Skip to content

Latest commit

 

History

History
544 lines (352 loc) · 23.2 KB

pd-configuration-file.md

File metadata and controls

544 lines (352 loc) · 23.2 KB
Raw
title summary aliases
PD Configuration File
Learn the PD configuration file.
/docs/dev/pd-configuration-file/
/docs/dev/reference/configuration/pd-server/configuration-file/

PD Configuration File

The PD configuration file supports more options than command-line parameters. You can find the default configuration file here.

This document only describes parameters that are not included in command-line parameters. Check here for the command line parameters.

Tip:

If you need to adjust the value of a configuration item, refer to Modify the configuration.

name

  • The unique name of a PD node
  • Default value: "pd"
  • To start multiply PD nodes, use a unique name for each node.

data-dir

  • The directory in which PD stores data
  • Default value: default.${name}"

client-urls

  • The list of client URLs to be listened to by PD
  • Default value: "http://127.0.0.1:2379"
  • When you deploy a cluster, you must specify the IP address of the current host as client-urls (for example, "http://192.168.100.113:2379"). If the cluster runs on Docker, specify the IP address of Docker as "http://0.0.0.0:2379".

advertise-client-urls

  • The list of advertise URLs for the client to access PD
  • Default value: "${client-urls}"
  • In some situations such as in the Docker or NAT network environment, if a client cannot access PD through the default client URLs listened to by PD, you must manually set the advertise client URLs.
  • For example, the internal IP address of Docker is 172.17.0.1, while the IP address of the host is 192.168.100.113 and the port mapping is set to -p 2380:2380. In this case, you can set advertise-client-urls to "http://192.168.100.113:2380". The client can find this service through "http://192.168.100.113:2380".

peer-urls

  • The list of peer URLs to be listened to by a PD node
  • Default value: "http://127.0.0.1:2380"
  • When you deploy a cluster, you must specify peer-urls as the IP address of the current host, such as "http://192.168.100.113:2380". If the cluster runs on Docker, specify the IP address of Docker as "http://0.0.0.0:2380".

advertise-peer-urls

  • The list of advertise URLs for other PD nodes (peers) to access a PD node
  • Default: "${peer-urls}"
  • In some situations such as in the Docker or NAT network environment, if the other nodes (peers) cannot access the PD node through the default peer URLs listened to by this PD node, you must manually set the advertise peer URLs.
  • For example, the internal IP address of Docker is 172.17.0.1, while the IP address of the host is 192.168.100.113 and the port mapping is set to -p 2380:2380. In this case, you can set advertise-peer-urls to "http://192.168.100.113:2380". The other PD nodes can find this service through "http://192.168.100.113:2380".

initial-cluster

  • The initial cluster configuration for bootstrapping

  • Default value: "{name}=http://{advertise-peer-url}"

  • For example, if name is "pd", and advertise-peer-urls is "http://192.168.100.113:2380", the initial-cluster is "pd=http://192.168.100.113:2380".

  • If you need to start three PD servers, the initial-cluster might be:

    pd1=http://192.168.100.113:2380, pd2=http://192.168.100.114:2380, pd3=192.168.100.115:2380

initial-cluster-state

  • The initial state of the cluster
  • Default value: "new"

initial-cluster-token

  • Identifies different clusters during the bootstrap phase
  • Default value: "pd-cluster"
  • If multiple clusters that have nodes with same configurations are deployed successively, you must specify different tokens to isolate different cluster nodes.

lease

  • The timeout of the PD Leader Key lease. After the timeout, the system re-elects a Leader.
  • Default value: 3
  • Unit: second

quota-backend-bytes

  • The storage size of the meta-information database, which is 8GiB by default
  • Default value: 8589934592

auto-compaction-mod

  • The automatic compaction modes of the meta-information database
  • Available options: periodic (by cycle) and revision (by version number).
  • Default value: periodic

auto-compaction-retention

  • The time interval for automatic compaction of the meta-information database when auto-compaction-retention is periodic. When the compaction mode is set to revision, this parameter indicates the version number for the automatic compaction.
  • Default value: 1h

force-new-cluster

  • Determines whether to force PD to start as a new cluster and modify the number of Raft members to 1
  • Default value: false

tso-update-physical-interval

  • The interval at which PD updates the physical time of TSO.
  • In a default update interval of TSO physical time, PD provides at most 262144 TSOs. To get more TSOs, you can reduce the value of this configuration item. The minimum value is 1ms.
  • Decreasing this configuration item might increase the CPU usage of PD. According to the test, compared with the interval of 50ms, the CPU usage of PD will increase by about 10% when the interval is 1ms.
  • Default value: 50ms
  • Minimum value: 1ms

pd-server

Configuration items related to pd-server

server-memory-limit New in v6.6.0

Warning:

This configuration is an experimental feature. It is not recommended to use it in production environments.

  • The memory limit ratio for a PD instance. The value 0 means no memory limit.
  • Default value: 0
  • Minimum value: 0
  • Maximum value: 0.99

server-memory-limit-gc-trigger New in v6.6.0

Warning:

This configuration is an experimental feature. It is not recommended to use it in production environments.

  • The threshold ratio at which PD tries to trigger GC. When the memory usage of PD reaches the value of server-memory-limit * the value of server-memory-limit-gc-trigger, PD triggers a Golang GC. Only one GC is triggered in one minute.
  • Default value: 0.7
  • Minimum value: 0.5
  • Maximum value: 0.99

enable-gogc-tuner New in v6.6.0

Warning:

This configuration is an experimental feature. It is not recommended to use it in production environments.

  • Controls whether to enable the GOGC Tuner.
  • Default value: false

gc-tuner-threshold New in v6.6.0

Warning:

This configuration is an experimental feature. It is not recommended to use it in production environments.

  • The maximum memory threshold ratio for tuning GOGC. When the memory exceeds this threshold, i.e. the value of server-memory-limit * the value of gc-tuner-threshold, GOGC Tuner stops working.
  • Default value: 0.6
  • Minimum value: 0
  • Maximum value: 0.9

flow-round-by-digit New in TiDB 5.1

  • Default value: 3
  • PD rounds the lowest digits of the flow number, which reduces the update of statistics caused by the changes of the Region flow information. This configuration item is used to specify the number of lowest digits to round for the Region flow information. For example, the flow 100512 will be rounded to 101000 because the default value is 3. This configuration replaces trace-region-flow.

Note:

If you have upgraded your cluster from a TiDB 4.0 version to the current version, the behavior of flow-round-by-digit after the upgrading and the behavior of trace-region-flow before the upgrading are consistent by default. This means that if the value of trace-region-flow is false before the upgrading, the value of flow-round-by-digit after the upgrading is 127; if the value of trace-region-flow is true before the upgrading, the value of flow-round-by-digit after the upgrading is 3.

min-resolved-ts-persistence-interval New in v6.0.0

  • Determines the interval at which the minimum resolved timestamp is persistent to the PD. If this value is set to 0, it means that the persistence is disabled.
  • Default value: Before v6.3.0, the default value is "0s". Starting from v6.3.0, the default value is "1s", which is the smallest positive value.
  • Minimum value: 0
  • Unit: second

Note:

For clusters upgraded from v6.0.0~v6.2.0, the default value of min-resolved-ts-persistence-interval does not change after the upgrade, which means that it will remain "0s". To enable this feature, you need to manually change the value of this configuration item.

security

Configuration items related to security

cacert-path

  • The path of the CA file
  • Default value: ""

cert-path

  • The path of the Privacy Enhanced Mail (PEM) file that contains the X509 certificate
  • Default value: ""

key-path

  • The path of the PEM file that contains the X509 key
  • Default value: ""

redact-info-log New in v5.0

  • Controls whether to enable log redaction in the PD log
  • Optional value: false, true, "marker"
  • Default value: false
  • For details on how to use it, see Log redaction in PD side.

log

Configuration items related to log

level

  • Specifies the level of the output log
  • Optional value: "debug", "info", "warn", "error", "fatal"
  • Default value: "info"

format

  • The log format
  • Optional value: "text", "json"
  • Default value: "text"

disable-timestamp

  • Whether to disable the automatically generated timestamp in the log
  • Default value: false

log.file

Configuration items related to the log file

max-size

  • The maximum size of a single log file. When this value is exceeded, the system automatically splits the log into several files.
  • Default value: 300
  • Unit: MiB
  • Minimum value: 1

max-days

  • The maximum number of days in which a log is kept
  • If the configuration item is not set, or the value of it is set to the default value 0, PD does not clean log files.
  • Default value: 0

max-backups

  • The maximum number of log files to keep
  • If the configuration item is not set, or the value of it is set to the default value 0, PD keeps all log files.
  • Default value: 0

metric

Configuration items related to monitoring

interval

  • The interval at which monitoring metric data is pushed to Prometheus
  • Default value: 15s

schedule

Configuration items related to scheduling

Note:

To modify these PD configuration items related to schedule, choose one of the following methods based on your cluster status:

  • For clusters to be newly deployed, you can modify the PD configuration file directly.
  • For existing clusters, use the command-line tool PD Control to make changes instead. Direct modifications to these PD configuration items related to schedule in the configuration file do not take effect on existing clusters.

max-merge-region-size

  • Controls the size limit of Region Merge. When the Region size is greater than the specified value, PD does not merge the Region with the adjacent Regions.
  • Default value: 54. Before v8.4.0, the default value is 20. Starting from v8.4.0, the default value is 54.
  • Unit: MiB

max-merge-region-keys

  • Specifies the upper limit of the Region Merge key. When the Region key is greater than the specified value, the PD does not merge the Region with its adjacent Regions.
  • Default value: 540000. Before v8.4.0, the default value is 200000. Starting from v8.4.0, the default value is 540000.

patrol-region-interval

  • Controls the running frequency at which the checker inspects the health state of a Region. The smaller this value is, the faster the checker runs. Normally, you do not need to adjust this configuration.
  • Default value: 10ms

patrol-region-worker-count New in v8.5.0

Warning:

Setting this configuration item to a value greater than 1 enables concurrent checks. This is an experimental feature. It is not recommended that you use it in the production environment. This feature might be changed or removed without prior notice. If you find a bug, you can report an issue on GitHub.

  • Controls the number of concurrent operators created by the checker when inspecting the health state of a Region. Normally, you do not need to adjust this configuration.
  • Default value: 1

split-merge-interval

  • Controls the time interval between the split and merge operations on the same Region. That means a newly split Region will not be merged for a while.
  • Default value: 1h

max-snapshot-count

  • Controls the maximum number of snapshots that a single store receives or sends at the same time. PD schedulers depend on this configuration to prevent the resources used for normal traffic from being preempted.
  • Default value value: 64

max-pending-peer-count

  • Controls the maximum number of pending peers in a single store. PD schedulers depend on this configuration to prevent too many Regions with outdated logs from being generated on some nodes.
  • Default value: 64

max-store-down-time

  • The downtime after which PD judges that the disconnected store cannot be recovered. When PD fails to receive the heartbeat from a store after the specified period of time, it adds replicas at other nodes.
  • Default value: 30m

max-store-preparing-time New in v6.1.0

  • Controls the maximum waiting time for the store to go online. During the online stage of a store, PD can query the online progress of the store. When the specified time is exceeded, PD assumes that the store has been online and cannot query the online progress of the store again. But this does not prevent Regions from transferring to the new online store. In most scenarios, you do not need to adjust this parameter.
  • Default value: 48h

leader-schedule-limit

  • The number of Leader scheduling tasks performed at the same time
  • Default value: 4

region-schedule-limit

  • The number of Region scheduling tasks performed at the same time
  • Default value: 2048

enable-diagnostic New in v6.3.0

  • Controls whether to enable the diagnostic feature. When it is enabled, PD records the state during scheduling to help diagnose. If enabled, it might slightly affect the scheduling speed and consume more memory when there are many stores.
  • Default value: Starting from v7.1.0, the default value is changed from false to true. If your cluster is upgraded from a version earlier than v7.1.0 to v7.1.0 or later, the default value does not change.

hot-region-schedule-limit

  • Controls the hot Region scheduling tasks that are running at the same time. It is independent of the Region scheduling.
  • Default value: 4

hot-region-cache-hits-threshold

  • The threshold used to set the number of minutes required to identify a hot Region. PD can participate in the hotspot scheduling only after the Region is in the hotspot state for more than this number of minutes.
  • Default value: 3

replica-schedule-limit

  • The number of Replica scheduling tasks performed at the same time
  • Default value: 64

merge-schedule-limit

  • The number of the Region Merge scheduling tasks performed at the same time. Set this parameter to 0 to disable Region Merge.
  • Default value: 8

high-space-ratio

  • The threshold ratio below which the capacity of the store is sufficient. If the space occupancy ratio of the store is smaller than this threshold value, PD ignores the remaining space of the store when performing scheduling, and balances load mainly based on the Region size. This configuration takes effect only when region-score-formula-version is set to v1.
  • Default value: 0.7
  • Minimum value: greater than 0
  • Maximum value: less than 1

low-space-ratio

  • The threshold ratio above which the capacity of the store is insufficient. If the space occupancy ratio of a store exceeds this threshold value, PD avoids migrating data to this store as much as possible. Meanwhile, to avoid the disk space of the corresponding store being exhausted, PD performs scheduling mainly based on the remaining space of the store.
  • Default value: 0.8
  • Minimum value: greater than 0
  • Maximum value: less than 1

tolerant-size-ratio

  • Controls the balance buffer size
  • Default value: 0 (automatically adjusts the buffer size)
  • Minimum value: 0

enable-cross-table-merge

  • Determines whether to enable the merging of cross-table Regions
  • Default value: true

region-score-formula-version New in v5.0

  • Controls the version of the Region score formula
  • Default value: v2
  • Optional values: v1 and v2. Compared to v1, the changes in v2 are smoother, and the scheduling jitter caused by space reclaim is improved.

Note:

If you have upgraded your cluster from a TiDB 4.0 version to the current version, the new formula version is automatically disabled by default to ensure consistent PD behavior before and after the upgrading. If you want to change the formula version, you need to manually switch through the pd-ctl setting. For details, refer to PD Control.

store-limit-version New in v7.1.0

  • Controls the version of the store limit formula
  • Default value: v1
  • Value options:
    • v1: In v1 mode, you can manually modify the store limit to limit the scheduling speed of a single TiKV.
    • v2: In v2 mode, you do not need to manually set the store limit value, as PD dynamically adjusts it based on the capability of TiKV snapshots. For more details, refer to Principles of store limit v2.

enable-joint-consensus New in v5.0

  • Controls whether to use Joint Consensus for replica scheduling. If this configuration is disabled, PD schedules one replica at a time.
  • Default value: true

hot-regions-write-interval New in v5.4.0

  • The time interval at which PD stores hot Region information.
  • Default value: 10m

Note:

The information about hot Regions is updated every three minutes. If the interval is set to less than three minutes, updates during the interval might be meaningless.

hot-regions-reserved-days New in v5.4.0

  • Specifies how many days the hot Region information is retained.
  • Default value: 7

replication

Configuration items related to replicas

max-replicas

  • The number of replicas, that is, the sum of the number of leaders and followers. The default value 3 means 1 leader and 2 followers. When this configuration is modified dynamically, PD will schedule Regions in the background so that the number of replicas matches this configuration.
  • Default value: 3

location-labels

isolation-level

strictly-match-label

  • Enables the strict check for whether the TiKV label matches PD's location-labels.
  • Default value: false

enable-placement-rules

label-property (deprecated)

Configuration items related to labels, which only support the reject-leader type.

Note:

Starting from v5.2, the configuration items related to labels are deprecated. It is recommended to use Placement Rules to configure the replica policy.

key (deprecated)

  • The label key for the store that rejected the Leader
  • Default value: ""

value (deprecated)

  • The label value for the store that rejected the Leader
  • Default value: ""

dashboard

Configuration items related to the TiDB Dashboard built in PD.

tidb-cacert-path

  • The path of the root CA certificate file. You can configure this path when you connect to TiDB's SQL services using TLS.
  • Default value: ""

tidb-cert-path

  • The path of the SSL certificate file. You can configure this path when you connect to TiDB's SQL services using TLS.
  • Default value: ""

tidb-key-path

  • The path of the SSL private key file. You can configure this path when you connect to TiDB's SQL services using TLS.
  • Default value: ""

public-path-prefix

  • When TiDB Dashboard is accessed behind a reverse proxy, this item sets the public URL path prefix for all web resources.
  • Default value: /dashboard
  • Do not modify this configuration item when TiDB Dashboard is accessed not behind a reverse proxy; otherwise, access issues might occur. See Use TiDB Dashboard behind a Reverse Proxy for details.

enable-telemetry

Warning:

Starting from v8.1.0, the telemetry feature in TiDB Dashboard is removed, and this configuration item is no longer functional. It is retained solely for compatibility with earlier versions.

  • Before v8.1.0, this configuration item controls whether to enable telemetry collection in TiDB Dashboard.
  • Default value: false

replication-mode

Configuration items related to the replication mode of all Regions. See Enable the DR Auto-Sync mode for details.

controller

This section describes the configuration items that are built into PD for Resource Control.

degraded-mode-wait-duration

  • Time to wait to trigger the degradation mode. Degradation mode means that when the Local Token Bucket (LTB) and Global Token Bucket (GTB) are lost, the LTB falls back to the default resource group configuration and no longer has a GTB authorization token, thus ensuring that the service is not affected in the event of network isolation or anomalies.
  • Default value: 0s
  • The degradation mode is disabled by default.

request-unit

The following are the configuration items about the Request Unit (RU).

read-base-cost

  • Basis factor for conversion from a read request to RU
  • Default value: 0.25

write-base-cost

  • Basis factor for conversion from a write request to RU
  • Default value: 1

read-cost-per-byte

  • Basis factor for conversion from read flow to RU
  • Default value: 1/(64 * 1024)
  • 1 RU = 64 KiB read bytes

write-cost-per-byte

  • Basis factor for conversion from write flow to RU
  • Default value: 1/1024
  • 1 RU = 1 KiB write bytes

read-cpu-ms-cost

  • Basis factor for conversion from CPU to RU
  • Default value: 1/3
  • 1 RU = 3 millisecond CPU time