MySQL 8.0 Reference Manual Including MySQL NDB Cluster 8.0
The management server is the process that reads the cluster configuration file and distributes this information to all nodes in the cluster that request it. It also maintains a log of cluster activities. Management clients can connect to the management server and check the cluster's status.
The following table includes options that are specific to the NDB Cluster management server program ndb_mgmd. Additional descriptions follow the table. For options common to most NDB Cluster programs (including ndb_mgmd), see Section 23.4.32, “Options Common to NDB Cluster Programs — Options Common to NDB Cluster Programs”.
Table 23.26 Command-line options for the ndb_mgmd program
Format | Description | Added, Deprecated, or Removed |
---|---|---|
Local bind address | (Supported in all MySQL 8.0 based releases) |
|
Override defaults group suffix when reading cluster_config sections in my.cnf file; used in testing | ADDED: NDB 8.0.24 |
|
Enable management server configuration cache; true by default | (Supported in all MySQL 8.0 based releases) |
|
Specify cluster configuration file; also specify --reload or --initial to override configuration cache if present | (Supported in all MySQL 8.0 based releases) |
|
Specify cluster management server configuration cache directory | (Supported in all MySQL 8.0 based releases) |
|
Run ndb_mgmd in daemon mode (default) | (Supported in all MySQL 8.0 based releases) |
|
Causes management server to reload configuration data from configuration file, bypassing configuration cache | (Supported in all MySQL 8.0 based releases) |
|
Used to install management server process as Windows service; does not apply on other platforms | (Supported in all MySQL 8.0 based releases) |
|
Run ndb_mgmd in interactive mode (not officially supported in production; for testing purposes only) | (Supported in all MySQL 8.0 based releases) |
|
Name to use when writing cluster log messages applying to this node | (Supported in all MySQL 8.0 based releases) |
|
Read cluster configuration data from my.cnf file | (Supported in all MySQL 8.0 based releases) |
|
Do not provide any node ID checks | (Supported in all MySQL 8.0 based releases) |
|
Do not run ndb_mgmd as a daemon | (Supported in all MySQL 8.0 based releases) |
|
Do not wait for management nodes specified when starting this management server; requires --ndb-nodeid option | (Supported in all MySQL 8.0 based releases) |
|
Print full configuration and exit | (Supported in all MySQL 8.0 based releases) |
|
Causes management server to compare configuration file with configuration cache | (Supported in all MySQL 8.0 based releases) |
|
Used to remove management server process that was previously installed as Windows service, optionally specifying name of service to be removed; does not apply on other platforms | (Supported in all MySQL 8.0 based releases) |
|
Write additional information to log | (Supported in all MySQL 8.0 based releases) |
Command-Line Format | --bind-address=host |
---|---|
Type | String |
Default Value | [none] |
Causes the management server to bind to a specific network interface (host name or IP address). This option has no default value.
Command-Line Format | --cluster-config-suffix=name |
---|---|
Introduced | 8.0.24-ndb-8.0.24 |
Type | String |
Default Value | [none] |
Override defaults group suffix when reading cluster
configuration sections in my.cnf
; used
in testing.
Command-Line Format | --config-cache[=TRUE|FALSE] |
---|---|
Type | Boolean |
Default Value | TRUE |
This option, whose default value is 1
(or
TRUE
, or ON
), can be
used to disable the management server's configuration
cache, so that it reads its configuration from
config.ini
every time it starts (see
Section 23.3.3, “NDB Cluster Configuration Files”). You can do
this by starting the ndb_mgmd process
with any one of the following options:
Using one of the options just listed is effective only if
the management server has no stored configuration at the
time it is started. If the management server finds any
configuration cache files, then the
--config-cache
option or the
--skip-config-cache
option is ignored.
Therefore, to disable configuration caching, the option
should be used the first time that the
management server is started. Otherwise—that is, if
you wish to disable configuration caching for a management
server that has already created a
configuration cache—you must stop the management
server, delete any existing configuration cache files
manually, then restart the management server with
--skip-config-cache
(or with
--config-cache
set equal to 0,
OFF
, or FALSE
).
Configuration cache files are normally created in a
directory named mysql-cluster
under the
installation directory (unless this location has been
overridden using the
--configdir
option). Each
time the management server updates its configuration data,
it writes a new cache file. The files are named sequentially
in order of creation using the following format:
ndb_node-id
_config.bin.seq-number
node-id
is the management
server's node ID; seq-number
is a sequence number, beginning with 1. For example, if the
management server's node ID is 5, then the first three
configuration cache files would, when they are created, be
named ndb_5_config.bin.1
,
ndb_5_config.bin.2
, and
ndb_5_config.bin.3
.
If your intent is to purge or reload the configuration cache
without actually disabling caching, you should start
ndb_mgmd with one of the options
--reload
or
--initial
instead of
--skip-config-cache
.
To re-enable the configuration cache, simply restart the
management server, but without the
--config-cache
or
--skip-config-cache
option that was used
previously to disable the configuration cache.
ndb_mgmd does not check for the
configuration directory
(--configdir
) or attempts
to create one when --skip-config-cache
is
used. (Bug #13428853)
--config-file=
,
filename
-f
filename
Command-Line Format | --config-file=file |
---|---|
Type | File name |
Default Value | [none] |
Instructs the management server as to which file it should
use for its configuration file. By default, the management
server looks for a file named
config.ini
in the same directory as the
ndb_mgmd executable; otherwise the file
name and location must be specified explicitly.
This option has no default value, and is ignored unless the
management server is forced to read the configuration file,
either because ndb_mgmd was started with
the --reload
or
--initial
option, or
because the management server could not find any
configuration cache. This option is also read if
ndb_mgmd was started with
--config-cache=OFF
. See
Section 23.3.3, “NDB Cluster Configuration Files”, for more
information.
Command-Line Format |
|
---|---|
Type | File name |
Default Value | $INSTALLDIR/mysql-cluster |
Specifies the cluster management server's configuration
cache directory. --config-dir
is an alias
for this option.
Command-Line Format | --daemon |
---|---|
Type | Boolean |
Default Value | TRUE |
Instructs ndb_mgmd to start as a daemon process. This is the default behavior.
This option has no effect when running ndb_mgmd on Windows platforms.
Command-Line Format | --initial |
---|---|
Type | Boolean |
Default Value | FALSE |
Configuration data is cached internally, rather than being
read from the cluster global configuration file each time
the management server is started (see
Section 23.3.3, “NDB Cluster Configuration Files”). Using the
--initial
option overrides this behavior,
by forcing the management server to delete any existing
cache files, and then to re-read the configuration data from
the cluster configuration file and to build a new cache.
This differs in two ways from the
--reload
option. First,
--reload
forces the server to check the
configuration file against the cache and reload its data
only if the contents of the file are different from the
cache. Second, --reload
does not delete any
existing cache files.
If ndb_mgmd is invoked with
--initial
but cannot find a global
configuration file, the management server cannot start.
When a management server starts, it checks for another management server in the same NDB Cluster and tries to use the other management server's configuration data. This behavior has implications when performing a rolling restart of an NDB Cluster with multiple management nodes. See Section 23.5.5, “Performing a Rolling Restart of an NDB Cluster”, for more information.
When used together with the
--config-file
option, the
cache is cleared only if the configuration file is actually
found.
Command-Line Format | --install[=name] |
---|---|
Platform Specific | Windows |
Type | String |
Default Value | ndb_mgmd |
Causes ndb_mgmd to be installed as a
Windows service. Optionally, you can specify a name for the
service; if not set, the service name defaults to
ndb_mgmd
. Although it is preferable to
specify other ndb_mgmd program options in
a my.ini
or my.cnf
configuration file, it is possible to use them together with
--install
. However, in such
cases, the --install
option
must be specified first, before any other options are given,
for the Windows service installation to succeed.
It is generally not advisable to use this option together
with the --initial
option,
since this causes the configuration cache to be wiped and
rebuilt every time the service is stopped and started. Care
should also be taken if you intend to use any other
ndb_mgmd options that affect the starting
of the management server, and you should make absolutely
certain you fully understand and allow for any possible
consequences of doing so.
The --install
option has no
effect on non-Windows platforms.
Command-Line Format | --interactive |
---|---|
Type | Boolean |
Default Value | FALSE |
Starts ndb_mgmd in interactive mode; that is, an ndb_mgm client session is started as soon as the management server is running. This option does not start any other NDB Cluster nodes.
Command-Line Format | --log-name=name |
---|---|
Type | String |
Default Value | MgmtSrvr |
Provides a name to be used for this node in the cluster log.
Command-Line Format | --mycnf |
---|---|
Type | Boolean |
Default Value | FALSE |
Read configuration data from the my.cnf
file.
Command-Line Format | --no-nodeid-checks |
---|---|
Type | Boolean |
Default Value | FALSE |
Do not perform any checks of node IDs.
Command-Line Format | --nodaemon |
---|---|
Type | Boolean |
Default Value | FALSE |
Instructs ndb_mgmd not to start as a daemon process.
The default behavior for ndb_mgmd on Windows is to run in the foreground, making this option unnecessary on Windows platforms.
Command-Line Format | --nowait-nodes=list |
---|---|
Type | Numeric |
Default Value | [none] |
Minimum Value | 1 |
Maximum Value | 255 |
When starting an NDB Cluster is configured with two management nodes, each management server normally checks to see whether the other ndb_mgmd is also operational and whether the other management server's configuration is identical to its own. However, it is sometimes desirable to start the cluster with only one management node (and perhaps to allow the other ndb_mgmd to be started later). This option causes the management node to bypass any checks for any other management nodes whose node IDs are passed to this option, permitting the cluster to start as though configured to use only the management node that was started.
For purposes of illustration, consider the following portion
of a config.ini
file (where we have
omitted most of the configuration parameters that are not
relevant to this example):
[ndbd] NodeId = 1 HostName = 198.51.100.101 [ndbd] NodeId = 2 HostName = 198.51.100.102 [ndbd] NodeId = 3 HostName = 198.51.100.103 [ndbd] NodeId = 4 HostName = 198.51.100.104 [ndb_mgmd] NodeId = 10 HostName = 198.51.100.150 [ndb_mgmd] NodeId = 11 HostName = 198.51.100.151 [api] NodeId = 20 HostName = 198.51.100.200 [api] NodeId = 21 HostName = 198.51.100.201
Assume that you wish to start this cluster using only the
management server having node ID 10
and
running on the host having the IP address 198.51.100.150.
(Suppose, for example, that the host computer on which you
intend to the other management server is temporarily
unavailable due to a hardware failure, and you are waiting
for it to be repaired.) To start the cluster in this way,
use a command line on the machine at 198.51.100.150 to enter
the following command:
shell> ndb_mgmd --ndb-nodeid=10 --nowait-nodes=11
As shown in the preceding example, when using
--nowait-nodes
, you must
also use the --ndb-nodeid
option to specify the node ID of this
ndb_mgmd process.
You can then start each of the cluster's data nodes in the usual way. If you wish to start and use the second management server in addition to the first management server at a later time without restarting the data nodes, you must start each data node with a connection string that references both management servers, like this:
shell> ndbd -c 198.51.100.150,198.51.100.151
The same is true with regard to the connection string used with any mysqld processes that you wish to start as NDB Cluster SQL nodes connected to this cluster. See Section 23.3.3.3, “NDB Cluster Connection Strings”, for more information.
When used with ndb_mgmd, this option
affects the behavior of the management node with regard to
other management nodes only. Do not confuse it with the
--nowait-nodes
option used with
ndbd or ndbmtd to
permit a cluster to start with fewer than its full
complement of data nodes; when used with data nodes, this
option affects their behavior only with regard to other data
nodes.
Multiple management node IDs may be passed to this option as a comma-separated list. Each node ID must be no less than 1 and no greater than 255. In practice, it is quite rare to use more than two management servers for the same NDB Cluster (or to have any need for doing so); in most cases you need to pass to this option only the single node ID for the one management server that you do not wish to use when starting the cluster.
When you later start the “missing” management server, its configuration must match that of the management server that is already in use by the cluster. Otherwise, it fails the configuration check performed by the existing management server, and does not start.
Command-Line Format | --print-full-config |
---|---|
Type | Boolean |
Default Value | FALSE |
Shows extended information regarding the configuration of
the cluster. With this option on the command line the
ndb_mgmd process prints information about
the cluster setup including an extensive list of the cluster
configuration sections as well as parameters and their
values. Normally used together with the
--config-file
(-f
) option.
Command-Line Format | --reload |
---|---|
Type | Boolean |
Default Value | FALSE |
NDB Cluster configuration data is stored internally rather than being read from the cluster global configuration file each time the management server is started (see Section 23.3.3, “NDB Cluster Configuration Files”). Using this option forces the management server to check its internal data store against the cluster configuration file and to reload the configuration if it finds that the configuration file does not match the cache. Existing configuration cache files are preserved, but not used.
This differs in two ways from the
--initial
option. First,
--initial
causes all cache files to be
deleted. Second, --initial
forces the
management server to re-read the global configuration file
and construct a new cache.
If the management server cannot find a global configuration
file, then the --reload
option is ignored.
When --reload
is used, the management
server must be able to communicate with data nodes and any
other management servers in the cluster before it attempts
to read the global configuration file; otherwise, the
management server fails to start. This can happen due to
changes in the networking environment, such as new IP
addresses for nodes or an altered firewall configuration. In
such cases, you must use
--initial
instead to force
the exsiting cached configuration to be discarded and
reloaded from the file. See
Section 23.5.5, “Performing a Rolling Restart of an NDB Cluster”, for
additional information.
Command-Line Format | --remove[=name] |
---|---|
Platform Specific | Windows |
Type | String |
Default Value | ndb_mgmd |
Remove a management server process that has been installed as a Windows service, optionally specifying the name of the service to be removed. Applies only to Windows platforms.
Command-Line Format | --verbose |
---|---|
Type | Boolean |
Default Value | FALSE |
Remove a management server process that has been installed as a Windows service, optionally specifying the name of the service to be removed. Applies only to Windows platforms.
It is not strictly necessary to specify a connection string when starting the management server. However, if you are using more than one management server, a connection string should be provided and each node in the cluster should specify its node ID explicitly.
See Section 23.3.3.3, “NDB Cluster Connection Strings”, for information about using connection strings. Section 23.4.4, “ndb_mgmd — The NDB Cluster Management Server Daemon”, describes other options for ndb_mgmd.
The following files are created or used by
ndb_mgmd in its starting directory, and are
placed in the DataDir
as
specified in the config.ini
configuration
file. In the list that follows,
node_id
is the unique node
identifier.
config.ini
is the configuration file
for the cluster as a whole. This file is created by the user
and read by the management server.
Section 23.3, “Configuration of NDB Cluster”, discusses how
to set up this file.
ndb_
is the cluster events log file. Examples of such events
include checkpoint startup and completion, node startup
events, node failures, and levels of memory usage. A
complete listing of cluster events with descriptions may be
found in Section 23.5, “Management of NDB Cluster”.
node_id
_cluster.log
By default, when the size of the cluster log reaches one
million bytes, the file is renamed to
ndb_
,
where node_id
_cluster.log.seq_id
seq_id
is the sequence
number of the cluster log file. (For example: If files with
the sequence numbers 1, 2, and 3 already exist, the next log
file is named using the number 4
.) You
can change the size and number of files, and other
characteristics of the cluster log, using the
LogDestination
configuration parameter.
ndb_
is the file used for node_id
_out.logstdout
and
stderr
when running the management server
as a daemon.
ndb_
is the process ID file used when running the management
server as a daemon.
node_id
.pid