Server Address⇑
Description
Local IP address and port used for replication.
Syntax
IP:port
DOCUMENTATION
Server Address | Replication Cluster | Heart Beat Interval (secs) | Heart Beat Timeout (secs) | Is Gzip Stream | Max Sequence Number Difference | Sync Interval | Data Master Priority | Enable Replication for Cached Data | File Cache Server Address
Name | Virtual IPs | Virtual IP ID | Virtual IP Token | NIC Name | Priority | Internal Device | Unicast Source Virtual IP | Unicast Peer Virtual IPs
Local IP address and port used for replication.
IP:port
List of replicators, by their IP Address and port number, that are a part of this replication cluster.
Comma separated list in the form IP:port
The time interval that "I am alive" messages will be sent.
Integer number
The time interval after which, if "I am alive" message is not sent, the replicator will be considered to have failed. Must be set higher than Heart Beat interval (secs).
Integer number
Enable Gzip compression for replicator traffic.
Select from radio box
Compares to the difference between the last sequence number sent and the last sequence number acknowledged by the master. If the difference is larger than maxAckDiff, individual SHM updates are paused until the next bulk update is complete.
This difference indicates the busy-ness level of the master. If multiple nodes are syncing a significant amount of data with the master, this value ensures that the data sync is still manageable and performs at an optimal level.
Default value: 100
Number between 1 and 65535.
The time in seconds that a node will remain in the upgrading state before transitioning to master. This interval is used to allow the group to communicate with one another to ensure that the correct node is selected as master. It is strongly recommended that this is set to a longer interval than the Heart Beat Interval.
Default value: 30
Number between 1 and 255.
Indicates the *data sync priority*. The node with the highest priority is selected to be the data sync 'master', when there are no masters selected yet. If a master is already selected, that node will remain master. Note that this is for data sync ONLY. The priority value in the HA Interfaces configuration group is for Virtual IP priority.
Default value: 0
Number between -32766 and 32767.
Enable file cache replication.
Select from radio box
Select file cache server address.
IP:port
Additions/Deletions/Changes made will not take effect until applied through HA -> Status page.
A unique number that identifies a virtual interface.
Integer number
A list of virtual IP addresses used to access the HA interface. These addresses can be either an IPv4 or IPv6 and must be on the same subnet as the real IP address.
Please verify proper usage with keepalived prior to using multiple Virtual IPs.
comma separated list
The ID of the virtual IP address. Please use an integer number between 1 and 10. Virtual IP, Virtual IP ID, and Virtual IP Token must be the same for the same HA group.
A virtual IP address token is required to authenticate interface(carp/vrrp) advertisements. Virtual IP, Virtual IP ID, and Virtual IP Token must be same for the same HA group.
The NIC name (such as eth0, eth1, etc..) that the virtual IP will bind to. Required for Linux platforms.
Note: Even though it is required, this setting will have no effect on FreeBSD platforms.
The priority for this load balancer server, where a higher number equals a higher priority. The highest priority load balancer will act as the master, with others in the cluster acting as slaves.
Note: This is reversed for FreeBSD platforms where a lower number equals a higher priority.
Integer number
A difference in priority of 50 or greater between two nodes will cause Keepalived (linux) to switch the Virtual IP back to the higher priority node upon recovery. For example, two nodes with priorities of 110 and 100 will start with the 110 priority node as the master. If that node dies and recovers, the 100 node will retain the Virtual IP. If on the other hand we have two nodes with priorities 200 and 100, the 200 node will reclaim the Virtual IP upon recovery.
The Internal Device (such as eth0, eth1, etc..) to use for internal communication by unicast/multicast. If this is different than the NIC Name configuration, the Keepalived configuration will be adjusted accordingly.
Default value: Configured NIC Name value
Select from drop down list
The local machine's virtual IP used to send Unicast messages.
Unicast Peer Virtual IPs must also be set for Unicast to be used. (Multicast is used by default)
Unicast is generally slower and more resource intensive than Multicast.
List of peer virtual IPs to send Unicast messages to. There should be one peer IP entry for each node expected to handle request processing.
Unicast Source Virtual IP must also be set for Unicast to be used. (Multicast is used by default)
comma separated list
Unicast is generally slower and more resource intensive than Multicast.
Displays the configuration and status of currently configured/live high availability interfaces by their unique Interface ID. The current configured settings for each interface are shown as well as the actual "live" settings in green. The "Status" field will contain 1-2 of the following values:
Configuration changes made to a high availability interface in the config section will not be applied until clicking "Apply Changes" next to the affected interface. The same applies to adding new interfaces and deleting existing interfaces, changes will not be applied until clicking "Add/Delete" respectively.
Displays the status of currently "live" replicators in the replication cluster by their server addresses, as well as the number of active sessions on record for each replicator. The source replicator (usually the first to come online) will have a highlighted SRC tag in front of it's Server Address. It is this replicators job to actively sync it's data to the other replicators in the replication cluster. These other replicators will be compared against the source replicator to determine if they are "in Sync" or "out of Sync".
SSL Session will show the number of currently cached SSL sessions for the current load balancer.
ID Session will show the session count in the session cache of each cluster defined in the load balancer.
Replication will happen in real-time as changes are written to shared memory.