High Availability commands

There are several commands used specifically for the HA environment.

Managing an HCL Traveler HA pool, in many ways, is very similar to managing a set of stand alone HCL Traveler servers. HCL Traveler servers are started and stopped individually. There is no command to start or shutdown all servers. When any server is shutdown in a pool, the other HCL Traveler servers detect this outage and handle the device requests.

The HCL Traveler tell commands are HA pool-aware. Where necessary, commands automatically route a request to another server or broadcast a request to all the servers in the pool without the administrator explicitly requesting that action. For example, tell traveler user <userName> can be used to show the details for a user. This command will automatically be forwarded to the appropriate server for the user. Another example is setting the log levels. The log levels are managed across the pool of servers, so to set log levels to fine, you would issue tell traveler log level fine and all servers will be notified that the log levels have been changed. Command helps have been updated, where applicable, to provide any new parameters that relate to working with a pool of servers.

The following commands are used for the HA environment:

tell traveler HADR

The HADR command is used for managing servers in the High Availability pool. The following commands are supported:

tell traveler HADR Show

This shows all servers in the High Availability pool along with the same information that is available in the Servers view from the web based administration client. The following information is returned:

Domino Name - Domino name of the server.

ID - The HCL Traveler identifier of the server.

Host Name - Host name of the server.

IP:SrvrPort,SrvltPort - IP Address in use for both the HCL Traveler task and the HCL Traveler servlet running in HTTP, Port number in use for the HCL Traveler task and Port number in use for the HCL Traveler servlet running in HTTP.

Alive - True if the server has updated its heartbeat entry in the HCL Traveler database within the last NTS_TSS_HEARTBEAT_TIMEOUT seconds; false otherwise.

Server Reachable - True if the HCL Traveler task responded to an HCL Traveler ping message; false otherwise.

Servlet Reachable - True if the HCL Traveler servlet running in the HTTP task responded to an HCL Traveler ping message, false if the servlet failed to respond, or n/a if the server is too old to support this check. False is not necessarily a problem, as HTTP only initializes the servlet as needed on the server. As a result, if there has not been an HTTP request to the HCL Traveler servlet yet, this will be false until the servlet is accessed the first time. However, the HTTP load should be spread such that this servlet is unused for only a short period of time. False is often an indication of a problem.

Last Heartbeat - The last time the server updated its heartbeat entry in the HCL Traveler database.

AI - The availability index of the server.

Users - The number of users assigned to the server. This is tied to the Push.Users.Total stat.

Devices - The number of devices assigned to the server. This is tied to the Push.Devices.Total stat.

Build Level - Build level of the server.

Site - Cloud site. Deprecated. Not applicable for on-premises environments.

Startup - Time the server was last started. Displayed when using Verbose flag.

DB Version - The DB Schema version that the server is using. Displayed when using Verbose flag.

Status - Status of the server. 0 is Green, 1 is Yellow and 2 is Red. Displayed when using Verbose flag.

tell traveler HADR Ping <server>

This command sends an internal ping request to the specified server. The response will indicate success if the server is active or a failure if the server cannot be contacted.

tell traveler HADR Delete <server>

Removes the server from the High Availability pool. You must take the server offline before executing this action.

tell traveler available

Specify on or off to enable or disable this server in the High Availability pool. If a server is available, then it will be considered for load balancing. Disabling a server means that the internal HCL Traveler load balancer will not consider this server for new work. For example:

tell traveler available on

tell traveler bind

The bind command allows you to view all of the registered users in the Traveler pool and to which servers they are currently bound. It also lets you force a user to always connect to a specific HCL Traveler server. This may be useful when debugging a single user so that all of their traffic is routed to a single server and ignores the availability index.

tell traveler bind show: Shows which users have been explicitly bound to a server. By default this list will be empty.
tell traveler bind showall: Shows all of the registered users in the High Availability pool, along with which server they are running on.
tell traveler bind enable <userid>: Binds the user to the local server.
tell traveler bind disable <userid>: Removes the binding for the user.
tell traveler bind move UserId: Moves the user to the server but does not bind them there (this is equivilant to doing both Enable and Disable).
tell traveler bind balanceUsers: Balances the users across the servers that are alive and reachable. The number of devices or other user characteristics are not taken into consideration.

Some commands can be issued to a specific server or broadcast to all servers. The tell commands have an optional command line parameter to send the command to a specific server or to broadcast a command. The parameter is -s <serverName | * | *->. When this parameter is not specified, the default is to process the command locally. Specifying -s <serverName> will send the command to the specified server. Specifying -s * will broadcast the command to all servers, while -s *- will broadcast the command to all servers except the local. For example, to check the status of all the HCL Traveler servers in the HA pool, issue tell traveler -s * status. The status command will be forwarded to each server in the pool and the response will be displayed on the local console where the command was originally issued.