ndb_desc provides a detailed description of one or more NDB tables.

Usage

ndb_desc -c connection_string tbl_name -d db_name [options]

ndb_desc -c connection_string index_name -d db_name -t tbl_name

Additional options that can be used with ndb_desc are listed later in this section.

Sample Output

MySQL table creation and population statements:

USE test;

CREATE TABLE fish (
    id INT NOT NULL AUTO_INCREMENT,
    name VARCHAR(20) NOT NULL,
    length_mm INT NOT NULL,
    weight_gm INT NOT NULL,

    PRIMARY KEY pk (id),
    UNIQUE KEY uk (name)
) ENGINE=NDB;

INSERT INTO fish VALUES
    (NULL, 'guppy', 35, 2), (NULL, 'tuna', 2500, 150000),
    (NULL, 'shark', 3000, 110000), (NULL, 'manta ray', 1500, 50000),
    (NULL, 'grouper', 900, 125000), (NULL ,'puffer', 250, 2500);

Output from ndb_desc:

shell> ./ndb_desc -c localhost fish -d test -p
-- fish --
Version: 2
Fragment type: HashMapPartition
K Value: 6
Min load factor: 78
Max load factor: 80
Temporary table: no
Number of attributes: 4
Number of primary keys: 1
Length of frm data: 337
Max Rows: 0
Row Checksum: 1
Row GCI: 1
SingleUserMode: 0
ForceVarPart: 1
PartitionCount: 2
FragmentCount: 2
PartitionBalance: FOR_RP_BY_LDM
ExtraRowGciBits: 0
ExtraRowAuthorBits: 0
TableStatus: Retrieved
Table options:
HashMap: DEFAULT-HASHMAP-3840-2
-- Attributes --
id Int PRIMARY KEY DISTRIBUTION KEY AT=FIXED ST=MEMORY AUTO_INCR
name Varchar(20;latin1_swedish_ci) NOT NULL AT=SHORT_VAR ST=MEMORY DYNAMIC
length_mm Int NOT NULL AT=FIXED ST=MEMORY DYNAMIC
weight_gm Int NOT NULL AT=FIXED ST=MEMORY DYNAMIC
-- Indexes --
PRIMARY KEY(id) - UniqueHashIndex
PRIMARY(id) - OrderedIndex
uk(name) - OrderedIndex
uk$unique(name) - UniqueHashIndex
-- Per partition info --
Partition       Row count       Commit count    Frag fixed memory       Frag varsized memory    Extent_space    Free extent_space
0               2               2               32768                   32768                   0               0
1               4               4               32768                   32768                   0               0


NDBT_ProgramExit: 0 - OK

Information about multiple tables can be obtained in a single invocation of ndb_desc by using their names, separated by spaces. All of the tables must be in the same database.

You can obtain additional information about a specific index using the --table (short form: -t) option and supplying the name of the index as the first argument to ndb_desc, as shown here:

shell> ./ndb_desc uk -d test -t fish
-- uk --
Version: 2
Base table: fish
Number of attributes: 1
Logging: 0
Index type: OrderedIndex
Index status: Retrieved
-- Attributes --
name Varchar(20;latin1_swedish_ci) NOT NULL AT=SHORT_VAR ST=MEMORY
-- IndexTable 10/uk --
Version: 2
Fragment type: FragUndefined
K Value: 6
Min load factor: 78
Max load factor: 80
Temporary table: yes
Number of attributes: 2
Number of primary keys: 1
Length of frm data: 0
Max Rows: 0
Row Checksum: 1
Row GCI: 1
SingleUserMode: 2
ForceVarPart: 0
PartitionCount: 2
FragmentCount: 2
FragmentCountType: ONE_PER_LDM_PER_NODE
ExtraRowGciBits: 0
ExtraRowAuthorBits: 0
TableStatus: Retrieved
Table options:
-- Attributes --
name Varchar(20;latin1_swedish_ci) NOT NULL AT=SHORT_VAR ST=MEMORY
NDB$TNODE Unsigned [64] PRIMARY KEY DISTRIBUTION KEY AT=FIXED ST=MEMORY
-- Indexes --
PRIMARY KEY(NDB$TNODE) - UniqueHashIndex

NDBT_ProgramExit: 0 - OK

When an index is specified in this way, the --extra-partition-info and --extra-node-info options have no effect.

The Version column in the output contains the table's schema object version. For information about interpreting this value, see NDB Schema Object Versions.

Three of the table properties that can be set using NDB_TABLE comments embedded in CREATE TABLE and ALTER TABLE statements are also visible in ndb_desc output. The table's FRAGMENT_COUNT_TYPE is always shown in the FragmentCountType column. READ_ONLY and FULLY_REPLICATED, if set to 1, are shown in the Table options column. You can see this after executing the following ALTER TABLE statement in the mysql client:

mysql> ALTER TABLE fish COMMENT='NDB_TABLE=READ_ONLY=1,FULLY_REPLICATED=1';
1 row in set, 1 warning (0.00 sec)

mysql> SHOW WARNINGS\G
+---------+------+---------------------------------------------------------------------------------------------------------+
| Level   | Code | Message                                                                                                 |
+---------+------+---------------------------------------------------------------------------------------------------------+
| Warning | 1296 | Got error 4503 'Table property is FRAGMENT_COUNT_TYPE=ONE_PER_LDM_PER_NODE but not in comment' from NDB |
+---------+------+---------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

The warning is issued because READ_ONLY=1 requires that the table's fragment count type is (or be set to) ONE_PER_LDM_PER_NODE_GROUP; NDB sets this automatically in such cases. You can check that the ALTER TABLE statement has the desired effect using SHOW CREATE TABLE:

mysql> SHOW CREATE TABLE fish\G
*************************** 1. row ***************************
       Table: fish
Create Table: CREATE TABLE `fish` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `name` varchar(20) NOT NULL,
  `length_mm` int(11) NOT NULL,
  `weight_gm` int(11) NOT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `uk` (`name`)
) ENGINE=ndbcluster DEFAULT CHARSET=latin1
COMMENT='NDB_TABLE=READ_BACKUP=1,FULLY_REPLICATED=1'
1 row in set (0.01 sec)

Because FRAGMENT_COUNT_TYPE was not set explicitly, its value is not shown in the comment text printed by SHOW CREATE TABLE. ndb_desc, however, displays the updated value for this attribute. The Table options column shows the binary properties just enabled. You can see this in the output shown here (emphasized text):

shell> ./ndb_desc -c localhost fish -d test -p
-- fish --
Version: 4
Fragment type: HashMapPartition
K Value: 6
Min load factor: 78
Max load factor: 80
Temporary table: no
Number of attributes: 4
Number of primary keys: 1
Length of frm data: 380
Max Rows: 0
Row Checksum: 1
Row GCI: 1
SingleUserMode: 0
ForceVarPart: 1
PartitionCount: 1
FragmentCount: 1
FragmentCountType: ONE_PER_LDM_PER_NODE_GROUP
ExtraRowGciBits: 0
ExtraRowAuthorBits: 0
TableStatus: Retrieved
Table options: readbackup, fullyreplicated
HashMap: DEFAULT-HASHMAP-3840-1
-- Attributes --
id Int PRIMARY KEY DISTRIBUTION KEY AT=FIXED ST=MEMORY AUTO_INCR
name Varchar(20;latin1_swedish_ci) NOT NULL AT=SHORT_VAR ST=MEMORY DYNAMIC
length_mm Int NOT NULL AT=FIXED ST=MEMORY DYNAMIC
weight_gm Int NOT NULL AT=FIXED ST=MEMORY DYNAMIC
-- Indexes --
PRIMARY KEY(id) - UniqueHashIndex
PRIMARY(id) - OrderedIndex
uk(name) - OrderedIndex
uk$unique(name) - UniqueHashIndex
-- Per partition info --
Partition       Row count       Commit count    Frag fixed memory       Frag varsized memory    Extent_space    Free extent_space

NDBT_ProgramExit: 0 - OK

For more information about these table properties, see Section 13.1.20.11, “Setting NDB_TABLE Options”.

The Extent_space and Free extent_space columns are applicable only to NDB tables having columns on disk; for tables having only in-memory columns, these columns always contain the value 0.

To illustrate their use, we modify the previous example. First, we must create the necessary Disk Data objects, as shown here:

CREATE LOGFILE GROUP lg_1
    ADD UNDOFILE 'undo_1.log'
    INITIAL_SIZE 16M
    UNDO_BUFFER_SIZE 2M
    ENGINE NDB;

ALTER LOGFILE GROUP lg_1
    ADD UNDOFILE 'undo_2.log'
    INITIAL_SIZE 12M
    ENGINE NDB;

CREATE TABLESPACE ts_1
    ADD DATAFILE 'data_1.dat'
    USE LOGFILE GROUP lg_1
    INITIAL_SIZE 32M
    ENGINE NDB;

ALTER TABLESPACE ts_1
    ADD DATAFILE 'data_2.dat'
    INITIAL_SIZE 48M
    ENGINE NDB;

(For more information on the statements just shown and the objects created by them, see Section 23.5.10.1, “NDB Cluster Disk Data Objects”, as well as Section 13.1.16, “CREATE LOGFILE GROUP Statement”, and Section 13.1.21, “CREATE TABLESPACE Statement”.)

Now we can create and populate a version of the fish table that stores 2 of its columns on disk (deleting the previous version of the table first, if it already exists):

DROP TABLE IF EXISTS fish;

CREATE TABLE fish (
    id INT NOT NULL AUTO_INCREMENT,
    name VARCHAR(20) NOT NULL,
    length_mm INT NOT NULL,
    weight_gm INT NOT NULL,

    PRIMARY KEY pk (id),
    UNIQUE KEY uk (name)
) TABLESPACE ts_1 STORAGE DISK
ENGINE=NDB;

INSERT INTO fish VALUES
    (NULL, 'guppy', 35, 2), (NULL, 'tuna', 2500, 150000),
    (NULL, 'shark', 3000, 110000), (NULL, 'manta ray', 1500, 50000),
    (NULL, 'grouper', 900, 125000), (NULL ,'puffer', 250, 2500);

When run against this version of the table, ndb_desc displays the following output:

shell> ./ndb_desc -c localhost fish -d test -p
-- fish --
Version: 1
Fragment type: HashMapPartition
K Value: 6
Min load factor: 78
Max load factor: 80
Temporary table: no
Number of attributes: 4
Number of primary keys: 1
Length of frm data: 1001
Max Rows: 0
Row Checksum: 1
Row GCI: 1
SingleUserMode: 0
ForceVarPart: 1
PartitionCount: 2
FragmentCount: 2
PartitionBalance: FOR_RP_BY_LDM
ExtraRowGciBits: 0
ExtraRowAuthorBits: 0
TableStatus: Retrieved
Table options: readbackup
HashMap: DEFAULT-HASHMAP-3840-2
Tablespace id: 16
Tablespace: ts_1
-- Attributes --
id Int PRIMARY KEY DISTRIBUTION KEY AT=FIXED ST=MEMORY AUTO_INCR
name Varchar(80;utf8mb4_0900_ai_ci) NOT NULL AT=SHORT_VAR ST=MEMORY
length_mm Int NOT NULL AT=FIXED ST=DISK
weight_gm Int NOT NULL AT=FIXED ST=DISK
-- Indexes --
PRIMARY KEY(id) - UniqueHashIndex
PRIMARY(id) - OrderedIndex
uk(name) - OrderedIndex
uk$unique(name) - UniqueHashIndex
-- Per partition info --
Partition       Row count       Commit count    Frag fixed memory       Frag varsized memory    Extent_space    Free extent_space
0               2               2               32768                   32768                   1048576         1044440
1               4               4               32768                   32768                   1048576         1044400


NDBT_ProgramExit: 0 - OK

This means that 1048576 bytes are allocated from the tablespace for this table on each partition, of which 1044440 bytes remain free for additional storage. In other words, 1048576 - 1044440 = 4136 bytes per partition is currently being used to store the data from this table's disk-based columns. The number of bytes shown as Free extent_space is available for storing on-disk column data from the fish table only; for this reason, it is not visible when selecting from the INFORMATION_SCHEMA.FILES table.

Tablespace id and Tablespace are displayed for Disk Data tables beginning with NDB 8.0.21.

For fully replicated tables, ndb_desc shows only the nodes holding primary partition fragment replicas; nodes with copy fragment replicas (only) are ignored. You can obtain such information, using the mysql client, from the table_distribution_status, table_fragments, table_info, and table_replicas tables in the ndbinfo database.

All options that can be used with ndb_desc are shown in the following table. Additional descriptions follow the table.

Table indexes listed in the output are ordered by ID.