5.2 GroupsThe Groups submenu is available only for administrator users, where the measurement server groups can be edited. The grouping of measurement servers has a double purpose: the pieces of equipment to be measured are distributed based on some algorithm among the simultaneously running servers of the same group. There are two types of distribution: · The distribution is based on the number of measurements on the equipment, which the system tries to balance. · If a primary server is selected and it is running, then it is going to measure all the equipments. In all other cases the equipments are distributed among the rest of the running servers based on the previous criteria. It is not recommended to divide measurement servers of the same type into different measurement server groups unless there are pieces of equipment that, for example, can only be reached by a measurement server located at a remote VPN. On the other hand, if for a measurement server group a flag is set indicating that its monitored equipment cannot be reached from the WEB server (see below), then the PVSR always selects one out of the currently running servers at random and queries the equipment through that (if no running server is found, then it will always try to access it directly). The currently existing server groups are displayed on the starting page with the following columns: · Editing links: o [add new]: creating a new group o [edit]: editing a group o [ · Name: Name of the group · Type: Group type (e.g. SNMP, ASCII, …) · Reachable from WEB server: Whether the pieces of equipment managed by the group are reachable from the WEB server. This counts only for active and discoverable types. · Primary measurement server: If it is set and the server is running, then it measures all equipments. · Number of servers: The number of servers currently in the given group. The same fields (name, type, availability, and primary server) are displayed when creating or modifying a server group, however the type field can only be set at creation time. The type means the type of the data collector registered in the database. New types cannot be added from the user interface since each new type means a new implementation of the collecting API. The types for the default installation are the following (if for all questions ‘yes’ was answered during installation): · ASCII: ASCII-based data collection, passive · IPSec: measurement of IPSec connections going through Cisco VPN3000, active and discoverable · Oracle: measurement of Oracle parameters, active and discoverable · QoS: measurement of Cisco QoS objects, active and discoverable · SNMP: arbitrary SNMP-based query, active and discoverable · SynthTrans: measurement of synthetic transactions, active and non-discoverable · Test: a test measurement server type provided for the documented API, active and discoverable · Unix/Linux: measurement of parameters of Unix and Linux machines, active and discoverable · NetFlow: Traffic measurement based on NetFlow, active and non-discoverable · Jaga: Measurement of ping based accessibility, active and non-discoverable · CiscoPING: pinging performed with the help of Cisco PING MIB (both configuration and query) |