of log contents. Logs can grow very quickly and they are rarely read before they need to be evicted from memory. Instead, drop them from memory as soon as they are flushed to disk.) type: bool default: true -log_backtrace_at (Emit a backtrace when logging at file:linenum.) type: string default: "" -log_dir (If specified, logfiles are written into this directory instead of the default logging directory.) type: string default: "" -log_link (Put additional links to the log files in this directory) type: string default: "" -log_prefix (Prepend the log prefix to the start of each log line) type: bool default: true -logbuflevel (Buffer log messages logged at this level or lower (-1 means don't buffer; 0 means buffer INFO only; ...)) type: int32 default: 0 -logbufsecs (Buffer log messages for at most this many seconds) type: int32 default: 30 -logemaillevel (Email log messages logged at this level or higher (0 means email all; 3 means email FATAL only; ...)) type: int32 default: 999 -logfile_mode (Log file mode/permissions.) type: int32 default: 436 -logmailer (Mailer used to send logging email) type: string default: "/bin/mail" -logtostderr (log messages go to stderr instead of logfiles) type: bool default: false -max_log_size (approx. maximum log file size (in MB). A value of 0 will be silently overridden to 1.) type: int32 default: 1800 -minloglevel (Messages logged at a lower level than this don't actually get logged anywhere) type: int32 default: 0 -stderrthreshold (log messages at or above this level are copied to stderr in addition to logfiles. This flag obsoletes --alsologtostderr.) type: int32 default: 2 -stop_logging_if_full_disk (Stop attempting to log to disk if the disk is full.) type: bool default: false Flags from src/utilities.cc: -symbolize_stacktrace (Symbolize the stack trace in the tombstone) type: bool default: true Flags from src/vlog_is_on.cc: -v (Show all VLOG(m) messages for m <= this. Overridable by --vmodule.) type: int32 default: 0 -vmodule (per-module verbose level. Argument is a comma-separated list of =. is a glob pattern, matched against the filename base (that is, name ignoring .cc/.h./-inl.h). overrides any value given by --v.) type: string default: ""