Documentation Home
MySQL NDB Cluster 7.5
Related Documentation Download this Excerpt

7.10.32 The ndbinfo restart_info Table

The restart_info table contains information about node restart operations. Each entry in the table corresponds to a node restart status report in real time from a data node with the given node ID. Only the most recent report for any given node is shown.

The following table provides information about the columns in the restart_info table. For each column, the table shows the name, data type, and a brief description. Additional information can be found in the notes following the table.

Column NameTypeDescription
node_idintegerNode ID in the cluster
node_restart_statusVARCHAR(256)Node status; see text for values. Each of these corresponds to a possible value of node_restart_status_int.
node_restart_status_intintegerNode status code; see text for values.
secs_to_complete_node_failureintegerTime in seconds to complete node failure handling
secs_to_allocate_node_idintegerTime in seconds from node failure completion to allocation of node ID
secs_to_include_in_heartbeat_protocolintegerTime in seconds from allocation of node ID to inclusion in heartbeat protocol
secs_until_wait_for_ndbcntr_masterintegerTime in seconds from being included in heartbeat protocol until waiting for NDBCNTR master began
secs_wait_for_ndbcntr_masterintegerTime in seconds spent waiting to be accepted by NDBCNTR master for starting
secs_to_get_start_permittedintegerTime in seconds elapsed from receiving of permission for start from master until all nodes have accepted start of this node
secs_to_wait_for_lcp_for_copy_meta_dataintegerTime in seconds spent waiting for LCP completion before copying meta data
secs_to_copy_meta_dataintegerTime in seconds required to copy metadata from master to newly starting node
secs_to_include_nodeintegerTime in seconds waited for GCP and inclusion of all nodes into protocols
secs_starting_node_to_request_local_recoveryintegerTime in seconds that the node just starting spent waiting to request local recovery
secs_for_local_recoveryintegerTime in seconds required for local recovery by node just starting
secs_restore_fragmentsintegerTime in seconds required to restore fragments from LCP files
secs_undo_disk_dataintegerTime in seconds required to execute undo log on disk data part of records
secs_exec_redo_logintegerTime in seconds required to execute redo log on all restored fragments
secs_index_rebuildintegerTime in seconds required to rebuild indexes on restored fragments
secs_to_synchronize_starting_nodeintegerTime in seconds required to synchronize starting node from live nodes
secs_wait_lcp_for_restartintegerTime in seconds required for LCP start and completion before restart was completed
secs_wait_subscription_handoverintegerTime in seconds spent waiting for handover of replication subscriptions
total_restart_secsintegerTotal number of seconds from node failure until node is started again

Defined values for node_restart_status_int and corresponding status names and messages (node_restart_status) are shown in the following table:

node_restart_status_int valueStatusMessage (node_restart_status)
0ALLOCATED_NODE_IDAllocated node id
1INCLUDED_IN_HB_PROTOCOLIncluded in heartbeat protocol
2NDBCNTR_START_WAITWait for NDBCNTR master to permit us to start
3NDBCNTR_STARTEDNDBCNTR master permitted us to start
4START_PERMITTEDAll nodes permitted us to start
5WAIT_LCP_TO_COPY_DICTWait for LCP completion to start copying metadata
6COPY_DICT_TO_STARTING_NODECopying metadata to starting node
7INCLUDE_NODE_IN_LCP_AND_GCPInclude node in LCP and GCP protocols
8LOCAL_RECOVERY_STARTEDRestore fragments ongoing
9COPY_FRAGMENTS_STARTEDSynchronizing starting node with live nodes
10WAIT_LCP_FOR_RESTARTWait for LCP to ensure durability
11WAIT_SUMA_HANDOVERWait for handover of subscriptions
12RESTART_COMPLETEDRestart completed
13NODE_FAILEDNode failed, failure handling in progress
14NODE_FAILURE_COMPLETEDNode failure handling completed
15NODE_GETTING_PERMITAll nodes permitted us to start
16NODE_GETTING_INCLUDEDInclude node in LCP and GCP protocols
17NODE_GETTING_SYNCHEDSynchronizing starting node with live nodes
19NODE_ACTIVERestart completed

Status numbers 0 through 12 apply on master nodes only; the remainder of those shown in the table apply to all restarting data nodes. Status numbers 13 and 14 define node failure states; 20 and 21 occur when no information about the restart of a given node is available.

See also Section 7.1, “Summary of NDB Cluster Start Phases”.

User Comments
Sign Up Login You must be logged in to post a comment.