storage_* server settings

These settings configure the ClickHouse server and are autogenerated from ClickHouse source.

storage_configuration

Allows for multi-disk configuration of storage. Storage configuration follows the structure:

<storage_configuration>
    <disks>
        <!-- configuration -->
    </disks>
    <policies>
        <!-- configuration -->
    </policies>
</storage_configuration>

Configuration of disks

Configuration of disks follows the structure given below:

<storage_configuration>
    <disks>
        <disk_name_1>
            <path>/mnt/fast_ssd/clickhouse/</path>
        </disk_name_1>
        <disk_name_2>
            <path>/mnt/hdd1/clickhouse/</path>
            <keep_free_space_bytes>10485760</keep_free_space_bytes>
        </disk_name_2>
        <disk_name_3>
            <path>/mnt/hdd2/clickhouse/</path>
            <keep_free_space_bytes>10485760</keep_free_space_bytes>
        </disk_name_3>
        ...
    </disks>
</storage_configuration>

The sub-tags above define the following settings for disks:

Setting	Description
`<disk_name_N>`	The name of the disk, which should be unique.
`path`	The path to which server data will be stored (`data` and `shadow` catalogues). It should end with `/`
`keep_free_space_bytes`	Size of the reserved free space on disk.

The order of the disks does not matter.

Configuration of policies

The sub-tags above define the following settings for policies:

Setting	Description
`policy_name_N`	Name of the policy. Policy names must be unique.
`volume_name_N`	The volume name. Volume names must be unique.
`disk`	The disk located inside the volume.
`max_data_part_size_bytes`	The maximum size of a chunk of data that can reside on any of the disks in this volume. If the merge results in a chunk size expected to be larger than max_data_part_size_bytes, the chunk will be written to the next volume. Basically this feature allows you to store new / small chunks on a hot (SSD) volume and move them to a cold (HDD) volume when they reach a large size. Do not use this option if the policy has only one volume.
`move_factor`	The share of available free space on the volume. If the space becomes less, the data will start transferring to the next volume, if there is one. For transfer, chunks are sorted by size from larger to smaller (descending) and chunks whose total size is sufficient to meet the `move_factor` condition are selected, if the total size of all chunks is insufficient, all chunks will be moved.
`perform_ttl_move_on_insert`	Disables moving data with expired TTL on insertion. By default (if enabled), if we insert a piece of data that has already expired according to the move on life rule, it is immediately moved to the volume / disk specified in the move rule. This can significantly slow down insertion in case the target volume / disk is slow (e.g. S3). If disabled, the expired portion of the data is written to the default volume and then immediately moved to the volume specified in the rule for the expired TTL.
`load_balancing`	Disk balancing policy, `round_robin` or `least_used`.
`least_used_ttl_ms`	Sets the timeout (in milliseconds) to update the available space on all disks (`0` - always update, `-1` - never update, default value is `60000`). Note, if the disk is only used by ClickHouse and will not be subject to file system resizing on the fly, you can use the `-1` value. In all other cases this is not recommended, as it will eventually lead to incorrect space allocation.
`prefer_not_to_merge`	Disables merging parts of the data on this volume. Note: this is potentially harmful and can cause slowdown. When this setting is enabled (don’t do this), merging data on this volume is prohibited (which is bad). This allows control of how ClickHouse interacts with slow disks. We recommend not to use this at all.
`volume_priority`	Defines the priority (order) in which volumes are filled. The smaller the value, the higher the priority. The parameter values must be natural numbers and cover the range from 1 to N (N is the largest parameter value specified) with no gaps.

For the volume_priority:

If all volumes have this parameter, they are prioritized in the specified order.
If only some volumes have it, volumes that do not have it have the lowest priority. Those that do have it are prioritized according to the tag value, the priority of the rest is determined by the order of description in the configuration file relative to each other.
If no volumes are given this parameter, their order is determined by the order of the description in the configuration file.
The priority of volumes may not be identical.

storage_metadata_write_full_object_key

Write disk metadata files with VERSION_FULL_OBJECT_KEY format. This is enabled by default. The setting is deprecated.

storage_shared_set_join_use_inner_uuid

If enabled, an inner UUID is generated during the creation of SharedSet and SharedJoin. ClickHouse Cloud only

SQL Reference

Data Types

Engines

Functions

Formats

Settings

System Tables

Data Lakes

storage_configuration

Configuration of disks

Configuration of policies

storage_metadata_write_full_object_key

storage_shared_set_join_use_inner_uuid

​storage_configuration

​Configuration of disks

​Configuration of policies

​storage_metadata_write_full_object_key

​storage_shared_set_join_use_inner_uuid

storage_configuration

Configuration of disks

Configuration of policies

storage_metadata_write_full_object_key

storage_shared_set_join_use_inner_uuid