ClickHouse cheatsheet

Client and HTTP

Command	Description
`clickhouse-client`	Interactive SQL shell (port 9000)
`clickhouse-client -q 'SELECT 1'`	Run single query
`clickhouse-client -h host -u user --password`	Connect to remote server
`curl 'http://127.0.0.1:8123/ping'`	HTTP health check
`curl 'http://127.0.0.1:8123/?query=SELECT+1'`	HTTP query

Service and logs

Command	Description
`systemctl status clickhouse-server`	Service status
`systemctl restart clickhouse-server`	Restart server
`tail -f /var/log/clickhouse-server/clickhouse-server.log`	Follow server log
`ss -tlnp \| grep -E '8123\|9000'`	Check listeners

Schema inspection

Query	Description
`SHOW DATABASES`	List databases
`SHOW TABLES FROM db`	List tables
`DESCRIBE TABLE db.table`	Column definitions
`SHOW CREATE TABLE db.table`	DDL statement

System table queries

-- Parts and size per table
SELECT database, table, count() AS parts, sum(bytes_on_disk) AS bytes
FROM system.parts WHERE active GROUP BY database, table ORDER BY bytes DESC;

-- Running queries
SELECT query_id, user, elapsed, memory_usage, query FROM system.processes;

-- Recent slow queries
SELECT type, query_duration_ms, memory_usage, query
FROM system.query_log
WHERE event_time > now() - INTERVAL 1 HOUR AND type = 'QueryFinish'
ORDER BY query_duration_ms DESC LIMIT 10;

-- Replication status
SELECT database, table, is_readonly, queue_size, absolute_delay
FROM system.replicas;

Operational commands

Query / command	Description
`KILL QUERY WHERE query_id = '...'`	Cancel a running query
`OPTIMIZE TABLE db.table FINAL`	Force merge (use carefully)
`SYSTEM FLUSH LOGS`	Flush log tables to disk
`du -sh /var/lib/clickhouse/*`	Disk usage by database

Pro tips

Check system.parts when inserts slow down — too many parts is a common cause
Use system.processes to find queries consuming memory before OOM kills
Batch inserts (thousands of rows per INSERT) — avoid one-row inserts at high rate
Read the server log for "Memory limit exceeded" and "Too many parts" messages

Practice scenarios

Hands-on ClickHouse scenarios on live Linux VMs: clickhouse