Some LDAPv3 servers are not quite as compliant as we thought. We need to add more configurable options so that different systems can be accommodated for the end user, with possible documentation on which configurations are expected to work for different directories/versions. For example, some servers don't support server side sort, but still work when the control isn't used (eDir); others support sort, but use a different OID (Oracle).

I have been trialing the use of Windows Event Log subscription emails as per my post here. Lately @ CSODBB these have proven very useful in early detection of issues such as unexpected 3rd party Identity Broker connector outages, or unexpected AD activity (scripted updates of AD accounts causing unexpected sync load). I mentioned this to Shane Day and he suggested I put in a feature request, so here it is.

I envisage an extension to Event Broker in several independent/complementary ways:

1. New change detection feature for Windows Event logs (potentially subscriber-push notifications) - these would complement the existing file changes detection for monitoring FIM and related logging;
2. Enhanced file logging to allow the specification of filtering in the log file in a similar way to what is possible with Event Logs - idea is that you may want to subscribe to only certain entry types in a file log (note that for Identity Broker this would be achievable already by specifying a Windows event logger instead, but other services that write file logs may not have this option);
3. an SMTP email operation (yes you can do this in PowerShell now, but with some thought you could use a wizard automate the creation of a PowerShell plugin with use of Cmdlets to provide basic parameter-driven email notifications);
4. enhanced log reading capabilities to improve searches;
5. enhanced log presentation capabilities (UI) to map occurrences of specified event filters over time for a range of log files - including Event Broker's own log files)

All of the above require more thought, but I'm really thinking of how to create something to improve the quality of our own UNIFY health checks, initially by providing an enhanced, more responsive service, but secondly improving the product's own capabilities to attract the market that was once there for NetPro's (now Quest's possibly defunct) Mission Control for ILM.

The Incoming Event for the Event Broker for File Changes Plugin continuously repeating the Run Profile Operations in it eventhouh the file was only changed once.

Created CSV MA in FIM

Create CSV MA in Event Broker, see attached image for details.

Make change to the specified file only once (closing the file or not seems to have no effect). The whole operation is repeating itself continuously. This occurs whether I added the "Commit File Changes Plugin, On Failure=Fail" or not.
See the image of log details.

This also happen when I only have two Rule Profile operation entries for the CSV MA.

The configuration for the File Changes Plugin is

<FileChanges> 
	<FileName><![CDATA[ 
		C:\Program Files\Microsoft Forefront Identity Manager\2010\Synchronization Service\MaData\FIM_CSV_pdt_AHG_MA-AHG\sample_data_csv_ahg.csv
	]]></FileName> 
	<GroupName> 
		CVSFileGroup
	</GroupName> 
</FileChanges>

The configuration for the Commit Changes Plugin is (see attached image for details)

<FileChanges>
	<FileName><![CDATA[
		C:\Program Files\Microsoft Forefront Identity Manager\2010\Synchronization Service\MaData\FIM_CSV_pdt_AHG_MA-AHG\sample_data_csv_ahg.csv
	]]></FileName>
	<GroupName>
		CSVFileGroup
	</GroupName>
</FileChanges>

Although this does not seems to correspond to the documentation on Jira File Changes .

Could you advise me what I did wrong with this configuration?

EvB service will not start. Manual control hangs and even server reboot will not allow for service to start. There are many entries in log for Error (-1): Error in MIIS Run Profile: call-failure:0x8007052E as well. If service start selected from menus it appears to start running but then only processes 1 or 2 operations before stopping again.

This resulted in having to spend almost 3 hours yesterday manually setting connectors for new user account of which there were over 124. Then once all connectors were set, ran export to AD to get accounts created. Normal operation of MA's does not appear to be working either.

Outgoing operation lists require a v3.0 equivalent to the "Outgoing Provisioning Pending" from former Event Broker versions. This likely will take the form of an outgoing check operation, checking the targeted FIM instance's database for pending export on particular MAs. In addressing this, the following questions will need to be addressed:

• Check if MAs have an ID that is consistent, even after management agent rename
• Check the WMI interface to see if you can trigger an MA by its ID rather than its name

The pending alerts message format looks strange in IE8. See attached screenshot. This is from the latest build, dated 28/06/11 5:06pm

The current status / FriendlyNextRun displayed against an operation list, after an operation has failed will be incorrect.

For example:

Create a powershell module Example.ps1 with the contents of:

New-Item C:\abcdefg.txt -type file

1. Execute this operation once so that the file exists.
2. Maintain the default retry count and wait time on the operation.
3. Execute the operation, and go to the home page.
4. After around 15-20 seconds the friendly next run will read '- Running (2 of 1)'

It will be useful to save information about operations that is not essential to their execution, but can be used for diagnostics and/or suggestions. The operation architecture should be extended to allow for the collection of information such as:

• Execution time of operations, for calculating averages or suggesting extensions
• Additional information required by/for plugin operations
• Operation list overlap for diagnostics

Estimate includes investigation of current setup, examining plugins where additional information may be required, further development, and testing.

Event Broker presently cannot detect changes in eDir, and a new plugin is required soon due to an increase in the number of Novell sites looking to migrate to ILM - such as the JPM Chase lead presently being followed up by PT. I understand that the eDir ODBC driver is supposed to be able to read the change log, but our resident Novell experts may be able to provide details on the best approach to use. With the spotlight on Event Broker now (especially with the advent of ILM2), it is important that we are able to add to the change detection plugin collection so as to maximise sales opportunities. I think eDir is at the top of the pile, but other candidates include the following:
. IBM Universe/DB2 (Cath Ed!)
. Exchange 5.5 (Cath Ed)
. Sun (QDETA)
. ???

In the EB Management Studio, the "Retry wait time" field in the configuration of the "Check Operation" for an operation uses a standard textbox instead of the usual selection of dropdowns menus for selecting a time. See the "Repeat every..."'s "Frequency" field for reference.