Tables

Viewing partition distribution for GridGain 9 tables and distribution zones, and recovering lost partitions from the Disaster Recovery tab.

The Tables screen is available for GridGain 9 clusters.

This screen enables you to see the distribution of partitions between the current cluster's nodes for a selected table or distribution zone.

Viewing Partitions for a Table

To view partition distribution for a table:

1. Select the Tables tab.

   

   The Tables tab
2. In the leftmost section, find and select the required table.

   The <Table name> panel displays the primary zone, secondary zone (if configured), and total partition count for the selected table.

3. In the <Table name> section, select the required cluster node(s).

4. In the Partitions on node: <Node name>, view the partitions and their state.

5. To view those nodes that have no selected table's partitions distributed to them, toggle on Show all nodes in the top right corner.

By default, the list includes the following columns:

Column	Description
Partition ID	The partition ID.
Node name	The name of the node.
Primary	Whether the partition copy on this node is primary.
State	The state of the partition on the node. Possible states are: - UNAVAILABLE - this state might be used when the partition is not yet started or is already stopping. - HEALTHY - alive partition with a healthy state machine. - INITIALIZING - partition is starting right now. - INSTALLING_SNAPSHOT- partition is installing a Raft snapshot from the leader. - CATCHING_UP - partition is catching up, meaning that it hasn't replicated part of the log yet. - BROKEN - partition is in a broken state, typically because its state machine threw an exception. - UNKNOWN - partition state is unknown for Control Center.

In all of the Tables tab sections, you can add or remove columns to/from the entity list by selecting or deselecting column names in the context menu.

Viewing Partitions for a Distribution Zone

To view partitions for a distribution zone:

1. Select the Distribution zones tab.

   

   The Distribution zones tab
2. In the leftmost section, find and select the required distribution zone.

   The <Zone name> panel displays the total partition count, replica count, and storage profiles for the selected zone.

3. In the <Zone name> section, select the required cluster node(s).

4. In the Partitions on node: <Node name>, view the partitions and their state.

5. To view those nodes that have no selected zone's partitions distributed to them, toggle on Show all nodes in the top right corner.

By default, the list includes the following columns:

Column	Description
Partition ID	The partition ID.
Records	The amount of records a partition contains.
Node name	The name of the node.
Primary	Whether the partition copy on this node is primary.
State	The state of the partition on the node. Possible states are: - UNAVAILABLE - this state might be used when the partition is not yet started or is already stopping. - HEALTHY - alive partition with a healthy state machine. - INITIALIZING - partition is starting right now. - INSTALLING_SNAPSHOT- partition is installing a Raft snapshot from the leader. - CATCHING_UP - partition is catching up, meaning that it hasn't replicated part of the log yet. - BROKEN - partition is in a broken state, typically because its state machine threw an exception. - UNKNOWN - partition state is unknown for Control Center.

In all the Distribution zones tab sections, you can add or remove columns to/from the entity list by selecting or deselecting column names in the context menu.

Disaster Recovery

Listing Lost Partitions

The Disaster Recovery tab enables you to identify and recover lost table partitions. By default, the tab displays partitions that are in the Degraded, Read-only, or Unavailable states. If all partitions are in Healthy state, the tab will remain empty.

Disaster recovery tab

The filter panel on the right lets you refine your search among all partitions by table, node, zone, and its local and global states:

Column	Description
Partition ID	Partition ID.
Node name	Name of the node hosting the partition.
Zone	Zone where partition is located
Local state	The local state of a partition. Possible states are: - UNAVAILABLE - this state might be used when the partition is not yet started or is already stopping. - HEALTHY - alive partition with a healthy state machine. - INITIALIZING - partition is starting right now. - INSTALLING_SNAPSHOT- partition is installing a Raft snapshot from the leader. - CATCHING_UP - partition is catching up, meaning that it hasn't replicated part of the log yet. - BROKEN - partition is in a broken state, typically because its state machine threw an exception. - UNKNOWN - partition state is unknown for Control Center.
Global state	The global state of a partition. If no information for a partition is available, it changes its status to Unknown. Possible global states are: - AVAILABLE - all replicas are healthy. - DEGRADED - there are healthy replicas, and they form a majority. - READ_ONLY - there are healthy replicas, but they don't form a majority. - UNAVAILABLE - there are no healthy replicas. - UNKNOWN - partition state is unknown for Control Center.

Recovering Lost Partitions

If the partition status indicates issues with its replicas, for example, when partitions are marked as DEGRADED, READ_ONLY, or UNAVAILABLE, you can attempt to repair it using either the Restart or Reset option.

For detailed instructions on disaster recovery and repair procedures, refer to the disaster recovery documentation.

It is advisable to use Restart first. If that does not resolve the issue, then consider using Reset. Keep in mind that Reset option may lead to data loss or cluster unavailability.

You can recover a single partition by using the row-level context menu:

Disaster recovery tab

Alternatively, you can recover multiple partitions by clicking the Recover button in the top-right corner of the screen. First, select the table and zone where the lost partitions are located. You can further refine your selection by partition global state.

Note, that by default, all partitions within the selected table and zone are chosen. If you need to recover only a subset, deselect the partitions you do not want to recover.

Disaster recovery tab

When performing a Restart operation, you can additionally select partitions by node. If no nodes are selected, the operation will restart partitions on all nodes.

Disaster recovery tab

During the Reset and Restart operations the state of the selected partitions may change. This change does not necessarily indicate a successful operation. If the operation fails or ends inconclusively, an error message will be displayed.

Disaster recovery tab

You also can perform recovery operations via Gridgain CLI commands.