Skip to content

Web Intelligence Processing Failures

What the "Webi Processing Failure" email means, why MetaMiner sends it, and how to restore access to the Web Intelligence Processing Servers.

How it works

MetaMiner reads the inner structure of Web Intelligence documents (data providers, queries, SQL, report elements, variables) during deep processing. By default deep processing runs every 5 minutes for each environment (Webi Metadata pull period on the MMA Pull settings page).

While it collects server metrics (every 5 minutes, when Pull Server Metrics is enabled), the MetaMiner Server (MMS) also checks every running Web Intelligence Processing Server registered in the CMS. It opens a TCP connection to the host name and request port that the server reports, with a 3-second timeout. If any check fails, MetaMiner:

Effect Detail
Sends an email Subject [<MetaMiner host>] <environment> Webi Processing Failure. At most one email every 6 hours per environment. Requires email notifications to be configured.
On XI 3.1 environments Skips Web Intelligence deep processing until the servers are reachable again, so the MetaMiner Client is not affected by repeated disconnects.
On BI 4.x environments Keeps processing documents. Documents that need the unreachable server can fail or time out and show no detail in the MetaMiner Client.

The rest of the pull (folders, users, schedules, universes, security) continues normally.

Image of a Webi Processing Failure notification email

../images/help/webi-processing-failure-email.png

Symptom

You receive an email with the subject "... Webi Processing Failure" and the body "One or more Webi Processing servers are not reachable." New or changed Web Intelligence documents may have no data providers, queries or SQL in the MetaMiner Client.

Cause

The MetaMiner server cannot open a connection to at least one Web Intelligence Processing Server. Common reasons:

  • A firewall between the MetaMiner server and the BusinessObjects processing tier blocks the request port.
  • The request port is assigned dynamically and changes when the server restarts, so a firewall rule no longer matches.
  • The host name the processing server registers in the CMS does not resolve from the MetaMiner server (for example a short name, an internal name, or a name in another DNS zone).
  • The processing server is running but overloaded or hung.

Fix

  1. In the CMC, open Servers, find each Web Intelligence Processing Server and note its host name and request port.
  2. On the MetaMiner server, test each one: 
    Test-NetConnection <wips-host> -Port <request-port>
  3. If the host name does not resolve, add a DNS entry, or set a fully qualified host name on the server in the CMC.
  4. If the port is blocked, set a fixed request port on each processing server in the CMC, restart the server, and open that port from the MetaMiner server in every firewall in between. See Ports.
  5. Wait for the next metrics cycle (about 5 minutes). Deep processing resumes on its own; you do not need to restart MetaMiner.

Image of the CMC server properties of a Web Intelligence Processing Server with the Request Port setting highlighted

../images/help/cmc-wips-request-port.png

What to send support

  • The notification email.
  • The Test-NetConnection output for each processing server.
  • A log bundle; server.log contains the reachability checks.