Skip to end of metadata
Go to start of metadata

Introduction

This page describes the way how to diagnose connector-related problems.

MidPoint is using the ConnId connector framework. Every connector is running in this framework. It means that midPoint calls the ConnId framework and the framework calls the connector. Therefore when it comes to troubleshooting connector problems there are several places where a problem can occur and also several places where you can get diagnostic data:

  1. MidPoint: the midpoint may invoke wrong operation at the first place. This may be caused by a misconfiguration or a bug. See Troubleshooting Mappings page for a guide how to diagnose these problems.
  2. ConnId: the framework may misinterpret the operation. The framework also simulates some operations and it may post-process the results.
  3. Connector: this is the tricky part of the story. Each connector is different. Very different. But there are ways. See below.
  4. Resource: it is possible that the problem is caused by resource misconfiguration. E.g. the connector is not allowed to see all data, there are some limits, etc. We will not go into details here. See the documentation that goes with the resource for troubleshooting details.

ConnId Framework

The ConnId connector framework stand between midPoint and the connectors. It knows about every operation that midPoint invokes on every connector and it knows about all the return values. Some time ago we have added ability to log all the operations that are passing through the connector framework. This can be easily enabled by using the following log configuration:

org.identityconnectors.framework: TRACE

Or more specifically (with less logging noise - applicable only in midPoint 3.9 and later):

org.identityconnectors.framework.api.operations: TRACE
org.identityconnectors.framework.spi.operations: TRACE
org.identityconnectors.framework.common.objects.ResultsHandler: TRACE

The ConnId operation traces look like this:

This is a very useful mechanism. It will log every operation of every connector. If you suspect that the connector is not executing the right operation this is the right place to check it. You can see what is the operation that the midPoint is passing to the connector. If that operation looks good then the problem is most likely in the connector (see below). If the operation does not make sense then the problem is usually in the provisioning (see above).

However, the operation is logged by the ConnId framework on relatively high level and the operation is still quite abstract. If you need more details about what really gets executed you have to rely on the connector logging.

Please note that the ConnId framework has two "faces": API and SPI. The API is facing midPoint. MidPoint invokes ConnId API operations. The SPI is facing the connector. Connector implements SPI operations and ConnId framework is invoking them. You can see the distinction in the class names that are written in the logfiles, e.g. SearchApiOp vs SearchOp (if there is no Api or Spi in the operation name then it is assumed to be SPI). There is also similar distinction in the package name of the logger. Most API and SPI operations are direct equivalents. But there may be subtle differences. E.g. The get API operation is executed as search (executeQuery) SPI operation.

Interpreting Request-Response Operation Logs

Most connector operations are "pure" request-response operations: there is one request and one response. These are operations such as create, modify, delete. In this case you will see one request in the log files and one response. And thats the whole operation. Like this:

The above example illustrates a very common create operation. It should be interpreted like this:

  1. ...api.operations.CreateApiOp Enter: MidPoint invokes ConnId API. The object to create is logged as an operation parameter. This is what midPoint sends.

  2. ...spi.operations.CreateOp Enter: ConnId invokes the connector. This is what the connector receives.

  3. Connector executes the operation. Logs from the connector will be here (if connector logging is enabled)
  4. ...spi.operations.CreateOp Return: Connector operation is finished. The connector returns the result to ConnId.
  5. ...api.operations.CreateApiOp Return: Operation is finished and post-processed by the framework. Framework returns the result to midPoint.

This is quite straightforward and it applies to vast majority of connector operations. However, there are some peculiarities.

There are three modify operations:

  • update(...) in UpdateOp: This replaces attribute values. It is (roughly) an equivalent to midPoint modify/replace deltas.
  • addAttributeValues(...) and removeAttributeValues(...) in UpdateAttributeValuesOp. This adds or deletes attribute values. It is (roughly) an equivalent to midPoint modify/add and modify/delete deltas.

Interpreting Search Operations

The connector search operations are quite different from other operations. These operations are not one request and one response. These are one request and any number of responses (including zero responses). The search operations are using the callback mechanisms to pass results to midPoint. The log of the search operations looks like this:

It should be interpreted like this:

  1. ...api.operations.SearchApiOp Enter: MidPoint invokes ConnId search operation. The parameters are recorded in the log.

  2. ...spi.operations.SearchOp executeQuery Enter: ConnId framework executes the connector executeQuery operation.
  3. Connector starts the search (sends search request to the resource). Resource will start sending back the search results.
  4. ...common.objects.ResultsHandler Enter: The connector has received a search result. The connector is invoking result hander to process it. The result handler is a midPoitn code that is processing the result. This is a method how the connector passes the search result to midPoint.
  5. MidPoint will process this search result.
  6. ...common.objects.ResultsHandler Return: MidPoint is done with processing of the search result. It returns control to the connector to get more search results (if there are any). The result value (true/false) indicates whether the search should continue or the operation should be stopped.
  7. The steps 4, 5 and 6 are repeated for every search result.
  8. ...spi.operations.SearchOp executeQuery Return: Search operation is done. There will be no more results. Connector returns control to the ConnId.

  9. ...api.operations.SearchApiOp Return: Search is done. ConnId returns control to midPoint.

Please note that get operations are also interpreted as search operations. The ConnId framework transforms them internally.

Also please note that there may be search operations inside a search. E.g. when midPoint processes the search result it may need to execute another search inside that processing. E.g. when listing all accounts, we may need to search for group membership for each account. This gives us search withing search. In that case you have to carefully watch for a start of new search operations inside the ResultsHandler, i.e. between the ResultsHandler Enter and ResultsHandler Return log messages.

Connectors

Each connector is slightly different. The connectors have to adapt to the resource communication protocol and therefore they are expected to use variety of client and protocol libraries. Each library may have its own method of troubleshooting. Therefore there is no universal way troubleshoot a connector. However, there is (almost) always some way. Please refer to the connector documentation for the details.

Even though there is no universal way how to troubleshoot a connector there are some general guidelines. Most connectors log at least some information about the operations. You just need to enable the correct logger. The logger name is usually the same as the package name of the connector classes. Look in the documentation or directly inside the connector JAR file to find out the package name. You may also need to enable logging of the libraries that come with the connector. You can examine these if you look in the lib directory inside the connector JAR file.

Some connectors have really good logging, such as the LDAP connector (and its subclasses). The LDAP connector will log all the LDAP operations if you set the com.evolveum.polygon.connector.ldap.OperationLog logger to DEBUG level (also see LDAP Connector Troubleshooting):

Some connectors will barely log anything. This is all connector-dependent. If the connector author did a good job you will get what you are looking for. If the author did a poor job you are mostly out of luck. But one way or another this is the best chance to learn what the connector is doing. If that fails you have to resort to packet sniffer and similar tools.

Each connector has its own logger names. The most reliable way to find out what it is is to look at the connector documentation or connector source code. But as rule of thumb the connectors usually use logger names matching their Java package. This is supposed to be the same as the package prefix used in the connectorType property (see ConnectorType), e.g. com.evolveum.polygon.connector.ldap.

I Suspect a Connector Bug

So, you think you have found a connector bug. That may happen. No practical software is completely bug-free and the connectors are no exception. But before going to report a bug please spend some time diagnosing the issue. Firstly, what looks like a connector bug may in fact be a misconfiguration. Secondly, if your bug report states justs "the connector is broken" then such a bug is very unlikely to ever get fixed. We need more data about the issue to fix it. Therefore this is the recommended procedure:

  1. Enable ConnId framework logging.
  2. Make sure that the operations that midPoint sends to the connector are correct. Make sure that the request makes sense. If the ConnId request is wrong then this is not a connector bug. It is most likely a midPoint bug.
  3. Look for error messages from the resource. Maybe the resource refused the operation due to insufficient access rights. Maybe you are trying to add many values to a single-valued attribute. Maybe you try to create an object that already exists. In that case this is most likely a midPoint misconfiguration. Not a connector bug.
  4. Look for objects that are returned from the search (ResultsHandler). Do these look OK? Is there correct number of objects? Is something missing? Are there all required attributes? May this be caused by the resource confifguration. E.g. does the resource allow to read all of these data?
  5. It is time to go deeper. Enable logging of connector operations. E.g. for LDAP connector set com.evolveum.polygon.connector.ldap.OperationLog logger to TRACE level.
  6. Check the requests that connector sends to the resource. If the ConnId operation looks OK but the request to the resource is not OK then this is very likely a connector bug.
  7. Check the response from the resource. Are there any errors? Do the data seems OK? If the resource returns the data correctly but the connector's response to ConnId is wrong then it is very likely a connector bug.
  8. If you still do not know what is going on it is time for heavy artillery. Enable full connector logging. And maybe logging of the underlying libraries. Try to figure out what is going on.
  9. If everything fails then you can report the issue. However at this point it is almost sure that the issue will be very tricky. It is unlikely that the midPoint team will have time to address that issue unless you have an active subscription.

Extra Tips

  • Enable Auditing to a log file, including the auditing details. This will log all the deltas that are requested to execute from the user interface and other interfaces. Sometimes the problem is in the request delta. The audit is the easiest way to check that.

See Also

{code}

  • No labels