Documentation Home
MySQL 8.4 Reference Manual
Related Documentation Download this Manual
PDF (US Ltr) - 40.0Mb
PDF (A4) - 40.1Mb
Man Pages (TGZ) - 259.1Kb
Man Pages (Zip) - 366.2Kb
Info (Gzip) - 4.0Mb
Info (Zip) - 4.0Mb


25.6.17.51 The ndbinfo processes Table

This table contains information about NDB Cluster node processes; each node is represented by the row in the table. Only nodes that are connected to the cluster are shown in this table. You can obtain information about nodes that are configured but not connected to the cluster from the nodes and config_nodes tables.

The processes table contains the following columns:

  • node_id

    The node's unique node ID in the cluster

  • node_type

    Type of node (management, data, or API node; see text)

  • node_version

    Version of the NDB software program running on this node.

  • process_id

    This node's process ID

  • angel_process_id

    Process ID of this node's angel process

  • process_name

    Name of the executable

  • service_URI

    Service URI of this node (see text)

Notes

node_id is the ID assigned to this node in the cluster.

The node_type column displays one of the following three values:

  • MGM: Management node.

  • NDB: Data node.

  • API: API or SQL node.

For an executable shipped with the NDB Cluster distribution, node_version shows the software Cluster version string, such as 8.4.0-ndb-8.4.0.

process_id is the node executable's process ID as shown by the host operating system using a process display application such as top on Linux, or the Task Manager on Windows platforms.

angel_process_id is the system process ID for the node's angel process, which ensures that a data node or SQL is automatically restarted in cases of failures. For management nodes and API nodes other than SQL nodes, the value of this column is NULL.

The process_name column shows the name of the running executable. For management nodes, this is ndb_mgmd. For data nodes, this is ndbd (single-threaded) or ndbmtd (multithreaded). For SQL nodes, this is mysqld. For other types of API nodes, it is the name of the executable program connected to the cluster; NDB API applications can set a custom value for this using Ndb_cluster_connection::set_name().

service_URI shows the service network address. For management nodes and data nodes, the scheme used is ndb://. For SQL nodes, this is mysql://. By default, API nodes other than SQL nodes use ndb:// for the scheme; NDB API applications can set this to a custom value using Ndb_cluster_connection::set_service_uri(). regardless of the node type, the scheme is followed by the IP address used by the NDB transporter for the node in question. For management nodes and SQL nodes, this address includes the port number (usually 1186 for management nodes and 3306 for SQL nodes). If the SQL node was started with the bind_address system variable set, this address is used instead of the transporter address, unless the bind address is set to *, 0.0.0.0, or ::.

Additional path information may be included in the service_URI value for an SQL node reflecting various configuration options. For example, mysql://198.51.100.3/tmp/mysql.sock indicates that the SQL node was started with the skip_networking system variable enabled, and mysql://198.51.100.3:3306/?server-id=1 shows that replication is enabled for this SQL node.