B.5. High Availability and Durability

Documentation

VoltDB Home » Documentation » Administrator's Guide

B.5. High Availability and Durability

The following tables describe the metrics related to Volt features that provide high availability and durability for the database, including snapshots, command logging, and Active(N) cross datacenter replication.

Table B.8. Snapshots

Metrics	Type	Description
voltdb_snapshot_site_summary_info	Metadata	Informational metric. One for every snapshot file in the recent snapshots performed on the cluster. Tags: SNAPSHOT_NONCE - The unique identifier for the snapshot. TABLE_NAME - The name of the database table whose data the file contains. SNAPSHOT_COLLECTION_ITERATIONS - . SNAPSHOT_COLLECTION_TIME - The length of time (in seconds) it took to complete the snapshot. SNAPSHOT_WRITE_TIME - The length of time (in seconds) it took to complete write stage of snapshot operation.
voltdb_snapshot_summary_info	Metadata	Informational metric. One for every snapshot file in the recent snapshots performed on the cluster. Tags: SNAPSHOT_NONCE - The unique identifier for the snapshot. SNAPSHOT_TXN_ID - The transaction ID of the snapshot. SNAPSHOT_TYPE - String value indicating how the snapshot was initiated. Possible values are: "Auto" - an automated snapshot as defined by the configuration file; "Commandlog" - a command log snapshot; "Manual" - a manual snapshot initiated by a user. SNAPSHOT_PATH - The directory path where the snapshot file resides. SNAPSHOT_START_TIME - The timestamp when the snapshot began (in milliseconds). SNAPSHOT_END_TIME - The timestamp when the snapshot was completed (in milliseconds). SNAPSHOT_BYTES_WRITTEN - Total number of bytes written to the file so far. SNAPSHOT_PROGRESS - For snapshots currently in progress, the percent complete at the time of the call (0-100). SNAPSHOT_RESULT - Value indicating whether the writing of the snapshot file was successful ("Success") or not ("Failure").
voltdb_snapshot_table_summary_info	Metadata	Informational metric. One for every snapshot file in the recent snapshots performed on the cluster. Tags: SNAPSHOT_NONCE - The unique identifier for the snapshot. SNAPSHOT_TXN_ID - The transaction ID of the snapshot. TABLE_NAME - The name of the database table whose data the file contains. TABLE_FILENAME - The file name. SNAPSHOT_BYTES_WRITTEN - The total size, in bytes, of the file. SNAPSHOT_RESULT - Value indicating whether the writing of the snapshot file was successful ("Success") or not ("Failure").

Table B.9. Command Logging

Metrics	Type	Description
voltdb_commandlog_fsync_interval_seconds	Gauge	The average interval between the last 10 fsync system calls.
voltdb_commandlog_in_use_segment_count_total	Gauge	The total number of segment files in use for command logging.
voltdb_commandlog_outstanding_bytes_bytes	Gauge	The size, in bytes, of pending command log data. For synchronous logging, this value is always zero.
voltdb_commandlog_outstanding_txns_total	Gauge	The number of transactions that have been initiated for which the log has yet to be written to disk. For synchronous logging, this value is always zero.
voltdb_commandlog_segment_count_total	Gauge	The number of segment files allocated, including currently unused segments.

Table B.10. Active(N) and XDCR

Metrics	Type	Description
voltdb_dr_conflicts_count_total	Counter	The total number of conflicts that have been recorded for this table in this partition.
voltdb_dr_constraint_violation_count_total	Counter	The number of constraint violation conflicts that occurred.
voltdb_dr_consumer_info	Metadata	Tags: DR_STATE - A text string indicating the current state of replication. Possible values are: UNINITIALIZED - DR has not begun yet or has stopped INITIALIZE - DR is enabled and the consumer is attempting to contact the producer SYNC - DR has started and the consumer is synchronizing by receiving snapshots of existing data from the master RECEIVE - DR is underway and the consumer is receiving binary logs from the master DISABLE - DR has been canceled for some reason and the consumer is stopping DR.
voltdb_dr_consumer_bytes_replicated_bytes	Counter	Total number of bytes this consumer received.
voltdb_dr_consumer_partition_info	Metadata	Tags: IS_COVERED - Boolean value indicating whether this partition is currently connected to and receiving data from a matching partition on the producer cluster. COVERING_HOST - The host name of the server in the producer cluster that is providing DR data to this partition. If IS_COVERED is "false", this label is empty. IS_PAUSED - Boolean indicating whether this partition is paused. A partition "pauses" when the schema of the DR tables on the producer change to no longer match the consumer and all binary logs prior to the change have been processed. CONSUMER_LIMIT_TYPE - The type of limit on the DR queue. The response is either BYTES or BUFFERS. LAST_APPLIED_DR_PROTOCOL - The current DR protocol version of binary logs being received and applied for this partition.
voltdb_dr_consumer_partition_available_buffers_total	Gauge	The number of free buffers left in the DR queue.
voltdb_dr_consumer_partition_available_buffer_bytes_bytes	Gauge	The number of free bytes left in the DR queue.
voltdb_dr_consumer_partition_duplicate_buffers_total	Gauge	The number of repeated buffers received after the initial packets were dropped because the queue was full.
voltdb_dr_consumer_partition_ignored_buffers_total	Gauge	The number of buffers received but dropped because the queue was full.
voltdb_dr_consumer_partition_last_applied_timestamp_seconds	Gauge	The timestamp of the last transaction successfully applied to this partition on the consumer.
voltdb_dr_consumer_partition_last_received_timestamp_seconds	Gauge	The timestamp of the last transaction received from the producer.
voltdb_dr_consumer_remote_creation_timestamp_seconds	Gauge	The timestamp when the remote cluster started for the first time.
voltdb_dr_divergence_count_total	Counter	The number of conflicts that may have resulted in divergence between the clusters, which is a subset of the total conflicts.
voltdb_dr_last_conflict_timestamp_seconds	Gauge	The timestamp of the last conflict.
voltdb_dr_missing_row_count_total	Counter	The number of missing row conflicts that occurred.
voltdb_dr_producer_cluster_info	Metadata	Informational metric, presents cluster level metadata. Tags: DR_STATE - The current state of the DR relationship. Possible values are the following: "Disabled", "Pending", "Active", "Stopped". LAST_APPLIED_DR_PROTOCOL - The current DR protocol version of binary logs being received and applied for this partition. SUPPORTED_DR_PROTOCOL - The highest version of DR protocol this cluster is capable of using to send data to consumers.
voltdb_dr_producer_node_info	Metadata	Informational metric, presents node level metadata. Tags: DR_STATE - The current state of the DR relationship. Possible values are the following: "Disabled", "Pending", "Active", "Stopped". DR_SYNC_SNAPSHOT_STATE - The current state of the synchronization snapshot that begins replication. During normal operation, this value is "None" indicating either that replication is not active or that transactions are actively being replicated. If a synchronization snapshot is in progress, this value provides additional information about the specific activity underway.
voltdb_dr_producer_node_remote_creation_timestamp_seconds	Gauge	The timestamp (in seconds) when the remote cluster started for the first time.
voltdb_dr_producer_node_rows_acked_for_sync_snapshot_total	Gauge
voltdb_dr_producer_node_rows_in_sync_snapshot_total	Gauge
voltdb_dr_producer_node_tasks_queue_depth_total	Gauge	The number of DR tasks waiting to be processed.
voltdb_dr_producer_partition_info	Metadata	Informational metric. Tags: DR_STREAM_TYPE - The type of stream, which can either be "Transactions" or "Snapshot". DR_LAST_QUEUED_ID - The ID of the last transaction queued for transmission to the consumer. DR_LAST_ACK_ID - The ID of the last transaction acknowledged by the consumer. DR_IS_SYNCED - Indicates whether the database is currently being replicated. If replication has not started, or the overflow capacity has been exceeded (that is, replication has failed), the value of ISSYNCED is "false". If replication is currently in progress, the value is "true". DR_MODE - Indicates whether this particular partition is replicating data to the consumer ("NOrmal") or not ("Paused"). Only one copy of each logical partition actually sends data during replication. So for clusters with a K-safety value greater than zero, not all physical partitions will report "Normal" even when replication is in progress. DR_CONNECTION_STATUS - Indicates whether the connection to the consumer is operational ("UP") or not ("DOWN"). CONSUMER_LIMIT_TYPE - The type of limit on the DR queue. The response is either BYTES or BUFFERS. CURRENT_DR_PROTOCOL - The DR protocol version currently in use when sending data to consumers. SUPPORTED_DR_PROTOCOL - The highest version of DR protocol this cluster is capable of using to send data to consumers.
voltdb_dr_producer_partition_available_to_send_buffers_total	Gauge	The number of buffers waiting to be sent to the consumer.
voltdb_dr_producer_partition_available_to_send_total_bytes	Gauge	The number of bytes waiting to be sent to the consumer.
voltdb_dr_producer_partition_buffers_waiting_for_ack_total	Gauge	The total number of buffers in this partition waiting for acknowledgement from the consumer.
voltdb_dr_producer_partition_last_ack_timestamp_seconds	Gauge	The total number of bytes currently queued for transmission to the consumer.
voltdb_dr_producer_partition_last_queued_timestamp_seconds	Gauge	The timestamp of the last transaction queued for transmission to the consumer.
voltdb_dr_producer_partition_queued_in_memory_total_bytes	Gauge	The total number of bytes of queued data currently held in memory. If the amount of total bytes is larger than the amount in memory, the remainder is kept in overflow storage on disk.
voltdb_dr_producer_partition_queued_total_bytes	Gauge	The total number of bytes currently queued for transmission to the consumer.
voltdb_dr_producer_partition_queue_gap_total	Gauge	The number of missing transactions between those already acknowledged by the consumer and the next available for transmission. Under normal operating conditions, this value is zero.
voltdb_dr_producer_partition_round_trip_time_seconds	Histogram	The distribution of time it took to receive acknowledgement from the consumer.
voltdb_dr_role_info	Metadata	Informational metric. Tags: DR_ROLE_NAME - The role of the current cluster in a DR relationship. Possible values are NONE, MASTER, REPLICA, and XDCR. DR_STATE - The current state of the DR relationship. DISABLED, PENDING, ACTIVE, STOPPED. REMOTE_CLUSTER_ID - The DR ID of the other DR cluster, or -1 if not available (for example, when DR is disabled or communication has not begun). SUPPORTED_DR_PROTOCOL - The highest version of DR protocol this cluster is capable of using to send data to consumers.
voltdb_dr_row_timestamp_mismatch_count_total	Counter	The number of timestamp mismatch conflicts that occurred.
voltdb_dr_schema_change_info	Metadata	Informational metric containing metadata. Tags: SITE_ID - Numeric ID of the execution site on the host node. TABLE_TYPE - The type of the table. E.g. "PersistentTable" for normal data tables. TABLE_NAME - The name of the database table for which schema was mismatched. CLUSTER_ID - The numeric ID of the current cluster. REMOTE_CLUSTER_ID - The numeric ID of the remote cluster. DR_SCHEMA_CHANGE_MATCH - A text string of "true" or "false" indicating whether the schema for the table matches on the two clusters.
voltdb_dr_schema_change_tuple_count_total	Counter	The total number of tuples exchanged for this tuple while the schema did not match.

< previous

table of contents

next >