Setting Cloudera Director Properties

This topic lists the configuration properties recognized by Cloudera Director. Upon installation, these properties are pre-configured with reasonable default values, and you can run either client or server versions without specifying any of them. However, you might want to customize one or more properties, depending on your environment and the Cloudera Director features you want to use.

Setting Configuration Properties

The Cloudera Director command line provides the simplest way to specify a configuration property. For example:

./bin/cloudera-director bootstrap aws.simple.conf \
--lp.pipeline.retry.maxWaitBetweenAttempts=60
./bin/cloudera-director-server --lp.security.disabled=false

For users upgrading Cloudera Director

If you modified the application.properties file in Cloudera Director, the result of an upgrade depends on the version of Linux you are using:
  • RHEL and CentOS - When new properties are introduced in Cloudera Director, they are added to application.properties.rpmnew. The original application.properties file functions as before and is not overwritten with the new Cloudera Director version properties. You do not need to copy the new properties from application.properties.rpmnew to the old application.properties file.
  • Ubuntu - The modified Cloudera Director application.properties file is backed up to a file named application.properties.dpkg-old. The original application.properties file is then overwritten by the new application.properties file containing new Cloudera Director properties. After upgrading, copy your changes from application.properties.dpkg-old to the new application.properties file.

All the new properties are commented, and they all use valid defaults, so you do not necessarily need to merge the two properties files. But you must merge the two files if you want to modify one of the newly introduced properties.

Property Types

Type Description
boolean Either true or false
char Single character
directory Valid directory path
enum Fixed set of string values; a list of each enumeration’s values is provided following the main property table below
enum list Comma-separated list of enums
file Valid file path
int Integer (32-bit)
long Long integer (64-bit)
string Ordinary character string
time unit Enumeration of time units: DAYS, HOURS, MICROSECONDS, MILLISECONDS, MINUTES, NANOSECONDS, SECONDS

Properties

Property Description
lp.access.logging.config.file
File for Cloudera Director server access log.

Type: string

Default: none; must be set if lp.access.logging.enabled is true.

lp.access.logging.enabled
Enable Cloudera Director server access logging.

Type: boolean

Default: false

lp.bootstrap.agents.maxNumberOfInstallAttempts
Maximum number of times to retry installing Cloudera Manager agent. Use -1 for unlimited.

Type: int

Default: -1

lp.bootstrap.parallelBatchSize
Parallelism for allocating and setting up cluster instances when bootstrapping a cluster.

Type: int

Default: 20

lp.bootstrap.parcels.distributeMaxConcurrentUploads
Maximum concurrent uploads of parcels across cluster.

Type: int

Default: 5

lp.bootstrap.parcels.distributeRateLimitKBs
Maximum rate of parcel upload, in KB/s.

Type: int

Default: 256000

lp.bootstrap.resume.policy
Action to take when resuming a previous bootstrap. Use RESTART to start from scratch. Use RESUME to resume from last known state. Use INTERACTIVE to prompt to ask.

Type: enum

Valid values: RESTART | RESUME | INTERACTIVE

Default: INTERACTIVE

lp.cache.health.expirationMultiplier
Multiplier applied to polling rate to find health cache expiration duration; negative = disable health polling.

Type: int

Default: 2

lp.cache.health.numberOfCacheExecutionThreads
Number of threads used to poll for service and cluster health.

Type: int

Default: 5

lp.cache.health.pollingRateInMilliseconds
Rate at which service and cluster health is polled, in milliseconds.

Type: long

Default: 30000

lp.cleanup.databases.intervalBetweenAttemptsInMs
Wait time between attempts to destroy external databases, in milliseconds.

Type: long

Default: 60000

lp.cleanup.databases.maxNumberOfDeleteAttempts
Maximum number of times to retry destroying external databases; -1 = unlimited.

Type: int

Default: 5

lp.cloud.databaseServers.allocate.timeoutInMinutes
Time to wait for allocated database server instances to begin running to have ports available.

Type: int

Default: 20

lp.cloud.databaseServers.destroy.timeoutInMinutes
Time to wait for terminated database server instances to stop running to have ports no longer available.

Type: int

Default: 20

lp.cloud.instances.allocate.numberOfRetriesOnConnectionError
Number of times to retry connecting to newly allocated instances over SSH.

Type: int

Default: 3

lp.cloud.instances.allocate.parallelBatchSize
Parallelism for waiting for SSH to become available on newly allocated instances.

Type: int

Default: 20

lp.cloud.instances.allocate.timeBetweenConnectionRetriesInSeconds
Time to wait between attempts to connect to newly allocated instances over SSH.

Type: int

Default: 1

lp.cloud.instances.allocate.timeoutInMinutes
Time to wait for allocated instances to begin running to have SSH ports available.

Type: int

Default: 20

lp.cloud.instances.terminate.timeoutInMinutes
Time to wait for terminated instances to stop running.

Type: int

Default: 20

lp.debug.collectDiagnosticDataOnFailure
Collect Cloudera Manager diagnostic data on unrecoverable bootstrap failure.

Type: boolean

Default: true

lp.debug.createDiagnosticDataDownloadDirectory
Create the download directory for Cloudera Manager diagnostic data if it does not already exist.

Type: boolean

Default: true

lp.debug.diagnosticDataDownloadDirectory
Destination directory for downloaded Cloudera Manager diagnostic data.

Type: string

Default: /tmp

lp.debug.downloadDiagnosticData
Download Cloudera Manager diagnostic data once it has been collected.

Type: boolean

Default: true

lp.debug.dumpClouderaManagerLogsOnFailure
Dump Cloudera Manager log entries into the Director logs on unrecoverable bootstrap failure.

Type: boolean

Default: false

lp.debug.dumpClusterLogsOnFailure
Dump cluster service logs, standard output, or standard error into the Cloudera Director logs on unrecoverable bootstrap failure.

Type: boolean

Default: false

lp.encryption.twoWayCipher
Cipher used to encrypt data. Possible values:
  • desede - Triple DES
  • passthrough - No encryption
  • transitional - Changing encryption

Type: string

Default: desede

lp.encryption.twoWayCipherConfig
The configuration string for the chosen cipher.

Type: string

Default: ZGVmYXVsdGRpcmVjdG9yZGVzZWRla2V5

Cloudera recommends that you configure a different triple DES key. A warning appears in the server log if the default key is detected.

lp.metrics.durationUnits
Time units for reporting durations in metrics.

Type: time unit

Valid values: DAYS | HOURS | MICROSECONDS | MILLISECONDS | MINUTES | NANOSECONDS | SECONDS

Default: MILLISECONDS

lp.metrics.enabled
Enable metrics gathering

Type: boolean

Default: false

lp.metrics.location
Directory for storing metrics reports.

Type: directory

Default: $LOG_DIR/metrics

lp.metrics.rateUnits
Time units for reporting rates in metrics.

Type: time unit

Valid values: DAYS | HOURS | MICROSECONDS | MILLISECONDS | MINUTES | NANOSECONDS | SECONDS

Default: SECONDS

lp.metrics.reportingRate
Frequency of metrics reporting, in minutes.

Type: long

Default: 1

lp.pipeline.retry.maxNumberOfAttempts
Maximum number of times to retry failed pipeline jobs; -1 = unlimited.

Type: int

Default: -1 for client, 16 for server

lp.pipeline.retry.maxWaitBetweenAttempts
Maximum wait time between pipeline retry attempts, in seconds.

Type: int

Default: 45

lp.proxy.http.domain
NT domain for HTTP proxy authentication; none = no domain.

Type: string

Default: none

lp.proxy.http.host
HTTP proxy host; none = no proxy.

Type: string

Default: none

lp.proxy.http.password
HTTP proxy password; none = no password.

Type: string

Default: none

lp.proxy.http.port
HTTP proxy port; -1 = no proxy.

Type: int

Default: -1

lp.proxy.http.preemptiveBasicProxyAuth
Whether to preemptively authenticate to HTTP proxy.

Type: boolean

Default: false

lp.proxy.http.username
HTTP proxy username; none = no username.

Type: string

Default: none

lp.proxy.http.workstation
Originating workstation in NT domain for HTTP proxy authentication; none = no workstation.

Type: string

Default: none

lp.remote.hostAndPort
Host and port of remote Cloudera Director server. Used by Cloudera Director client to connect to a running Cloudera Director server.

Type: string

Default: localhost:7189

lp.remote.password
Remote Cloudera Director server password (client only).

Type: string

Default:

lp.remote.username
Remote Cloudera Director server username (client only).

Type: string

Default: none

lp.remote.terminate.assumeYes
Whether to skip prompting user to confirm termination for client terminate-remote command.

Type: boolean

Default: false

lp.security.enabled
Whether to enable Cloudera Director server security (server only).

Type: boolean

Default: true

lp.security.userSource
Source for user account information (server only).

Type: enum

Default: internal

lp.ssh.connectTimeoutInSeconds
SSH connection timeout.

Type: int

Default: 30

lp.ssh.heartbeatIntervalInSeconds
SSH heartbeat interval.

Type: int

Default: 45

lp.ssh.readTimeoutInSeconds
SSH read timeout.

Type: int

Default: 30

lp.task.evictionRate
Rate of execution of database eviction, in milliseconds.

Type: long

Default: 600000

lp.terminate.assumeYes
Whether to skip prompting user to confirm termination for client terminate command.

Type: boolean

Default: false

lp.terminate.deployment.clouderaManagerServerStopWaitTimeInMs
Time to wait for Cloudera Manager to stop when terminating a deployment, in milliseconds.

Type: long

Default: 300000

lp.terminate.deployment.timeBetweenConnectionRetriesInMs
Time to wait between checks for whether Cloudera Manager has been terminated.

Type: int

Default: 10000

lp.update.parallelBatchSize
Parallelism for allocating and setting up cluster instances when bootstrapping a cluster.

Type: int

Default: 20

lp.update.redeployClientConfigs.numberOfRetries
Maximum number of times to retry deploying Cloudera Manager client configurations; -1 = unlimited.

Type: int

Default: 5

lp.update.redeployClientConfigs.sleepAfterFailureInSeconds
Wait time between attempts to deploy Cloudera Manager client configurations, in seconds.

Type: int

Default: 10

lp.update.restartCluster.numberOfRetries
Maximum number of times to retry a Cloudera Manager rolling restart; -1 = unlimited.

Type: int

Default: 5

lp.update.restartCluster.rollingRestartSlaveBatchSize
Number of instances with Cloudera Manager worker roles to restart at a time.

Type: int

Default: 20

lp.update.restartCluster.rollingRestartSlaveFailCountThreshold
Threshold for number of worker host batches that are allowed to fail to restart before the entire command is considered failed (advanced use only).

Type: int

Default: 0

lp.update.restartCluster.rollingRestartSleepSeconds
Number of seconds to sleep between restarts of Cloudera Manager worker host batches.

Type: int

Default: 0

lp.update.restartCluster.sleepAfterFailureInSeconds
Wait time between attempts to perform a Cloudera Manager rolling restart, in seconds.

Type: int

Default: 10

lp.validate.dumpTemplates
Whether to output validated configuration data as JSON.

Type: boolean

Default: false

lp.webapp.anonymousUsageDataAllowed
Allow Cloudera Director to send anonymous usage information to help Cloudera improve the product.

Type: boolean

Default: true

lp.webapp.documentationType
Whether Cloudera Director opens the latest help from the Cloudera web site (online) or locally installed help (embedded).

Type: enumerated string {ONLINE, EMBEDDED}

Default: ONLINE

port
Cloudera Director server port (server only).

Type: int

Default: 7189

server.sessionTimeout
Cloudera Director server session timeout (server only).

Type: int

Default: 18000