Skip to main content

Usage Recommendations


This page is not applicable to ClickHouse Cloud. The procedure documented here is automated in ClickHouse Cloud services.

CPU Scaling Governor

Always use the performance scaling governor. The on-demand scaling governor works much worse with constantly high demand.

$ echo 'performance' | sudo tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor

CPU Limitations

Processors can overheat. Use dmesg to see if the CPU’s clock rate was limited due to overheating. The restriction can also be set externally at the datacenter level. You can use turbostat to monitor it under a load.


For small amounts of data (up to ~200 GB compressed), it is best to use as much memory as the volume of data. For large amounts of data and when processing interactive (online) queries, you should use a reasonable amount of RAM (128 GB or more) so the hot data subset will fit in the cache of pages. Even for data volumes of ~50 TB per server, using 128 GB of RAM significantly improves query performance compared to 64 GB.

Do not disable overcommit. The value cat /proc/sys/vm/overcommit_memory should be 0 or 1. Run

$ echo 0 | sudo tee /proc/sys/vm/overcommit_memory

Use perf top to watch the time spent in the kernel for memory management. Permanent huge pages also do not need to be allocated.


If your system has less than 16 GB of RAM, you may experience various memory exceptions because default settings do not match this amount of memory. The recommended amount of RAM is 32 GB or more. You can use ClickHouse in a system with a small amount of RAM, even with 2 GB of RAM, but it requires additional tuning and can ingest at a low rate.

Storage Subsystem

If your budget allows you to use SSD, use SSD. If not, use HDD. SATA HDDs 7200 RPM will do.

Give preference to a lot of servers with local hard drives over a smaller number of servers with attached disk shelves. But for storing archives with rare queries, shelves will work.


When using HDD, you can combine their RAID-10, RAID-5, RAID-6 or RAID-50. For Linux, software RAID is better (with mdadm). When creating RAID-10, select the far layout. If your budget allows, choose RAID-10.

LVM by itself (without RAID or mdadm) is ok, but making RAID with it or combining it with mdadm is a less explored option, and there will be more chances for mistakes (selecting wrong chunk size; misalignment of chunks; choosing a wrong raid type; forgetting to cleanup disks). If you are confident in using LVM, there is nothing against using it.

If you have more than 4 disks, use RAID-6 (preferred) or RAID-50, instead of RAID-5. When using RAID-5, RAID-6 or RAID-50, always increase stripe_cache_size, since the default value is usually not the best choice.

$ echo 4096 | sudo tee /sys/block/md2/md/stripe_cache_size

Calculate the exact number from the number of devices and the block size, using the formula: 2 * num_devices * chunk_size_in_bytes / 4096.

A block size of 64 KB is sufficient for most RAID configurations. The average clickhouse-server write size is approximately 1 MB (1024 KB), and thus the recommended stripe size is also 1 MB. The block size can be optimized if needed when set to 1 MB divided by the number of non-parity disks in the RAID array, such that each write is parallelized across all available non-parity disks. Never set the block size too small or too large.

You can use RAID-0 on SSD. Regardless of RAID use, always use replication for data security.

Enable NCQ with a long queue. For HDD, choose the mq-deadline or CFQ scheduler, and for SSD, choose noop. Don’t reduce the ‘readahead’ setting. For HDD, enable the write cache.

Make sure that fstrim is enabled for NVME and SSD disks in your OS (usually it's implemented using a cronjob or systemd service).

File System

Ext4 is the most reliable option. Set the mount options noatime. XFS works well too. Most other file systems should also work fine.

FAT-32 and exFAT are not supported due to lack of hard links.

Do not use compressed filesystems, because ClickHouse does compression on its own and better. It's not recommended to use encrypted filesystems, because you can use builtin encryption in ClickHouse, which is better.

While ClickHouse can work over NFS, it is not the best idea.

Linux Kernel

Don't use an outdated Linux kernel.


If you are using IPv6, increase the size of the route cache. The Linux kernel prior to 3.2 had a multitude of problems with IPv6 implementation.

Use at least a 10 GB network, if possible. 1 Gb will also work, but it will be much worse for patching replicas with tens of terabytes of data, or for processing distributed queries with a large amount of intermediate data.

Huge Pages

If you are using old Linux kernel, disable transparent huge pages. It interferes with memory allocator, which leads to significant performance degradation. On newer Linux kernels transparent huge pages are alright.

$ echo 'madvise' | sudo tee /sys/kernel/mm/transparent_hugepage/enabled

If you want to modify the transparent huge pages setting permanently, editing the /etc/default/grub to add the transparent_hugepage=madvise to the GRUB_CMDLINE_LINUX_DEFAULT option:

$ GRUB_CMDLINE_LINUX_DEFAULT="transparent_hugepage=madvise ..."

After that, run the sudo update-grub command then reboot to take effect.

Hypervisor configuration

If you are using OpenStack, set


in nova.conf.

If you are using libvirt, set

<cpu mode='host-passthrough'/>

in XML configuration.

This is important for ClickHouse to be able to get correct information with cpuid instruction. Otherwise you may get Illegal instruction crashes when hypervisor is run on old CPU models.

ClickHouse Keeper and ZooKeeper

ClickHouse Keeper is recommended to replace ZooKeeper for ClickHouse clusters. See the documentation for ClickHouse Keeper

If you would like to continue using ZooKeeper then it is best to use a fresh version of ZooKeeper – 3.4.9 or later. The version in stable Linux distributions may be outdated.

You should never use manually written scripts to transfer data between different ZooKeeper clusters, because the result will be incorrect for sequential nodes. Never use the “zkcopy” utility for the same reason:

If you want to divide an existing ZooKeeper cluster into two, the correct way is to increase the number of its replicas and then reconfigure it as two independent clusters.

You can run ClickHouse Keeper on the same server as ClickHouse in test environments, or in environments with low ingestion rate. For production environments we suggest to use separate servers for ClickHouse and ZooKeeper/Keeper, or place ClickHouse files and Keeper files on to separate disks. Because ZooKeeper/Keeper are very sensitive for disk latency and ClickHouse may utilize all available system resources.

You can have ZooKeeper observers in an ensemble but ClickHouse servers should not interact with observers.

Do not change minSessionTimeout setting, large values may affect ClickHouse restart stability.

With the default settings, ZooKeeper is a time bomb:

The ZooKeeper server won’t delete files from old snapshots and logs when using the default configuration (see autopurge), and this is the responsibility of the operator.

This bomb must be defused.

The ZooKeeper (3.5.1) configuration below is used in a large production environment:



# The number of milliseconds of each tick
# The number of ticks that the initial
# synchronization phase can take
# This value is not quite motivated
# The number of ticks that can pass between
# sending a request and getting an acknowledgement


# It is the maximum value that client may request and the server will accept.
# It is Ok to have high maxSessionTimeout on server to allow clients to work with high session timeout if they want.
# But we request session timeout of 30 seconds by default (you can change it with session_timeout_ms in ClickHouse config).
# the directory where the snapshot is stored.
dataDir=/opt/zookeeper/{{ '{{' }} cluster['name'] {{ '}}' }}/data
# Place the dataLogDir to a separate physical disc for better performance
dataLogDir=/opt/zookeeper/{{ '{{' }} cluster['name'] {{ '}}' }}/logs


# To avoid seeks ZooKeeper allocates space in the transaction log file in
# blocks of preAllocSize kilobytes. The default block size is 64M. One reason
# for changing the size of the blocks is to reduce the block size if snapshots
# are taken more often. (Also, see snapCount).

# Clients can submit requests faster than ZooKeeper can process them,
# especially if there are a lot of clients. To prevent ZooKeeper from running
# out of memory due to queued requests, ZooKeeper will throttle clients so that
# there is no more than globalOutstandingLimit outstanding requests in the
# system. The default limit is 1000.
# globalOutstandingLimit=1000

# ZooKeeper logs transactions to a transaction log. After snapCount transactions
# are written to a log file a snapshot is started and a new transaction log file
# is started. The default snapCount is 100000.

# If this option is defined, requests will be will logged to a trace file named

# Leader accepts client connections. Default value is "yes". The leader machine
# coordinates updates. For higher update throughput at thes slight expense of
# read throughput the leader can be configured to not accept clients and focus
# on coordination.

dynamicConfigFile=/etc/zookeeper-{{ '{{' }} cluster['name'] {{ '}}' }}/conf/zoo.cfg.dynamic

Java version:

openjdk 11.0.5-shenandoah 2019-10-15
OpenJDK Runtime Environment (build 11.0.5-shenandoah+10-adhoc.heretic.src)
OpenJDK 64-Bit Server VM (build 11.0.5-shenandoah+10-adhoc.heretic.src, mixed mode)

JVM parameters:

NAME=zookeeper-{{ '{{' }} cluster['name'] {{ '}}' }}

# TODO this is really ugly
# How to find out, which jars are needed?
# seems, that log4j requires the file to be in the classpath

JAVA_OPTS="-Xms{{ '{{' }} cluster.get('xms','128M') {{ '}}' }} \
-Xmx{{ '{{' }} cluster.get('xmx','1G') {{ '}}' }} \
-verbose:gc \
-XX:+UseG1GC \
-Djute.maxbuffer=8388608 \

Salt initialization:

description "zookeeper-{{ '{{' }} cluster['name'] {{ '}}' }} centralized coordination service"

start on runlevel [2345]
stop on runlevel [!2345]


limit nofile 8192 8192

pre-start script
[ -r "/etc/zookeeper-{{ '{{' }} cluster['name'] {{ '}}' }}/conf/environment" ] || exit 0
. /etc/zookeeper-{{ '{{' }} cluster['name'] {{ '}}' }}/conf/environment
[ -d $ZOO_LOG_DIR ] || mkdir -p $ZOO_LOG_DIR
end script

. /etc/zookeeper-{{ '{{' }} cluster['name'] {{ '}}' }}/conf/environment
[ -r /etc/default/zookeeper ] && . /etc/default/zookeeper
if [ -z "$JMXDISABLE" ]; then
exec start-stop-daemon --start -c $USER --exec $JAVA --name zookeeper-{{ '{{' }} cluster['name'] {{ '}}' }} \
-- -cp $CLASSPATH $JAVA_OPTS -Dzookeeper.log.dir=${ZOO_LOG_DIR} \
-Dzookeeper.root.logger=${ZOO_LOG4J_PROP} $ZOOMAIN $ZOOCFG
end script

Antivirus software

If you use antivirus software configure it to skip folders with ClickHouse datafiles (/var/lib/clickhouse) otherwise performance may be reduced and you may experience unexpected errors during data ingestion and background merges.