Documentation PI Nexus+ Documentation

PI Nexus+ / User Guide

Servers and Sites

This chapter shows how to check whether PI Nexus+ can still reach a server, find which server holds the bad objects, and how sites group servers.

Overview

This chapter shows how to check whether PI Nexus+ can still reach a server, find which server holds the bad objects, and how sites group servers.

Check whether a server is reachable

The Servers page lists every PI Data Archive, AF Server, AF Database and PI Vision instance with two kinds of health side by side:

  • Service health: can PI Nexus+ connect to the server and keep its scans current? Good, Warning, Error, Unknown (never tested and never scanned) or Not monitored (disabled, or no longer a scan target).
  • Content health: how healthy the PI Points, AF Attributes or AF Analyses on it are.
Servers page with the tabs, the filter cards, search and Filters, a row with its service health, and the download button
  1. Choose the tab: PI Data Archives, AF Servers, AF Databases or PI Vision.
  2. Click Service Error or Service Warning to show only those servers.
  3. Search for a name, or choose Filters, for example Show: Also not monitored.
  4. Read the row's service health and its first reason, for example a failed connection test or failing scans. Click the row to open the dialog with every reason.
  5. Operators: the download button saves the tab as CSV.

How to act on the result:

  • Error service health: the content figures may be out of date. Ask an administrator to fix the connection or the failing scans first.
  • Good service health but many Bad objects: the server is fine; the problem is in the data. Continue on the inventory pages.

An AF Server's service health also counts its AF databases, because the databases are what PI Nexus+ scans. Hover a reason, or open the dialog, to see which database fails. The rules for each state are under Service health rules in the Reference.

Find which server holds the bad objects

  1. On the Servers page, click With Bad PI Points (PI Data Archives), With Bad AF objects (AF Servers, AF Databases) or With Display Errors (PI Vision).
  2. Open the row's action menu and choose, for example, Show PI Points, Show Point Sources or Show AF Analyses. The inventory page opens filtered to that server.

The counts on a row match the inventory pages: an archive's PI Points equal the PI Points page filtered to that archive.

Open a server's detail

Click a row to open the server's dialog. It shows the service health with every reason, the content health, and a Scanning section with the last inventory scan, the last health refresh and the change-feed state.

AF Server dialog on the Analytics tab with the Analysis Service Load and the Costliest Analyses

The Analytics tab of an AF Server or AF Database estimates the load on the Analysis Service and lists the Costliest Analyses. Click an analysis to open it. The load is an estimate from the Analysis Service statistics, not a measurement.

When the Windows server under a PI server is monitored, the Overview has a Host section with CPU, memory and the fullest disk. The Hosts tab of the Servers page and the host dialog are described in Host Monitoring.

How sites group servers

A site groups servers, usually by plant, for example Hamburg and Lyon. An administrator creates sites and assigns servers (see Sites in the Administration Guide).

  • Every PI Data Archive, AF Server and PI Vision instance belongs to one site, and everything on it follows its server. New servers go to the Default site.
  • Sites stay hidden until servers are spread over two or more sites.
  • A site has no health of its own: its content health covers its servers' objects, and its service health is that of its worst server.
  • If a central PI Data Archive receives copies of plant PI Points, for example through PI to PI, the copies belong to the central site. A dead interface at the plant then shows as Bad at both sites, which is correct, because the copies are stale too.

Reference: Servers and Sites Reference