Skip to end of metadata
Go to start of metadata

Troubleshooting Guide for PI Monitoring



 


 

 

Back

 

Central PI Monitoring Home Page

 

 

 

 


 

Define Scope 

No PI Domain available

You are trying to create a new PI monitoring scenario. When you create the scenario you cannot select the PI domain you want to set up PI Monitoring for.

Proposed Solution:

  1. Make sure your PI system meets the software prerequisites for PI Monitoring with Solution Manager. To be able to use PI Monitoring with SAP Solution Manager your PI systems needs to be at least on release SAP NW PI 7.11 SP06. Provided that your LMDB is up to date you can check the SW levels of your PI system as described in section PIMon_FAQs#WhatPIreleaseisrequiredtousePIMonitoringwithSolutionManager.
  2. If your PI system is a Java only system and you run on Solution Manager 7.1 SP05 or SP06 please check SAP note 1755427.
  3. Make sure that your PI domain was synchronized into LMDB of Solution Manager. In section PIMon_FAQs#HowcanIsynchronizePIDomaindataintoLMDBofSolutionManager you find different options how to synch your PI domain in the LMDB. The check PIMon_Checks#CheckifthePIdomainandallthecomponentsareavailableinLMDB describes how to find the PI domain in LMDB.

Back


 

The correct SLD is not showing in the dropdown list

During the creation of the scenario for PI Monitoring you have to select the SLD for the PI domain. This is the SLD used by the PI system in this domain to find business systems etc.

The correct SLD is not available in the dropdown list:

 

Proposed Solution:

Check 1: Check in the SLD of the PI system, if there is an instance for the PI SLD, under:

Home -> Development -> CIM Instances -> All Classes -> System Landscape Directory

Or in older SLD versions:
Administration -> CIM Instances -> All Classes -> System Landscape Directory

Check 2 If this instance does not exist there, please restart SLD -> this should create the entry

Check 3 If SLD restart does not help, check the destination "SLD_Client" of the AS Java (follow this link Configuring the SLD Client) Make sure, that you restart SLD after making changes here!

Check 4 If this SLD instance exists in PI SLD, the program from SAP Note 1631346 should usually transfer it to the SLD which would synchronize with LMDB. Check the Log of the program, for exceptions.

The SLD entry selected here during PI Monitoring Setup is only used to model the SLD in the system group belonging to PI Domain. This way you can later on also monitor SLD availability. There is no information read from the PI SLD at this point in time.

Back


 

ABAP business system not visible during scope selection

When you create the PI Monitoring scenario the ABAP business systems are not visible during the scope selection in the scenario wizard.

 

Reason:

Most likely the PI Domain is not properly/completely transferred from PI SLD into central SLD and LMDB. There seem to be an association missing in the Integration Server. The hierarchy is usually (very simplified hierarchy): PI Domain

  • Adapter Engine -> Technical System of AE
  • Integration Directory -> Technical System
  • ESR -> Technical System.
  • Integration Server -> Business System of IS -> Business Systems that exchange Messages in this Domain -> Technical Systems of Business Systems -> Technical System of IS

If the last association is missing the Integration Server has no association to a Business System or the Business System has no association to a technical System. Then the PI domain will not get copied completely into LMDB.

Proposed Solution:

Check the PI domain in the PI SLD to make sure it is correct there. Then follow SAP note 1631346 - Solution Manager 7.1: SLD configuration for PI monitoring to synchronize the PI domain fully to central SLD and LMDB. Please also refer to: PIMon_FAQs#HowcanIsynchronizePIDomaindataintoLMDBofSolutionManager

Back


 

ABAP Proxy is wrongly displayed as Business System in the scenario configuration

When you create the PI scenario you see an ABAP system with an ABAP proxy showing up under the business systems section. 

Reason:

Based only on SLD or LMDB data, we cannot distinguish between “ABAP Proxy” (which sends PI Messages) and an “ABAP Business System” (a system connected via ALE/IDoc).
The PI itself does not distinguish, but Solution Manager needs to know, if it needs to schedule an extractor for PI messages or not. This extractor is only scheduled for ABAP proxies. During the creation of the PI domain we determine whether a system is an ABAP proxy or not by checking a local PI parameter in the system. If this parameter is set, it is an ABAP Proxy. If this parameter is not set (or cannot be checked), it is a Business System.

If the ABAP proxy shows up as Business System unexpectedly it most likely means the parameter couldn't be checked, e.g. if the RFC Destination is not working. In this case Solution Manager handles the system as ABAP Business System, because it cannot collect message data anyway (since the RFC is not working).

Proposed Solution:

Please check in Solution Manager the READ RFC Destination for the client XXX of the ABAP System which is displayed as Business System. Please check if the ping and the authorization test are successful.

Back


 

Load Metadata

HTTP Code 307 when executing Configuration Step "Load Metadata"

In PI Monitoring Setup step "Load Metadata" there are several functions executed, that read data via Web Service Execution form the Integration Server. There can be exceptions in the log, after executing this step. In the log details you see for all activities the same error message:

SOAP:1.007 SRT: Unsupported xstream found: ("HTTP Code 307 : TEMPORARY REDIRECT")

Proposed Solution:

Please check in Managed System Setup for the Technical System of the Integration Server the entries in Step "Enter System Data". In case of HTTP 307, usually the Message Server Port as Load balancer Port is entered. The entries for Loadbalancer Host and Port are used to create the URLs for HTTP based communication. Also to create logical ports for WS communication, the endpoint URLs of web services is created based on these data. When the WS Call is sent to the Message Server port, it gets redirected to an instance. These redirects are blocked in the WS Runtime for security reasons.

The only possible workaround is the usage of a Webdispatcher or the usage of a http(s) port of any instance in the Managed System setup step. If these data are changed, the step for logical port creation in the automatic setup steps needs to be re-executed.

Then the errors will disappear in the PI Monitoring setup.

Back


 

No logical port <XXX> exists for the proxy class <XXX>

During the step "Load Meta Data" in the PI Monitoring setup you receive the error message that the logical port is missing.

Proposed Solution:

Make sure the managed system setup for the PI system is completed successfully. The logical ports are created during the managed system setup in the activity "WEB Service Logical Port Creation" in the step "Finalize Configuration".

If this activity fails check the log to find out why. A common problem is that the user SM_COLL_<SID> is not created correctly in the previous step.

Once the activity ran successful the log will contain the logical ports created.

You can also check the logical ports directly in SOAMANAGER as described here: PIMon_Checks#CheckifthelogicalportstoPIexistinSOAMANAGER

Back


 

Error: SoapFaultCode.5javax.ejb.EJBException

During the activity Load Meta Data with SAP Solution Manager 7.1 SP 07 the error: SoapFaultCode: 5 javax.ejb.EJBException: (Failed in component: sap.com/com.sap.xi.directory.webservices640, BC-XI-IBD) ... occurs.

Proposed Solution:

The user SM_COLL_<SID> created in the previous step doesn't have the correct authorizations.

Please assign the following Java Roles:

  • SAP_XI_API_DEVELOP_J2EE
  • SAP_XI_API_DISPLAY_J2EE

Back


 

Error: No Logical Port defined in EXTRACTORCLASS

The activity Load Meta data finished with the error message: "Unexpected exception while initializing PI domain <domain>". In the error details you find errors like this:

Reason:

The reason for the errors is that the PI domain, e.g. domain.24.sappi6pv, does not contain an Integration Directory. We need to use the connection to the Integration Directory to read the master data of channels, iFlow and integrated configurations. Since the PI domain does not have an Integration Directory, the connection cannot be found. The root cause of the problem is, that there is an PI Domain in LMDB, that has an association to an Integration Directory, but no further association to a technical system (e.g. I6P).

Proposed Solution: 

Check that your PI domain is correct in SLD. Resynchronize the PI domain to LMDB as described in PIMon_FAQs#HowcanIsynchronizePIDomaindataintoLMDBofSolutionManager and check if it is correct in LMDB now as described in PIMon_Checks#CheckifthePIdomainandallthecomponentsareavailableinLMDB.

Once the PI domain is correct go to the step "Define Scope" in the PI Monitoring setup and open the scenario with "Maintain Scenario". Click thru the wizard and save. This will fix the scenario.

Back


 

ERROR RFC call: 2

You get an error "ERROR RFC call: 2" during the step "Load Meta-Data".

Reason:

This happens due to an authorization problem.

Proposed Solution:

Please follow the instructions of SAP Note 1572183 and update your profiles of the Solution Manager RFC user in your managed PI system.

Back


 

Error: DESTINATION_NOT_FOUND

During the step Load Metadata you get the error: Unexpected exception while initializing PI domain <domain>.
In the message details you find the following error messages:

PI iFlow Attributes

 HTTP Error code 002 with message: DESTINATION_NOT_FOUND   

PI iConf Attributes  

HTTP Error code 002 with message: DESTINATION_NOT_FOUND  

Reason:

The error code occures, the type G RFC destination E2E_SM_DIRQUERY_<SID of central Adapter Engine> is missing.

Proposed Solution:

The missing destination is usually created during the Managed System Setup for the Java stack of the Integration Server in step "Configure Automatically" with activity "Create Logical Ports". Please rerun this step/activity.

Back


 

SOAP:1.023 SRT: Processing Error in Internet Communication Framework

During the step "Load Meta-Data" you receive the error:

SOAP:1.023 SRT: Processing error in Internet Communication Framework: ("ICF Error when receiving the response: HTTPIO_PLG_CANCELED")

Check the ICM log via transaction SMICM and check if there are any obvious errors. HTTPIO_PLG_CANCELED usually indicates a problem with the SSL configuration or with missing, expired or invalid certificates.

Correct the SSL configuration and make sure all certificates are correct and valid before continuing with the configuration.

Back


 

Message Alerting Setup

No Alert Rules are available in Solution Manager

You want to set up message-based alerting in SAP Solution Manager. When you come to the step "Message Alerting" where you would select the alert rules from PI. But in this step you don't see any message alert rules.

You check in the PI system and there the alert rules are available.

Check 1: Check if the release of the PI system is equal or higher than PI 7.31

Message-based alerting is only supported for PI systems with release 7.31 or higher. Please check the release of your PI system as described here: PIMon_FAQs#WhatPIreleaseisrequiredtousePIMonitoringwithSolutionManager

Check 2: Check if all logical ports for the system exist in Solution Manager

The logical port needed for the collection of the messages is located in consumer proxy CO_E2E_PI_ALERT_RULE_IN. Check if the logical port exists as described here: PIMon_Checks#CheckifthelogicalportstoPIexistinSOAMANAGER

  • If the logical ports exist then perform the next check.
  • If the logical ports do not exist then you can either rerun the activity "Create Logical Ports" in the managed system setup of the PI system in step "Finalize Configuration". Or you can create the logical ports using function module "E2E_WS_CONN_LOG_PORT_CREATION" in SE37.

Input:

    • SYSTEM_ID = <SID of the MANAGED SYSTEM>
    • CLIENT = <blank> SYSTEM_TYPE = JAVA
    • USERNAME = SM_COLL_<solMan SID>
    • PASSWORD = <password of SM_COLL_<SID> User >
    • CHECK_ONLY = <blank>

NOTE : Set the flag for "Upper/Lower cases" in SE37 transaction (to execute the above FM) because Password usually contains small letters

Check 3: Check if the Webservice ping is successful after creating the logical ports in transaction SOAMANAGER as described here: PIMon_Checks#CheckifthelogicalportstoPIexistinSOAMANAGER. If the Webservice ping is not successful double-check that the user SM_COLL_<SolMan SID>. Check that it is not locked and the password is correct and that has the required authorizations as described here: PIMon_Authorizations.

Back


 

Error "SoapFaultCode.5 Authentication failed."

While setting up message alerting for PI you get an error message: “Error in Message Alerting Web Service for domain <domain>: SoapFaultCode:5 Authentication failed. For details see log entry logID=<log ID> in security log.”

Check 1: Does the setup user exist in both systems?

For the message-based alerting setup SSO is used to connect to the PI system with the setup user. Solution Manager registers itself as consumer for an alert rule, hence changes a configuration in PI. The PI admin should be able to see who changed some configuration, which would not be possible with a generic anonymous user. For SSO to work, the user used in the setup has to exist in both systems.

Check 2: Does the user have all required roles in the PI system?

Check the link to make sure the setup user has all the necessary authorizations: PIMon_Authorizations

Check 3: Is there any SSO configuration issue?

This error message also could indicate an SSO setup issue. The SSO setup is done during the managed system setup for the system in step "Finalize Configuration". Check the log of this activity for any errors in SOLMAN_SETUP. If there are any errors also check the following link for SSO setup issues: SSOIssueonABAPandJavaPIComponent.

Back


 

Runtime

Authentication and Authorization Issues 

PI Message Alerting - /SDF/E2E_PI_MESSAGE_ALERTING User not authorized.

You are using PI Component-based Message Alerting. There are no alerts for failed PI messages on the different PI components. The DPC Pull Extractor for the relevant technical system and client of the PI component shows an error:

Data Provider /SDF/E2E_PI_MESSAGE_ALERTING : User SM_xxx is not authorized

The xxx in SM_xxx contains the System ID of SAP Solution Manager.

The error message can be found in the Solution Manager Administration Workcenter for Alerting Framework:

Proposed Solution:

Please logon to the related client of the managed system and assign the role SAP_XI_ALERT_CONSUMER to the user SM_xxx (this is the user used in the READ RFC Connection from SAP Solution Manager into the client of the Managed System).

Back


 

Authentication Issues in PI System of type ABAP for Manage Functions.

You are using the Manage function (Ping or Selftest) in the Component Monitor and you get a result with an authentication error in the manages system. The result in the monitor looks as below:

 

 

All Manage functions are executed directly at the Managed System. Therefore an authentication at the Manages System is required. For all PI Components running on Managed systems of Type ABAP, we are using the TRUSTED RFC connection, created during Manages System Configuration. This is using SSO, means we are logging on with the current user name working in PI monitoring ins solution Manager.

Proposed Solution:

Please check, if the user name, that you are suing in SAP Solution Manager for executing the Manage function is also existing in the Managed System of the PI Component. To allow the logon as trusted user (SSO), an authorization for authority object S_RFCACL must be assigned to the user in the managed system.

For details regarding the usage of the authorization object S_RFCACL, please check SAP note 128447.

Back


 

Authentication Issues in PI System of type JAVA for Manage Functions.

You are using the Manage function (Ping or Selftest) in the Component Monitor and you get a result with an authentication error in the manages system. The result in the monitor looks as below:

This text appears in the upper section of the WD Screen, which is reserved for exceptions.   All Manage functions are executed directly at the Managed System. Therefore an authentication at the Manages System is required. For all PI Components running on Managed systems of Type JAVA, we are using the ticket based authentication. The SSO is configured during Managed System Configuration automatically.

Proposed Solution:

Please check, if the user name that you are using in SAP Solution Manager for executing the Manage function is also existing in the Managed System of the PI Component. To allow the logon as trusted user (SSO), both systems need to use same user names.

Additionally the called applications need to accept EvaluateAssertionTicketLoginModule. The normal EvaluateTicketLoginModule is not sufficient, since the Manage functions are implemented as ABAP Coding executed at the backend. This means the SSO is established between Solution Manager Backend and Managed System (not between Browser and Managed System), which requires EvaluateAssertionTicketLoginModule (instead of or additionally to EvaluateTicketLoginModule).

Login Modules are assigned automatically during Managed System Configuration. You can check this in the Log of this configuration step of directly in the Managed System.

To check locally in the Managed System, please go to NWA -> Configuration Management -> Security -> Authentication and Single Sign-On -> Authentication and enter the application and check the assigned Login Modules:

Back


 

Missing Authorization on Java PI Component.

The SSO user in the managed system is missing an authorization in Java. This leads to the following exception in the Monitoring UI:

Proposed Solution:

Please check, if the user has all required authorizations ( List of Authorizations, required in Managed PI System ).

Back


 

SSO Issue on ABAP and Java PI Component.

When using Manage Functions or cUDS, you get authorization issue. This can affect ABAP and JAVA Components, but the reason is different on ABAP and Java (see Proposed Solution).

Proposed Solution:

After Manages System Configuration, the SSO should be setup properly. But there might be still issues. In ABAP and JAVA there are different things to do:

ABAP: The user in the Managed System (same user name as in Solution Manager) is missing the authorization for authorization object S_RFCACL (For details see SAP note 128447.). There is a role SAP_SM_S_RFCACL in the managed system, that contains the authorization Object. But this role needs to be adjusted and generated. Then you can assign it to the users, that need to logon via SSO.

Java: In SAP Systems, sometimes there are issues with the certificates. They are self signed by SAP and some systems do not accept this. Please logon to NWA and go to Configuration -> Trusted Systems. If there is an issue mentioned in tab SingleSign-On with SAP Login Tickets, you can press the button I agree to accept the certificates. Then the SSO should work.

Back


 

No Data collected on PI Components of Type Java (Authorization Issue).

You find in the following log entries (example), when using MAI Servlet (http://<SolManHost>:<Port>/tc~smd~agent~application~e2emai/servlet/MAI) for the agent, that is collectong the Metrics:

Statistics for          : MAIJOB:ctx=00000000000000000024:T_SYSTEM Metric Context type     : T_SYSTEM Metric Collector        : PUSH_GENERIC(implementation class:com.sap.smd.mai.model.collector.SAPGrmgCollector) Connected for           :  0 hour(s) 0 minute(s) 0 second(s), from: Jul 13, 2011 10:49:15 AM to Jul 13, 2011 10:49:15 AM       | Total execution count  : 1       | Last  average duration : 378 ms       | 1.  Elapsed:    378 ms      TimeOut: false    Err: false Last Collection data: Data: MetricRef [0050568A3F5B02EF85EC0A88C9905C0A,PI_MRT_METRIC_SELFT] Exception[0]---------------------------- Type         : EXCEPTION_TYPE_COLL_RUNTIME Message      : Error processing HTTP request: HTTP [401] : HTTP/1.1 401 Unauthorized {{------------------------------------------------ }}
Data: MetricRef [0050568A3F5B02EF85EC0A87F2B49C08,PI_AE_METRIC_SELFT] {Exception[0]---------------------------- Type         : EXCEPTION_TYPE_COLL_RUNTIME Message      : Error processing HTTP request: HTTP [401] : HTTP/1.1 401 Unauthorized ------------------------------------------------
Data: MetricRef [0050568A721F02DF85BCFF4D431D1572,PI_ESR_METRIC_SELFT] Exception[0]---------------------------- Type         : EXCEPTION_TYPE_COLL_RUNTIME Message      : cannot parse GRMG response Cause        : Fatal Error: com.sap.engine.lib.xml.parser.ParserException: XMLParser: No data allowed here(:main:, row:6, col:2) ------------------------------------------------
Data: MetricRef [0050568A721F02DF85BCFF4DD5FD1573,PI_INTEGR_DIR_METRIC_SELFTEST] Exception[0]---------------------------- Type         : EXCEPTION_TYPE_COLL_RUNTIME Message      : Error processing HTTP request: HTTP [401] : HTTP/1.1 401 Unauthorized ------------------------------------------------

Usually all PI Components of type Java are affected by this issue.

Proposed Solution:

Please check the user in the Managed System of type JAVA. Usually this user is locked.

Just unlocking this user will not help, because usually the user is locked again after a view minutes. To understand the root cause for this issue, you have to understand the user management in a PI domain. Usually the Users are managed centrally in the ABAP stack of the integration server. If you create a user there, this user is distributed to all systems in the PI Domain.

This happens also to the user SM_COLL_<SolManSID>, that is usually created for the Java Stack of the Integration Server and then distributed to all Adapter Engines. This user creation is initiated during Managed System Configuration in SAP Solution Manager for the Integration Server.

If you then start Managed System Configuration for any of the Adapter Engines, and you did not specify, that the users are managed in a remote ABAP stack, you may create the same user SM_COLL_<SolManSID> again, which means, you actually change the existing user centrally in the ABAP Stack of Integration Server. If you specify a different Password or let the system generate a Password, the PAssword saved for Integration Server in Solution Manager will not work anymore. After a view tries of Logon from Solution Manager or from Diagnostic Agent to the Integration Server, the user will be locked. In this case also the Adapter Engine cannot logon any longer. You can imagine the complexity of this issue in case you have a larger number of Adapter Engines.

Since you cannot stop the data collection centrally and correct the issue, it is hard to fix. The only way is to change to a new user with a different name. The following section describes first, how to correct the issue and then how to avoid it at all.

How to fix this issue:

The only way to fix this issue is to create a new user on all affected Managed Systems. This can be done in Managed System Configuration in Step "Create Users". Just change the User Name of the SM_COLL_XXX User to a new value and press the execute button.

How to avoid the issue:

The issue can be avoided in general, if you know and pay attention to the user handling in the PI domain. If you know, that all user on the Adapter Engine are managed from the ABAP stack of the Integration Server, please set the following flag in the Managed System Configuration of the decentral Adapter Engines and enter the System Data of ABAP Stack in the appearing entry fields:

 

Pick the SID and client of the ABAP Stack and enter the Admin user credentials here.

The user is then always checked centrally and will not be created again from Solution Manager.

Back


 

PI Monitoring Alerts in the Alert Inbox with "HTTP [401] : HTTP/1.1 401 Unauthorized"

In the alert inbox for PI Monitoring you find alerts for PING and self-test metrics that have the error "HTTP [401] : HTTP/1.1 401 Unauthorized" in the alert details.

Please refer to SAP note 1838687 - HTTP 401 PI PING and SELF TEST.

Back



Unauthorized Error in Overview Monitor and Message Monitor

 You get an "Unauthorized" error when you open the PI Monitoring Overview or the PI Message monitor. Also all message counts are on 0.

Reason:

In sufficient authorizations for the logon user.

Proposed Solution:

Assign the following roles to the login user:

  • SAP_SM_BI_BILO
  • SAP_SM_BI_DISP

Back



Missing Data in the PI Monitoring UI and Alert Inbox

Grey LEDs for Java Components

After successfully setting up PI Monitoring you notice grey LEDs for the Java Component monitor in the PI Monitoring Overview.

There are several possible reasons for this issue. Please follow the single checks.

Check 1: Are the metrics active?

Check the tool tip: A) If the metric is inactive you will get a tool tip as below

Resolution: In this case, go to PI Monitoring setup and re-run the setup.B) If the metrics are active you will get a tooltip as shown belowIn this case, perform the next check

Check 2: Are there error messages in the activation logs in the final step of PI Monitoring setup in SOLMAN_SETUP?

For more detailed logs please also check SLG1.

  • Go to transaction SLG1 in your Solution Manager system and enter the details as shown in the following screenshot.
  • Then click on execute and you can see detailed logs.
     

Check 3: Is the SMD agent connected?

Check if the SMD agent is connected and has the correct role assignment as described here: http://wiki.scn.sap.com/wiki/display/TechOps/PIMon_Checks#PIMon_Checks-CheckiftheDiagnosticsAgentisassignedandrunning

Check 4: Check the metric in Alerting Directory Browser

Open the Alerting Directory Browser via transaction MAI_TOOLS -> Alerting Directory Browser.

Navigate to the metric and check if it is activated. Note down the metric ID (needed for the next check)

Check 5: Is the SMD agent reporting the metric?

At first you have to check if the metric is collected by the SMD agent. The metrics collected by the SMD agent are maintained in an XML file that is generated/updated during the monitoring activation. To check this XML file open the E2E MAI Support Servlet via transaction MAI_TOOLS -> E2E MAI Support Servlet.

Select the correct agent and click the "Refresh" button. Then switch to "Show Current Configuration"

Then search for the metric name (you noted down in the step before)

If the metric does not exist, check the activation logs in the final step of PI Monitoring setup. Check if there is any error. Please check the time stamp of the XML (see screenshot below) and the time stamp of the last PI Monitoring activation in the Complete Setup step. Check if there is any mismatch. The XML is generated in the activation step and if there is a mismatch this means nit most likely is not up-to-date.

If you find the metric switch to the view "Show Collection Statistics" in the E2E MAI Support Servlet and search for the same metric.

If you don't find the metric check the view "Show Last Collection Errors" in the E2E MAI Support Servlet. Seach for “com.sap.smd.mai.model.collector.SAPGrmgCollector” to check if there is any collection error. In case of channel data search for “com.sap.smd.nw.mai.pi.ChannelMonCollectorBuilder”.

If the problem is not solved even after following all the checks, please raise an OSS message on the component SV-SMG-MON-PI.

Back 


 

Monitoring for PI Channels shows 0 for Green Channels or only a Grey Rating

You set up PI Monitoring for PI channels. Even though channels are running successful the monitoring shows 0 for the green LED or even just a grey LED.

Reason:

When the channel data is collected by the data collector on the PI system, per default only channels in error status (any not successful status) are collected. So successful channel runs are not reported to Solution Manager.

You can change this setting and then all channels will be collected. Remove the flag in the setting and re-activate the content in the final step.

The grey LED shows channels that are either stopped or "not administered".

If you have channels in error status but still see only zeros, you have to investigate further.

Proposed Solution:

If you see all zeros and are not sure if this is correct (because of the above setting) or if something is wrong you have to check the data collection for the channel monitor. The data is collected by the SMD agent. To check if the data is collected successfully please follow the checks in PIMon_Checks#ChecktheMetricConfigurationonSMDAgentLevel.

The below screen shot is an sample for a case where the data collection is successful but no data is reported as there are no channels in error status.

Back 


 

No data is reported for the Message Monitor

You have set up PI Monitoring and also want to monitor PI messages via the Message Overview monitor. In the monitoring UI all counters for the messages show 0.  

Also in the Message Monitor you don't see any data and the result of the latest data collection shows errors.

Check 1: Check if UnauthorizedErrorinOverviewMonitorandMessageMonitor applies to you.

Check 2: Check the Extractor Framework Administration for Extractor Errors

a) When you check the PI MESSAGE FLOW MONITORING extractor in the Extractor Framework administration UI you see the extractor is in error state.

As the error message of the extractor suggests the logical port might be missing.

Proposed Solution:

Check if the logical ports were created correctly for this system as described here: PIMon_Checks#CheckifthelogicalportstoPIexistinSOAMANAGER

b) You check the PI Message* extractors and there is no error. The extractors show green LEDs.

Scroll down and check the frequency of the extractor runs:

 If the interval is not 15 minutes as configured, we have an issue.

Solution 1: The Resource Cap for the NONE RFC destination is too low

Usually the extractors are started by the Resource Manager (a background job) if resources are available. Resource Management is done for the different RFC destinations to avoid issues in the Managed systems. Therefore you can maintain a resource cap for each RFC Destination used by Extractor Framework. The Message Extractor is started locally in Solution Manager and then it checks resources for Managed Systems and starts the extractors for the different PI components. So the initial resource used is the NONE RFC destination.

In case an extractor cannot be started, the priority is increased to prefer this extractor in the next run. The message extractor has a quite high priority to begin with so this means, if it could not be started several times in the past, you should consider increasing the resource cap for the NONE RFC destination.

Follow this link to check and adjust the resource cap for the NONE RFC destination: EFWK_SETUP#ChangingtheResourceCapforRFCresources

Solution 2 (if Solution 1 didn't help): There is a too high backlog for Worklist items (WLIs) now

After increasing the resource cap there can still be a huge list of WLIs that have to be processed in the extractor manager run. The parameter "Number of worklist items" restricts the number of WLIs that can be processed in one run. The standard setting is 51. If the backlog of extractors is higher than 51 the extractors are executed by priority. To process the backlog you could temporarily increase this parameter to e.g. 500. Please note that it can still take quite some time to process the backlog. Follow the following link to change the parameter: EFWK_SETUP#ChangingtheNumberofWLIperEFWKRMRun

c) If you see any other error in the extractor log that is not self-explaining please let us know in the comments.

Back 


 

Missing Alerts in Message Alerting

You set up Message Alerting but you are not getting any notifications for the alert rule even though messages are failing and there are alerts in PI.

Check 1: Check if you find alerts in the alert inbox

If there are alerts in alert inbox but you don't get auto-notifications, please check the notification settings as described here: PIMon_Checks#CheckNotificationSetupforPIMonitoring

If there are no alerts in alert inbox, please proceed to the next check.

Check 2: Check if the alert rule is marked as active in Solution Manager:

Check that the alert rule is marked as active for monitoring and alerting and that all necessary components are included in the alert rule.

If yes, check the activation log in step "Complete Setup"

If the node is missing, go to SLG1 and check the activation logs for PI Monitoring with object E2E_ALERTING and sub-object PI_CONFIG.

Check 3: Check if Solution Manager is registered as consumer in the PI system

Follow the following check to make sure Solution Manager is correctly registered in PI as consumer: PIMon_Checks#IsSolutionManagerregisteredasAlertConsumerforanAlertRule

If not rerun the activation of PI Monitoring in step "Complete Configuration"

Check 4: Check if the metric is collected and reported correctly by the SMD agent

Use the description PIMon_Checks#ChecktheMetricConfigurationonSMDAgentLevel to find out if the metric is collected by the agent and reported to Solution Manager. Check for any collection errors in the agent log.

If none of the above checks help please raise an OSS ticket for component SV-SMG-MON-IPO-PI.

Back


 

The Status of a Channel is not updated after Starting/Stopping it from Solution Manager

You perform a Start/Stop of a channel from SAP Solution Manager monitoring UI. But the status of the channel in the UI doesn't reflect the status change. 

Reason:

The channel status data in SAP Solution Manager is collected in frequent intervals. If you start or stop a channel you need to wait until the next data collection is done. You can check the data collection time stamp in the monitoring UI.

Proposed Solution:

At first check the current data collection frequency and when the latest data collection took place. To do this click on the status symbol for the "Data Collection Details"

In the pop-up you see the frequency and the date of the last data collection. The default data collection period is 10 minutes.

If the data collection frequency is too low you can change the data collection in the channel monitoring setup.

Go to the step "Channel Availability Monitor" in the PI Monitoring setup. Select the component for which you want to change the settings and click the button "Change SAP Settings".

Change the collection period to the desired frequency by selecting a frequency from the drop-down list.

For performance reason the smallest possible frequency is 5 minutes.

Back


 

Central user-defined Message Search doesn't return Data

You use the central user-defined message search in SAP Solution Manager to find a message, but the search is not returning any data. The iser-defined message search is set up as described here: Using central user-defined message search

At first make sure that your user has authorizations in the PI system to perform a payload-specific message search.

Check if the Message Monitor in the PI system would have returned something with the same search criteria:

Switch to the "Advanced" view to search with user-defined search:

Do the same user-defined search done in Solution Manager and check if data is returned.

If data is returned in local message search in the PI system, and data is not returned in SAP Solution Manager PI message search, open an OSS ticket on SV-SMG-MON-IPO-PI.

If data is not returned in the local message search in PI system, but you expect data for the search please open a ticket on BC-NWA-XPI.

Back


 

Other Exceptions and Errors

Error "EX: ERROR when executing RFC to SXMS_AGG_GET_XPI_STATUS_DELTA." in EFWK for PI Message Flow Extractor 

When you check the Extractor Framework administration you find the following error for the PI MESSAGE FLOW MONITORING extractor:
"EX: ERROR when executing RFC to SXMS_AGG_GET_XPI_STATUS_DELTA."

When you check the application log (SLG1) for object E2E_PI_MESSAGE and sub-object EXTRACTION you find erroneous logs with the error message:
"Aggregation profile XPI_SEND_RECV_01 not available"

Reason:

The profile parameter for the above indicated parameter has not been set and hence does not receive aggregated data.

Proposed Solution:

Please implement SAP note 1031773 OR execute the report RSXMB_AGG_XPI_ADMIN directly in the affected system to fill the necessary tables.

Back 



Exception of Data Collector SAPPIMessageAlertCollector in Agent log.

You find in the following log entries (example), when looking in the SMA Agent log (http://<SolManHost>:<Port>/tc~smd~agent~application~e2emai/servlet/MAI) for the agent, tha collects Message Alerts from PI Java components:

[EXCEPTION]          com.sap.smd.mai.model.collector.ConfigCollectorException: Consumer queue s not available: SOLMAN_<SID of SolMan>       at com.sap.smd.mai.model.collector.LoggingHelper.createAndLogConfigurationException(LoggingHelper.java:212)       at com.sap.smd.mai.model.collector.SAPPIMessageAlertCollector#MessageReader.readNextChunk... ...

Proposed Solution:

This exception is raised by the Message alert data collector, if the defined consumer queue cannot be found. Usually this exception is raised after you initially registred SolMan as Consumer if an Alert Rule. The Error event queue is created on the PI runtime component, if there is the first error event for the queue. If no PI message in Error applies to any rule, the queue is not yet existing.

This means, you can ignore the exception, als long as there are no erroneous messages, that apply to any rule, the Solution Manager has registered for.

Back


 

 

Back

 

Central PI Monitoring Home Page

 

 

 

 

 

 

  • No labels