This topic explains how to use gemfire.properties and gfsecurity.properties in VMware Tanzu GemFire to manage access to Tanzu GemFire security configuration data.
You use the gemfire.properties settings to join a cluster and configure system member behavior. Cluster members include applications, the cache server, the locator, and other Tanzu GemFire processes.
You can place any security-related (properties that begin with security-*) configuration properties in gemfire.properties into a separate gfsecurity.properties file. Placing these configuration settings in a separate file allows you to restrict access to security configuration data. This way, you can still allow read or write access for your gemfire.properties file.
You can also define provider-specific properties (“ssl” properties) in gfsecurity.properties instead of defining them at the command-line or in your environment.
You can specify non-ASCII text in your properties files by using Unicode escape sequences. See Using Non-ASCII Strings in VMware Tanzu GemFire Property Files for more details.
Note: Unless otherwise indicated, these settings only affect activities within this cluster - not activities between clients and servers or between a gateway sender and gateway receiver in a multi-site installation.
| Setting | Definition | Applies to (Server Locator, Client) | Default |
|---|---|---|---|
| ack-severe-alert-threshold | Number of seconds the cluster will wait after the ack-wait-threshold for a message to be acknowledged before it issues an alert at severe level. A value of zero deactivate this feature. |
S, L | 0 |
| ack-wait-threshold | Number of seconds a distributed message can wait for acknowledgment before it sends an alert to signal that something might be wrong with the system member that is unresponsive. The waiter continues to wait. The alerts are logged in the system member's log as warnings. Valid values are in the range 0...2147483647 |
S, L | 15 |
| archive-disk-space-limit | Maximum size (in megabytes) of all inactive statistic archive files combined. If this limit is exceeded, inactive archive files are deleted, oldest first, until the total size is within the limit. If set to zero, disk space use is unlimited. | S, L | 0 |
| archive-file-size-limit | The maximum size (in megabytes) of a single statistic archive file. Once this limit is exceeded, a new statistic archive file is created, and the current archive file becomes inactive. If set to zero, file size is unlimited. | S, L | 0 |
| async-distribution-timeout | The number of milliseconds a process that is publishing to this process should attempt to distribute a cache operation before switching over to asynchronous messaging for this process. The switch to asynchronous messaging lasts until this process catches up, departs, or some specified limit is reached, such as async-queue-timeout or async-max-queue-size. To enable asynchronous messaging, the value must be set above zero. Valid values are in the range 0...60000.
Note:
This setting controls only peer-to-peer communication and does not apply to client/server or multi-site communication. |
S | 0 |
| async-max-queue-size | Affects non-conflated asynchronous queues for members that publish to this member. This is the maximum size the queue can reach (in megabytes) before the publisher asks this member to leave the cluster. Valid values are in the range 0...1024.
Note:
This setting controls only peer-to-peer communication and does not apply to client/server or multi-site communication. |
S | 8 |
| async-queue-timeout | Affects asynchronous queues for members that publish to this member. This is the maximum milliseconds the publisher should wait with no distribution to this member before it asks this member to leave the cluster. Used for handling slow receivers.
Note:
This setting controls only peer-to-peer communication and does not apply to client/server or multi-site communication. |
S, L | 60000 |
| bind-address | Relevant only for multi-homed hosts - machines with multiple network interface cards. Specifies the adapter card the cache binds to for peer-to-peer communication. Also specifies the default location for Tanzu GemFire servers to listen on, which is used unless overridden by the server-bind-address. An empty string causes the member to listen on the default card for the machine. This is a machine-wide attribute used for system member and client/server communication. It has no effect on locator location, unless the locator is embedded in a member process. Specify the IP address, not the hostname, because each network card may not have a unique hostname. An empty string (the default) causes the member to listen on the default card for the machine. |
S, L | not set |
| cache-xml-file | Declarative initialization file for the member's cache. | S | cache.xml |
| cluster-configuration-dir | Deprecated. Use the gfsh import/export cluster-configuration command instead. This property specifies the directory in which the cluster configuration related disk-store and artifacts are stored. This property is only applicable to dedicated locators that have "enable-cluster-configuration" set to true. | L | not set |
| conflate-events | Used only by clients in a client/server installation. This is a client-side property that is passed to the server. Affects subscription queue conflation in this client's servers. Specifies whether to conflate (true setting), not conflate (false), or to use the server's conflation setting (server). | S | server |
| conserve-sockets | Specifies whether sockets are shared by the system member's threads. If true, threads share, and a minimum number of sockets are used to connect to the cluster. If false, every application thread has its own sockets for distribution purposes. You can override this setting for individual threads inside your application. WAN deployments increase the messaging demands on a Tanzu GemFire system. To avoid hangs related to WAN messaging, always use the default setting of conserve-sockets=false for Tanzu GemFire members that participate in a WAN deployment. |
S, L | false |
| delta-propagation | Specifies whether to distribute the deltas for entry updates, instead of the full values, between clients and servers and between peers. | S | true |
| deploy-working-dir | Working directory used when deploying JAR application files to cluster members. This directory can be local and unique to the member or a shared resource. See Deploying Application JARs to VMware Tanzu GemFire Members for more information. | S | . (current directory) |
| disable-auto-reconnect | By default, a Tanzu GemFire member (both locators and servers) will attempt to reconnect and reinitialize the cache after it has been forced out of the cluster by a network partition event or has otherwise been shunned by other members. Use this property to turn off the auto-reconnect behavior. See Handling Forced Cache Disconnection Using Auto-reconnect for more details. | S, L | false |
| disable-jmx | By default, Tanzu GemFire automatically creates JMX MBeans. This boolean, when true, prevents the creation of JMX MBeans. Both gfsh and Pulse require JMX MBeans. | S, L | false |
| distributed-system-id | Identifier used to distinguish messages from different clusters. This is required for Portable Data eXchange (PDX) data serialization. Set distributed-system-id to different values for different systems in a multi-site (WAN) configuration, and to different values for production vs. development environments. This setting must be the same for every member of a given cluster and unique to each cluster within a WAN installation. Valid values are integers in the range -1...255. -1 means no setting. |
S, L | -1 |
| durable-client-id | Used only for clients in a client/server installation. If set, this indicates that the client is durable and identifies the client. The ID is used by servers to reestablish any messaging that was interrupted by client downtime. | C | not set |
| durable-client-timeout | Used only for clients in a client/server installation. Number of seconds this client can remain disconnected from its server and have the server continue to accumulate durable events for it. | C | 300 |
| enable-cluster-configuration | A value of "true" causes the creation of cluster configuration on dedicated locators. The cluster configuration service on dedicated locators with this property set to "true" serves the configuration to new members joining the cluster and also saves the configuration changes caused by the gfsh commands. This property is applicable only to dedicated locators. |
L | true |
| enable-management-rest-service | A value of "true" enables the cluster management REST service. This service requires that the cluster configuration service also be enabled. If enable-cluster-configuration is false and this is true, the management REST service would do nothing but report an error if you attempt to use it. Also, even if this property is set to true, The management REST service will be started only if http-service-port is not "0". |
L | true |
| enable-network-partition-detection | Boolean instructing the system to detect and handle splits in the cluster, typically caused by a partitioning of the network (split brain) where the cluster is running. You must set this property to the same value across all your cluster members. In addition, this property must be set to true if you are using persistent regions and configure your regions to use DISTRIBUTED_ACK or GLOBAL scope to avoid potential data conflicts. |
S, L | true |
| enable-time-statistics | Boolean instructing the system to track time-based statistics for the cluster and caching. Deactivated by default for performance reasons and not recommended for production environments. You must also configure statistic-sampling-enabled to true and specify a statistic-archive-file. |
S, L | true |
| enforce-unique-host | Whether partitioned regions will put redundant copies of the same data in different members running on the same physical machine. By default, Tanzu GemFire tries to put redundant copies on different machines, but it will put them on the same machine if no other machines are available. Setting this property to true prevents this and requires different machines for redundant copies. | S | false |
| groups | Defines the list of groups that this member belongs to. Use commas to separate group names. Note that anything defined by the roles GemFire property will also be considered a group. See Using Member Groups for more information. | S | not set |
| http-service-bind-address | If set, then the Tanzu GemFire member binds the embedded HTTP service to the specified address. If this property is not set but the HTTP service is enabled using http-service-port, then Tanzu GemFire binds the HTTP service to the member's local address. Used by the Tanzu GemFire Pulse Web application and the developer REST API service. |
S | not set |
| http-service-port | If non-zero, then Tanzu GemFire starts an embedded HTTP service that listens on this port. The HTTP service is used to host the Tanzu GemFire Pulse Web application and the development REST API service. If you are hosting the Pulse web app on your own Web server and are not using the development REST API service, then deactivate this embedded HTTP service by setting this property to zero. Ignored if jmx-manager and start-dev-rest-api are both set to false. |
S | 7070 |
| jmx-manager | If true then this member is willing to be a JMX Manager. All the other JMX Manager properties will be used when it does become a manager. If this property is false then all other jmx-manager-* properties are ignored. |
S, L | false (except on locators) |
| jmx-manager-bind-address | By default the jmx-manager (when configured with a port) will listen on all the local host's addresses. You can use this property to configure what IP address or host name the JMX Manager will listen on for non-HTTP connections. Ignored if JMX Manager is false or jmx-manager-port is zero. |
S, L | not set |
| jmx-manager-hostname-for-clients | Lets you control what hostname will be given to clients that ask the locator for the location of a JMX Manager. By default the IP address that the jmx-manager reports is used. But for clients on a different network this property allows you to configure a different hostname that will be given to clients. Ignored if jmx-manager is false or jmx-manager-port is zero. |
S, L | not set |
| jmx-manager-http-port | Deprecated. Use http-service-port instead. |
S, L | 7070 |
| jmx-manager-port | The port this JMX Manager will listen to for client connections. If this property is set to zero then Tanzu GemFire will not allow remote client connections but you can alternatively use the standard system properties supported by the JVM for configuring access from remote JMX clients. Ignored if jmx-manager is false. |
S, L | 1099 |
| jmx-manager-start | If true then this member will start a jmx manager when it creates a cache. Management tools like gfsh can be configured to connect to the jmx-manager. In most cases you should not set this because a jmx manager will automatically be started when needed on a member that sets "jmx-manager" to true. Ignored if jmx-manager is false. | S, L | false |
| jmx-manager-update-rate | The rate, in milliseconds, at which this member will push updates to any JMX Managers. Currently this value should be greater than or equal to the statistic-sample-rate. Setting this value too high will cause stale values to be seen by gfsh and Tanzu GemFire Pulse. | S, L | 2000 |
| load-cluster-configuration-from-dir | Setting this property to "true" causes loading of cluster configuration from "cluster_config" directory in the locator. This property is only applicable to dedicated locators that have "enable-cluster-configuration" set to true. | L | false |
| locator-wait-time | The number of seconds that a member should wait for a locator to start if a locator is not available when attempting to join the cluster. Use this setting when you are starting locators and peers all at once. This timeout allows peers to wait for the locators to finish starting up before attempting to join the cluster. | S | 0 |
| locators | The list of locators used by system members. The list must be configured consistently for every member of the cluster. If the list is empty, locators are not used. For each locator, provide a host name and/or address (separated by '@', if you use both), followed by a port number in brackets. Examples:
Note:
On multi-homed hosts, this last notation will use the default address. If you use bind addresses for your locators, explicitly specify the addresses in the locators list—do not use just the hostname.
Note:
If you specify invalid DNS hostnames in this property, any locators or servers started with |
S, L | not set |
| lock-memory | When true, locks heap and off-heap memory into RAM to prevent the operating system from paging the memory out to disk. |
S, L | false |
| log-disk-space-limit | Maximum size in megabytes of all inactive log files combined. If this limit is exceeded, inactive log files are deleted, oldest first, until the total size is within the limit. If set to zero, disk space use is unlimited. | S, L | 0 |
| log-file | File to which a running system member writes log messages. If set to null, the default is used. Each member type has its own default output:
|
S, L | null |
| log-file-size-limit | Maximum size in megabytes of a log file before it is closed and logging rolls on to a new (child) log file. If set to 0, log rolling is deactivated. | S, L | 0 |
| log-level | Level of detail of the messages written to the system member's log. Setting log-level to one of the ordered levels causes all messages of that level and greater severity to be printed. Valid values from lowest to highest are fine, config, info, warning, error, severe, and none. |
S, L | config |
| max-wait-time-reconnect | Maximum number of milliseconds to wait for the cluster to reconnect on each reconnect attempt. | S, L | 60000 |
| member-timeout | Tanzu GemFire uses the member-timeout server configuration, specified in milliseconds, to detect unresponsive members. Tanzu GemFire members use a combination of heartbeat messages and TCP sockets to monitor the health of other members. If a member is suspected to be unreachable, two active checks of the member are performed. Each of those checks waits a maximum of one member-timeout. If those health checks receive a connection failure or no response within the timeout, the member is considered dead and is removed from the list of members. This setting also controls how frequently heartbeat messages are sent. They are sent at a frequency of half the member-timeout. The total time it takes to remove a member that is not responding to any network traffic is therefore 2 to 3 times the member-timeout. This is not the minimum time it takes to remove a member from the list of members in the cluster. If the Java process has crashed, but the operating system can still return a connection failure response to the health checks, the crashed member may be removed from the membership list immediately. Valid values are in the range 1000...600000. |
S, L | 5000 |
| membership-port-range | The range of ports available for membership messaging and for TCP failure detection. This is specified as two integers separated by a hyphen. Different members can use different ranges. Tanzu GemFire randomly chooses at least two unique integers from this range for the member, one for membership messaging and the other for TCP failure detection messaging. If tcp-port is configured to 0, it will also randomly select a port from this range for TCP sockets used for peer-to-peer communication only. Therefore, the specified range must include at least three available port numbers. The system uniquely identifies the member using the combined host IP address and membership port number. You may want to restrict the range of ports that Tanzu GemFire uses so the product can run in an environment where routers only allow traffic on certain ports. |
S, L | 41000-61000 |
| memcached-port | If specified and is non-zero, sets the port number for an embedded Gemcached server and starts the Gemcached server. | S | 0 |
| memcached-protocol | Sets the protocol used by an embedded Gemcached server. Valid values are BINARY and ASCII. If you omit this property, the ASCII protocol is used. |
S | ASCII |
| name | Symbolic name used to identify this system member. | S, L | not set |
| off-heap-memory-size | Specifies the size of off-heap memory in megabytes (m) or gigabytes (g). For example: |
S | not set |
| redundancy-zone | Defines this member's redundancy zone. Used to separate member's into different groups for satisfying partitioned region redundancy. If this property is set, Tanzu GemFire will not put redundant copies of data in members with the same redundancy zone setting. See Configure High Availability for a Partitioned Region for more details. | S | not set |
| remote-locators | Used to configure the locators that a cluster will use in order to connect to a remote site in a multi-site (WAN) configuration. To use locators in a WAN configuration, you must specify a unique distributed system ID (distributed-system-id) for the local cluster and remote locators for the remote clusters to which you will connect. For each remote locator, provide a host name and/or address (separated by '@', if you use both), followed by a port number in brackets. Examples: |
L | not set |
| remove-unresponsive-client | When this property is set to true, the primary server drops unresponsive clients from all secondaries and itself. Clients are deemed unresponsive when their messaging queues become full on the server. While a client's queue is full, puts that would add to the queue block on the server. | S | false |
| security-* | Any security-related (properties that begin with security-) configuration properties that are normally configured in gemfire.properties can be moved to a separate gfsecurity.properties file. Placing these configuration settings in a separate file allows you to restrict access to security configuration data. This way, you can still allow read or write access for your gemfire.properties file. |
S, L | not set |
| security-client-accessor | Deprecated. Used for authorization. Static creation method returning an AccessControl object, which determines authorization of client-server cache operations. This specifies the callback that should be invoked in the pre-operation phase, which is when the request for the operation is received from the client. |
S, L | not set |
| security-auth-token-enabled-components | A comma-delimited list of component names which works in conjunction with the security-manager property. If security-manager is enabled, this property determines which components will use token-based authentication instead of basic (username/password) authentication. Valid values are: "all", "management", "pulse" "all": shorthand for all the security components that support token authentication. "management": the management REST service. "pulse": the Pulse web app |
L | "" |
| security-client-accessor-pp | Deprecated. Used for authorization. The callback that should be invoked in the post-operation phase, which is when the operation has completed on the server but before the result is sent to the client. The post-operation callback is also invoked for the updates that are sent from server to client through the notification channel. | S, L | not set |
| security-client-auth-init | Used for authentication. Defines the fully-qualified static method that instantiates and returns the AuthInitialize object. AuthInitialize is responsible for producing the necessary credentials for client authentication. |
S, L | not set |
| security-client-authenticator | Deprecated. Used for authentication. Static creation method returning an Authenticator object, which is used by a peer to verify the credentials of the connecting peer. |
S, L | not set |
| security-client-dhalgo | Deprecated. Use ssl-enabled-components instead. Used for authentication. For secure transmission of sensitive credentials like passwords, you can encrypt the credentials using the Diffie-Hellman key-exchange algorithm. Do this by setting the security-client-dhalgo system property on the clients to the name of a valid, symmetric key cipher supported by the JDK. |
S, L | not set |
| security-log-file | Used with authentication. The log file for security log messages. If not specified, the member's regular log file is used. | S, L | not set |
| security-log-level | Used with authentication. Logging level detail for security log messages. Valid values from lowest to highest are fine, config, info, warning, error, severe, and none. |
S, L | config |
| security-manager | Specifies the implementation of the SecurityManager interface that implements the callbacks that do authentication and authorization. |
S, L | not set |
| security-peer-auth-init | Used for authentication. Defines the fully-qualified static method that instantiates and returns the AuthInitialize object. AuthInitialize is responsible for producing the necessary credentials for peer authentication. |
S, L | not set |
| security-peer-authenticator | Deprecated. Used with authentication. Static creation method returning an Authenticator object, which is used by a peer to verify the credentials of the connecting peer. |
S, L | not set |
| security-peer-verifymember-timeout | Used with authentication. Timeout in milliseconds used by a peer to verify membership of an unknown authenticated peer requesting a secure connection. | S, L | 1000 |
| security-post-processor | Specifies the implementation of the PostProcessor interface that implements user-defined callbacks that can change the returned results of region get operations. |
S, L | not set |
| serializable-object-filter | A semicolon-separated list of items that become full class names of objects that the system will serialize when the property validate-serializable-objects is set to true. The list is expanded using the patterns specified in the createFilter method at ObjectInputFilter.Config. |
S, C | "!*" |
| server-bind-address | Relevant only for multi-homed hosts - machines with multiple network interface cards. Network adapter card a Tanzu GemFire server binds to for client/server communication. You can use this to separate the server's client/server communication from its peer-to-peer communication, spreading the traffic load. This is a machine-wide attribute used for communication with clients in client/server and multi-site installations. This setting has no effect on locator configuration. Specify the IP address, not the hostname, because each network card may not have a unique hostname. An empty string causes the servers to listen on the same card used for peer-to-peer communication. This is either the |
S | not set |
| socket-buffer-size | Receive buffer sizes in bytes of the TCP/IP connections used for data transmission. To minimize the buffer size allocation needed for distributing large, serializable messages, the messages are sent in chunks. This setting determines the size of the chunks. Larger buffers can handle large messages more quickly, but take up more memory. | S, L | 32768 |
| socket-lease-time | Time, in milliseconds, a thread can have exclusive access to a socket it is not actively using. A value of zero causes socket leases to never expire. This property is ignored if conserve-sockets is true. Valid values are in the range |
S, L | 1800000 (30 minutes) |
| ssl-enabled-components | Components for which to enable SSL. Comma-separated list of one or more of (cluster, gateway, jmx, locator, server, web) or "all". | C, S, L | all |
| ssl-endpoint-identification-enabled | Boolean. When set to true, causes clients to validate the server's hostname using the server's certificate. | C, S, L | false |
| ssl-use-default-context | Boolean. When set to true, allows Tanzu GemFire to use the default SSL context. When enabled, also sets ssl-endpoint-identification-enabled to true. | C, S, L | false |
| ssl-require-authentication | Boolean. Require two-way authentication for SSL-enabled components. Applies to all components except web. | S, L | true |
| ssl-default-alias | String. Default certificate name. If empty, use first certificate in key store. | S, L | |
| ssl-ciphers | Comma-separated list of SSL ciphers or "any" | S, L | any |
| ssl-parameter-extension | Specifies the implementation of the SSLParameterExtension interface that implements the SSL parameter extensions. |
S, L | not set |
| ssl-protocols | Comma-separated list of SSL protocols or "any" | S, L | any |
| ssl-keystore, ssl-keystore-password | Strings. Path to key store, key store password. | S, L | |
| ssl-truststore, ssl-truststore-password | Strings. Path to trust store, trust store password. | S, L | |
| ssl-keystore-type, ssl-truststore-type | Strings. Type of key store or trust store. "JKS" indicates Java. One common alternative is "pkcs12". | S, L | JKS |
| standard-output-always-on | Boolean value specifying to always log to standard out even if a log file exists. | S, L | false |
| start-dev-rest-api | If set to true, then the developer REST API service will be started when cache is created. REST service can be configured using http-service-port and http-service-bind-address properties. |
S | false |
| start-locator | If set, automatically starts a locator in the current process when the member connects to the cluster and stops the locator when the member disconnects. To use, specify the locator with an optional address or host specification and a required port number, in one of these formats: If not already there, this locator is automatically added to the list of locators in this set of |
S | not set |
| statistic-archive-file | The file to which the running system member writes statistic samples. For example: "StatisticsArchiveFile.gfs". Adding .gz suffix to the file name causes it to be compressed. To deactivate archiving set the filename to /dev/null. |
S, L | not set |
| statistic-sample-rate | How often to sample statistics, in milliseconds. Valid values are in the range 1000...60000. |
S, L | 1000 |
| statistic-sampling-enabled | Whether to collect and archive statistics on the member. Statistics sampling provides valuable information for ongoing system tuning and troubleshooting purposes. Sampling statistics at the default sample rate does not impact system performance. We recommend enabling statistics sampling in production environments.
Note:
This setting does not apply to partitioned regions, where statistics are always enabled. |
S, L | true |
| tcp-port | The TCP port to listen on for cache communications. If set to zero, the operating system selects an available port. Each process on a machine must have its own TCP port. Note that some operating systems restrict the range of ports usable by non-privileged users, and using restricted port numbers can cause runtime errors in Tanzu GemFire startup. Valid values are in the range 0...65535. |
S, L | 0 |
| thread-monitor-enabled | Boolean. When true, enables monitoring of Tanzu GemFire-created operational threads. Informational messages are written to the log file. |
S | true |
| thread-monitor-interval-ms | The time interval (in milliseconds) with which thread monitoring is scheduled to run. |
S | 60000 |
| thread-monitor-time-limit-ms | The time period (in milliseconds) after which the monitored thread is considered to be stuck. |
S | 30000 |
| use-cluster-configuration | This property is only applicable for data members (non-client and non-locator). A value of "true" causes a member to request and use the configuration from cluster configuration services running on dedicated locators. Setting this property to "false" causes a member to not request the configuration from the configuration services running on the locators. | S | true |
| use-udp-membership-messenger | Old versions of GemFire used UDP for low-level membership communication. Now it defaults to TCP instead. If you need the old behavior, set this configuration parameter to "true". | S,L | false |
| user-command-packages | A comma separated list of Java packages that contain classes implementing the CommandMarker interface. Matching classes will be loaded when the VM starts and will be available in the GFSH command-line utility. |
S | not set |
| validate-serializable-objects | A boolean that defaults to false. When true, instances of classes that are not internal to Tanzu GemFire and whose class name is not allowed by the list defined in the serializable-object-filter property will not be permitted to be deserialized. An IncompatibleClassException is thrown for objects not listed. JDK 8 build 121 or a later build must be installed to use this property. Servers and clients that do not meet this requirement will throw an exception upon startup. |
S, C | false |
-
Using Non-ASCII Strings in VMware Tanzu GemFire Property Files
You can specify Unicode (non-ASCII) characters in VMware Tanzu GemFire property files by using a
\uXXXXescape sequence.
