Operational Changes (MapR 6.0.1)

Note the following functional changes to existing commands in Version 6.0.1.

MapR Installer

Unlike earlier MapR Installer versions, MapR Installer 1.9 requires some new user ID and password information for installing Grafana and ElasticSearch/Kibana:

  • When installing a secure cluster using the MapR Installer web-based interface:
    • The installer does not request a user ID or password for Grafana because the installer uses the cluster admin authentication information (requested separately) for security with Grafana.
    • The installer requires you to supply the Elasticsearch/Kibana password. Enter the password for the admin user ID.
  • When installing a nonsecure cluster using the MapR Installer web-based interface:
    • The installer requires you to provide the Grafana password. Enter the password for the admin user.
    • No Elasticsearch/Kibana user ID or password is required.

MapR-FS

The -cid option is not supported with mrdirectorystats utility. Instead, use -c to specify the container ID. See mrdirectorystats for more information.

MapR-DB

Starting in the MapR 6.0.1 release, you do not have to enable a separate query service to use secondary indexes. The Operational Applications with MapR-DB template installs and configures the replication gateways needed to update secondary indexes in MapR-DB JSON and includes the components needed to run OJAI queries.

However, you must enable the OJAI Distributed Query Service to use the following features:

  • Advanced secondary index selection
  • Sorts on large data sets
  • Parallel query execution

For complete information, see Preparing Clusters for Querying using Secondary Indexes on JSON Tables

MapR-ES

As of MapR 6.0.1, along with support of Apache Kafka 1.0, the java.util.Collection interface is being used. This impacts applications using certain APIs. See the Backward Compatibility section in MapR-ES Java API Library for detailed information.

For MapR-ES Java applications, the behavior has changed for max.partition.fetch.bytes. If the first record batch in the first non-empty partition of the fetch is larger than this configuration, the record batch is still returned to ensure that the consumer can make progress. See Configuration Parameters for more information.

For MapR-ES Java applications, the fs.mapr.rpc.timeout parameter is valid for MapR 6.0.0, but has been replaced with streams.rpc.timeout.ms in MapR 6.0.1.

For MapR-ES Java applications, the streams.rpc.timeout.ms parameter replaces the following Apache Kafka parameters: request.timeout.ms, reconnect.backoff.ms, and retry.backoff.ms.

MapR monitoring uses 2 MB disk space per minute per node when MapR-ES metrics is enabled. This is approximately 3 GB per day on a single node or 7 GB per node per day with a 3X replication. This stream metrics data is automatically deleted every 3 days.