All Products
Search
Document Center

E-MapReduce:Configuration parameters

Last Updated:Mar 26, 2026

Configure the ClickHouse service in an E-MapReduce (EMR) ClickHouse cluster using Extensible Markup Language (XML) configuration files. Parameters are organized across three configuration tabs in the EMR console: client-config, server-config, and server-metrika. User permission parameters are managed separately.

Prerequisites

Before you begin, make sure that you have:

Usage notes

All three configuration tabs generate XML files. Take note of these rules when adding custom parameters:

  • If a parameter belongs to the yandex tag, add it directly — no console action is needed.

  • For nested parameters, separate each layer with a period (.). For example, to set a password for a user named aliyun on the server-users tab, use users.aliyun.password as the parameter name.

  • Do not specify parameter names or values in XML format when adding custom parameters through the console.

Configuration tabs reference

Tab Generated file Description
client-config config.xml ClickHouse client configuration
server-config config.xml ClickHouse server configuration
server-metrika metrika.xml Extended parameters, including ZooKeeper, shards, replicas, and compression. Referenced by the server config.xml by default.
User permissions See Configure user permissions.

client-config

Parameters on the client-config tab configure the ClickHouse client. Navigate to the ClickHouse service page in the EMR console, click the Configure tab, and then click client-config.

Parameter Default Description
user default Username for logging on to the ClickHouse client.
password (blank) Password for logging on to the ClickHouse client.
prompt_by_server_display_name.production Custom prompt for the ClickHouse client. The prompt changes based on the display_name value set on the server-config tab. If display_name is set to default, the client uses prompt_by_server_display_name.default. For prompt color options, see Color prompts with readline and tip_colors_and_formatting.
prompt_by_server_display_name.default Prompt used when display_name is default.
prompt_by_server_display_name.test Prompt used when display_name is test.

server-config

Parameters on the server-config tab configure the ClickHouse server. Navigate to the ClickHouse service page in the EMR console, click the Configure tab, and then click server-config.

Network

Parameter Default Description
tcp_port 9000 TCP port for ClickHouse client communication.
http_port 8123 HTTP port for ClickHouse server communication. Open source ClickHouse Java Database Connectivity (JDBC) drivers (see clickhouse-jdbc) also use this port.
listen_host 0.0.0.0 IP address the ClickHouse server listens on. Accepts IPv4 or IPv6 addresses. Set to :: to listen on all interfaces. Separate multiple addresses with commas, for example: 127.0.0.1,localhost.
keep_alive_timeout 10 Seconds ClickHouse waits for incoming HTTP requests before closing an idle connection.

Logging

All three logging parameters control the log rotation policy.

Parameter Default Description
logger.level information Log level. Valid values in order of urgency: none (logging disabled), fatal, critical, error, warning, notice, information, debug, trace.
logger.size 1000M Maximum size of a log file. When this limit is reached, ClickHouse archives and renames the file and creates a new one.
logger.count 10 Maximum number of archived log files to keep. When the limit is reached, the oldest archived files are deleted.

Sessions and queries

Parameter Default Unit Description
max_session_timeout 3600 seconds Maximum session timeout.
default_session_timeout 60 seconds Default session timeout.
max_concurrent_queries 0 Maximum number of queries processed in parallel. 0 means no limit.

Caching

Parameter Default Description
uncompressed_cache_size 0 Cache size for decompressed blocks when using the MergeTree table engine. 0 disables caching.
mark_cache_size 5368709120 Approximate size (in bytes) of the mark index cache for MergeTree table engine tables.

Database

Parameter Default Description
default_database default Default database name.
default_profile default Default settings profile name.
timezone Asia/Shanghai Time zone of the ClickHouse server.

Distributed DDL

Parameter Default Description
distributed_ddl.path /clickhouse/task_queue/ddl ZooKeeper path for the DDL query queue. By default, CREATE, DROP, ALTER, and RENAME statements affect only the node that processes the query. Setting distributed_ddl parameters enables these statements to run across the entire ClickHouse cluster. Requires ZooKeeper to be enabled.

Replication

Parameter Default Description
merge_tree.allow_remote_fs_zero_copy_replication Set to true to enable metadata replication for the Replicated\*MergeTree engine type. With this enabled, metadata pointing to the Hadoop Distributed File System (HDFS) disk is replicated to generate multiple metadata replicas for the same shard.
transaction.enable_public_ip By default, ClickHouse uses a private IP address to identify transactions. Set to true to use a public IP address instead. All nodes must have public IP addresses assigned.

server-metrika

Parameters on the server-metrika tab generate the metrika.xml file, which the ClickHouse server config.xml references by default. Navigate to the ClickHouse service page in the EMR console, click the Configure tab, and then click server-metrika.

Parameter Default Description
clickhouse_compression (blank) Data compression settings for MergeTree engine tables. By default, this parameter is left empty. Set this parameter to enable data compression. For more information, see Server Settings.
storage_configuration Custom disk configuration.
zookeeper_servers ZooKeeper server created at cluster creation ZooKeeper server information for the ClickHouse cluster. Separate multiple servers with commas, for example: emr-header-1.cluster-12345:2181,emr-worker-1.cluster-12345:2181,emr-worker-2.cluster-12345:2181.
quotas_default Quota settings for resource limits. Configures the quota named default. Add custom quotas as needed.
clickhouse_remote_servers Topology based on shard and replica count at cluster creation Shard and replica configuration for the ClickHouse cluster.
Important

Change only when necessary. Manually modifying the shard count, replica count, or topology may cause errors when writing data to or querying data from the cluster.

References

What to do next

To modify or add configuration parameters, see Manage configuration items.