Apache Impala 3.4.0 —— 集群可调参数

FlagDescriptionDefaultExperimental
use_local_catalog (bool)Use experimental implementation of a local catalog. If this is set, the catalog service is not used and does not need to be started.FALSETRUE
ssl_client_ca_certificate (string)(Advanced) The full path to a certificate used by Thrift clients to check the validity of a server certificate. May either be a certificate for a third-party Certificate Authority, or a copy of the certificate the client expects to receive from the server.FALSE
webserver_password_file (string)(Optional) Location of .htpasswd file containing user names and hashed passwords for debug webserver authenticationFALSE
s3a_access_key_cmd (string)A Unix command whose output returns the access key to S3, i.e. “fs.s3a.access.key”.FALSE
ssl_private_key_password_cmd (string)A Unix command whose output returns the password used to decrypt the certificate private key file specified in --ssl_private_key. If the .PEM key file is not password-protected, this command will not be invoked. The output of the command will be truncated to 1024 bytes, and then all trailing whitespace will be trimmed before it is used to decrypt the private keyFALSE
webserver_private_key_password_cmd (string)A Unix command whose output returns the password used to decrypt the Webserver’s certificate private key file specified in --webserver_private_key_file. If the .PEM key file is not password-protected, this command will not be invoked. The output of the command will be truncated to 1024 bytes, and then all trailing whitespace will be trimmed before it is used to decrypt the private keyFALSE
s3a_secret_key_cmd (string)A Unix command whose output returns the secret key to S3, i.e. “fs.s3a.secret.key”.FALSE
keytab_file (string)Absolute path to Kerberos keytab fileFALSE
krb5_conf (string)Absolute path to Kerberos krb5.conf if in a non-standard location. Does not normally need to be set.FALSE
redaction_rules_file (string)Absolute path to sensitive data redaction rules. The rules will be applied to all log messages and query text shown in the Web UI and audit records. Query results will not be affected. Refer to the documentation for the rule file format.FALSE
sasl_path (string)Colon separated list of paths to look for SASL security library plugins.FALSE
blacklisted_tables (string)Comma separated full names (in format: .) of blacklisted tables. Configure which tables to be skipped for loading (in startup and reseting metadata of the table). Users can’t access, create, or drop tables which are blacklisted
FALSE
undefok (string)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 formatFALSE
env_inject_lock_failure_globs (string)Comma-separated list of glob patterns specifying files on which attempts to obtain a file lock will fail. By default, no files will fail.FALSE
query_event_hook_classes (string)Comma-separated list of java QueryEventHook implementation classes to load and register at Impala startup. Class names should be fully-qualified and on the classpath. Whitespace acceptable around delimiters.FALSE
webserver_authentication_domain (string)Domain used for debug webserver authenticationFALSE
log_backtrace_at (string)Emit a backtrace when logging at file:linenum.FALSE
disk_reserved_override_prefix_1_path_for_testing (string)For testing only! Specifies a prefix to override the visible ‘bytes free’ on. Use --disk_reserved_override_prefix_1_bytes_free_for_testing to set the number of bytes free for this path prefix. Set to empty string to disable.FALSE
disk_reserved_override_prefix_2_path_for_testing (string)For testing only! Specifies a prefix to override the visible ‘bytes free’ on. Use --disk_reserved_override_prefix_2_bytes_free_for_testing to set the number of bytes free for this path prefix. Set to empty string to disable.FALSE
debug_actions (string)For testing only. Uses the same format as the debug action query options, but allows for injection of debug actions in code paths where query options are not available.FALSE
hdfs_zone_info_zip (string)HDFS/S3A/ADLS path to a zip archive of the IANA time-zone database to use.FALSE
hdfs_zone_alias_conf (string)HDFS/S3A/ADLS path to config file defining non-standard time-zone aliases.FALSE
hostname (string)Hostname to use for this daemon, also used as part of the Kerberos principal, if enabled. If not set, the system default will be usedFALSE
tab_completion_word (string)If non-empty, HandleCommandLineCompletions() will hijack the process and attempt to do bash-style command line flag completion on this value.FALSE
ldap_bind_pattern (string)If set, Impala will try to bind to LDAP with a name of <ldap_bind_pattern>, but where the string #UID is replaced by the user ID. Use to control the bind name precisely; do not set --ldap_domain or --ldap_baseDN with this optionFALSE
ldap_domain (string)If set, Impala will try to bind to LDAP with a name of the form @<ldap_domain>FALSE
ldap_baseDN (string)If set, Impala will try to bind to LDAP with a name of the form uid=,<ldap_baseDN>FALSE
asm_module_dir (string)if set, saves disassembly for generated IR modules to the specified directory.FALSE
opt_module_dir (string)if set, saves optimized generated IR modules to the specified directory.FALSE
unopt_module_dir (string)if set, saves unoptimized generated IR modules to the specified directory.FALSE
log_dir (string)If specified, logfiles are written into this directory instead of the default logging directory.FALSE
webserver_interface (string)Interface to start debug webserver on. If blank, webserver binds to 0.0.0.0FALSE
local_ip_for_outbound_sockets (string)IP to bind to when making outgoing socket connections. This must be an IP address of the form A.B.C.D, not a hostname. Advanced parameter, subject to change.FALSE
be_principal (string)Kerberos principal for backend network connections only,overriding --principal if set. Must not be set if --principal is not set.FALSE
principal (string)Kerberos principal. If set, both client and backend network connections will use Kerberos encryption and authentication and the daemon will acquire a Kerberos TGT (i.e. do the equivalent of the kinit command) and keep it refreshed for the lifetime of the daemon. If this is not set the TGT ticket will not be acquired and incoming connections will not be authenticated or encrypted using Kerberos. However, the TGT and other settings may be inherited from the environment and used by client libraries in certain cases.FALSE
default_query_options (string)key=value pair of default query options for impalad, separated by ‘,’FALSE
executor_groups (string)List of executor groups, separated by comma. Each executor group specification can optionally contain a minimum size, separated by a ‘:’, e.g. --executor_groups default-pool-1:3. Default minimum size is 1. Only when the cluster membership contains at least that number of executors for the group will it be considered healthy for admission. Currently only a single group may be specified.FALSE
flagfile (string)load flags from fileFALSE
sentry_config (string)Local path to a sentry-site.xml configuration file. If set, authorization will be enabled.FALSE
alsologtoemail (string)log messages go to these email addresses in addition to logfilesFALSE
default_pool_mem_limit (string)Maximum amount of memory that all outstanding requests in this pool may use before new requests to this pool are queued. Specified as number of bytes (’[bB]?’), megabytes (’[mM]’), gigabytes (’[gG]’), or percentage of the physical memory (’%’). Defaults to bytes if no unit is given. Ignored if fair_scheduler_config_path and llama_site_path are set.FALSE
heap_profile_dir (string)Output directory to store heap profiles. If not set profiles are stored in the current working directory.FALSE
fair_scheduler_allocation_path (string)Path to the fair scheduler allocation file (fair-scheduler.xml).FALSE
llama_site_path (string)Path to the Llama configuration file (llama-site.xml). If set, fair_scheduler_allocation_path must also be set.FALSE
vmodule (string)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.FALSE
log_filename (string)Prefix of log filename - full path is <log_dir>/<log_filename>.[INFOWARNERROR
log_link (string)Put additional links to the log files in this directoryFALSE
fromenv (string)set flags from the environment [use ‘export FLAGS_flag1=value’]FALSE
tryfromenv (string)set flags from the environment if presentFALSE
helpmatch (string)show help on modules whose name contains the specified substrFALSE
helpon (string)show help on the modules named by this flag valueFALSE
authorization_factory_class (string)Specifies the class name that implements the authorization provider. This will override the authorization_provider flag if both are specified.FALSE
kudu_master_hosts (string)Specifies the default Kudu master(s). The given value should be a comma separated list of hostnames or IP addresses; ports are optional.FALSE
ranger_app_id (string)Specifies the Ranger application ID. Ranger application ID is an ID to uniquely identify the application that communicates with Ranger. This flag is required when authorization with Ranger is enabled.FALSE
authorized_proxy_group_config (string)Specifies the set of authorized proxy groups (users who can delegate to other users belonging to the specified groups during authorization) and whom they are allowed to delegate. Input is a semicolon-separated list of key=value pairs of authorized proxy users to the group(s) they can delegate to. These groups are specified as a list of groups separated by a delimiter (which defaults to comma and may be changed via --authorized_proxy_group_config_delimiter), or '’ to indicate all users. For example: hue=group1,group2;admin=FALSE
authorized_proxy_user_config (string)Specifies the set of authorized proxy users (users who can delegate to other users during authorization) and whom they are allowed to delegate. Input is a semicolon-separated list of key=value pairs of authorized proxy users to the user(s) they can delegate to. These users are specified as a list of short usernames separated by a delimiter (which defaults to comma and may be changed via --authorized_proxy_user_config_delimiter), or '’ to indicate all users. For example: hue=user1,user2;admin=FALSE
ssl_cipher_list (string)The cipher suite preferences to use for TLS-secured Thrift RPC connections. Uses the OpenSSL cipher preference list format. See man (1) ciphers for more information. If empty, the default cipher list for your platform is usedFALSE
data_cache (string)The configuration string for IO data cache. Default to be an empty string so it’s disabled. The configuration string is expected to be a list of directories, separated by ‘,’, followed by a ‘:’ and a capacity quota per directory. For example /data/0,/data/1:1TB means the cache may use up to 2TB, with 1TB max in /data/0 and /data/1 respectively. Please note that each Impala daemon on a host must have a unique caching directory.FALSE
audit_event_log_dir (string)The directory in which audit event log files are written. Setting this flag will enable audit event logging.FALSE
lineage_event_log_dir (string)The directory in which lineage event log files are written. Setting this flag with enable lineage logging.FALSE
profile_log_dir (string)The directory in which profile log files are written. If blank, defaults to <log_file_dir>/profilesFALSE
ldap_ca_certificate (string)The full path to the certificate file used to authenticate the LDAP server’s certificate for SSL / TLS connections.FALSE
ssl_private_key (string)The full path to the private key used as a counterpart to the public key contained in --ssl_server_certificate. If --ssl_server_certificate is set, this option must be set as well.FALSE
webserver_private_key_file (string)The full path to the private key used as a counterpart to the public key contained in --webserver_certificate_file. If --webserver_certificate_file is set, this option must be set as well.FALSE
ssl_server_certificate (string)The full path to the SSL certificate file used to authenticate Impala to clients. If set, both Beeswax and HiveServer2 ports will only accept SSL connectionsFALSE
webserver_certificate_file (string)The location of the debug webserver’s SSL certificate file, in .pem format. If empty, webserver SSL support is not enabledFALSE
server_name (string)The name to use for securing this impalad server during authorization. Set to enable authorization.FALSE
ldap_uri (string)The URI of the LDAP server to authenticate users againstFALSE
trace_to_console (string)Trace pattern specifying which trace events should be dumped directly to the consoleFALSE
krb5_debug_file (string)Turn on Kerberos debugging and output to this fileFALSE
env_inject_eio_globs (string)Comma-separated list of glob patterns specifying files on which I/O will fail. By default, all files may cause a failure.*FALSE
env_inject_full_globs (string)Comma-separated list of glob patterns specifying which paths return with space errors.*FALSE
authorized_proxy_group_config_delimiter (string)Specifies the delimiter used in authorized_proxy_group_config.FALSE
authorized_proxy_user_config_delimiter (string)Specifies the delimiter used in authorized_proxy_user_config.FALSE
disk_reserved_override_prefix_1_bytes_free_for_testing (int64)For testing only! Set number of bytes free on the path prefix specified by --disk_reserved_override_prefix_1_path_for_testing. Set to -1 to disable.-1FALSE
disk_reserved_override_prefix_2_bytes_free_for_testing (int64)For testing only! Set number of bytes free on the path prefix specified by --disk_reserved_override_prefix_2_path_for_testing. Set to -1 to disable.-1FALSE
disk_reserved_bytes_free_for_testing (int64)For testing only! Set to number of bytes free on each filesystem. Set to -1 to disable this test-specific override-1FALSE
rpc_inject_cancellation_state (int32)If this flag is not -1, it is the state in which a cancellation request will be injected. Should use values in OutboundCall::State only-1FALSE
default_pool_max_requests (int64)Maximum number of concurrent outstanding requests allowed to run before queueing incoming requests. A negative value indicates no limit. 0 indicates no requests will be admitted. Ignored if fair_scheduler_config_path and llama_site_path are set.-1FALSE
logmailer (string)Mailer used to send logging email/bin/mailFALSE
webserver_doc_root (string)Files under <webserver_doc_root>/www are accessible via the debug webserver. Defaults to $IMPALA_HOME, or if $IMPALA_HOME is not set, disables the document root/home/q/apache-impala-3.4.0FALSE
local_library_dir (string)Scratch space for local fs operations. Currently used for copying UDF binaries locally from HDFS and also for initializing the timezone db/tmpFALSE
scratch_dirs (string)Writable scratch directories. This is a comma-separated list of directories. Each directory is specified as the directory path and an optional limit on the bytes that will be allocated in that directory. If the optional limit is provided, the path and the limit are separated by a colon. E.g. ‘/dir1:10G,/dir2:5GB,/dir3’ will allow allocating up to 10GB of scratch in /dir1, 5GB of scratch in /dir2 and an unlimited amount in /dir3./tmpFALSE
krb5_ccname (string)Absolute path to the file based credentials cache that we pass to the KRB5CCNAME environment variable./tmp/krb5cc_impala_internalFALSE
tcmalloc_max_total_thread_cache_bytes (int64)(Advanced) Bound on the total amount of bytes allocated to TCMalloc thread caches. If left at 0 (default), use the default value in TCMalloc library.0FALSE
num_cores (int32)(Advanced) If > 0, it sets the number of cores available to Impala. Setting it to 0 means Impala will use all available cores on the machine according to /proc/cpuinfo.0FALSE
non_impala_java_vlog (int32)(Advanced) The log level (equivalent to --v) for non-Impala Java classes (0: INFO, 1 and 2: DEBUG, 3: TRACE)0FALSE
admission_control_slots (int32)(Advanced) The maximum degree of parallelism to run queries with on this backend. This determines the number of slots available to queries in admission control for this backend. The degree of parallelism of the query determines the number of slots that it needs. Defaults to number of cores / -num_cores for executors, and 8x that value for dedicated coordinators).0FALSE
catalog_client_rpc_timeout_ms (int32)(Advanced) The underlying TSocket send/recv timeout in milliseconds for a catalog client RPC.0FALSE
kudu_client_rpc_timeout_ms (int32)(Advanced) Timeout (milliseconds) set for individual Kudu client rpcs. An operation may consist of several rpcs, so this is expected to be less than kudu_operation_timeout_ms. This must be a positive value or it will be ignored and Kudu’s default of 10s will be used. There is no way to disable timeouts.0FALSE
max_concurrent_queries (int32)(Deprecated) This has been replaced with --admission_control_slots, which better accounts for the higher parallelism of queries with mt_dop > 1. If --admission_control_slots is not set, the value of --max_concurrent_queries is used instead for backward compatibility.0FALSE
logbuflevel (int32)Buffer log messages logged at this level or lower (-1 means don’t buffer; 0 means buffer INFO only; …)0FALSE
hms_event_polling_interval_s (int32)Configure catalogd to invalidate cached table metadata based on metastore events. These metastore events could be generated by external systems like Apache Hive or a different Impala cluster using the same Hive metastore server as this one. A non-zero value of this flag sets the polling interval of catalogd in seconds to fetch new metastore events. A value of zero disables this feature. When enabled, this flag has the same effect as “INVALIDATE METADATA” statement on the table for certain metastore event types. Additionally, in case of events which detect creation or removal of objects from metastore, catalogd adds or removes such objects from its cached metadata. This feature is independent of time and memory based automatic invalidation of tables. Note that this is still an experimental feature and not recommended to be deployed on production systems until it is made generally available.0FALSE
env_inject_eio (double)Fraction of the time that operations on certain files will fail with the posix code EIO.0FALSE
env_inject_full (double)Fraction of the time that space checks on certain paths will yield the posix code ENOSPC.0FALSE
invalidate_tables_timeout_s (int32)If a table has not been referenced in a SQL statement for more than the configured amount of time, the catalog server will automatically evict its cached metadata about this table. This has the same effect as a user-initiated “INVALIDATE METADATA” statement on the table. Configuring this to 0 disables time-based automatic invalidation of tables. This is independent from memory-based invalidation configured by invalidate_tables_on_memory_pressure. To enable this feature, a non-zero flag must be applied to both catalogd and impalad.0FALSE
rpc_negotiation_inject_delay_ms (int32)If enabled, injects the given number of milliseconds delay into the RPC negotiation process on the server side.0FALSE
log_mem_usage_interval (int32)If non-zero, impalad will output memory usage every log_mem_usage_interval’th fragment completion.0FALSE
rpc_inject_invalid_authn_token_ratio (double)If set higher than 0, AuthenticateByToken() randomly injects errors replying with FATAL_INVALID_AUTHENTICATION_TOKEN code. The flag’s value corresponds to the probability of the fault injection event. Used for only for tests.0FALSE
memory_limit_hard_bytes (int64)Maximum amount of memory this daemon should use, in bytes. A value of 0 autosizes based on the total system memory. A value of -1 disables all memory limiting.0FALSE
max_audit_event_log_files (int32)Maximum number of audit event log files to retain. The most recent audit event log files are retained. If set to 0, all audit event log files are retained.0FALSE
minloglevel (int32)Messages logged at a lower level than this don’t actually get logged anywhere0FALSE
num_disks (int32)Number of disks on data node.0FALSE
num_threads_per_disk (int32)Number of I/O threads per disk0FALSE
num_io_threads_per_rotational_disk (int32)Number of I/O threads per rotational disk. Has priority over num_threads_per_disk. If neither is set, defaults to 1 thread(s) per rotational disk0FALSE
num_io_threads_per_solid_state_disk (int32)Number of I/O threads per solid state disk. Has priority over num_threads_per_disk. If neither is set, defaults to 8 thread(s) per solid state disk0FALSE
inject_latency_on_kernel_stack_lookup_ms (int32)Number of milliseconds of latency to inject when reading a thread’s kernel stack0FALSE
thread_inject_start_latency_ms (int32)Number of ms to sleep when starting a new thread. (For tests).0FALSE
num_reactor_threads (int32)Number of threads dedicated to managing network IO for RPC services. If left at default value 0, it will be set to number of CPU cores.0FALSE
control_service_num_svc_threads (int32)Number of threads for processing control service’s RPCs. if left at default value 0, it will be set to number of CPU cores. Set it to a positive value to change from the default.0FALSE
datastream_service_num_svc_threads (int32)Number of threads for processing datastream services’ RPCs. If left at default value 0, it will be set to number of CPU cores. Set it to a positive value to change from the default.0FALSE
v (int32)Show all VLOG(m) messages for m <= this. Overridable by --vmodule.0FALSE
max_row_batches (int32)the maximum number of batches to queue in multithreaded HDFS scans0FALSE
kudu_max_row_batches (int32)The maximum size of the row batch queue, for Kudu scanners.0FALSE
env_inject_short_read_bytes (int32)The number of bytes less than the requested bytes to read0FALSE
env_inject_short_write_bytes (int32)The number of bytes less than the requested bytes to write0FALSE
rpc_inject_invalid_channel_bindings_ratio (double)The ratio of injection of invalid channel bindings during connection negotiation. This is a test-only flag.0FALSE
idle_query_timeout (int32)The time, in seconds, that a query may be idle for (i.e. no processing work is done and no updates are received from the client) before it is cancelled. If 0, idle queries are never expired. The query option QUERY_TIMEOUT_S overrides this setting, but, if set, --idle_query_timeout represents the maximum allowable timeout.0FALSE
idle_session_timeout (int32)The time, in seconds, that a session may be idle for before it is closed (and all running queries cancelled) by Impala. If 0, idle sessions are never expired. It can be overridden by the query option ‘idle_session_timeout’ for specific sessions0FALSE
cache_memtracker_approximation_ratio (double)The MemTracker associated with a cache can accumulate error up to this ratio to improve performance. For tests.0.01FALSE
min_filter_reject_ratio (double)(Advanced) If the percentage of rows rejected by a runtime filter drops below this value, the filter is disabled.0.1FALSE
max_filter_error_rate (double)(Advanced) The maximum probability of false positives in a runtime filter before it is disabled.0.75FALSE
umask (string)The umask that will be used when creating files and directories. Permissions of top-level data directories will also be modified at start-up to conform to the given umask. Changing this value may enable unauthorized local users to read or modify data stored by Kudu.77FALSE
data_cache_write_concurrency (int32)(Advanced) Number of concurrent threads allowed to insert into the cache per partition.1FALSE
query_event_hook_nthreads (int32)Number of threads to use for QueryEventHook execution. If this number is >1 then hooks will execute concurrently.1FALSE
num_check_authorization_threads (int32)The number of threads used to check authorization for the user when executing show tables/databases. This configuration is applicable only when authorization is enabled. A value of 1 disables multi-threaded execution for checking authorization.However, a small value of larger than 1 may limit the parallism of FE requests when checking authorization with a high concurrency. The value must be in the range of 1 to 128.1FALSE
statestore_num_heartbeat_threads (int32)(Advanced) Number of threads used to send heartbeats in parallel to all registered subscribers.10FALSE
statestore_num_update_threads (int32)(Advanced) Number of threads used to send topic updates in parallel to all registered subscribers.10FALSE
statestore_num_priority_update_threads (int32)(Advanced) Number of threads used to send prioritized topic updates in parallel to all registered subscribers.10FALSE
statestore_max_missed_heartbeats (int32)Maximum number of consecutive heartbeat messages an impalad can miss before being declared failed by the statestore.10FALSE
max_log_files (int32)Maximum number of log files to retain per severity level. The most recent log files are retained. If set to 0, all log files are retained.10FALSE
max_profile_log_files (int32)Maximum number of profile log files to retain. The most recent log files are retained. If set to 0, all log files are retained.10FALSE
tcp_keepalive_retry_count (int32)The maximum number of keepalive probes sent before declaring the remote end as dead. Effective only if --tcp_keepalive_probe_period_s is not 0.10FALSE
catalog_client_connection_num_retries (int32)The number of times connections or RPCs to the catalog should be retried.10FALSE
statestore_subscriber_cnxn_attempts (int32)The number of times to retry an RPC connection to the statestore. A setting of 0 means retry indefinitely10FALSE
buffer_pool_clean_pages_limit (string)(Advanced) Limit on bytes of clean pages that will be accumulated in the buffer pool. Specified as number of bytes (’[bB]?’), megabytes (’[mM]’), gigabytes (’[gG]’), or percentage of the buffer pool limit (’%’). Defaults to bytes if no unit is given…10%FALSE
statestore_priority_update_frequency_ms (int32)(Advanced) Frequency (in ms) with which the statestore sends prioritized topic updates to subscribers.100FALSE
statestore_heartbeat_frequency_ms (int32)(Advanced) Frequency (in ms) with which the statestore sends heartbeat heartbeats to subscribers.1000FALSE
data_cache_max_opened_files (int32)(Advanced) The maximum number of allowed opened files. This must be at least the number of specified partitions.1000FALSE
runtime_filter_wait_time_ms (int32)(Advanced) the maximum time, in ms, that a scan node will wait for expected runtime filters to arrive.1000FALSE
result_tracker_gc_interval_ms (int64)Interval at which the result tracker will look for entries to GC.1000FALSE
rpc_duration_too_long_ms (int32)Threshold (in milliseconds) above which a RPC is considered too long and its duration and method name are logged at INFO level. The time measured is between when a RPC is accepted and when its call handler completes.1000FALSE
accepted_cnxn_queue_depth (int32)(Advanced) The size of the post-accept, pre-setup connection queue in each thrift server set up to service Impala internal and external connections.10000FALSE
pause_monitor_warn_threshold_ms (int64)If the pause monitor sleeps more than this time period, a warning is logged. If set to 0 or less, pause monitor is disabled.10000FALSE
fs_slow_read_log_threshold_ms (int64)Log diagnostics about I/Os issued via the HDFS client that take longer than this threshold.10000FALSE
datastream_service_deserialization_queue_size (int32)Number of deferred RPC requests that can be enqueued before being processed by a deserialization thread.10000FALSE
memory_maintenance_sleep_time_ms (int32)Sleep time in milliseconds between memory maintenance iterations10000FALSE
statestore_max_subscribers (int32)Used to control the maximum size of the pending topic-update queue. There is at most one entry per subscriber.10000FALSE
max_result_cache_size (int64)Maximum number of query results a client may request to be cached on a per-query basis to support restarting fetches. This option guards against unreasonably large result caches requested by clients. Requests exceeding this maximum will be rejected.100000FALSE
rpc_callback_max_cycles (int64)The maximum number of cycles for which an RPC callback should be allowed to run without emitting a warning. (Advanced debugging option)100000000FALSE
data_source_batch_size (int32)Batch size for calls to GetNext() on external data sources.1024FALSE
webserver_max_post_length_bytes (int32)The maximum length of a POST request that will be accepted by the embedded web server.1048576FALSE
web_log_bytes (int64)The maximum number of bytes to display on the debug webserver’s log page1048576FALSE
exchg_node_buffer_size_bytes (int64)(Advanced) Maximum size of per-query receive-side buffer10485760FALSE
kudu_mutation_buffer_size (int32)The size (bytes) of the Kudu client buffer for mutations.10485760FALSE
kudu_error_buffer_size (int32)The size (bytes) of the Kudu client buffer for returning errors, with a min of 1KB.If the actual errors exceed this size the query will fail.10485760FALSE
data_cache_file_max_size_bytes (int64)(Advanced) The maximum size which a cache file can grow to before data stops being appended to it.1.09951E+12FALSE
initial_hms_cnxn_timeout_s (int32)Number of seconds catalogd will wait to establish an initial connection to the HMS before exiting.120FALSE
shutdown_grace_period_s (int64)Shutdown startup grace period in seconds. When the shutdown process is started for this daemon, it will wait for at least the startup grace period before shutting down. This gives time for updated cluster membership information to propagate to all coordinators and for fragment instances that were scheduled based on old cluster membership to start executing (and therefore be reflected in the metrics used to detect quiescence).120FALSE
datastream_sender_timeout_ms (int32)(Advanced) The time, in ms, that can elapse before a plan fragment will time-out trying to send the initial row batch.120000FALSE
impala_slow_rpc_threshold_ms (int64)(Advanced) Threshold for considering Impala internal RPCs to be unusually slow. Slow RPCs trigger additional logging and other diagnostics. Lowering this value may result in false positivesThis overrides KRPC’s --rpc_duration_too_long_ms setting.120000FALSE
metrics_retirement_age_ms (int32)The minimum number of milliseconds a metric will be kept for after it is no longer active. (Advanced option)120000FALSE
trusted_subnets (string)A trusted subnet whitelist. If set explicitly, all unauthenticated or unencrypted connections are prohibited except the ones from the specified address blocks. Otherwise, private network (127.0.0.0/8, etc.) and local subnets of all local network interfaces will be used. Set it to ‘0.0.0.0/0’ to allow unauthenticated/unencrypted connections from all remote IP addresses. However, if network access is not otherwise restricted by a firewall, malicious users may be able to gain unauthorized access.127.0.0.0/8,10.0.0.0/8,172.16.0.0/12,192.168.0.0/16,169.254.0.0/16FALSE
rpc_acceptor_listen_backlog (int32)Socket backlog parameter used when listening for RPC connections. This defines the maximum length to which the queue of pending TCP connections inbound to the RPC server may grow. If a connection request arrives when the queue is full, the client may receive an error. Higher values may help the server ride over bursts of new inbound connection requests.128FALSE
kudu_scanner_keep_alive_period_sec (int32)The period at which Kudu Scanners should send keep-alive requests to the tablet server to ensure that scanners do not time out.15FALSE
num_metadata_loading_threads (int32)(Advanced) The number of metadata loading threads (degree of parallelism) to use when loading catalog metadata.16FALSE
num_hdfs_worker_threads (int32)(Advanced) The number of threads in the global HDFS operation pool16FALSE
num_abfs_io_threads (int32)Number of ABFS I/O threads16FALSE
num_adls_io_threads (int32)Number of ADLS I/O threads16FALSE
num_ozone_io_threads (int32)Number of Ozone I/O threads16FALSE
num_file_handle_cache_partitions (uint64)Number of partitions used by the file handle cache.16FALSE
num_s3_io_threads (int32)Number of S3 I/O threads16FALSE
datastream_service_num_deserialization_threads (int32)Number of threads for deserializing RPC requests deferred due to the receiver not ready or the soft limit of the receiver is reached.16FALSE
data_stream_sender_buffer_size (int64)(Advanced) Max size in bytes which a row batch in a data stream sender’s channel can accumulate before the row batch is sent over the wire.16384FALSE
max_queued_row_batch_bytes (int64)(Advanced) the maximum bytes of queued rows per multithreaded scan node.16777216FALSE
max_log_size (int32)approx. maximum log file size (in MB). A value of 0 will be silently overridden to 1.1800FALSE
kudu_operation_timeout_ms (int32)Timeout (milliseconds) set for all Kudu operations. This must be a positive value, and there is no way to disable timeouts.180000FALSE
accepted_cnxn_setup_thread_pool_size (int32)(Advanced) The size of the thread pool that is used to process the post-accept, pre-setup connection queue in each thrift server set up to service Impala internal and external connections.2FALSE
stderrthreshold (int32)log messages at or above this level are copied to stderr in addition to logfiles. This flag obsoletes --alsologtostderr.2FALSE
num_acceptor_threads (int32)Number of threads dedicated to accepting connection requests for RPC services2FALSE
concurrent_scratch_ios_per_device (int32)Set this to influence the number of concurrent write I/Os issues to write data to scratch files. This is multiplied by the number of active scratch directories to obtain the target number of scratch write I/Os per query.2FALSE
max_nonhdfs_partitions_parallel_load (int32)(Advanced) Number of threads used to load block metadata for tables that do not support the notion of blocks/storage IDs. Currently supported for S3/ADLS.20FALSE
status_report_cancellation_padding (int32)(Advanced) The coordinator will wait --status_report_max_retry_s * (1 + --status_report_cancellation_padding / 100) without receiving a status report before deciding that a backend is unresponsive and the query should be cancelled. This must be > 0.20FALSE
num_expected_executors (int32)The number of executors that are expected to be available for the execution of a single query. This value is used during planning if no executors have started yet. Once a healthy executor group has started, its size is used instead.20FALSE
default_pool_max_queued (int64)Maximum number of requests allowed to be queued before rejecting requests. A negative value or 0 indicates requests will always be rejected once the maximum number of concurrent requests are executing. Ignored if fair_scheduler_config_path and llama_site_path are set.200FALSE
hung_task_check_interval_ms (int32)Number of milliseconds in between checks for hung threads200FALSE
statestore_update_frequency_ms (int32)(Advanced) Frequency (in ms) with which the statestore sends topic updates to subscribers.2000FALSE
max_cached_file_handles (uint64)Maximum number of HDFS file handles that will be cached. Disabled if set to 0.20000FALSE
ipki_server_key_size (int32)the number of bits for server cert’s private key. The server cert is used for TLS connections to and from clients and other servers.2048FALSE
minidump_size_limit_hint_kb (int32)Size limit hint for minidump files in KB. If a minidump exceeds this value, then breakpad will reduce the stack memory it collects for each thread from 8KB to 2KB. However it will always include the full stack memory for the first 20 threads, including the thread that crashed.20480FALSE
log_async_buffer_bytes_per_level (int32)The number of bytes of buffer space used by each log level. Only relevant when --log_async is enabled.2097152FALSE
inc_stats_size_limit_bytes (int64)Maximum size of incremental stats the catalog is allowed to serialize per table. This limit is set as a safety check, to prevent the JVM from hitting a maximum array limit of 1GB (or OOM) while building the thrift objects to send to impalads. By default, it’s set to 200MB209715200FALSE
beeswax_port (int32)port on which Beeswax client requests are served.If 0 or less, the Beeswax server is not started.21000FALSE
hs2_port (int32)port on which HiveServer2 client requests are served.If 0 or less, the HiveServer2 server is not started.21050FALSE
unused_file_handle_timeout_sec (uint64)Maximum time, in seconds, that an unused HDFS file handle will remain in the file handle cache. Disabled if set to 0.21600FALSE
be_port (int32)port on which thrift based ImpalaInternalService is exported22000FALSE
state_store_subscriber_port (int32)port where StatestoreSubscriberService should be exported23000FALSE
state_store_port (int32)port where StatestoreService is running24000FALSE
query_log_size (int32)Number of queries to retain in the query log. If -1, the query log has unbounded size.25FALSE
webserver_port (int32)Port to start debug webserver on25000FALSE
catalog_service_port (int32)port where the CatalogService is running26000FALSE
krpc_port (int32)port on which KRPC based ImpalaInternalService is exported27000FALSE
hs2_http_port (int32)port on which HiveServer2 HTTP(s) client requests are served. If 0 or less, the HiveServer2 http server is not started.28000FALSE
statestore_heartbeat_tcp_timeout_seconds (int32)(Advanced) The time after which a heartbeat RPC to a subscriber will timeout. This setting protects against badly hung machines that are not able to respond to the heartbeat RPC in short order3FALSE
num_threads_per_core (int32)Number of threads per core.3FALSE
backend_client_connection_num_retries (int32)Retry backend connections.3FALSE
tcp_keepalive_retry_period_s (int32)The duration in seconds between successive keepalive probes from an outbound connection if the previous probes are not acknowledged. Effective only if --tcp_keepalive_probe_period_s is not 0.3FALSE
reserved_words_version (string)Reserved words compatibility version. Reserved words cannot be used as identifiers in SQL. This flag determines the impala version from which the reserved word list is taken. The value must be one of [“2.11.0”, “3.0.0”].3.0.0FALSE
logbufsecs (int32)Buffer log messages for at most this many seconds30FALSE
statestore_subscriber_timeout_seconds (int32)The amount of time (in seconds) that may elapse before the connection with the statestore is considered lost.30FALSE
idle_client_poll_period_s (int32)The poll period, in seconds, after no activity from an Impala client which an Impala service thread (beeswax and HS2) wakes up to check if the connection should be closed. If --idle_session_timeout is also set, a client connection will be closed if all the sessions associated with it have become idle. Set this to 0 to disable the polling behavior and clients’ connection will remain opened until they are explicitly closed.30FALSE
statestore_update_tcp_timeout_seconds (int32)(Advanced) The time after which an update RPC to a subscriber will timeout. This setting protects against badly hung machines that are not able to respond to the update RPC in short order.300FALSE
hdfs_operation_timeout_sec (uint64)Maximum time, in seconds, that an HDFS operation should wait before timing out and failing.300FALSE
catalog_client_rpc_retry_interval_ms (int32)(Advanced) The time to wait before retrying when the catalog RPC client fails to connect to catalogd or when RPCs to the catalogd fail.3000FALSE
statestore_subscriber_cnxn_retry_interval_ms (int32)The interval, in ms, to wait between attempts to make an RPC connection to the statestore.3000FALSE
accepted_client_cnxn_timeout (int64)(Advanced) The amount of time in milliseconds an accepted connection will wait in the post-accept, pre-setup connection queue before it is timed out and the connection request is rejected. A value of 0 means there is no timeout.300000FALSE
backend_client_rpc_timeout_ms (int32)(Advanced) The underlying TSocket send/recv timeout in milliseconds for a backend client RPC.300000FALSE
sasl_connect_tcp_timeout_ms (int32)(Advanced) The underlying TSocket send/recv timeout in milliseconds for the initial SASL handeshake.300000FALSE
rpc_negotiation_timeout_ms (int32)Time in milliseconds of waiting for a negotiation to complete before timing out.300000FALSE
shutdown_deadline_s (int64)Default time limit in seconds for the shut down process. If this duration elapses after the shut down process is started, the daemon shuts down regardless of any running queries.3600FALSE
remember_clients_ttl_ms (int64)Maximum amount of time, in milliseconds, the server “remembers” a client for the purpose of caching its responses. After this period without hearing from it, the client is no longer remembered and the memory occupied by its responses is reclaimed. Retries of requests older than ‘remember_clients_ttl_ms’ are treated as new ones.3600000FALSE
max_hdfs_partitions_parallel_load (int32)(Advanced) Number of threads used to load block metadata for HDFS based partitioned tables. Due to HDFS architectural limitations, it is unlikely to get a linear speed up beyond 5 threads.5FALSE
cancellation_thread_pool_size (int32)(Advanced) Size of the thread-pool processing cancellations due to node failure5FALSE
rpc_retry_interval_ms (int32)Time in millisecond of waiting before retrying an RPC when remote is busy5FALSE
datastream_service_queue_mem_limit (string)(Advanced) Limit on RPC payloads consumption for DataStreamService. Specified as number of bytes (’[bB]?’), megabytes (’[mM]’), gigabytes (’[gG]’), or percentage of the process memory limit (’%’). Defaults to bytes if no unit is given.5%FALSE
periodic_counter_update_period_ms (int32)Period to update rate counters and sampling counters in ms500FALSE
pause_monitor_sleep_time_ms (int64)Sleep time in milliseconds for pause monitor thread.500FALSE
status_report_interval_ms (int32)(Advanced) Interval between profile reports in milliseconds. If set to <= 0, periodic reporting is disabled and only the final report is sent.5000FALSE
max_audit_event_log_file_size (int32)The maximum size (in queries) of the audit event log file before a new one is created (if event logging is enabled)5000FALSE
max_lineage_log_file_size (int32)The maximum size (in queries) of the lineage event log file before a new one is created (if lineage logging is enabled)5000FALSE
max_profile_log_file_size (int32)The maximum size (in queries) of the profile log file before a new one is created5000FALSE
control_service_queue_mem_limit (string)(Advanced) Limit on RPC payloads consumption for ControlService. Specified as number of bytes (’[bB]?’), megabytes (’[mM]’), gigabytes (’[gG]’), or percentage of the process memory limit (’%’). Defaults to bytes if no unit is given.50MBFALSE
jvm_deadlock_detector_interval_s (int64)(Advanced) Interval b
  • 1
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 1
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 1
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值