In this document, a description about the opHA 3 features when integrating with other products will be made.
opHA will synchronise all the NMIS 9 database information. That means:
- Log files from the pollers will not be synchronise
- RRD files will not be moved
The functional differences between a primary, poller and standalone server are as follows:
- It will poll all the local nodes
- All the local nodes can be managed locally.
- It can poll local nodes
- It will have information about the remote nodes
- It will create network metrics for local and remove nodes
- Just Local nodes can be edited
- It will redirect to the poller to read information for remote nodes
Remote node is redirected from the primary.
- It will poll local nodes.
- Nodes must be edited from primary server.
- All the information can be read from the poller.
Node configuration view in the poller: Not possible edition.
|View local Nodes|
|View Remote Nodes|
|Edit Local Nodes|
|Edit remote Nodes|
|Poll Local Nodes|
|Poll Remote Nodes|
The views from the poller and the primary will change.
We will access the main screen, just the local information will be available.
We will see all the actions and all the pollers:
We will see a message that indicate us that it is a poller:
As opCharts operates with NMIS 9 data, we will be able to see and use the remote nodes as local nodes from opCharts, and use them in Charts or maps.
We just need:
- opHA keeps synchronising the data in order to be updated.
- The nodes should be activated for opCharts in order to see them.
- Have opHA urls configured correctly.
Remote node from the primary.
As RRD data is not synchronised, opCharts makes a remote call to visualise the graph information. If you cannot see the graphs, please review this information.
opHA just synchronises the information collected by NMIS9. But there are different approaches regarding opEvents data integration between servers:
Depending on the followed approach, the opEvents rules can be set up in the poller or the primary.
They should be set up in the primary in case os using the syslog approach.
If we are using the API they can be set up in the poller or the primary. It is a common approach to just send the important events to the primary.
opHA does not provide integration at the moment for opConfig.
opConfig will work just with local nodes.
opHA does not provide integration at the moment for opReports.
As opReports uses rrd data and this is not synchronised.
The reports must be created from the pollers, and use local data. Then the reports can be synchronised with the primary server.