SCOM to SCSM Connector Stops

I had a situation that the SCOM to SCSM (both 2012 SP1) had stopped several days ago yet the CI Connector showed that it was still working. There was no obvious reason why this connector was still working when the alert connector was not.Looking at the SCOM side and showing the column Forwarding Status then I could see that all the alerts that met the criteria were showing as Forwarding Pending. This showed that the SCOM side was putting the alerts up for SCSM to pick up but SCSM was not picking them up. When I clicked the Synchronise Now button on the connector in SCSM the following event was generated.

Warning event 34070
Source – Operations Manager Connector

Error of type System.Xml.XmlException while reading configuration. The following information may help take corrective action:
Message: An error occurred while parsing EntityName. Line 1, position 958.
Subsequent error messages in the event log will indicate the affected connector. In order to correct this error, export the  ServiceManager.LinkingFramework.Configuration management pack, correct the connector configuration, and then reimport the management pack.

The named MP (ServiceManager.LinkingFramework.Configuration) was exported and I looked at it in an editor. Line 1 was the normal MP start line and certainly did not have 958 characters. But down in the XML there was this section:- 

<ComputerCriteriaRule>

                      <Name>Customer A &amp; B Incidents</Name>

                      <TemplateId>8f5bfff8-2e46-a477-58a0-8e60d4b3b59d</TemplateId>

                      <AlertSeverity>Error</AlertSeverity>

                      <CriteriaType>Computer</CriteriaType>

                      <ComputerCriteria computerGroup=”Customer A &amp; B” />

            </ComputerCriteriaRule>

While the line number and character number in the event was no help it did point me to the correct MP to look in. I was wary as soon as I saw the ampersand which is a reserved character in XML and even though it was encased properly as &amp; I decided to try that first.

There were a number of Alert Routing Rules that had been created. These used groups to determine which template a particular alert should use when it comes in. One of these groups and the template had an ampersand (&) in the name. This Alert Routing Rule was deleted and the connector sprung back into life. The group and template were renamed without the ampersand and put back into the Alert Routing Rule table and all is fine now. 

 

About these ads

Comments are closed.

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: