(zos)

Application server custom properties for z/OS

Some of the application server custom properties provided with the product can only be used with z/OS . This topic describes how to use these properties.

Deprecated feature: Setting these custom properties at the server level is deprecated. However, we can specify them as WebSphere variables with a scope of either a specific server, specific node, or specific cell. Server scoped WebSphere variables still override any settings specified at the node scope, or higher, and are added to the was.env file.depfeat

To set one of these custom properties for either an application server or a deployment manager, in the console, click Environment > WebSphere variables, select the appropriate node or cell from the list of available servers, nodes and cells, and then click New.

We can use the custom properties page to define the following application server custom properties for z/OS:


adjunct_jvm_direct_options

Specifies options that we need to pass directly to the JVM launch in the adjunct. This property is typically used for JVM options that the JVM cannot read from the options file specified as the value of the control_region_jvm_properties_file property. For example, the JVM cannot read the value specified for the -memorycheck option in the options file.

If we specify multiple options, use a semicolon to separate the options.

We can use the servant_jvm_direct_options and control_jvm_direct_options custom properties to specify options that we need to pass directly to the JVM launch in the servant and controller, respectively.

Information Value
Data Type String
Default Empty string
Used by Daemon No


(v8552) adjunct_region_start_sib_abend

Specifies what action should take place when a critical messaging engine takes too long to start.

If we set this property to 0, warning message BBOO0418W is issued.

If we set this property to 1, the entire application server abnormally terminates.

Information Value
Data Type Integer
Valid Values 1 or 0
Default 0
Used by Daemon No


(v8552) adjunct_region_start_sib_waittime

How many seconds the server startup process should wait for a critical messaging engine to become active.

This variable only applies if the adjunct_region_start_synchronized environment variable is set to 1. This property controls how long the startup process waits for both the adjunct region and the critical messaging engine to start. A value of 0 disables the timeout, and the startup process waits indefinitely for the adjunct region and the critical messaging engine to start.

Information Value
Data Type Integer
Default 300
Used by Daemon No


(v8552) adjunct_region_start_synchronized

Whether the startup process waits for all critical messaging engines to start before opening for e-business.

If we set this property to 1, the startup process waits for all critical messaging engines to start before starting to process e-business applications.

If we set this property to 0, the startup process does not wait for the critical messaging engines to start.

Information Value
Data Type Integer
Valid Values 1 or 0
Default 0
Used by Daemon No


allow_large_SAF_groups

Specifies to allow the application server to do lookups on large SAF groups.

If we set this property to 1, the size of the buffer used to do lookups is tripled from 8192 bytes to 24576 bytes.

We can also set this property to a specific number off bytes up to and including 2147483647. If we specify an integer, other than 1, as the value for this property, the buffer size becomes that number of bytes. For example, if specified allow_large_SAF_groups=21400000, the size of the buffer used to do lookups on SAF groups is 21400000 bytes.

If we do not specify a value for this property, or specify a value of 0, the buffer size is 8192 bytes.

Information Value
Data Type Integer
Range 1 - 2147483647
Default 0


com.ibm.ws.sib.ra.inbound.impl.MessageLockExpiry

When a message arrives on the queue that a message-driven bean (MDB) is consuming from, the message is locked and passed to the MDB in the servant region. If the servant region is disabled, or if there is an error processing the message on the servant region, this property defines how long the messaging engine waits before unlocking the message so that it can be delivered again.

Information Value
Data Type Integer
Units milliseconds
Default 300000
Range A positive integer. The value 0 indicates that the message lock never expires and the messaging engine waits indefinitely for the servant region to process and unlock the message.


(v8552) condense_java_trace

Whether to condense java trace or to show it in the old style which includes category information. The default is to condense the trace.

When this property is set to 1, java trace will be condensed.

When this property is set to 0, java trace will be expanded.

Information Value
Data Type Boolean
Acceptable values 0 or 1
Default 1


control_region_confirm_recovery_on_ no_srs

Whether requests are dispatched to servants following the detection of a no-servants situation. This property is ignored if the control_region_dreg_on_no_srs custom property is set to 0.

When this property is set to 1, the controller does not dispatch requests to the servants until it receives a response to message BBOO0297A. This message is issued following a no-servant situation when the server detects that the required minimal number of servants are available to process requests.

When this property is set to 0 (zero), the controller determines when to allow requests to be dispatched to the servants after a no-servant condition is detected.

Information Value
Data Type Integer
Acceptable values 0 or 1
Default 0


control_region_dreg_on_no_srs

Whether the controller rejects requests for dispatch within a servant when it detects that no servants are available to process requests.

When this property is set to 1, if the controller detects that there are no servants available to process requests, it rejects requests for dispatch within the servants. It also removes the application server from the registry of servers that workload management (WLM) uses to assign work, and stops the HTTP and message-driven bean (MDB) listeners. If 0, then the function is disabled.

When the minimum number of servants become available, the controller registers the application server with WLM again, starts the HTTP and MDB listeners, and allows requests to be dispatched to the servants.

Information Value
Data Type Integer
Acceptable values 0 or 1
Default 0


control_region_http_queue_timeout_percent

Percentage of the HTTP dispatch time limit used as the maximum amount of time that an HTTP request can spend on the workload management (WLM) queue. The protocol_http_timeout_output custom property is used to specify the maximum amount of time that an HTTP request can spend on the queue and in dispatch before an error message is issued, which indicates that an HTTP dispatch timeout has occurred.

The queue_timeout_percent request-level Reliability Availability and Serviceability (RAS) attribute overrides the control_region_http_queue_timeout_percent custom property for HTTP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Integer
Range 0 - 99
Default 0


control_region_https_queue_timeout_percent

Percentage of the HTTPS dispatch time limit used as the maximum amount of time that an HTTPS request can spend on the workload management (WLM) queue. The protocol_https_timeout_output custom property is used to specify the maximum amount of time that an HTTPS request can spend on the queue and in dispatch before an error message is issued, which indicates that an HTTPS dispatch timeout has occurred.

The queue_timeout_percent request-level RAS attribute overrides the control_region_https_queue_timeout_percent custom property for HTTPS requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Integer
Range 0 - 99
Default 0


control_region_http_requeue_enabled

This property applies only to a Websphere Application Server configured with a servant. Whether HTTP requests on the WLM Queue can be requeued to any available servant.

When this property is set to 1, HTTP requests waiting on the WLM queue with an affinity to a servant that is marked for termination can be requeued to any available servant when the servant with affinity has terminated.

When this property is set to 0 (zero), the server will fail any HTTP requests on the WLM queue with an affinity to a servant that is marked for termination immediately. Also, the server will fail any new HTTP request with affinity to the unhealthy servant.

Information Value
Data Type Integer
Acceptable values 0 or 1
Default 0


control_region_iiop_queue_timeout_percent

Percentage of the IIOP dispatch time limit used as the maximum amount of time that an IIOP request can spend on the workload management (WLM) queue.

This property only applies to the time that the request spends on the WLM queue. Use the control_region_wlm_dispatch_timeout custom property to limit the amount of time that the request spends on both the WLM queue and in dispatch,

The queue_timeout_percent request-level RAS attribute overrides the control_region_iiop_queue_timeout_percent custom property for IIOP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Integer
Range 0 - 99
Default 0


control_region_mdb_queue_timeout_percent

Percentage of the MDB dispatch time limit used as the maximum amount of time that an MDB request can spend on the workload management (WLM) queue.

This property only applies to the time that the request spends on the WLM queue. Use the control_region_mdb_request_timeout custom property to limit the amount of time that the request spends on both the WLM queue and in dispatch,

The queue_timeout_percent request-level RAS attribute overrides the control_region_mdb_queue_timeout_percent custom property for MDB requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Integer
Range 0 - 99
Default 0


control_region_mdb_request_timeout

Time, in seconds, that the server waits for a message-driven bean (MDB) request to receive a response. If the response is not received within the specified amount of time, the server removes the MDB request, and issues an error message that indicates that an MDB dispatch timeout has occurred.

The request_timeout request-level RAS attribute overrides the control_region_mdb_request_timeout custom property for MDB requests. You define the request-level RAS attributes in the workload classification file.

Set this value to 0 to disable the function.

Information Value
Data Type Integer
Units Seconds
Default 120


control_region_sip_queue_timeout_percent

Percentage of the Session Initiation protocol (SIP) dispatch time limit used as the maximum amount of time that a SIP request can spend on the workload management (WLM) queue.

This property only applies to the time that the request spends on the WLM queue. Use the protocol_sip_timeout_output custom property to limit the amount of time that the request spends on both the WLM queue and in dispatch.

The queue_timeout_percent request-level RAS attribute overrides the control_region_sip_queue_timeout_percent custom property for MDB requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Integer
Range 0 - 99
Default 0


control_region_sips_queue_timeout_percent

Percentage of the SIP Secure Sockets Layer (SSL) dispatch time limit used as the maximum amount of time that a SIP SSL request can spend on the workload management (WLM) queue.

This property only applies to the time that the request spends on the WLM queue. Use the protocol_sips_timeout_output custom property to limit the amount of time that the request spends on both the WLM queue and in dispatch.

The queue_timeout_percent request-level RAS attribute overrides the control_region_sips_queue_timeout_percent custom property for MDB requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Integer
Range 0 - 99
Default 0


control_region_timeout_delay

Number of seconds a controller waits after detecting a timeout before it terminates the servant. This time delay gives work currently running in the servant a chance to complete before the servant is terminated.

The specified length of time period starts when a timeout occurs. When a servant thread completes its current work item and determines that the servant is being terminated, the servant thread waits for the specified length of time instead of selecting a new work item.

When this field is set to 0 the controller terminates a servant as soon as the controller detects a timeout.

Information Value
Data Type Integer
Units Seconds
Default 0

This property is affected by the setting for the server_use_wlm_to_queue_work property:


control_region_timeout_dump_action

Type of dump that is taken whenever a timeout occurs for work that has been dispatched to a servant. This property only applies if the control_region_timeout_delay custom property is set to a non-zero value.

Valid values for this property are SVCDUMP, JAVACORE, HEAPDUMP, TRACEBACK, JAVATDUMP, and NONE. JAVACORE generates a Java core dump. SVCDUMP generates an SVC dump. JAVATDUMP generates a JVM-initiated TDUMP.

Information Value
Default TRACEBACK


control_region_timeout_dump_action_session

Type of dump that is taken whenever a timeout occurs for an HTTP, HTTPS, SIP, or SIPS request that has been dispatched to a servant.

This property only applies if the following corresponding variable is set to SESSION:

Valid values for this property are SVCDUMP, JAVACORE, HEAPDUMP, TRACEBACK, JAVATDUMP, and NONE. JAVACORE generates a Java core dump. SVCDUMP generates an SVC dump. JAVATDUMP generates a JVM-initiated TDUMP.

Information Value
Default TRACEBACK


control_region_timeout_save_last_servant

Whether the controller terminates the last available servant when a timeout situation occurs. If the controller does not terminate the last available servant when a timeout situation occurs, other work continues to be processed until a new servant is initialized. However, not terminating the last available servant might cause the loss of system resources if the dispatched servant thread that encountered the timeout situation continues to loop or stops functioning. For example, if timeouts keep occurring, the system might consume a high percentage of the available servant threads.

The functionality of this property depends on the values specified for other custom properties:

Information Value
Data Type Boolean
Default 0


controller_jvm_direct_options

Specifies options that we need to pass directly to the JVM launch in the controller. This property is typically used for JVM options that the JVM cannot read from the options file specified as the value of the control_region_jvm_properties_file property. For example, the JVM cannot read the value specified for the -memorycheck option in the options file.

If we specify multiple options, use a semicolon to separate the options.

We can use the servant_jvm_direct_options to specify options that we need to pass directly to the JVM launch in the servant. We can use the adjunct_jvm_direct_options custom properties to specify options that we need to pass directly to the JVM launch in the adjunct.

Information Value
Data Type String
Default Empty string
Used by Daemon No


Daemon_ras_trace_ctraceParms

Identity of the CTRACE PARMLIB member. The value can be either a two-character suffix, which is added to the CTIBBO string to form the name of the PARMLIB member, or the fully specified name of the PARMLIB member. For example, we can use the 01 suffix, which the system resolves to CTIBBO01. A fully specified name must conform to the naming requirements for a CTRACE PARMLIB member. For details, see z/OS MVS Diagnosis: Tools and Service Aids, GA22-7589.

If the specified PARMLIB member is not found, then tracing is defined to CTRACE, but no connection is available to a CTRACE external writer.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


default_internal_work_transaction_class

Default transaction class for internally processed work within the server.

If an internal classification element is not listed in the wlm_classification_file, or if the wlm_classification_file is not specified, then the default_internal_work_transaction_class setting is used. If a value is specified for internal classification setting listed in the wlm_classification_file, The value specified for the default_internal_work_transaction_class custom property is ignored.

Information Value
Data Type String
Default null (empty string)
Used by Daemon No


dynapplenv_wlm_select_policy

Specifies that value you want passed to the z/OS Workload Manager (WLM) as the SELECT_POLICY parameter of the IWMAEDEF service.

The IWMAEDEF service is used to create the dynamic application environment for the selection of work from the WLM queue. For more information about this service, see the z/OS WLM documentation for your version of the z/OS operating system.

We can specify values of 0, 1 or 2 for this property.

Information Value
Data Type Integer
Default 1
Used by Daemon No


iiop_max_msg_megsize

Specifies, in megabytes, the maximum size for IIOP requests. For example, if set the property to 35, then any requests over 35 MB are rejected. The minimum value for this property is 10, and the maximum value is 2048. Set the value to 0 or omit this property if we do not want to limit the size of IIOP requests.

Avoid trouble: This custom property only applies to application servers that are running in 64-bit mode. The maximum size for IIOP requests that are being handled by application servers running in 31-bit mode is 10 MB.gotcha

Information Value
Data Type Integer
Default 0
Used by Daemon No


iiop_max_send_queue_megsize

Specifies, in megabytes, the maximum amount of data that can be queued up to send asynchronously over a single IIOP connection. If the amount of data queued exceeds the specified value, future IIOP requests over this connection fail with a C9C26A4D minor code. The minimum value for this property is 0, which indicates that there is no limit to the amount of data that can be queued for sending. The maximum value is 2048.

Information Value
Data Type Integer
Default 0
Used by Daemon No


internal_work_thread_pool_size

Number of threads available for the servant region internal work thread pool. Set this property to the number of threads you would like available for the servant region internal work thread pool.

Information Value
Data Type Integer
Range 1-100
Default 3
Used by Daemon No


local_comm_max_msg_megsize

Specifies, in megabytes, the maximum size of locally connected communications requests. For example, if set the property to 35, then any requests over 35 MB are rejected. The minimum value for this property is 10 and the maximum value is 2048. Set the value to 0 or omit this property if we do not want to limit the size of locally connected communications requests.

Avoid trouble: This custom property only applies to application servers that are running in 64-bit mode. The maximum size for IIOP requests that are being handled by application servers running in 31-bit mode is 10 MB.gotcha

Information Value
Data Type Integer
Default 0
Used by Daemon No


ola_cicsuser_identity_propagate

Specifies permission for CICS application level identities to be used for authentication when the registration request is made.

When an application requests an optimized local adapter registration, the programmer can use two security-related bits in the registration flags structure to control identity propagation. For transactions that are inbound from CICS to WebSphere Application Server, bit 29 (x'00',x'00',x'00',x'04') controls how the identity is determined. When this property is turned on, the CICS application identity is used in WebSphere server authentication. When this property is turned off , the CICS region identity is used.

WebSphere Application Server administrators must set the environment variable to permit CICS application level identities to be used for authentication when the registration request is made. Set the value of to 1 to allow the use of CICS task level identity. If it is undefined or set to 0 (zero), registration requests can only request CICS region level authentication, otherwise if registration flag bit 29 is set to 1, the BBOA1REG registration request fails with a return code 8 and reason code 21. Bit 21 (x'00',x'00',x'00',x'01') controls outbound transaction security propagation. See outbound transactions for information.

Information Value
Data Type Integer
Default
Used by Daemon


protocol_accept_http_work_after_min_srs

Whether the application server waits until a minimum number of servants are ready to accept work before the application server starts the HTTP transport channels. If true, then when the minimum number of servants are ready for work, the HTTP transport channels start accepting work. If false, the HTTP transport channels start when the controller starts.

When true, the value specified for the Minimum number of instances property determines the number of servants that must be ready before the HTTP transport channels start. To change the setting of the Minimum number of instances property for an application server, in the console, click Servers > Server Types > WebSphere application servers > server_name > Java and process management > Server instance. To change the setting of this property for a deployment manager, in the console, click System Administration > dmgr > Java and process management > Server instance.

The job output indicates protocol_accept_http_work_after_min_srs: 1, if this property is set to true, or protocol_accept_http_work_after_min_srs: 0, if this property is set to false.

Information Value
Data Type Boolean
Default true
Used by Daemon No


protocol_accept_iiop_work_after_min_srs

Determines when the daemon starts to send requests to the IIOP transport channels. If true, the daemon starts to send requests to the IIOP transport channels when the minimum number of servants, as specified for the Minimum Number of Instances property, are ready to accept work, If false, the daemon starts to sent requests to the IIOP transport channels when the controller starts.

Avoid trouble: Even if this property is set to true, if you are running client applications that do caching, such as bean caching, it is possible for requests from these applications to be sent directly to the open IIOP port before the minimum number of servants are available. This situation might occur because the IIOP listeners start early in the server startup process, thereby possibly opening the IIOP port before the specified minimum number of servants are initialized. gotcha

To change the setting of the Minimum Number of Instances property for an application server, in the console, click Application servers > server > Java and Process Management > Server Instance.

To change the setting of this property for a deployment manager, in the console, click System administration > Deployment manager > Java and Process Management > Server Instance.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


protocol_bboc_log_response_failure

Specifies that if the BBOO0168W message is issued, then the failure that is detected when attempting to send a response to a client is recorded. The message is sent to the error log. The message text contains the request method name, the reply status, and the routing information identifying client.

Information Value
Data Type Boolean
Default false
Used by Daemon Yes


protocol_bboc_log_return_exception

Specifies that if the BBOO0169W message is issued, the response containing the SystemException is recorded. The message is sent to the error log. The message text contains the exception identifier and minor code, the request method name, and the routing information identifying client.

Avoid trouble: This property only applies to application servers. This property is ignored if it is specified for a deployment manager.gotcha

Information Value
Data Type Boolean
Default false
Used by Daemon Yes


protocol_giop_level_highest

Specifies the CORBA General Inter-ORB Protocol (GIOP) protocol version level used by the application server object request broker (ORB). Valid values are 1.1 and 1.2. Interoperable object references (IORs) that are exported from this server use the GIOP level indicated.

You might need to change the setting of this property from the default value if you use a client ORB that is not shipped as part of the product, and that supports a previous version of the CORBA standard. For example, we might need to change from the default protocol version level of 1.2 to 1.1 to support a client ORB that supports the 1.1 CORBA standard instead of the 1.2 CORBA standard.

The maximum GIOP level that the daemon address space supports is 1.1. The GIOP level has of the daemon has no effect on the GIOP levels of the application servers that connect to the daemon.gotcha

Information Value
Data Type String
Default 1.2
Used by Daemon Yes


protocol_http_backlog

Maximum length for the queue of pending connections using HTTP. The value specified can be limited by the specification of the SOMAXCONN statement in the TCP/IP profile.

Information Value
Data Type Integer
Default 10
Used by Daemon No


protocol_http_large_data_inbound_buffer

Specifies, in bytes, the limit of the size of incoming HTTP requests when inbound HTTP chunking is disabled. For example, if set the property to 15728640, any requests over 15 MB are rejected. Specify 0 to reject any requests that are larger than 10 MB.

Avoid trouble: Use this custom property only for an application server running in 31-bit mode. If the application server is running in 64-bit mode, use the protocol_http_large_data_inbound_buffer_64bit custom property to set a limit for this inbound buffer.gotcha

Information Value
Data Type Integer
Default 0
Used by Daemon No


protocol_http_large_data_inbound_buffer_64bit

Specifies, in megabytes, the size limit for incoming HTTP requests when inbound HTTP chunking is disabled. For example, if set the property to 35, any HTTP requests over 35 MB are rejected. Specify 0 for this property if we do not want to limit the size of unchunked HTTP requests.

Avoid trouble: Use this custom property only for an application server running in 64-bit mode. If the application server is running in 31-bit mode, use the protocol_http_large_data_inbound_buffer custom property to set a limit for this inbound buffer.gotcha

Information Value
Data Type Integer
Default 0
Used by Daemon No


protocol_http_large_data_response_buffer

Specifies, in bytes, the maximum length of the response buffer used for HTTP requests. Responses larger than this value are rejected. Specify a value of 0 if we do not need a large response buffer because all of the HTTP responses are less than 10 MB.

Avoid trouble: This property only applies to application servers. It does not apply to a deployment manager.gotcha

Information Value
Data Type Integer
Default 104857600
Used by Daemon No


protocol_http_resolve_foreign_hostname

Whether to perform Domain Name Server (DNS) resolution of the IP address of a foreign client to a DNS registered host name for each established HTTP session. If 1, then the DNS host name resolution is performed. If 0, then the DNS host name resolution is not performed, and a textual representation of the IP address of the foreign client is used instead of the DNS host name.

Information Value
Data Type Boolean
Default 0
Used by Daemon Yes


protocol_http_timeout_output_recovery

Recovery action that is taken when an HTTP request does not complete within a designated length of time. Setting this property to SERVANT allows servants to terminate when a timeout occurs. If an HTTP request is under dispatch in a servant when its timeout value is reached, the servant terminates with an ABEND EC3 RSN=04130007. The HTTP request and socket are then cleaned up. If SESSION, no attempt is made to disrupt the processing of a dispatched HTTP request within a servant. However, the HTTP request and socket are still cleaned up. Using the SESSION setting might result in a loss of resources if the dispatched HTTP request loops or becomes inactive.

The timeout_recovery request-level RAS attribute overrides the protocol_http_timeout_output_recovery custom property for HTTP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Default SERVANT
Used by Daemon No


protocol_https_backlog

Maximum length for the queue of pending connections using HTTPS. The value specified can be limited by the specification of the SOMAXCONN statement in the TCP/IP Profile.

Information Value
Data Type Integer
Default 10
Used by Daemon No


protocol_https_cert_mapping_file

Name of a file containing entries that map IP addresses to server certificate labels. We can set this property at the cell, node, or server level.

Deprecated feature: The protocol_https_cert_mapping_file property is deprecated. depfeat

When an HTTP SSL connection request is received, the application server checks the IP address against entries in the file specified for this property. If the application server finds a match, the certificate mapped to the IP address is used for the connection. If the application server does not find a match, it checks the protocol_https_default_cert_label property for the name of a certificate. If a certificate name is specified, the application server uses that certificate to establish the connection. If a certificate name is not specified, the default server certificate specified in the RACF SSL keyring, that is owned by the application server, is used to establish the HTTP SSL connection.


protocol_https_default_cert_label

Label of the server certificate that the application server uses when establishing HTTP SSL connections with the application server. We can set this property at the cell, node, or server level.

Deprecated feature: The protocol_https_default_cert_label property is deprecated. depfeat

If the name of a certificate is note specified for this property, the default server certificate specified in the RACF SSL keyring, that is owned by the application server, is used to establish the HTTP SSL connection.


protocol_https_timeout_output_recovery

Recovery action that is taken when an HTTPS request does not complete within a designated length of time. Setting this property to SERVANT allows servants to terminate when a timeout occurs. If an HTTPS request is under dispatch in a servant when its timeout value is reached, the servant terminates with an ABEND EC3 RSN=04130007. The HTTPS request and socket are then cleaned up. If SESSION, no attempt is made to disrupt the processing of a dispatched HTTPS request within a servant. However, the HTTPS request and socket are still cleaned up. Using the SESSION setting might result in a loss of resources if the dispatched HTTPS request loops or becomes inactive.

The timeout_recovery request-level RAS attribute overrides the protocol_https_timeout_output_recovery custom property for HTTPS requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Default SERVANT
Used by Daemon No


protocol_iiop_backlog

Maximum length for the queue of pending connections using the CORBA Internet Inter-ORB protocol (IIOP). The value specified might be limited by the specification of the SOMAXCONN statement in the TCP/IP profile.

Information Value
Data Type Integer
Default 10
Used by Daemon Yes


protocol_iiop_backlog_ssl

Maximum length for the queue of pending connections using IIOP SSL. The value specified can be limited by the specification of the SOMAXCONN statement in the TCP/IP profile.

Information Value
Data Type Integer
Default 10
Used by Daemon Yes


protocol_iiop_enable_wlm_routing

Whether WebSphere Application Server should enable the process that allows it to receive input from the Workload Manager (WLM) to make IIOP routing decisions.

When this property is enabled, WLM sends WebSphere Application Server input that WAS uses to make IIOP routing decisions.

IBM recommends that you leave this property enabled. However, because there is a limit to the number of servers than can use this feature on a given LPAR, we might need to disable this property for some of the servers on an LPAR if the servers start experiencing the following error:

.

Information Value
Data Type Boolean
Default 1
Used by Daemon No


protocol_iiop_local_propagate_wlm_enclave

Whether to propagate the workload management (WLM) enclave that is associated with the currently dispatched request on an outbound IIOP request that was made to another server on the same z/OS system over local interaddress space communication protocols.

Information Value
Data Type Boolean
Default 1
Used by Daemon No


protocol_iiop_resolve_foreign_hostname

Whether to perform Domain Name Server (DNS) resolution of the IP address of a foreign client to a DNS registered host name for each established IIOP session. If 1, then the DNS host name resolution is performed. If o, then the DNS host name resolution is not performed, and a textual representation of the IP address of the foreign client is used instead of the DNS host name.

Information Value
Data Type Boolean
Default 1
Used by Daemon Yes


protocol_iiop_resolve_object_url_local

Whether, if possible, the application server resolves the object URLs to a server that is local to the LPAR. By default, the object URL is processed in order and the first server that can be reached is chosen. When this property is set to 1, any servers local to the LPAR are checked first and only if none of those servers are reachable is a resolution attempted to a remote server.

Information Value
Data Type Integer
Default 0
Acceptable values 0 or 1


protocol_jfap_queue_limit

Number of messages that can reside on the JFAP protocol message queue that is associated with a servant process. This queue is used to hold pending messages that are being sent between the servant process and the controller.

We can specify any positive integer as the value for this property. However, whenever possible, you should use the default value 0. A value of 0 indicates that the number of messages that can reside in the JFAP protocol message queue is four times the number of dispatch threads in the servant. Therefore, this value provides a calculated limit based on workload, and removes the arbitrary restraints imposed by specifying a specific value for this property.

The number of dispatch threads in the servant is controlled by the ORB services Workload profile setting. See the topic ORB services advanced settings on the z/OS platform for more information about this setting.

You should also review the topic Configure MDB throttling for the default messaging provider for a description of how to tune message-driven beans.

Information Value
Data Type Integer
Default 0
Used by Daemon No


protocol_sip_timeout_output_recovery

Recovery action that is taken when a SIP request does not complete within a designated length of time. Setting this property to SERVANT allows servants to terminate when a timeout occurs. If a SIP request is under dispatch in a servant when its timeout value is reached, the servant terminates with an ABEND EC3 RSN=04130007. The SIP request and socket are then cleaned up. If SESSION, no attempt is made to disrupt the processing of a dispatched SIP request within a servant. However, the SIP request and socket are still cleaned up. Using the SESSION setting might result in a loss of resources if the dispatched SIP request loops or becomes inactive.

The timeout_recovery request-level RAS attribute overrides the protocol_sip_timeout_output_recovery custom property for SIP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Default SERVANT
Used by Daemon No


protocol_sips_timeout_output_recovery

Recovery action that is taken when a SIPS request does not complete within a designated length of time. Setting this property to SERVANT allows servants to terminate when a timeout occurs. If a SIPS request is under dispatch in a servant when its timeout value is reached, the servant terminates with an ABEND EC3 RSN=04130007. The SIPS request and socket are then cleaned up. If SESSION, no attempt is made to disrupt the processing of a dispatched SIPS request within a servant. However, the SIPS request and socket are still cleaned up. Using the SESSION setting might result in a loss of resources if the dispatched SIPS request loops or becomes inactive.

The timeout_recovery request-level RAS attribute overrides the protocol_sips_timeout_output_recovery custom property for SIPS requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Default SERVANT
Used by Daemon No


ras_debugEnabled

Specifies to use an external debugger tool with the application server for tracing and debugging client and server application components such as JSP files, servlets, and enterprise beans.

Information Value
Data Type Boolean
Default false
Used by Daemon Yes


ras_default_msg_dd

Whether to redirect write-to-operator (WTO) messages that use the default routing to SYSPRINT. These messages are redirected to the location identified through the DD card on the JCL start procedure for the server. These WTO messages are typically issued during initialization.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


ras_dumpoptions_dumptype

Default dump used by the signal handler. Do not change this property unless directed to do so by IBM Support personnel.

Information Value
0 No dump is generated.
1 A ctrace dump is taken.
2 A cdump dump is taken.
3 A csnap dump is taken.
4 A CEE3DMP dump is taken.

Information Value
Data Type Integer
Default 3
Used by Daemon Yes


ras_dumpoptions_ledumpoptions

Dump options to use with a CEE3DMP dump. If we want more than one option, then separate each option with a blank space. Do not change this property unless directed to do so by IBM service personnel.

Information Value
Data Type String
Default THREAD(ALL) BLOCKS
Used by Daemon Yes


ras_enhanced_serviceability_level

Level of diagnostic information that the application server collects. Reducing the value of this property increases the performance of the application server. However, less data is available to IBM Support personnel when diagnosing a problem. The opposite is true when increasing the value of this property. The default value of 5 provides a balance of performance and diagnostic information.

Information Value
Data type Integer
Range 0 - 11
Default 5
Used by Daemon Yes


ras_error_log_version=n

Version of trace log to display. Valid values are 1 through 3. Version 3 displays a message tag defined in the classification file if an application request is associated with the current thread at the time the error occurs, in addition to previous version information. Version 2 displays the server name, cluster name, and cell name, in addition to the base information. For example, ras_error_log_version=2.

Information Value
Data Type String
Default 3
Used by Daemon Yes


ras_hardcopy_msg_dd

Specifies to redirect write-to-operator (WTO) messages that are routed to hardcopy. These messages are redirected to the location that is identified through the DD card on the server JCL start procedure. These WTO messages are primarily audit messages that are issued from Java code during initialization.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


ras_java_oom_action

Type of diagnostic action performed when a Java Virtual Machine out-of-memory condition occurs.

The possible values are:

NONE

No diagnostic action is performed.

WTO

A BBOO0404E error message is written to hardcopy.

SVCDUMP

An SVCDUMP of the affected address space is taken, with minor code C9C2704B.

Information Value
Data Type String
Default NONE
Used by Daemon Yes


ras_java_oom_interval

Specifies, in seconds, the length of time during which the diagnostic action specified by the ras_java_oom_action environment variable is not repeated. This property helps to limit the number of diagnostic actions that occur for related out-of-memory conditions.

Information Value
Data Type integer
Range 0 - 65535
Default 600
Used by Daemon Yes


ras_log_logstreamName

Log stream that the product uses for error information. If the specified log stream is not found or not accessible, a message is issued and errors are written to the server job log. If this variable is not specified, the product uses the SYSOUT stream.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


ras_message_routing_console

Specifies a comma-delimited list of WAS message IDs to be rerouted to the operator console (route code 2). Each message displays in the operator console, instead of in the default location.

Information Value
Data Type String
Default None
Used by Daemon Yes


(v8552) ras_message_routing_copy_console

Specifies a comma-delimited list of WAS message IDs to be copied to the operator console. Each message displays in both the operator console and the default location.

Information Value
Data Type String
Default None
Used by Daemon Yes


(v8552) ras_message_routing_copy_errorlog

Specifies a comma-delimited list of WAS message IDs to be copied to the errorlog. Each message displays in both the errorlog and the default location.

Information Value
Data Type String
Default None
Used by Daemon Yes


(v8552) ras_message_routing_copy_hardcopy

Specifies a comma delimited list of WebSphere message IDs to be rerouted to hardcopy (route code 11). Each message shows up in the hardcopy as well as in its default location.

Information Value
Data Type String
Default None
Used by Daemon Yes


ras_message_routing_errorlog

Specifies a comma-delimited list of WAS message IDs to be rerouted to the error log. Each message displays in the error log, rather than its default location.

Information Value
Data Type String
Default None
Used by Daemon Yes


ras_message_routing_hardcopy

Specifies a comma delimited list of WebSphere message IDs to be rerouted to hardcopy (route code 11). Each message shows up in the hardcopy instead of in its default location.

Information Value
Data Type String
Default None
Used by Daemon Yes


ras_message_routing_none

Specifies a comma-delimited list of WAS message IDs to be ignored. Each message is hidden from one or more normal output locations.

Information Value
Data Type String
Default None
Used by Daemon Yes


ras_minorcode_action

Default behavior for gathering documentation about system exception minor codes.

Information Value
Data Type String
Default NODIAGNOSTICDATA
Used by Daemon Yes

We can also specify the following values.

Information Value
CEEDUMP Captures callback and offsets. Taking a CEE dumps is a lengthy process, and transaction timeouts can occur during this process.
TRACEBACK Captures Language Environment and UNIX traceback data for the z/OS operating system.
SVCDUMP Captures an MVS dump, but does not produce a dump in the client.


ras_stderr_ff_interval

Interval of time, in minutes, that the system waits before writing the next form-feed character to standard error (SYSOUT).

If we are running on z/OS Version 1.13 or later, and using JES2, we can use JES2 DD keywords to segment output using the periodic writing of form-feed characters to the output streams.

If we are running on z/OS Version 1.12 or earlier, and using JES2, to segment the output, include the SEGMENT= parameter, along with this environment setting, on the SYSPRINT DD statement. The value of the SEGMENT= parameter is the number of form-feeds required before the first segment is closed and a new segment is allocated. The SEGMENT= parameter is not supported on JES3.

Information Value
Data Type Integer
Default 0
Used by Daemon Yes


ras_stderr_ff_line_interval

Number of lines of output that is written between the writing of form-feed characters to standard error (SYSOUT).

Because of uncontrollable factors, such as line wrapping, the product can only approximate the number of lines of output it has written. Therefore the actual number of lines written between the writing of form-feed characters might be plus or minus 5 percent of the value specified for the property.

If we are running on z/OS Version 1.13 or later, and using JES2, we can use JES2 DD keywords to segment output using the periodic writing of form-feed characters to the output streams.

If we are running on z/OS Version 1.12 or earlier, and using JES2, to segment the output, include the SEGMENT= parameter, along with this environment setting, on the SYSPRINT DD statement. The value of the SEGMENT= parameter is the number of form-feeds required before the first segment is closed and a new segment is allocated. The SEGMENT= parameter is not supported on JES3.

Information Value
Data Type Integer
Default 0
Used by Daemon Yes


ras_stdout_ff_interval

Interval of time, in minutes, between the writing of a form-feed character to standard output(SYSPRINT).

If we are running on z/OS Version 1.13 or later, and using JES2, we can use JES2 DD keywords to segment output using the periodic writing of form-feed characters to the output streams.

If we are running on z/OS Version 1.12 or earlier, and using JES2, to segment the output, include the SEGMENT= parameter, along with this environment setting, on the SYSPRINT DD statement. The value of the SEGMENT= parameter is the number of form-feeds required before the first segment is closed and a new segment is allocated. The SEGMENT= parameter is not supported on JES3.

Information Value
Data Type Integer
Default 0
Used by Daemon Yes


ras_stdout_ff_line_interval

Number of lines of output between the writing of form-feed characters to standard output (SYSPRINT).

Because of uncontrollable factors, such as line wrapping, the product can only approximate the number of lines of output it has written. Therefore the actual number of lines written between the writing of form-feed characters might be plus or minus 5 percent of the value specified for the property.

If we are running on z/OS Version 1.13 or later, and using JES2, we can use JES2 DD keywords to segment output using the periodic writing of form-feed characters to the output streams.

If we are running on z/OS Version 1.12 or earlier, and using JES2, to segment the output, include the SEGMENT= parameter, along with this environment setting, on the SYSPRINT DD statement. The value of the SEGMENT= parameter is the number of form-feeds required before the first segment is closed and a new segment is allocated. The SEGMENT= parameter is not supported on JES3.

Information Value
Data Type Integer
Default 0
Used by Daemon Yes


ras_tag_wto_messages=n

Whether messages that are written to the operator console will be tagged, if appropriate. Valid values are 0 and 1. For example, ras_tag_wto_messages=0.

Information Value
Data Type Integer
Default 1
Used by Daemon Yes


ras_time_local

Whether time stamps in the error log display are in local time or Greenwich Mean Time (GMT). The time stamp is in GMT if this property is set to false.

Information Value
Data Type Boolean
Default false
Used by Daemon Yes


ras_trace_basic

Specifies tracing overrides for particular product subcomponents. Subcomponents, specified by numbers, receive basic and exception traces. If we specify more than one subcomponent, use parentheses and separate the numbers with commas. Do not change this property unless directed to do so by IBM service personnel.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


ras_trace_BufferCount

Number of trace buffers to allocate.

Information Value
Data Type Integer
Valid values 4 through 8
Default 4
Used by Daemon Yes


ras_trace_BufferSize

Specifies, in bytes, the size of a single trace buffer. We can use the letters K, for kilobytes, or M, for megabytes.

Information Value
Data Type String
Valid values 128K through 4M
Default 1M
Used by Daemon Yes


ras_trace_defaultTracingLevel

Default tracing level for the product. Use this variable with the ras_trace_basic and ras_trace_detail variables to set tracing levels for product subcomponents. Do not change this property unless directed by IBM Support personnel.

Information Value
0 No tracing
1 Exception tracing
2 Basic and exception tracing
3 Detailed tracing, including basic and exception tracing

Information Value
Data Type Integer
Default 1
Used by Daemon Yes


ras_trace_detail

Specifies tracing overrides for particular product subcomponents. Subcomponents, specified by numbers, receive detailed traces. If we specify more than one subcomponent, use parentheses and separate the numbers with commas. Do not change this property unless directed to do so by IBM Support personnel.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


ras_trace_exclude_specific

Specifies product trace points to exclude from tracing activity.

Trace points are specified by 8-digit, hexadecimal numbers. Do not use this property unless directed to do so by IBM service personnel. If IBM service personnel directs you to specify more than one trace point, use parentheses and separate the numbers with commas. You also can specify a variable name by enclosing the name in single quotation marks.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes

Avoid trouble: Results sometimes depend on the value specified for the ras_trace_minorCodeDefault environment variable. If we specify ras_trace_minorCodeTraceBacks=ALL and ras_minorcode_action=NODIAGNOSTICDATA, you get a traceback. However, if specified ras_trace_minorCodeTraceBacks=(null value) and ras_minorcode_action=TRACEBACK, you also get a traceback. Specifying ras_trace_minorCodeTraceBacks=(null value) causes TRACEBACK data to be collected, instead of canceled.gotcha


ras_trace_log_version=n

Version of trace log to display. Valid values are 1 and 2. Version 2 displays a message tag defined in the classification file if an application request is associated with the current thread at the time the trace is issued. For example, ras_trace_log_version=1.

Information Value
Data Type String
Default 2
Used by Daemon Yes


ras_trace_outputLocation

Specifies where to send trace records. We can specify:

For servers, we can specify one or more values, separated by a space.

Information Value
Data Type String
Default SYSPRINT BUFFER
Used by Daemon Yes


ras_trace_specific

Specifies tracing overrides for specific product trace points. Trace points are indicated by 8-digit, hexadecimal numbers. To specify more than one trace point, use parentheses and separate the numbers with commas. We can also specify tracing on a specific environment variable using the name enclosed in single quotation marks. Do not use this property unless directed to do so by IBM Support personnel.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


(V8502) redirect_server_output_dir

Specifies the output directory on the Unix Systems Services(USS) file system to which redirected servant output files are to be written.

This value must be an absolute path to an existing directory with appropriate write permissions. If a valid value is not specified for this variable, by default, output is redirected to DD cards.

Information Value
Data Type String
Default Empty string
Used by Daemon Yes


register_ifaedreg_also

Whether you want z/OS to create SMF Type 89 Subtype 2 records, in addition to SMF Type 89 Subtype 1 records. In previous releases of the product, after the product registered with z/OS, z/OS created SMF Type 89 Subtype 2 records to collect product usage data.

transition: Now, after the product registers with z/OS, z/OS produces SMF Type 89 Subtype 1 records instead of SMF Type 89 Subtype 2 records. To have SMF Type 89 Subtype 2 records created in addition to the SMF Type 89 Subtype 1 records, add the register_ifaedreg_also variable to the WebSphere variables and set this property to 1. To turn off the creation of SMF Type 89 Subtype 2 records, set this variable to 0.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


security_SMF_record_first_auth_user

Whether to record the first authenticated user under request dispatch in the SM120CRE field in the System Management Facility (SMF) server activity record.

If 1, then the first authenticated user under request dispatch is written to the SM120CRE field. If 0, then the ID of the user under which the server activity began is written to the SM120CRE field.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


servant_jvm_direct_options

Specifies options that we need to pass directly to the JVM launch in the servant. This property is typically used for JVM options that the JVM cannot read from the options file specified as the value of the control_region_jvm_properties_file property. For example, the JVM cannot read the value specified for the -memorycheck option in the options file.

If we specify multiple options, use a semicolon to separate the options.

We can use the controller_jvm_direct_options and adjunct_jvm_direct_options custom properties to specify options that we need to pass directly to the JVM launch in the controller and adjunct, respectively.

Information Value
Data Type String
Default Empty string
Used by Daemon No


servant_region_custom_thread_count

Number of application threads used in each of the servants that are running in an application server.

If we specify a value for this custom property, set the Workload profile property on the ORB services z/OS additional settings page in the console to CUSTOM before this setting becomes effective. To navigate to this page, in the console, click Servers > Server Types > WebSphere application servers > server_name > Container services > ORB service > z/OS additional settings.

Best practice: Before specified a value greater than 100 in a production environment, you should try the value in a test environment to ensure we do not negatively affect performance.bprac

Information Value
Data Type Integer
Range 1 - 500
Used by Daemon No


server_dlls_in_hfs

Whether product DLLs are loaded from the Hierarchical File System (HFS) or from the STEPLIB, LPA, or link list.

Information Value
Data type Boolean
Default 1
Used by Daemon Yes

Set this property to 1 (the default) indicates that the DLLs should be loaded form the HFS. Setting this property to 0 indicates that the DLLs have been put into the STEPLIB, LPA, or link list.

We can use the switchModules.sh script to load the DLLs from the HFS into a dataset.


server_region_connect_to_wlm_early

Whether the servant should connect to Workload Manager (WLM) at the beginning or the end of servant initialization.

Information Value
Data Type Boolean
Default 0
Used by Daemon No

Set this property to 1 enables the servant to connect to WLM at the beginning of servant initialization. After the servant connects to WLM, WLM is able to classify asynchronous work that is started during the remainder of the servant initialization process. A consequence of enabling the servant to connect to WLM at the beginning of servant initialization is that when there are multiple servants defined, the servants all start almost concurrently. Concurrent initialization of multiple servants might cause high CPU usage that needs to be considered.

Set this property to 0 prevents any asynchronous work that is started during servant initialization from being classified by WLM because the servant does not connect to WLM until the end of the servant initialization process.


server_region_cputimeused_dump_action

Dump action that the server takes when requests exceed their CPU timeout specified on the server_region_request_cputimeused_limit custom property.

Information Value
Valid values SVCDUMP, JAVACORE, HEAPDUMP, TRACEBACK, JAVATDUMP, and NONE
Range TRACEBACK
Used by Daemon No

The cputimeused_dump_action request-level RAS attribute overrides the server_region_cputimeused_dump_action custom property for HTTP, IIOP, SIP, MDB, and optimized local adapter requests. You define the request-level RAS attributes in the workload classification file.


server_region_dpm_dump_action

Type of dump that is taken when the dispatch progress monitor (DPM) interval has expired for a request. After the dump is taken the DPM interval is reset.

The dpm_dump_action request-level RAS attribute overrides the server_region_dpm_dump_action custom property. You define the request-level RAS attributes in the workload classification file.

Information Value
Valid values none, svcdump, javacore, heapdump, javatdump, and traceback
Default traceback
Used by Daemon No


server_region_http_stalled_thread_dump_action

Type of dump that is taken when a request is considered unresponsive. After the dump is taken the controller is notified of the unresponsive request. The controller might then terminate the servant based on the values specified for other custom properties, such as server_region_stalled_thread_threshold_percent, control_region_timeout_delay, and control_region_timeout_save_last_servant.

The stalled_thread_dump_action request-level RAS attribute overrides the server_region_http_stalled_thread_dump_action custom property. You define the request-level RAS attributes in the workload classification file.

Information Value
Valid values none, svcdump, javacore, heapdump, and traceback.
Default traceback
Used by Daemon No


server_region_https_stalled_thread_dump_action

Type of dump that is taken when a request is considered unresponsive. After the dump is taken the controller is notified of the unresponsive request. The controller might then terminate the servant based on the values specified for other custom properties, such as server_region_stalled_thread_threshold_percent, control_region_timeout_delay, and control_region_timeout_save_last_servant.

The stalled_thread_dump_action request-level RAS attribute overrides the server_region_ https_stalled_thread_dump_action custom property. You define the request-level RAS attributes in the workload classification file.

Information Value
Valid values none, svcdump, javacore, heapdump, and traceback.
Default traceback
Used by Daemon No


server_region_iiop_stalled_thread_dump_action

Type of dump that is taken when a request is considered unresponsive. After the dump is taken the controller is notified of the unresponsive request. The controller might then terminate the servant based on the values specified for other custom properties, such as server_region_stalled_thread_threshold_percent, control_region_timeout_delay, and control_region_timeout_save_last_servant.

The stalled_thread_dump_action request-level RAS attribute overrides the server_region_ iiop_stalled_thread_dump_action custom property. You define the request-level RAS attributes in the workload classification file.

Information Value
Valid values none, svcdump, javacore, heapdump, and traceback.
Default traceback
Used by Daemon No


server_region_jvm_localrefs

Do not use this property unless directed to do so by IBM Support personnel.

Information Value
Data Type Integer
Default 128
Used by Daemon No


server_region_jvm_logfile

Specifies the Hierarchical File System (HFS) file in which Java Native Interface (JNI) and class debug messages from the JVM are logged. Use this variable only in a single-server environment. If we use this property in a multiple-server environment, then all of the servers write to the same file, and we might have difficulty using the file for diagnostic purposes.

Information Value
Data Type String (file name)
Default Empty string
Used by Daemon No


server_region_mdb_stalled_thread_dump_action

Type of dump that is taken when a request is considered unresponsive. After the dump is taken the controller is notified of the unresponsive request. The controller might then terminate the servant based on the values specified for other custom properties, such as server_region_stalled_thread_threshold_percent, control_region_timeout_delay, and control_region_timeout_save_last_servant.

The stalled_thread_dump_action request-level RAS attribute overrides the server_region_mdb_stalled_thread_dump_action custom property for MDB requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Valid values none, svcdump, javacore, heapdump, and traceback.
Default traceback
Used by Daemon No


server_region_recycle_count

Number of transactions that are processed by a servant process after which the servant process is recycled. Workload management (WLM) ends the servant after all affinity requirements are met. Specify a nonzero value to enable recycling.

You might want to enable recycling if, after running for an extended period, the application is experiencing out-of-memory exceptions. Out-of-memory exceptions can result from memory leakage by the application.

Information Value
Data Type Integer
Default 0
Used by Daemon No


server_region_request_cputimeused_limit

Specifies, in milliseconds, the amount of CPU time that an application request can consume.

The server_region_request_cputimeused_limit custom property helps you to prevent a single application request from monopolizing the available CPU time because it allows you to limit the amount of CPU time that a single request can use. A CPU monitor is invoked when a request is dispatched. If the request exceeds the specified amount of CPU time, the controller considers the request unresponsive. The controller then issues message BBOO0327, to let the requesting application know that the request was unresponsive.

The monitor, that monitors the amount of CPU time that a request is using, typically sends a signal to the dispatched thread when the amount of CPU time used exceeds the specified amount. However, there are situations when this signal cannot be delivered, and the request remains pending. For example, if the thread goes native and invokes a PC routine, the signal remains pending until the PC routine returns.

The cputimeused_limit request-level RAS attribute overrides the server_region_request_cputimeused_limit custom property for HTTP, IIOP, SIP, MDB, and optimized local adapter requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data type integer
Default 0
Used by Daemon No


server_region_sip_stalled_thread_dump_action

Type of dump that is taken when a request is considered unresponsive. After the dump is taken the controller is notified of the unresponsive request. The controller might then terminate the servant based on the values specified for other custom properties, such as server_region_stalled_thread_threshold_percent, control_region_timeout_delay, and control_region_timeout_save_last_servant.

The stalled_thread_dump_action request-level RAS attribute overrides the server_region_sip_stalled_thread_dump_action custom property. You define the request-level RAS attributes in the workload classification file.

Information Value
Valid values none, svcdump, javacore, heapdump, and traceback.
Default traceback
Used by Daemon No


server_region_sips_stalled_thread_dump_action

Type of dump that is taken when a request is considered unresponsive. After the dump is taken the controller is notified of the unresponsive request. The controller might then terminate the servant based on the values specified for other custom properties, such as server_region_stalled_thread_threshold_percent, control_region_timeout_delay, and control_region_timeout_save_last_servant.

The stalled_thread_dump_action request-level RAS attribute overrides the server_region_sips_stalled_thread_dump_action custom property. You define the request-level RAS attributes in the workload classification file.

Information Value
Valid values none, svcdump, javacore, heapdump, and traceback.
Default traceback
Used by Daemon No


server_region_stalled_thread_threshold_percent

Percentage of threads that can become unresponsive before the controller terminates the servant.

If 0 is specified, the controller terminates the servant as soon as the controller determines that at least one thread has become unresponsive.

Information Value
Data Type Integer
Default 0
Used by Daemon No


server_SMF_outbound_enabled

Whether you want a z/OS System Management Facility (SMF) 120 Subtype 10 record created for WebSphere Optimized Local Adapter outbound requests.

If we specify 1 for this property, an SMF 120 Subtype 10 record is created.

Collecting the data for this record could impact performance; do not enable this property unless you have a specific reason for collecting the data included in this record.

Information Value
Data type Boolean
Default 0
Used by Daemon No


server_SMF_request_activity_async

Whether you want the asynchronous data section included in SMF 120 Subtype 9 records created.

If true, the asynchronous data section is included in any SMF 120 Subtype 9 record created during asynchronous work.

The setting for this property is ignored if the server_SMF_request_activity_enabled property is set to false.

Information Value
Data type Boolean
Default 0
Used by Daemon No


server_SMF_request_activity_CPU_detail

Whether you want the CPU usage breakdown section included in any SMF 120 Subtype 9 record created.

If true, the CPU usage breakdown section is included in any SMF 120 Subtype 9 record created.

The setting for this property is ignored if the server_SMF_request_activity_enabled property is set to false.

The SMF_request_activity_CPU_detail request-level RAS attribute overrides the server_SMF_request_activity_CPU_detail custom property for HTTP requests and IIOP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


server_SMF_request_activity_enabled

Whether you want the z/OS System Management Facility (SMF) to create an SMF 120 Subtype 9 record.

If we specify true for this property, an SMF 120 Subtype 9 record is created. Because the record is a relatively large, and collecting the data for this record could have an impact on performance, do not enable this property unless you have a specific reason for collecting the data included in this record.

We can also lower the performance impact of creating this record by disabling one or more of the following properties.

The SMF_request_activity_enabled request-level RAS attribute overrides the server_SMF_request_activity_enabled custom property for HTTP requests and IIOP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


server_SMF_request_activity_security

Whether you want the Security data section included in any SMF 120 Subtype 9 record created.

If true, the Security data section is included in any SMF 120 Subtype 9 record created.

The setting for this property is ignored if the server_SMF_request_activity_enabled property is set to false.

The SMF_request_activity_security request-level RAS attribute overrides the server_SMF_request_activity_security custom property for HTTP requests and IIOP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


server_SMF_request_activity_timestamps

Whether you want the z/OS formatted timestamps section included in any SMF 120 Subtype 9 record created.

If true, the z/OS formatted timestamps section is included in any SMF 120 Subtype 9 record created.

The setting for this property is ignored if the server_SMF_request_activity_enabled property is set to false.

The SMF_request_activity_timestamps request-level RAS attribute overrides the server_SMF_request_activity_timestamps custom property for HTTP requests and IIOP requests. You define the request-level RAS attributes in the workload classification file.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


server_start_wait_for_initialization_Timeout

How long the startServer.sh command processing waits for the product initialization process to complete. By default, startServer.sh command processing waits indefinitely until initialization is complete.

Complete one of the following actions:

Information Value
Data Type Integer
Default 0
Used by Daemon No


server_use_wlm_to_queue_work

Whether WLM is used for workload queuing.

Set this property to 1 if you are using stateless application models. With these models, application objects, such as EJB and HTTP sessions, are only resident in memory for the life of an individual request. In this situation, you want WLM to dynamically balance individual requests. This configuration allows linear scalability and consistent, repeatable response times.

Set this property to 0 if you are using conversational application models. With these models, a client might hold, and periodically interact with, a reference to a stateful object which is pinned in the memory of one of the JVMs for time that is greater than the duration of an individual request. For example, the client might be using HTTP sessions, stateful session beans, or entity beans that are maintained in memory instead of being stored in a database or file system between requests, as is done in stateless application models.

Conversational application models prevent WLM from dynamically routing individual requests in a clustered environment because the client has affinity to a specific JVM. In this situation, a round robin algorithm is used to handle the initial request from the client. This algorithm evenly distributes the creation of long-term affinities and is the best technique for achieving a balanced utilization of system resources in this type of environment. If we set this property to 0 for conversational application models, you must also set the server_work_distribution_algorithm property to 1.

If we prefer, we can use the round robin capability that WLM provides, instead of the previously described product round robin capability. The differences between the round robin capability that WLM provides and the product round robin capability is explained in the following scenario.

Scenario: A customer starts two clients to talk to a server. The server has two servants, and each servant has multiple threads. The customer expects one client to go to one servant, and the second client to go to the other servant. The behavior of product-provided round robin, that is initiated by specifying server_use_wlm_to_queue_work=0, and server_work_distribution_algorithm=1, adheres to this expectation. However, the WLM-provided round robin uses up all the threads in the first servant before it starts to use the threads in the next servant. Therefore, in this situation, both clients go to the same servant and the second servant remains idle.

When the server_use_wlm_to_queue_work property is set to 0, the wlm_minimumSRCount and wlm_maximumSRCount properties are set to the same value. Because the work is not going through WLM, WLM only starts the number of servants specified for the wlm_minimumSRCount property.

Information Value
0 The z/OS WLM function is not used.
1 The z/OS WLM function is used.
Default 1
Used by Daemon No


server_work_distribution_algorithm

Type of work distribution algorithm that the application server uses for workload balancing. This property is only used if the server_use_wlm_to_queue_work property is set to 0. If the server_use_wlm_to_queue_work property is set to 1, then the value specified for this property is ignored.

Information Value
0 The hot thread algorithm is used.

When the hot thread algorithm is used, each new work request is assigned to the first servant that has a thread available to process the request. If none of the servants have an available thread, the request is queued into the global work queue that is shared by all servants. The request is then selected from the global work queue when the next thread becomes available, regardless of which servant owns that thread.

The goal of the hot thread algorithm is to route requests to the fewest number of servants possible. If the number of concurrent requests is not greater than the number of threads in a servant, all requests are processed by the same servant.

1 The round robin algorithm is used.

When the round robin algorithm is used, new work requests are distributed evenly across all servants. If all of the servant threads are already processing other work requests, the new request is added to the request queue for a specific servant. The queued request is then selected when it becomes the top request in the queue and a thread becomes available in that servant.

2 The hot robin algorithm is used.

When the hot robin algorithm is used, new work requests are distributed evenly across all servants. If the assigned servant does not have a thread available to process the request, the request is reassigned to another servant that has an available thread. If none of the servants have an available thread, the new request is queued into the global work queue that all of the servants share. The request is then selected from the global work queue when the next thread becomes available, regardless of which servant owns that thread.

Default 0
Used by Daemon No


server_wto_on_write_error

Indicates whether the error message BBOO0384I ERROR OCCURRED WRITING TO {0} is written to the SYSLOG when an error occurs while writing to SYSPRINT or SYSOUT.

This property can be set to 0 or 1. When this property is set to 1, this error message is written to the SYSLOG. When this property is set to 0 no message is issued.

Information Value
Data Type Integer
Default 0
Used by Daemon No


suppress_hung_thread_abend

Controls whether an ABENDDC3 or an ABENDSDC3 code is issued when a hung thread is found during the server stop process. This property disables the detection of a hung thread.

When a STOP command is issued for a server, a hung thread might fail to terminate. By default, the application server issues an ABENDDC3 or ABENDSDC3 code with a 000C000B reason code when this condition is encountered to avoid any delays and to provide a diagnostic dump earlier in the shutdown process. If we set this property to 1, the application server does not abend the address space for this reason.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


suppress_hung_thread_dump

Controls whether a dump is taken when an ABENDDC3 or an ABENDSDC3 occurs because a hung thread did not terminate in response to a STOP command.

When a STOP command is issued for a server, a hung thread might fail to terminate, which causes an ABENDDC3 or an ABENDSDC3 to occur with reason code 000C000B. Setting this property to 1 prevents a dump from being taken if this abend occurs.

Information Value
Data Type Boolean
Default 0
Used by Daemon No


transaction_recoveryTimeout

Time, in minutes, that this controller uses to attempt to complete all restarted transactions before issuing a write-to-operator-with-reply (WTOR) message to the console, requesting whether to complete one of the following options:

If the operator replies to continue the recovery process, then the controller attempts recovery for the specified amount of time before reissuing the write-to-operator message. After all the transactions are resolved, the controller terminates. This variable applies only to controllers that are running in peer restart and recovery mode.

Information Value
Data Type Integer
Default 15
Used by Daemon No


WAS_DAEMON_ONLY_adapter_max_conn

Maximum number of connections supported for a daemon group. A connection is defined as a local communication connection between a client address space and a server within the daemon group.

As client address spaces register with the optimized local adapter APIs, the number of physical connections that the client address spaces want to make with the WebSphere address space is defined. To prevent the connectors from over-running the WAS with requests, the number of connections across the entire daemon group can be set to a maximum number by specifying the adapter_max_conn variable.

If a subsystem connects to the WebSphere server and requests a number of connections that is the maximum for the daemon group, the register request fails with a reason code that indicates that the maximum number of connections has been reached.

The maximum number of physical connections between the WAS and external address spaces is set on the connection factory.

Information Value
Data Type Integer
Default
Used by Daemon Yes


WAS_DAEMON_ONLY_adapter_max_serv

Specifies support for an optimized local adapter. Support is enabled when this variable is used to start the daemon.

The WebSphere daemon needs to know whether this daemon group is going to support the optimized local adapters, since it manages the shared storage used by the WebSphere server and the associated address spaces. All daemon configuration exists in the was.env file for the daemon. The daemon does not have a JVM, therefore it does not have access to the WCCM model.

This variable sets the maximum number of adapter outbound services that can be active for a single registration. When the value is not specified, it defaults to 100.

Information Value
Data Type Integer
Default 100
Used by Daemon Yes


WAS_DAEMON_ONLY_adapter_max_shremem

Specifies support for an optimized local adapter. Support is enabled when this variable is used to start the daemon.

The WebSphere daemon needs to know whether this daemon group is going to support the optimized local adapters, since it manages the shared storage used by the WebSphere server and the associated address spaces. All daemon configuration exists in the was.env file for the daemon. The daemon does not have a JVM, therefore it does not have access to the WCCM model.

This variable sets the maximum size of the adapters shared 64-bit memory for adapters control structures for the daemon group. When this value is not specified, it defaults to 32M (33554432).

Information Value
Data Type Integer
Default 32M
Used by Daemon Yes


WAS_DAEMON_ONLY_enable_adapter

Specifies support for an optimized local adapter. Support is enabled when this variable is used to start the daemon.

The WebSphere daemon needs to know whether or not this daemon group is going to support the optimized local adapters, since it manages the shared storage used by the WebSphere server and the associated address spaces. All daemon configuration exists in the was.env file for the daemon. The daemon does not have a JVM, therefore it does not have access to the WCCM model.

Set this property to true to start the daemon with this variable and enable support for the optimized local adapters. Set this property to false if we do not want to start the daemon with this variable and enable support for the optimized local adapters.

Information Value
Data Type Boolean
Default false
Used by Daemon Yes


wlm_ae_spreadmin

Setting of the AE_SPREADMIN parameter on the IWM4SLIP service, which you use for servant address spaces. The IWM4SLIP service is the application environment limit service for workload management (WLM) on the z/OS system. The AE_SPREADMIN parameter indicates how WLM distributes the minimum number of servants across the service classes being used to process work requests.

Read the topic on controlling the minimum and maximum number of servants for more information about setting the Minimum Number of Instances field.

Information Value
Data Type Boolean
Default 1
Used by Daemon No


wlm_classification_file

Location of the workload classification document. We can set this property at the cell, node, or server level.

The workload classification document is a common XML file that classifies inbound HTTP, IIOP, and message-driven bean (MDB) work. Any rules defined in this XML file override the old format HTTP classification. The rules in the common XML file also override any rules that are in the MDB classification file defined by the endpoint_config_file property.

For example, the configuration has the common XML file defined in a server that also has the old format HTTP classification document specified. There are no HTTP classification rules defined in our common XML file, so the old format file is used to classify inbound HTTP requests. However, if the common XML file contains HTTP rules, these classification rules are used instead of classification rules definedd in the old style HTTP classification document.


wlm_enclavecreate_exstartdefer

Value you want used for the Workload Management Service EXSTARTDEFER parameter when you use the IWMECREA Workload Management Service to create an Enclave.

The supported values are YES and NO.

See the z/OS Information Center for the version of z/OS running on the system for more information about the IWMECREA Workload Management Service and the EXSTARTDEFER parameter.

Information Value
Data Type String
Default YES
Used by Daemon No


wlm_health_increment

Increment amount provided to the z/OS Workload Managers Health API (IWM4HLTH) at the interval specified for the wlm_health_interval custom property.

If the value specified for the wlm_health_interval custom property is 0, a value of 100 is used for the wlm_health_increment custom property.

Information Value
Data Type Integer
Valid Values 1 - 100
Default 10
Used by Daemon No


wlm_health_interval

Amount of time, in seconds, between calls to the z/OS Workload Managers Health API (IWM4HLTH).

The health value provided increments for each call to the WLM health API by the value specified for the wlm_health_increment custom property. If the value for the wlm_health_interval custom property is 0, an increment value of 100 is used when updating the servers health value. The initial health value of the server is set to 0 during initialization. The servers health value is updated at the intervals defined by the value specified for the wlm_health_interval custom property. The first interval starts when the deferred listeners for the server are started. See the description of the protocol_accept_http_work_after_min_srs custom property for additional information.

Information Value
Data Type Integer
Units Seconds
Default 0
Used by Daemon No


wlm_servant_start_parallel

How a server, configured to start more than one servant address space, starts these address spaces.

Avoid trouble: This property requires z/WLM support. Therefore, before specifying this property, verify that you are running the product on z/OS Version 1, Release 9 or higher. If we are running the product on a lower-level z/OS, specifying this property has no effect.gotcha

Information Value
Data Type Boolean
Default 0
Used by Daemon No


wlm_stateful_session_placement_on

Whether round robin queuing of HTTP sessions is enabled among servants. We can set this property at the cell, node, or server level.

Set this property to true if you want round robin queuing of HTTP sessions enabled among servants. Set this property to false if we do not want round robin queuing of HTTP sessions enabled among servants.

See the use of the custom property, WLMStatefulSession, for more information on the effect of using wlm_stateful_session_placement_on.

Information Value
Data Type Boolean
Default false
Used by Daemon No


wsadmin_dumpthreads_enable_heapdump

Whether a Java heap dump is generated, in addition to a Java core dump when you issue the wsadmin dumpThreads command.

Set this property to 1 if you want a Java heap dump generated when you issue the wsadmin dumpThreads command. Set this property to 0 if we do not want a Java heap dump generated when you issue the wsadmin dumpThreads command.

Information Value
Data Type Boolean
Default 1
Used by Daemon No


wsadmin_dumpthreads_enable_javatdump

Whether a system TDUMP is generated, in addition to a Java core dump when you issue the wsadmin dumpThreads command.

Set this property to 1 if you want a system TDUMP generated when you issue the wsadmin dumpThreads command. Set this property to 0 if we do not want a system TDUMP generated when you issue the wsadmin dumpThreads command.

Information Value
Data Type Boolean
Default 1
Used by Daemon No


Subtopics


Related concepts

  • Workload management (WLM) for z/OS


    Related tasks

  • Use the console to enable properties for specific SMF record types
  • Classifying z/OS workload
  • Controlling the minimum and maximum number of servants
  • Configure an application server to use the WLM even distribution of HTTP requests function

  • Repository service custom properties
  • HTTP transport channel custom properties
  • Workload classification file