conf/pools
The conf/pools directory contains configuration files for backend node pools. The name of a file is the name of the pool it defines. Pools can be configured under the Services > Pools section of the Admin Server UI or by using functions under the Pool section of the SOAP API and CLI.
Key | Description |
|---|---|
autoscale!addnode_delaytime | The time in seconds from the creation of the node which the traffic manager should wait before adding the node to the autoscaled pool. Set this to allow applications on the newly created node time to intialize before being sent traffic. autoscale!enabled' is set to "yes"' Value type: seconds Default value: "0" |
autoscale!cloudcredentials | The Cloud Credentials object containing authentication credentials to use in cloud API calls. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!cluster | The ESX host or ESX cluster name to put the new virtual machine instances on. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!datacenter | The name of the logical datacenter on the vCenter server. Virtual machines will be scaled up and down under the datacenter root folder. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!datastore | The name of the datastore to be used by the newly created virtual machine. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!enabled | Are the nodes of this pool subject to autoscaling? If yes, nodes will be automatically added and removed from the pool by the chosen autoscaling mechanism. Value type: Yes / No Default value: "No" |
autoscale!external | Whether or not autoscaling is being handled by an external system. Set this value to Yes if all aspects of autoscaling are handled by an external system, such as RightScale. If set to No, the traffic manager will determine when to scale the pool and will communicate with the cloud provider to create and destroy nodes as necessary. autoscale!enabled' is set to "yes"' Value type: Yes / No Default value: "Yes" |
autoscale!extraargs | Any extra arguments to the autoscaling API. Each argument can be separated by comma. E.g in case of EC2, it can take extra parameters to the Amazon's RunInstance API say DisableApiTermination=false,Placement.Tenancy=default. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!hysteresis | The time period in seconds for which a change condition must persist before the change is actually instigated. Value type: unsigned integer Default value: "20" |
autoscale!imageid | The identifier for the image of the instances to create. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!ipstouse | Which type of IP addresses on the node to use. Choose private IPs if the traffic manager is in the same cloud as the nodes, otherwise choose public IPs. autoscale!enabled' is set to "yes"' Value type: enumeration Default value: "publicips" Permitted values: publicips: Public IP addresses privateips: Private IP addresses |
autoscale!lastnode_idletime | The time in seconds for which the last node in an autoscaled pool must have been idle before it is destroyed. This is only relevant if min_nodes is 0. Value type: unsigned integer Default value: "3600" |
autoscale!max_nodes | The maximum number of nodes in this autoscaled pool. autoscale!enabled' is set to "yes"' Value type: unsigned integer Default value: "4" |
autoscale!min_nodes | The minimum number of nodes in this autoscaled pool. autoscale!enabled' is set to "yes"' Value type: unsigned integer Default value: "1" |
autoscale!name | The beginning of the name of nodes in the cloud that are part of this autoscaled pool. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!port | The port number to use for each node in this autoscaled pool. autoscale!enabled' is set to "yes"' Value type: unsigned integer Default value: "80" |
autoscale!refractory | The time period in seconds after the instigation of a re-size during which no further changes will be made to the pool size. autoscale!enabled' is set to "yes"' Value type: unsigned integer Default value: "180" |
autoscale!response_time | The expected response time of the nodes in ms. This time is used as a reference when deciding whether a node's response time is conforming. All responses from all the nodes will be compared to this reference and the percentage of conforming responses is the base for decisions about scaling the pool up or down. autoscale!enabled' is set to "yes"' Value type: unsigned integer Default value: "1000" |
autoscale!scaledown_level | The fraction, in percent, of conforming requests above which the pool size is decreased. If the percentage of conforming requests exceeds this value, the pool is scaled down. autoscale!enabled' is set to "yes"' Value type: unsigned integer Default value: "95" |
autoscale!scaleup_level | The fraction, in percent, of conforming requests below which the pool size is increased. If the percentage of conforming requests drops below this value, the pool is scaled up. autoscale!enabled' is set to "yes"' Value type: unsigned integer Default value: "40" |
autoscale!securitygroupids | List of security group IDs to associate to the new EC2 instance. autoscale!enabled' is set to "yes"' Value type: list Default value: <none> |
autoscale!sizeid | The identifier for the size of the instances to create. autoscale!enabled' is set to "yes"' Value type: string Default value: <none> |
autoscale!subnetids | List of subnet IDs where the new EC2-VPC instance(s) will be launched. Instances will be evenly distributed among the subnets. If the list is empty, instances will be launched inside EC2-Classic. autoscale!enabled' is set to "yes"' Value type: list Default value: <none> |
bandwidth_class | The Bandwidth Management Class this pool uses, if any. Value type: string Default value: <none> |
disabled | A list of nodes in the pool that are in the 'disabled' state. Value type: list Default value: <none> |
dns_autoscale!enabled | When enabled, the Traffic Manager will periodically resolve the hostnames in the "hostnames" list using a DNS query, and use the results to automatically add, remove or update the IP addresses of the nodes in the pool. Value type: Yes / No Default value: "No" |
dns_autoscale!hostnames | A list of hostnames which will be used for DNS-derived autoscaling dns_autoscale!enabled' is set to "yes"' Value type: list Default value: <none> |
dns_autoscale!port | The port number to use for each node when using DNS-derived autoscaling Value type: unsigned integer Default value: "80" |
draining | A list of nodes in the pool that are in the 'draining' state. Value type: list Default value: <none> |
failpool | If all of the nodes in this pool have failed, then requests can be diverted to another pool. Value type: string Default value: <none> |
ftp_support_rfc_2428 | Whether or not the backend IPv4 nodes understand the EPRT and EPSV command from RFC 2428. It is always assumed that IPv6 nodes support these commands. Value type: Yes / No Default value: "No" |
keepalive | Whether or not the pool should maintain HTTP keepalive connections to the nodes. Value type: Yes / No Default value: "Yes" |
keepalive!non_idempotent | Whether or not the pool should maintain HTTP keepalive connections to the nodes for non-idempotent requests. Value type: Yes / No Default value: "No" |
kerberos_protocol_transition!principal | The Kerberos principal the traffic manager should use when performing Kerberos Protocol Transition. Value type: string Default value: <none> |
kerberos_protocol_transition!target | The Kerberos principal name of the service this pool targets. Value type: string Default value: <none> |
load_balancing!algorithm | The load balancing algorithm that this pool uses. Value type: enumeration Default value: "roundrobin" Permitted values: roundrobin: Round Robin wroundrobin: Weighted Round Robin cells: Perceptive connections: Least Connections wconnections: Weighted Least Connections responsetimes: Fastest Response Time random: Random Node |
load_balancing!weighting!* | Weights for each node in the pool. The actual values in isolation do not matter, as long as they are valid integers, the per-node weightings are calculated on their relative values between the nodes. The key should be specified once-per-node with the node identifier (<ip>:<port>) replacing the * in each instance. Value type: int Default value: <none> |
max_connect_time | How long the pool should wait for a connection to a node to be established before giving up and trying another node. Value type: seconds Default value: "4" |
max_connection_attempts | The maximum number of nodes to which the traffic manager will attempt to send a request before returning an error to the client. Requests that are non-retryable will be attempted against only one node. Zero signifies no limit. Value type: unsigned integer Default value: "0" |
max_connections_pernode | The maximum number of concurrent connections allowed to each back-end node in this pool per machine. A value of 0 means unlimited connections. Value type: unsigned integer Default value: "0" |
max_idle_connections_pernode | The maximum number of unused HTTP keepalive connections that should be maintained to an individual node. Zero signifies no limit. Value type: unsigned integer Default value: "50" |
max_queue_size | The maximum number of connections that can be queued due to connections limits. A value of 0 means unlimited queue size. Value type: unsigned integer Default value: "0" |
max_reply_time | How long the pool should wait for a response from the node before either discarding the request or trying another node (retryable requests only). Value type: seconds Default value: "30" |
max_timed_out_connection_attempts | The maximum number of connection attempts the traffic manager will make where the server fails to respond within the time limit defined by the max_reply_time setting. Zero signifies no limit. Value type: unsigned integer Default value: "2" |
max_transactions_per_node | The maximum number of concurrent transactions allowed to each back-end node in this pool per machine. A value of 0 means unlimited transactions. Idle connections kept alive for reuse do not count against this limit. A transaction begins by allocating a connection for sending the request, and ends (for the purposes of queuing) after a complete response has been received from the node. Value type: unsigned integer Default value: "0" |
monitors | A list of monitors assigned to this pool. Value type: list Default value: <none> |
node_close_with_rst | Whether or not connections to the back-end nodes should be closed with a RST packet, rather than a FIN packet. This avoids the TIME_WAIT state, which on rare occasions allows wandering duplicate packets to be safely ignored. Value type: Yes / No Default value: "No" |
node_connclose | Close all connections to a node once we detect that it has failed. Value type: Yes / No Default value: "No" |
node_connection_attempts | The number of times the software will attempt to connect to the same back-end node before marking it as failed. This is only used when passive_monitoring is enabled. Value type: unsigned integer Default value: "3" |
node_delete_behavior | Specify the deletion behavior for nodes in this pool. Value type: enumeration Default value: "immediate" Permitted values: immediate: All connections to the node are closed immediately. drain: Allow existing connections to the node to finish before deletion. |
node_drain_to_delete_timeout | The maximum time that a node will be allowed to remain in a draining state after it has been deleted. A value of 0 means no maximum time. Value type: seconds Default value: "0" |
node_fail_time | The amount of time, in seconds, that a traffic manager will wait before re-trying a node that has been marked as failed by passive monitoring. Value type: seconds Default value: "60" |
node_so_nagle | Whether or not Nagle's algorithm should be used for TCP connections to the back-end nodes. Value type: Yes / No Default value: "Yes" |
nodes | A list of nodes in this pool. A node should be specified as a <ip>:<port> pair. Value type: list Default value: <none> |
note | A description of the pool. Value type: string Default value: <none> |
passive_monitoring | Whether or not the software should check that 'real' requests (i.e. not those from monitors) to this pool appear to be working. This should normally be enabled, so that when a node is refusing connections, responding too slowly, or sending back invalid data, it can mark that node as failed, and stop sending requests to it. If this is disabled, you should ensure that suitable health monitors are configured to check your servers instead, otherwise failed requests will not be detected and subsequently retried. Value type: Yes / No Default value: "Yes" |
persistence | The default Session Persistence class this pool uses, if any. Value type: string Default value: <none> |
priority!enabled | Enable priority lists. Value type: Yes / No Default value: "No" |
priority!nodes | Minimum number of highest-priority active nodes. Value type: unsigned integer Default value: "1" |
priority!values | A list of node priorities, higher values signify higher priority. Priorities are specified using the format <ip>:<port>:<priority>, if a priority is not specified for a node it is assumed to be 1. Value type: list Default value: <none> |
queue_timeout | The maximum time to keep a connection queued in seconds. Value type: seconds Default value: "10" |
service_discovery!enabled | Are the nodes of this pool determined by a Service Discovery plugin? If yes, nodes will be automatically added and removed from the pool by the traffic manager. Value type: Yes / No Default value: "No" |
service_discovery!interval | The minimum time before rerunning the Service Discovery plugin Value type: unsigned integer Default value: "10" |
service_discovery!plugin | The plugin script a Service Discovery autoscaled pool should use to retrieve the list of nodes. Value type: string Default value: <none> |
service_discovery!plugin_args | The arguments for the script specified in "service_discovery!plugin", e.g. a common instance tag, or name of a managed group of cloud instances. Value type: string Default value: <none> |
service_discovery!timeout | The maximum time a plugin should be allowed to run before timing out. Set to 0 for no timeout. Value type: unsigned integer Default value: "0" |
smtp!send_starttls | If we are encrypting traffic for an SMTP connection, should we upgrade to SSL using STARTTLS. Value type: Yes / No Default value: "Yes" |
ssl_cipher_suites | The SSL/TLS cipher suites to allow for connections to a back-end node. Leaving this empty will make the pool use the globally configured cipher suites, see configuration key ssl!cipher_suites in the Global Settings section of the System tab. See there for how to specify SSL/TLS cipher suites. Value type: string Default value: <none> |
ssl_client_auth | Whether or not a suitable certificate and private key from the SSL Client Certificates catalog be used if the back-end server requests client authentication. Value type: Yes / No Default value: "No" |
ssl_common_name_match | A list of names against which the 'common name' of the certificate is matched; these names are used in addition to the node's hostname or IP address as specified in the config file or added by the autoscaler process. Value type: list Default value: <none> |
ssl_elliptic_curves | The SSL elliptic curve preference list for SSL connections from this pool using TLS version 1.0 or higher. Leaving this empty will make the pool use the globally configured preference list, ssl!elliptic_curves in the Global Settings section of the System tab. See there for how to specify SSL elliptic curves. Value type: string Default value: <none> |
ssl_encrypt | Whether or not the pool should encrypt data before sending it to a back-end node. Value type: Yes / No Default value: "No" |
ssl_enhance | SSL protocol enhancements allow your traffic manager to prefix each new SSL connection with information about the client. This enables Virtual Traffic Manager virtual servers referenced by this pool to discover the original client's IP address. Only enable this if you are using nodes for this pool which are vTMs, whose virtual servers have the ssl_trust_magic setting enabled. Value type: Yes / No Default value: "No" |
ssl_fixed_client_certificate | An entry in the SSL client certificates catalog, containing a certificate and private key to be used whenever client authentication is requested. If set, this overrides server request parameters. Value type: string Default value: <none> |
ssl_middlebox_compatibility | Whether or not TLS 1.3 middlebox compatibility mode as described in RFC 8446 appendix D.4 will be used in connections to pool nodes. Choosing the global setting means the value of configuration key ssl!middlebox_compatibility from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for use of middlebox compatibility enabled: Enable use of middlebox compatibility disabled: Disable use of middlebox compatibility |
ssl_send_close_alerts | Whether or not to send an SSL/TLS "close alert" when initiating a socket disconnection. Value type: Yes / No Default value: "Yes" |
ssl_server_name | Whether or not the software should use the TLS 1.0 server_name extension, which may help the back-end node provide the correct certificate. Enabling this setting will force the use of at least TLS 1.0. Value type: Yes / No Default value: "No" |
ssl_session_cache_enabled | Whether or not the SSL client cache will be used for this pool. Choosing the global setting means the value of the configuration key ssl!client_cache!enabled from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for use of the session cache enabled: Enable use of the session cache disabled: Disable use of the session cache |
ssl_session_tickets_enabled | Whether or not SSL session tickets, including TLS >= 1.3 PSKs, will be used for this pool if the session cache is also enabled. Choosing the global setting means the value of the configuration key ssl!client_cache!tickets_enabled from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for use of session tickets enabled: Enable use of session tickets disabled: Disable use of session tickets |
ssl_signature_algorithms | The SSL signature algorithms preference list for SSL connections from this pool using TLS version 1.2 or higher. Leaving this empty will make the pool use the globally configured preference list, ssl!signature_algorithms in the Global Settings section of the System tab. See there for how to specify SSL signature algorithms. Value type: string Default value: <none> |
ssl_strict_verify | Whether or not strict certificate verification should be performed. This will turn on checks to disallow server certificates that don't match the server name or a name in the ssl_common_name_match list, are self-signed, expired, revoked, or have an unknown CA. Value type: Yes / No Default value: "No" |
ssl_support_ssl3 | Whether or not SSLv3 is enabled for this pool. Choosing the global setting means the value of the configuration key ssl!support_ssl3 from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for SSLv3 enabled: Enable SSLv3 disabled: Disable SSLv3 |
ssl_support_tls1 | Whether or not TLSv1.0 is enabled for this pool. Choosing the global setting means the value of the configuration key ssl!support_tls1 from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for TLSv1.0 enabled: Enable TLSv1.0 disabled: Disable TLSv1.0 |
ssl_support_tls1_1 | Whether or not TLSv1.1 is enabled for this pool. Choosing the global setting means the value of the configuration key ssl!support_tls1_1 from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for TLSv1.1 enabled: Enable TLSv1.1 disabled: Disable TLSv1.1 |
ssl_support_tls1_2 | Whether or not TLSv1.2 is enabled for this pool. Choosing the global setting means the value of the configuration key ssl!support_tls1_2 from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for TLSv1.2 enabled: Enable TLSv1.2 disabled: Disable TLSv1.2 |
ssl_support_tls1_3 | Whether or not TLSv1.3 is enabled for this pool. Choosing the global setting means the value of the configuration key ssl!support_tls1_3 from the Global Settings section of the System tab will be enforced. Value type: enumeration Default value: "use_default" Permitted values: use_default: Use the global setting for TLSv1.3 enabled: Enable TLSv1.3 disabled: Disable TLSv1.3 |
transparent | Whether or not connections to the back-ends appear to originate from the source client IP address. Value type: Yes / No Default value: "No" |
udp_accept_from | The IP addresses and ports from which responses to UDP requests should be accepted. If set to accept responses from a specific set of IP addresses, you will need to enter a CIDR Mask (such as 10.100.0.0/16). Value type: enumeration Default value: "dest_only" Permitted values: dest_only: Only the IP address and port to which the request was sent. dest_ip_only: Only the IP address to which the request was sent, but from any port. ip_mask: Only a specific set of IP addresses, but from any port. all: Any IP address and any port. |
udp_accept_from_mask | The CIDR mask that matches IPs we want to receive responses from. Requires: udp_accept_from is set to "ip_mask" Value type: string Default value: <none> |
udp_response_timeout | The maximum length of time that a node is permitted to take after receiving a UDP request packet before sending a reply packet. Zero indicates that there is no maximum, preventing a node that does not send replies from being presumed to have failed. Value type: seconds Default value: "0" |