Configuration File

HarperDB is configured through a YAML file called harperdb-config.yaml located in the operations API root directory (by default this is a directory named hdb located in the home directory of the current user).

All available configuration will be populated by default in the config file on install, regardless of whether it is used.

Using the Configuration File and Naming Conventions

The configuration elements in harperdb-config.yaml use camelcase: operationsApi.

To change a configuration value edit the harperdb-config.yaml file and save any changes. HarperDB must be restarted for changes to take effect.

Alternately, configuration can be changed via environment and/or command line variables or via the API. To access lower level elements, use underscores to append parent/child elements (when used this way elements are case insensitive):

- Environment variables: `OPERATIONSAPI_NETWORK_PORT=9925`
- Command line variables: `--OPERATIONSAPI_NETWORK_PORT 9925`
- Calling `set_configuration` through the API: `operationsApi_network_port: 9925`

_Note: Component configuration cannot be added or updated via CLI or ENV variables.

Importing installation configuration

To use a custom configuration file to set values on install, use the CLI/ENV variable HDB_CONFIG and set it to the path of your custom configuration file.

To install HarperDB overtop of an existing configuration file, set HDB_CONFIG to the root path of your install <ROOTPATH>/harperdb-config.yaml

Configuration Options

`http`

sessionAffinity - Type: string; Default: null

HarperDB is a multi-threaded server designed to scale to utilize many CPU cores with high concurrency. Session affinity can help improve the efficiency and fairness of thread utilization by routing multiple requests from the same client to the same thread. This provides a fairer method of request handling by keeping a single user contained to a single thread, can improve caching locality (multiple requests from a single user are more likely to access the same data), and can provide the ability to share information in-memory in user sessions. Enabling session affinity will cause subsequent requests from the same client to be routed to the same thread.

To enable sessionAffinity, you need to specify how clients will be identified from the incoming requests. If you are using HarperDB to directly serve HTTP requests from users from different remote addresses, you can use a setting of ip. However, if you are using HarperDB behind a proxy server or application server, all the remote ip addresses will be the same and HarperDB will effectively only run on a single thread. Alternately, you can specify a header to use for identification. If you are using basic authentication, you could use the "Authorization" header to route requests to threads by the user's credentials. If you have another header that uniquely identifies users/clients, you can use that as the value of sessionAffinity. But be careful to ensure that the value does provide sufficient uniqueness and that requests are effectively distributed to all the threads and fully utilizing all your CPU cores.

http:
  sessionAffinity: ip

compressionThreshold - Type: number; Default: 1200 (bytes)

For HTTP clients that support (Brotli) compression encoding, responses that are larger than than this threshold will be compressed (also note that for clients that accept compression, any streaming responses from queries are compressed as well, since the size is not known beforehand).

http:
  compressionThreshold:  1200

cors - Type: boolean; Default: true

Enable Cross Origin Resource Sharing, which allows requests across a domain.

corsAccessList - Type: array; Default: null

An array of allowable domains with CORS

headersTimeout - Type: integer; Default: 60,000 milliseconds (1 minute)

Limit the amount of time the parser will wait to receive the complete HTTP headers with.

keepAliveTimeout - Type: integer; Default: 30,000 milliseconds (30 seconds)

Sets the number of milliseconds of inactivity the server needs to wait for additional incoming data after it has finished processing the last response.

port - Type: integer; Default: 9926

The port used to access the component server.

securePort - Type: integer; Default: null

The port the HarperDB component server uses for HTTPS connections. This requires a valid certificate and key.

timeout - Type: integer; Default: Defaults to 120,000 milliseconds (2 minutes)

The length of time in milliseconds after which a request will timeout.

http:
    cors: true
    corsAccessList:
      - null
    headersTimeout: 60000
    https: false
    keepAliveTimeout: 30000
    port: 9926
    securePort: null
    timeout: 120000

`threads`

threads - Type: number; Default: One less than the number of logical cores/ processors

The threads option specifies the number of threads that will be used to service the HTTP requests for the operations API and custom functions. Generally, this should be close to the number of CPU logical cores/processors to ensure the CPU is fully utilized (a little less because HarperDB does have other threads at work), assuming HarperDB is the main service on a server.

threads: 11

`clustering`

The clustering section configures the clustering engine, this is used to replicate data between instances of HarperDB.

Clustering offers a lot of different configurations, however in a majority of cases the only options you will need to pay attention to are:

clustering.enabled Enable the clustering processes.
clustering.hubServer.cluster.network.port The port other nodes will connect to. This port must be accessible from other cluster nodes.
clustering.hubServer.cluster.network.routesThe connections to other instances.
clustering.nodeName The name of your node, must be unique within the cluster.
clustering.user The name of the user credentials used for Inter-node authentication.

enabled - Type: boolean; Default: false

Enable clustering.

Note: If you enabled clustering but do not create and add a cluster user you will get a validation error. See user description below on how to add a cluster user.

clustering:
  enabled: true

clustering.hubServer.cluster

Clustering’s hubServer facilitates the HarperDB mesh network and discovery service.

clustering:
  hubServer:
    cluster:
      name: harperdb
      network:
        port: 9932
        routes:
          - host: 3.62.184.22
            port: 9932
          - host: 3.735.184.8
            port: 9932

name - Type: string, Default: harperdb

The name of your cluster. This name needs to be consistent for all other nodes intended to be meshed in the same network.

port - Type: integer, Default: 9932

The port the hub server uses to accept cluster connections

routes - Type: array, Default: null

An object array that represent the host and port this server will cluster to. Each object must have two properties port and host. Multiple entries can be added to create network resiliency in the event one server is unavailable. Routes can be added, updated and removed either by directly editing the harperdb-config.yaml file or by using the cluster_set_routes or cluster_delete_routes API endpoints.

host - Type: string

The host of the remote instance you are creating the connection with.

port - Type: integer

The port of the remote instance you are creating the connection with. This is likely going to be the clustering.hubServer.cluster.network.port on the remote instance.

clustering.hubServer.leafNodes

clustering:
  hubServer:
    leafNodes:
      network:
        port: 9931

port - Type: integer; Default: 9931

The port the hub server uses to accept leaf server connections.

clustering.hubServer.network

clustering:
  hubServer:
    network:
      port: 9930

port - Type: integer; Default: 9930

Use this port to connect a client to the hub server, for example using the NATs SDK to interact with the server.

clustering.leafServer

Manages streams, streams are ‘message stores’ that store table transactions.

clustering:
  leafServer:
    network:
      port: 9940
      routes:
        - host: 3.62.184.22
          port: 9931
        - host: node3.example.com
          port: 9931
    streams:
      maxAge: 3600
      maxBytes: 10000000
      maxMsgs: 500
      path: /user/hdb/clustering/leaf

port - Type: integer; Default: 9940

Use this port to connect a client to the leaf server, for example using the NATs SDK to interact with the server.

routes - Type: array; Default: null

An object array that represent the host and port the leaf node will directly connect with. Each object must have two properties port and host. Unlike the hub server, the leaf server will establish connections to all listed hosts. Routes can be added, updated and removed either by directly editing the harperdb-config.yaml file or by using the cluster_set_routes or cluster_delete_routes API endpoints.

host - Type: string

The host of the remote instance you are creating the connection with.

port - Type: integer

The port of the remote instance you are creating the connection with. This is likely going to be the clustering.hubServer.cluster.network.port on the remote instance.

clustering.leafServer.streams

maxAge - Type: integer; Default: null

The maximum age of any messages in the stream, expressed in seconds.

maxBytes - Type: integer; Default: null

The maximum size of the stream in bytes. Oldest messages are removed if the stream exceeds this size.

maxMsgs - Type: integer; Default: null

How many messages may be in a stream. Oldest messages are removed if the stream exceeds this number.

path - Type: string; Default: <ROOTPATH>/clustering/leaf

The directory where all the streams are kept.

logLevel - Type: string; Default: error

Control the verbosity of clustering logs.

clustering:
  logLevel: error

There exists a log level hierarchy in order as trace, debug, info, warn, and error. When the level is set to trace logs will be created for all possible levels. Whereas if the level is set to warn, the only entries logged will be warn and error. The default value is error.

nodeName - Type: string; Default: null

The name of this node in your HarperDB cluster topology. This must be a value unique from the rest of the cluster node names.

Note: If you want to change the node name make sure there are no subscriptions in place before doing so. After the name has been changed a full restart is required.

clustering:
  nodeName: great_node

tls

Transport Layer Security default values are automatically generated on install.

clustering:
  tls:
    certificate: ~/hdb/keys/certificate.pem
    certificateAuthority: ~/hdb/keys/ca.pem
    privateKey: ~/hdb/keys/privateKey.pem
    insecure: true
    verify: true

certificate - Type: string; Default: <ROOTPATH>/keys/certificate.pem

Path to the certificate file.

certificateAuthority - Type: string; Default: <ROOTPATH>/keys/ca.pem

Path to the certificate authority file.

privateKey - Type: string; Default: <ROOTPATH>/keys/privateKey.pem

Path to the private key file.

insecure - Type: boolean; Default: true

When true, will skip certificate verification. For use only with self-signed certs.

republishMessages - Type: boolean; Default: false

When true, all transactions that are received from other nodes are republished to this node's stream. When subscriptions are not fully connected between all nodes, this ensures that messages are routed to all nodes through intermediate nodes. This also ensures that all writes, whether local or remote, are written to the NATS transaction log. However, there is additional overhead with republishing, and setting this is to false can provide better data replication performance. When false, you need to ensure all subscriptions are fully connected between every node to every other node, and be aware that the NATS transaction log will only consist of local writes.

verify - Type: boolean; Default: true

When true, hub server will verify client certificate using the CA certificate.

user - Type: string; Default: null

The username given to the cluster_user. All instances in a cluster must use the same clustering user credentials (matching username and password).

Inter-node authentication takes place via a special HarperDB user role type called cluster_user.

The user can be created either through the API using an add_user request with the role set to cluster_user, or on install using environment variables CLUSTERING_USER=cluster_person CLUSTERING_PASSWORD=pass123! or CLI variables harperdb --CLUSTERING_USER cluster_person --CLUSTERING_PASSWORD pass123!

clustering:
  user: cluster_person

`localStudio`

The localStudio section configures the local HarperDB Studio, a simplified GUI for HarperDB hosted on the server. A more comprehensive GUI is hosted by HarperDB at https://studio.harperdb.io. Note, all database traffic from either localStudio or HarperDB Studio is made directly from your browser to the instance.

enabled - Type: boolean; Default: false

Enabled the local studio or not.

localStudio:
  enabled: false

`logging`

The logging section configures HarperDB logging across all HarperDB functionality. This includes standard text logging of application and database events as well as structured data logs of record changes. Logging of application/database events are logged in text format to the ~/hdb/log/hdb.log file (or location specified by logging.root).

In addition, structured logging of data changes are also available:

auditLog - Type: boolean; Default: false

Enabled table transaction logging.

logging:
  auditLog: false

To access the audit logs, use the API operation read_audit_log. It will provide a history of the data, including original records and changes made, in a specified table.

{
  "operation": "read_audit_log",
  "schema": "dev",
  "table": "dog"
}

file - Type: boolean; Default: true

Defines whether or not to log to a file.

logging:
  file: true

auditRetention - Type: string|number; Default: 3d

This specifies how long audit logs should be retained.

level - Type: string; Default: error

Control the verbosity of text event logs.

logging:
  level: error

There exists a log level hierarchy in order as trace, debug, info, warn, error, fatal, and notify. When the level is set to trace logs will be created for all possible levels. Whereas if the level is set to fatal, the only entries logged will be fatal and notify. The default value is error.

root - Type: string; Default: <ROOTPATH>/log

The path where the log files will be written.

logging:
  root: ~/hdb/log

rotation

Rotation provides the ability for a user to systematically rotate and archive the hdb.log file. To enable interval and/or maxSize must be set.

Note: interval and maxSize are approximates only. It is possible that the log file will exceed these values slightly before it is rotated.

logging:
  rotation:
    enabled: true
    compress: false
    interval: 1D
    maxSize: 100K
    path: /user/hdb/log

enabled - Type: boolean; Default: false

Enables logging rotation.

compress - Type: boolean; Default: false

Enables compression via gzip when logs are rotated.

interval - Type: string; Default: null

The time that should elapse between rotations. Acceptable units are D(ays), H(ours) or M(inutes).

maxSize - Type: string; Default: null

The maximum size the log file can reach before it is rotated. Must use units M(egabyte), G(igabyte), or K(ilobyte).

path - Type: string; Default: <ROOTPATH>/log

Where to store the rotated log file. File naming convention is HDB-YYYY-MM-DDT-HH-MM-SSSZ.log.

stdStreams - Type: boolean; Default: false

Log HarperDB logs to the standard output and error streams.

logging:
  stdStreams: false

`authentication`

The authentication section defines the configuration for the default authentication mechanism in HarperDB.

authentication:
  authorizeLocal: true
  cacheTTL: 30000
  enableSessions: true
  operationTokenTimeout: 1d
  refreshTokenTimeout: 30d

authorizeLocal - Type: boolean; Default: true

This will automatically authorize any requests from the loopback IP address as the superuser. This should be disabled for any HarperDB servers that may be accessed by untrusted users from the same instance. For example, this should be disabled if you are using a local proxy, or for general server hardening.

cacheTTL - Type: number; Default: 30000

This defines the length of time (in milliseconds) that an authentication (a particular Authorization header or token) can be cached.

enableSessions - Type: boolean; Default: true

This will enable cookie-based sessions to maintain an authenticated session. This is generally the preferred mechanism for maintaining authentication in web browsers as it allows cookies to hold an authentication token securely without giving JavaScript code access to token/credentials that may open up XSS vulnerabilities.

operationTokenTimeout - Type: string; Default: 1d

Defines the length of time an operation token will be valid until it expires. Example values: https://github.com/vercel/ms.

refreshTokenTimeout - Type: string; Default: 1d

Defines the length of time a refresh token will be valid until it expires. Example values: https://github.com/vercel/ms.

`operationsApi`

The operationsApi section configures the HarperDB Operations API. All the operationsApi configuration is optional. Any configuration that is not provided under this section will default to the http configuration section.

network

operationsApi:
  network:
    cors: true
    corsAccessList:
      - null
    headersTimeout: 60000
    keepAliveTimeout: 5000
    port: 9925
    securePort: null
    timeout: 120000

cors - Type: boolean; Default: true

Enable Cross Origin Resource Sharing, which allows requests across a domain.

corsAccessList - Type: array; Default: null

An array of allowable domains with CORS

headersTimeout - Type: integer; Default: 60,000 milliseconds (1 minute)

Limit the amount of time the parser will wait to receive the complete HTTP headers with.

keepAliveTimeout - Type: integer; Default: 5,000 milliseconds (5 seconds)

Sets the number of milliseconds of inactivity the server needs to wait for additional incoming data after it has finished processing the last response.

port - Type: integer; Default: 9925

The port the HarperDB operations API interface will listen on.

securePort - Type: integer; Default: null

The port the HarperDB operations API uses for HTTPS connections. This requires a valid certificate and key.

timeout - Type: integer; Default: Defaults to 120,000 milliseconds (2 minutes)

The length of time in milliseconds after which a request will timeout.

tls

This configures the Transport Layer Security for HTTPS support.

operationsApi:
  tls:
    certificate: ~/hdb/keys/certificate.pem
    certificateAuthority: ~/hdb/keys/ca.pem
    privateKey: ~/hdb/keys/privateKey.pem

certificate - Type: string; Default: <ROOTPATH>/keys/certificate.pem

Path to the certificate file.

certificateAuthority - Type: string; Default: <ROOTPATH>/keys/ca.pem

Path to the certificate authority file.

privateKey - Type: string; Default: <ROOTPATH>/keys/privateKey.pem

Path to the private key file.

`componentsRoot`

componentsRoot - Type: string; Default: <ROOTPATH>/components

The path to the folder containing the local component files.

componentsRoot: ~/hdb/components

`rootPath`

rootPath - Type: string; Default: home directory of the current user

The HarperDB database and applications/API/interface are decoupled from each other. The rootPath directory specifies where the HarperDB application persists data, config, logs, and Custom Functions.

rootPath: /Users/jonsnow/hdb

`storage`

writeAsync - Type: boolean; Default: false

The writeAsync option turns off disk flushing/syncing, allowing for faster write operation throughput. However, this does not provide storage integrity guarantees, and if a server crashes, it is possible that there may be data loss requiring restore from another backup/another node.

storage:
  writeAsync: false

caching - Type: boolean; Default: true

The caching option enables in-memory caching of records, providing faster access to frequently accessed objects. This can incur some extra overhead for situations where reads are extremely random and don't benefit from caching.

storage:
  caching: true

compression - Type: boolean; Default: false

The compression option enables compression of records in the database. This can be helpful for very large databases in reducing storage requirements and potentially allowing more data to be cached. This uses the very fast LZ4 compression algorithm, but this still incurs extra costs for compressing and decompressing.

storage:
  compression: false

noReadAhead - Type: boolean; Default: true

The noReadAhead option advises the operating system to not read ahead when reading from the database. This provides better memory utilization, except in situations where large records are used or frequent range queries are used.

storage:
  noReadAhead: true

prefetchWrites - Type: boolean; Default: true

The prefetchWrites option loads data prior to write transactions. This should be enabled for databases that are larger than memory (although it can be faster to disable this for smaller databases).

storage:
  prefetchWrites: true

path - Type: string; Default: <rootPath>/schema

The path configuration sets where all database files should reside.

storage:
  path: /users/harperdb/storage

Note: This configuration applies to all database files, which includes system tables that are used internally by HarperDB. For this reason if you wish to use a non default path value you must move any existing schemas into your path location. Existing schemas is likely to include the system schema which can be found at <rootPath>/schema/system.

`tls`

Transport Layer Security

tls:
    certificate: ~/hdb/keys/certificate.pem
    certificateAuthority: ~/hdb/keys/ca.pem
    privateKey: ~/hdb/keys/privateKey.pem

certificate - Type: string; Default: <ROOTPATH>/keys/certificate.pem

Path to the certificate file.

certificateAuthority - Type: string; Default: <ROOTPATH>/keys/ca.pem

Path to the certificate authority file.

privateKey - Type: string; Default: <ROOTPATH>/keys/privateKey.pem

Path to the private key file.

`databases`

The databases section is an optional configuration that can be used to define where database files should reside down to the table level. This configuration should be set before the database and table have been created. The configuration will not create the directories in the path, that must be done by the user.\

To define where a database and all its tables should reside use the name of your database and the path parameter.

databases:
  nameOfDatabase:
    path: /path/to/database

To define where specific tables within a database should reside use the name of your database, the tables parameter, the name of your table and the path parameter.

databases:
  nameOfDatabase:
    tables:
      nameOfTable:
        path: /path/to/table

This same pattern can be used to define where the audit log database files should reside. To do this use the auditPath parameter.

databases:
  nameOfDatabase:
    auditPath: /path/to/database

Setting the database section through the command line, environment variables or API

When using command line variables,environment variables or the API to configure the databases section a slightly different convention from the regular one should be used. To add one or more configurations use a JSON object array.

Using command line variables:

--DATABASES [{\"nameOfSchema\":{\"tables\":{\"nameOfTable\":{\"path\":\"\/path\/to\/table\"}}}}]

Using environment variables:

DATABASES=[{"nameOfSchema":{"tables":{"nameOfTable":{"path":"/path/to/table"}}}}]

Using the API:

{
  "operation": "set_configuration",
  "databases": [{
    "nameOfDatabase": {
      "tables": {
        "nameOfTable": {
          "path": "/path/to/table"
        }
      }
    }
  }]
}

PreviousEnable Mixed Content NextHarperDB CLI

Last updated 1 year ago