TPO and Turret connection

In order to use TPOs, turrets need first to be able to locate TPOs. In all users and turret configuration referring to a TPO is done using a logical name, never with an IP address. This logical name may hide a standalone TPO or a resilient TPO, it does not matter for the turret.

As a general principle, a single turret can access multiple TPOs over multiple clusters at the same time.

Everywhere in documentation we talk about "DNS name". However, true DNS requests and DNS server requirements only depend of the TPO resiliency mode chosen. This "DNS name" is a way to provide a common way to configure TPOs and turrets whatever the chosen resilency mode.

1) Thanks to TPO installation procedure, TPO has registered with TSS, TSS knows about TPOs IP addresses

2) At boot time, turret learn about TPOs names & clusters from TSS through the terminal register process.

3) With the information obtain at step 2 Turrets know how to match names with final IP address, this match can be done "once for all" or dynamic (true DNS request used) this will depends on the TPO redundancy mode chosen

Whatever the TPO redundancy mode, this high level description does not mean that TSS is required in order to proceed to TPOs name resolution each time turret need to contact TPO, turrets only learn basic information at boot time on how they have to proceed.

Notes:

•Turret reboot is required if a new TPO is defined in TSS in order to restart  step 2. This way turret will learn about this new TPO or TPO name change.

•The same remark apply if TPO clusters deeply change by adding/deleting/moving a node or if cluster type change.

•Prior to version 8.0, administrators need to declare TPOs to the turret via Boot Setting "TPO Groups" (application.setting.TPO.groups). This declaration is not anymore required as turrets learn about TPO existance through the TSS Terminal register service automatically.