Puppet Class: pulsevtm::global_settings
- Defined in:
- manifests/global_settings.pp
Overview
class: pulsevtm::global_settings
Global Settings General settings that apply to every machine in the cluster.
Parameters
- basic__accepting_delay
-
How often, in milliseconds, each traffic manager child process (that isn’t listening for new connections) checks to see whether it should start listening for new connections.
- basic__afm_cluster_timeout
-
How long in seconds to wait for the application firewall control script to complete clustering operations for the application firewall.
- basic__afm_control_timeout
-
How long in seconds to wait for the application firewall control script to complete operations such as starting and stopping the application firewall.
- basic__afm_enabled
-
Is the application firewall enabled.
- basic__child_control_command_timeout
-
Timeout for waiting for child processes to respond to parent control requests If a child process (zeus.zxtm, zeus.eventd, zeus.autoscaler, etc) takes longer than this number of seconds to respond to a parent control command, error messages will be logged for every multiple of this number of seconds, for example, if set to 10 seconds, a log message will be logged every 10 seconds, until the child responds or is automatically killed (see the child_control_kill_timeout setting).
- basic__child_control_kill_timeout
-
Timeout for waiting for child processes to respond to parent control requests If a child process (zeus.zxtm, zeus.eventd, zeus.autoscaler, etc) takes longer than this number of seconds to respond to a parent control command, then the parent zeus.zxtm process will assume this process is stuck in an unresponsive loop and will kill it, log the termination event, and wait for a new process of the same type to restart. Set this to 0 to disable killing unresponsive child processes.
- basic__chunk_size
-
The default chunk size for reading/writing requests.
- basic__client_first_opt
-
Whether or not your traffic manager should make use of TCP optimisations to defer the processing of new client-first connections until the client has sent some data.
- basic__cluster_identifier
-
Cluster identifier. Generally supplied by Services Director.
- basic__cpu_starvation_check_interval
-
How frequently should child processes check for CPU starvation? A value of 0 disables the detection.
- basic__cpu_starvation_check_tolerance
-
How much delay in milliseconds between starvation checks do we allow before we assume that the machine or its HyperVisor are overloaded.
- basic__http2_no_cipher_blacklist_check
-
Disable the cipher blacklist check in HTTP2 (mainly intended for testing purposes)
- basic__internal_config_logging
-
Whether or not messages pertaining to internal configuration files should be logged to the event log.
- basic__license_servers
-
A list of license servers for FLA licensing. A license server should be specified as a “<ip/host>:<port>” pair. Type:array Properties:
- basic__max_fds
-
The maximum number of file descriptors that your traffic manager will allocate.
- basic__monitor_memory_size
-
The maximum number of each of nodes, pools or locations that can be monitored. The memory used to store information about nodes, pools and locations is allocated at start-up, so the traffic manager must be restarted after changing this setting.
- basic__rate_class_limit
-
The maximum number of Rate classes that can be created. Approximately 100 bytes will be pre-allocated per Rate class.
- basic__shared_pool_size
-
The size of the shared memory pool used for shared storage across worker processes (e.g. bandwidth shared data).This is specified as either a percentage of system RAM, “5%” for example, or an absolute size such as “10MB”.
- basic__slm_class_limit
-
The maximum number of SLM classes that can be created. Approximately 100 bytes will be pre-allocated per SLM class.
- basic__so_rbuff_size
-
The size of the operating system’s read buffer. A value of “0” (zero) means to use the OS default; in normal circumstances this is what should be used.
- basic__so_wbuff_size
-
The size of the operating system’s write buffer. A value of “0” (zero) means to use the OS default; in normal circumstances this is what should be used.
- basic__socket_optimizations
-
Whether or not the traffic manager should use potential network socket optimisations. If set to “auto”, a decision will be made based on the host platform.
- basic__storage_shared
-
Whether the storage for the traffic managers’ configuration is shared between cluster members.
- basic__tip_class_limit
-
The maximum number of Traffic IP Groups that can be created.
- admin__honor_fallback_scsv
-
Whether or not the admin server, the internal control port and the config daemon honor the Fallback SCSV to protect connections against downgrade attacks.
- admin__ssl3_allow_rehandshake
-
Whether or not SSL3/TLS re-handshakes should be supported for admin server and internal connections.
- admin__ssl3_ciphers
-
The SSL ciphers to use for admin server and internal connections. For information on supported ciphers see the online help.
- admin__ssl3_diffie_hellman_key_length
-
The length in bits of the Diffie-Hellman key for ciphers that use Diffie-Hellman key agreement for admin server and internal connections.
- admin__ssl3_min_rehandshake_interval
-
If SSL3/TLS re-handshakes are supported on the admin server, this defines the minimum time interval (in milliseconds) between handshakes on a single SSL3/TLS connection that is permitted. To disable the minimum interval for handshakes the key should be set to the value “0”.
- admin__ssl_elliptic_curves
-
The SSL elliptic curve preference list for admin and internal connections. The named curves P256, P384 and P521 may be configured. Type:array Properties:
- admin__ssl_insert_extra_fragment
-
Whether or not SSL3 and TLS1 use one-byte fragments as a BEAST countermeasure for admin server and internal connections.
- admin__ssl_max_handshake_message_size
-
The maximum size (in bytes) of SSL handshake messages that the admin server and internal connections will accept. To accept any size of handshake message the key should be set to the value “0”.
- admin__ssl_prevent_timing_side_channels
-
Take performance degrading steps to prevent exposing timing side-channels with SSL3 and TLS used by the admin server and internal connections.
- admin__ssl_signature_algorithms
-
The SSL signature algorithms preference list for admin and internal connections using TLS version 1.2 or higher. For information on supported algorithms see the online help.
- admin__support_ssl3
-
Whether or not SSL3 support is enabled for admin server and internal connections.
- admin__support_tls1
-
Whether or not TLS1.0 support is enabled for admin server and internal connections.
- admin__support_tls1_1
-
Whether or not TLS1.1 support is enabled for admin server and internal connections.
- admin__support_tls1_2
-
Whether or not TLS1.2 support is enabled for admin server and internal connections.
- appliance__bootloader_password
-
The password used to protect the bootloader. An empty string means there will be no protection.
- appliance__return_path_routing_enabled
-
Whether or not the traffic manager will attempt to route response packets back to clients via the same route on which the corresponding request arrived. Note that this applies only to the last hop of the route - the behaviour of upstream routers cannot be altered by the traffic manager.
- aptimizer__cache_entry_lifetime
-
The period of time (in seconds) that unaccessed cache entries will be retained by Web Accelerator.
- aptimizer__cache_entry_limit
-
The maximum number of cache entries that will be retained by Web Accelerator before removing old entries to make room for new ones.
- aptimizer__default_profile
-
The Profile to use by default if no mappings are configured (or if Web Accelerator is licensed in Express mode)
- aptimizer__default_scope
-
The Scope to use by default if no mappings are configured (or if Web Accelerator is licensed in Express mode)
- aptimizer__dependent_fetch_timeout
-
How long to wait for dependent resource fetches (default 30 seconds).
- aptimizer__enable_state_dump
-
Whether or not the Web Accelerator state will be dumped if “/aptimizer-state-dump” is appended to an optimized URL.
- aptimizer__ipc_timeout
-
The time after which connections between the traffic manager and Web Accelerator processes will be closed, should an optimization job take considerably longer than expected.
- aptimizer__max_concurrent_jobs
-
How many direct jobs can be in progress before optimization jobs start getting rejected by Web Accelerator.
- aptimizer__max_dependent_fetch_size
-
The maximum size of a dependent resource that can undergo Web Accelerator optimization. Any content larger than this size will not be optimized. Units of KB and MB can be used, no postfix denotes bytes. A value of 0 disables the limit.
- aptimizer__max_original_content_buffer_size
-
The maximum size of unoptimized content buffered in the traffic manager for a single backend response that is undergoing Web Accelerator optimization. Responses larger than this will not be optimized. Note that if the backend response is compressed then this setting pertains to the compressed size, before Web Accelerator decompresses it. Units of KB and MB can be used, no postfix denotes bytes. Value range is 1 - 128MB.
- aptimizer__queue_buffer_size
-
The size in bytes of the operating system buffer which is used to send request URLs and data to Web Accelerator and return optimized resources from Web Accelerator. A larger buffer will allow a greater number of simultaneous resources to be optimized, particularly if a large number of requests are made at the same time, for example an HTML page containing hundreds of images to optimize. If this is set to zero, the default operating system buffer size will be used.
- aptimizer__resource_lifetime
-
The period of time (in seconds) that resource data is retained by Web Accelerator after it is no longer actively in use.
- aptimizer__resource_memory_limit
-
The maximum amount of memory the cache is allowed to have pinned. Once it goes over that limit, it starts releasing resource data in LRU order.
- aptimizer__watchdog_interval
-
The period of time (in seconds) after which a previous failure will no longer count towards the watchdog limit.
- aptimizer__watchdog_limit
-
The maximum number of times the Web Accelerator sub-process will be started or restarted within the interval defined by the aptimizer!watchdog_interval setting. If the process fails this many times, it must be restarted manually from the Diagnose page. Zero means no limit.
- auditlog__via_eventd
-
Whether to mirror the audit log to EventD.
- auditlog__via_syslog
-
Whether to output audit log message to the syslog.
- auth__saml_key_lifetime
-
Lifetime in seconds of cryptographic keys used to decrypt SAML SP sessions stored externally (client-side).
- auth__saml_key_rotation_interval
-
Rotation interval in seconds for cryptographic keys used to encrypt SAML SP sessions stored externally (client-side).
- autoscaler__slm_interval
-
The interval at which the parent sends new SLM status to the autoscaler.
- autoscaler__verbose
-
Whether or not detailed messages about the autoscaler’s activity are written to the error log.
- bandwidth__license_sharing
-
For the global BW limits, how the bandwidth allocation should be shared between consumers. In ‘pooled’ mode, the allocation is shared between all consumers, who can write as much data as they want until the pool of data is exhausted. In ‘quota’ mode, bandwidth is divided between consumers, who can write only as much as they are allocated. Any unused bandwidth will be lost.
- bandwidth__pooled_min_write
-
For the global BW limits using ‘pooled’ bandwidth allocation sharing between consumers, when the license limit is reached the allowance will be evenly distributed between the remaining consumers. Each consumer will, however be permitted to write at least this much data.
- bgp__as_number
-
The number of the BGP AS in which the traffic manager will operate. Must be entered in decimal.
- bgp__enabled
-
Whether BGP Route Health Injection is enabled
- cluster_comms__allow_update_default
-
The default value of “allow_update” for new cluster members. If you have cluster members joining from less trusted locations (such as cloud instances) this can be set to “false” in order to make them effectively “read-only” cluster members.
- cluster_comms__allowed_update_hosts
-
The hosts that can contact the internal administration port on each traffic manager. This should be a list containing IP addresses, CIDR IP subnets, and “localhost”; or it can be set to “all” to allow any host to connect. Type:array Properties:
- cluster_comms__state_sync_interval
-
How often to propagate the session persistence and bandwidth information to other traffic managers in the same cluster. Set this to “0” (zero) to disable propagation.<br /> Note that a cluster using “unicast” heartbeat messages cannot turn off these messages.
- cluster_comms__state_sync_timeout
-
The maximum amount of time to wait when propagating session persistence and bandwidth information to other traffic managers in the same cluster. Once this timeout is hit the transfer is aborted and a new connection created.
- connection__idle_connections_max
-
The maximum number of unused HTTP keepalive connections with back-end nodes that the traffic manager should maintain for re-use. Setting this to “0” (zero) will cause the traffic manager to auto-size this parameter based on the available number of file-descriptors.
- connection__idle_timeout
-
How long an unused HTTP keepalive connection should be kept before it is discarded.
- connection__listen_queue_size
-
The listen queue size for managing incoming connections. It may be necessary to increase the system’s listen queue size if this value is altered. If the value is set to “0” then the default system setting will be used.
- connection__max_accepting
-
Number of processes that should accept new connections. Only this many traffic manager child processes will listen for new connections at any one time. Setting this to “0” (zero) will cause your traffic manager to select an appropriate default value based on the architecture and number of CPUs.
- connection__multiple_accept
-
Whether or not the traffic manager should try to read multiple new connections each time a new client connects. This can improve performance under some very specific conditions. However, in general it is recommended that this be set to ‘false’.
- dns__checktime
-
How often to check the DNS configuration for changes.
- dns__hosts
-
The location of the “hosts” file.
- dns__hostsfirst
-
Whether or not to try reading the “dns!hosts” file before calling gethostbyname(). This config key exists for testing purposes only.
- dns__max_ttl
-
Maximum Time To Live (expiry time) for entries in the DNS cache.
- dns__maxasynctries
-
How often to send DNS request packets before giving up.
- dns__min_ttl
-
Minimum Time To Live (expiry time) for entries in the DNS cache.
- dns__negative_expiry
-
Expiry time for failed lookups in the DNS cache.
- dns__resolv
-
The location of the “resolv.conf” file.
- dns__size
-
Maximum number of entries in the DNS cache.
- dns__timeout
-
Timeout for receiving a response from a DNS server.
- dns_autoscale__resolver
-
The IP address and port number of the DNS server to use for DNS-derived autoscaling, in the form addr:port. This is intended for test and debug purposes, and will override the configuration of the system resolver, which is usually defined in /etc/resolv.conf
- ec2__access_key_id
-
Amazon EC2 Access Key ID.
- ec2__action_timeout
-
How long, in seconds, the traffic manager should wait while associating or disassociating an Elastic IP to the instance.
- ec2__awstool_timeout
-
The maximum amount of time requests to the AWS Query API can take before timing out.
- ec2__metadata_server
-
URL for the EC2 metadata server, “169.254.169.254/latest/meta-data” for example.
- ec2__metadata_timeout
-
The maximum amount of time requests to the EC2 Metadata Server can take before timing out.
- ec2__query_server
-
URL for the Amazon EC2 endpoint, “ec2.amazonaws.com/” for example.
- ec2__secret_access_key
-
Amazon EC2 Secret Access Key.
- ec2__verify_query_server_cert
-
Whether to verify Amazon EC2 endpoint’s certificate using CA(s) present in SSL Certificate Authorities Catalog.
- ec2__vpc_decluster_on_stop
-
Whether to decluster the traffic manager running inside vpc when the instance stops.
- eventing__mail_interval
-
The minimum length of time that must elapse between alert emails being sent. Where multiple alerts occur inside this timeframe, they will be retained and sent within a single email rather than separately.
- eventing__max_attempts
-
The number of times to attempt to send an alert email before giving up.
- fault_tolerance__arp_count
-
The number of ARP packets a traffic manager should send when an IP address is raised.
- fault_tolerance__auto_failback
-
Whether or not traffic IPs automatically move back to machines that have recovered from a failure and have dropped their traffic IPs.
- fault_tolerance__autofailback_delay
-
Configure the delay of automatic failback after a previous failover event. This setting has no effect if autofailback is disabled.
- fault_tolerance__child_timeout
-
How long the traffic manager should wait for status updates from any of the traffic manager’s child processes before assuming one of them is no longer servicing traffic.
- fault_tolerance__frontend_check_ips
-
The IP addresses used to check front-end connectivity. The text “%gateway%” will be replaced with the default gateway on each system. Set this to an empty string if the traffic manager is on an Intranet with no external connectivity. Type:array Properties:
- fault_tolerance__heartbeat_method
-
The method traffic managers should use to exchange cluster heartbeat messages.
- fault_tolerance__igmp_interval
-
The interval between unsolicited periodic IGMP Membership Report messages for Multi-Hosted Traffic IP Groups.
- fault_tolerance__monitor_interval
-
The frequency, in milliseconds, that each traffic manager machine should check and announce its connectivity.
- fault_tolerance__monitor_timeout
-
How long, in seconds, each traffic manager should wait for a response from its connectivity tests or from other traffic manager machines before registering a failure.
- fault_tolerance__multicast_address
-
The multicast address and port to use to exchange cluster heartbeat messages.
- fault_tolerance__routing_sw_run_ribd
-
Whether the ribd routing daemon is to be run. The routing software needs to be restarted for this change to take effect.
- fault_tolerance__routing_sw_watchdog_interval
-
The period of time in seconds after which a failure will no longer count towards the watchdog limit.
- fault_tolerance__routing_sw_watchdog_limit
-
The maximum number of times the routing software suite of processes will be started or restarted within the interval defined by the flipper!routing_sw_watchdog_interval setting. If the routing software fails this many times within the interval, it will be stopped and can only be restarted manually from the Diagnose page or by switching OSPF off and on again. Zero means no limit.
- fault_tolerance__tipv6_raise_deprecated
-
Mark Traffic IPv6 addresses as “deprecated” to prevent their use during IPv6 source selection.
- fault_tolerance__unicast_port
-
The unicast UDP port to use to exchange cluster heartbeat messages.
- fault_tolerance__use_bind_ip
-
Whether or not cluster heartbeat messages should only be sent and received over the management network.
- fault_tolerance__verbose
-
Whether or not a traffic manager should log all connectivity tests. This is very verbose, and should only be used for diagnostic purposes.
- fips__enabled
-
Enable FIPS Mode (requires software restart).
- ftp__data_bind_low
-
Whether or not the traffic manager should permit use of FTP data connection source ports lower than 1024. If “No” the traffic manager can completely drop root privileges, if “Yes” some or all privileges may be retained in order to bind to low ports.
- glb__verbose
-
Write a message to the logs for every DNS query that is load balanced, showing the source IP address and the chosen datacenter.
- historical_activity__keep_days
-
Number of days to store historical traffic information, if set to “0” the data will be kept indefinitely.
- http__max_chunk_header_length
-
The maximum length the header line of an HTTP chunk can have in an upload from the client. Header lines exceeding this length will be considered invalid. The traffic manager buffers the header line before it can read any payload data in the chunk; the limit exists to protect against malicious clients that send very long lines but never any payload data.
- ip__appliance_returnpath
-
A table of MAC to IP address mappings for each router where return path routing is required. Type:array Properties:MAC address of a router the software is connected to.”, “type”=>“string”, “ipv4”=>MAC address to IPv4 address mapping of a router the software is connected to. The "*" (asterisk) in the key name is the MAC address, the value is the IP address.”, “type”=>“string”, “default”=>“”, “ipv6”=>MAC address to IPv6 address mapping of a router the software is connected to. The "*" (asterisk) in the key name is the MAC address, the value is the IP address.”, “type”=>“string”, “default”=>“”}
- java__classpath
-
CLASSPATH to use when starting the Java runner.
- java__command
-
Java command to use when starting the Java runner, including any additional options.
- java__enabled
-
Whether or not Java support should be enabled. If this is set to “No”, then your traffic manager will not start any Java processes. Java support is only required if you are using the TrafficScript “java.run()” function.
- java__lib
-
Java library directory for additional jar files. The Java runner will load classes from any “.jar” files stored in this directory, as well as the * jar files and classes stored in traffic manager’s catalog.
- java__max_connections
-
Maximum number of simultaneous Java requests. If there are more than this many requests, then further requests will be queued until the earlier requests are completed. This setting is per-CPU, so if your traffic manager is running on a machine with 4 CPU cores, then each core can make this many requests at one time.