Known issues

Workaround To import the custom SAI file into the SAI Editor properly,

1. Close the Software Application Index (SAI) Editor if it is open.
2. Navigate to the <DataFlowProbe>\runtime\probeManager\
   discoverySaiResources\saiRuntime directory, locate and remove the auto.zsai file.
3. Launch the Software Application Index (SAI) Editor again.
4. Import the customer SAI file (in this case, User.zsai).

Workaround To resolve the issue, you may use one of the following workarounds:

• Use the admin or sysadmin user.

• Go to UCMDB UI and manually set for CM integration user the tenant owner and consumers.

• Go to the UCMDB server JMX Console,

  1. Locate the listResourceTypes operation in the UCMDB:service=URM Services category.
  2. Provide customer ID (default: 1) and click Invoke.
  3. On the returned page, select Auth_ROLE_ASSIGNMENT, then select the CM user.

  4. In the Resource XML field, add the following tags:

     <xmlSecurityDomainType2Id>
        <xmlSecurityDomainType2IdAny>
           <xmlAccessLevel>FULL</xmlAccessLevel>
           <xmlTenantAssociationType>OWNED_BY</xmlTenantAssociationType>
           <name>Any Tenant Security Domain with Full permissions on 
     Owned resources</name>
        </xmlSecurityDomainType2IdAny>
        <xmlSecurityDomainType2IdAny>
           <xmlAccessLevel>VIEW</xmlAccessLevel>
           <xmlTenantAssociationType>USED_BY</xmlTenantAssociationType>
           <name>Any Tenant Security Domain with View permissions on 
     Used resources</name>
        </xmlSecurityDomainType2IdAny>
     </xmlSecurityDomainType2Id>
     

  5. Click Save resource.

Workaround To resolve the issue, configure Configuration Manager by following the instructions in How to Configure Configuration Manager 10.23 to Connect to UCMDB Server.

Workaround To use PostgreSQL 9.6 with UCMDB 10.32 (or later), follow the upgrade path below:

1. Install UCMDB version 10.30 with a supported database (for example, embedded PostgreSQL 9.4.8).

2. Upgrade UCMDB to version 10.32 (or later).

3. Reconfigure UCMDB to connect to PostgreSQL 9.6.

   For detailed instructions, see How to Reconfigure UCMDB to Use Another Schema/Database.

Workaround Stop the Data Flow Probe before restarting the UCMDB server. After the server is up, start the Data Flow Probe. The issue will not occur.

Workaround The shortcut problem is a Microsoft issue in Windows Server 2016. If you want to uninstall UCMDB server or Data Flow Probe, go to the UninstallerData folder to uninstall it.

Workaround Restart the UCMDB server and the exceptions will disappear.

Workaround In addition to the fix provided in UCMDB 10.3x, for views with huge number of nodes and relations, you may also increase the memory of the UCMDB server and the applet. For detailed instructions, see How to Increase the Java Heap Memory Used by the UCMDB UI Java Applet.

Workaround If you encounter such an issue, contact HPE Software Support to obtain new wrapper license keys, and then manually deploy the provided License keys on probe.

Workaround When LDAP authentication is enabled, users need to set the Passwords must use default policy infrastructure setting to False.

Workaround Make sure that the Solr process, a Java process that points to <UCMDB_Server_Home>/bin/jre, is stopped before uninstalling the UCMDB server.

To stop the Solr process,

1. Access http://localhost:<port>/solr.
2. Go to <UCMDB_Server_Home>/bin, and then run solr.bat on Windows and solr.sh on Linux to stop the Solr server.

3. If the Solr process is still available,

   Windows: End the Solr process in Task Manager.

   Linux: Use console/terminal to terminate the SOLR PID.

Workaround If the problem still appears, manually delete the index folder and restart the server.

Workaround In case you experience the above behavior and there is no other solution that can be used:

1. Go to the UCMDB JMX Console, locate the setGlobalSettingValue JMX method in the UCMDB:service=Settings Services category.

2. Provide the following parameter values:

   name: pbv.max.perspectives.to.start.use.remove.on.update

   value: Specify the number of perspectives used in the perspective based view which crashed the system.

3. Click Invoke.

4. Restart the UCMDB Sever.

Workaround This is a Windows OS platform behavior and it is not linked to Solr process or UCMDB. However, you can restart the UCMDB server, which will then start the Solr, and there will be no command line visible in Windows.

Workaround When the attribute definition is overwritten in a descendent class, if you want to delete the attribute from the parent class, do the following:

1. Remove all attribute overriding from the descendent classes.

2. Delete the attribute from the parent class.

Workaround None.

Workaround In order for all changes in Global Settings that require the server’s restart to take effect, you must restart all nodes from the cluster (one by one or with the restartCluster JMX method).

Workaround In case you encounter an internal error when trying to launch the UCMDB UI, delete the C:\Users\<user>\AppData\Local\Temp\UcmdbAppletJars folder and try to launch the UCMDB UI again.

Workaround If you did not install any PostgreSQL database before, you can simply click Ignore to proceed. It is just a warning message that has no impact on the usage of UCMDB server and probes. The UCMDB server and probes can be normally used.

Workaround Select English or any other supported language provided in the installation wizard.

Workaround If the log folder is missing (for example, it was deleted accidentally), before starting up the UCMDB server, make sure you create the log folder manually.

Note: If you want to delete the logs, delete the content in the log folder only, and never delete the folder itself. Make sure the <UCMDBServer>\runtime\log folder always exists.

Workaround To resolve this issue, follow these steps:

1. Go to the UCMDB server JMX Console.
2. Locate the rebuildModelDBSchemaAndViews operation in the UCMDB:service=DAL services category.

3. Enter the following parameters:

   • customer ID. Enter your customer ID (default: 1).
   • previewMode. Select False.
   • skipIndexRebuild. Select False.
   • skipColumnsUpdate. Select False.
4. Click Invoke.

Workaround The SOLR version in UCMDB 10.31 is upgraded from 5.2.1 to 6.2.1. When you upgrade UCMDB from version 10.31 (or earlier) to version 2018.05, the Solr version used in old UCMDB does not upgrade automatically, and the configuration files do not match anymore. Therefore, to resolve the issue, do one of the following depending on the Solr edition in your environment:

• Embedded Solr. After the upgrade, make sure you manually remove the search folder under the <UCMDB_server> directory.

• Standalone Solr.

  1. Upgrade Solr to version 7.2.1.
  2. Manually remove the search folder under the <UCMDB_server> directory.

Workaround None.

Workaround Close all UCMDB Server folders and files before the upgrade, and DO NOT open or access any of those folders and files during the upgrade.

In case of such upgrade failure, to restore the server, copy the entire content of the C:\hp\UCMDB\UCMDBServer\old folder into the C:\hp\UCMDB\UCMDBServer folder. Then you can continue to use the server or perform another upgrade.

Workaround After upgrading UCMDB to version 10.22 (or later), manually re-deploy the AutoDiscoveryInfra package.

Workaround After upgrading UCMDB from version 10.21 to 10.22 (or later), and you have created and changed the database schema, perform the following to align the password from the database with the one from the super integration credentials file. This requires calling two JMX methods.

1. Reset the DB password.

   1. On the UCMDB server, go to JMX console > UCMDB-UI:name=UCMDB Integration > setCMDBSuperIntegrationUser.
   2. In the User Name filed, enter UISysadmin, and in the Password field, enter the desired password value.
   3. Click Invoke

2. Reset the super integration user credentials file.

   1. Go to JMX console > UCMDB:service=Authorization Services > resetPassword.

   2. In the Value box for customerID enter 1, in User Name field enter UISysadmin, and in the Password field, enter the password value to match the one you entered in step 1.

   3. Click Invoke

Workaround In order to find instances of the Tags business service attribute when searching, you must change the indexer and ranking configuration in the JMX console.

• In UCMDB:Service=Topology Search Services. Method: editIndexerConfiguration, for the business_element class, the Tags attribute must be lower case in order to be searchable.
• In UCMDB:Service=Topology Search Services. Method: editRankingConfiguration, the Tags element should be tags in order to be searchable.

Workaround Reinstall the latest Content Pack manually after the upgrade.

Workaround Invoke the JMX method updateClassModel to upgrade custom class models. To do so,

1. Launch the Web browser and enter the following address:

   http://<machine name or IP address>.<domain_name>:1977/

   where <machine name or IP address> is the machine on which the Data Flow Probe is installed. You may have to log in with the user name and password.

2. In the Search field, type updateClassModel and press Enter.

3. Locate the updateClassModel method, in the customerID Value field, type 1, and click Invoke.

Workaround To resolve the credentials related issues, do the following:

• key.bin file. Check and make sure that the new probe uses the same key.bin file as the UCMDB server.

• Time. Check and make sure that the probe time is the same as the UCMDB server time. If the probe time is different, it may cause credentials error.

• Credentials fail on the upgraded environment.

  Install a new probe and copy specified files to override problematic files. To do this:

  1. Install a clean probe on the same machine of UCMDB server, connect it to UCMDB server. Make sure that the new probe can get credentials normally.

  2. Go to the clean probe, copy the Cmcache.bin and secured_storage.bin files from C:\hp\UCMDB\DataFlowProbe\conf\security directory.

  3. Go to the probe that has get credentials errors, override the two files with yours.

  4. Restart the probe and try to get credentials again.

• Export credentials from the upgraded environment and import them to the fresh install environment, the JMX operation fails. When the exported credential xml file contains customized class models, or the customized class models are not defined on the UCMDB server that you upgraded, you may encounter credentials import issue.

  • Remove the customized protocol from the exported credential xml file.
  • Add the customized class models on the UCMDB server manually via CIT.

  Note: Make sure that the export CIT or import CIT works well. Otherwise it may cause the workaround not working.

Workaround To fix the issue, perform either of the following:

• Deploy a new Content Pack package.

• Manually delete the <Data Flow Probe Installation>\runtime\probeGateway\maxupdatetime.txt file and then restart the Data Flow Probe.

Workaround If you use ASM, to workaround this issue, make sure you upgrade both your UCMDB Server and Content Pack to the latest versions, that is to upgrade UCMDB Server to version 10.21 (or later) and Content Pack to version 16.00 (or later).

Workaround Manually upgrade data flow probes using the deployProbePatch JMX method. To do so,

1. Launch the Web browser and navigate to: https://<server_name>:8443/jmx-console, where <server_name> is the name of the machine on which Universal CMDB is installed.

2. Go to UCMDB:service=Discovery Manager.

3. Locate deployProbePatch.

4. In the Value box for the parameter customerId, enter the <customer id>. The default value is 1.

5. In the Value box for the parameter filePath, enter the full file path of the patch file.

6. Click Invoke.

Workaround If you use Cisco VPN switches, make sure you configure one of the following:

• Exclude these VPN MACs from the arp_mac attribute of IP Address and connected interface to IP of interface with that MAC in reconciliation auto exclude.

• Exclude these VPN MACs from the arp_mac attribute of IP Address and connected interface to IP of interface with that MAC in the reconcilation rule of IP Address.

Workaround Currently UCMDB only supports users with a maximum username length of 50 characters.

Workaround Make sure you clear your web browser cache, or use the UCMDB Local Client available in 2018.05 as an alternative option. For more details, see Launch UCMDB UI as a Desktop Application from UCMDB Local Client.

Workaround Make sure to use Firefox version 51 or lower, or use the UCMDB Local Client available in 2018.05 as an alternative option. For more details, see Launch UCMDB UI as a Desktop Application from UCMDB Local Client.

Workaround Pattern-based models must be created or edited only from the model instead of from the TQL query. The same TQL query cannot be used by both the model and the view.

To fix a corrupted PBM, use one of the following two ways:

• Delete the model and then recreate it

  1. Log in to UCMDB JMX Console and URM Services.
  2. Run method listResources for Resources of type: Topology_ENRICHMENT_BUSINESS_VIEW.
  3. Locate the EnrichmentBusinessView that is mentioned in the error message and delete it. Then the model is deleted.
  4. Recreate the model from Modeling Studio using the same TQL.

• Edit the unsynchronized resources of the problematic model

  1. Log in to UCMDB JMX Console and URM Services.
  2. Run method listResources for Resources of type: Topology_ENRICHMENT_BUSINESS_VIEW and Topology_TQL.
  3. Retrieve the xml definition of the Topology_ENRICHMENT_BUSINESS_VIEW and Topology_TQL of the problematic model.

  4. Open the EnrichmentBusinessView and locate the ID that is mentioned in the error message.

     <CmdbProperty>
         <Key>nodeNumberEnd2</Key>
         <Type>integer</Type>
         <Value>12</Value>
     </CmdbProperty>

  5. Do one of the following:

     • Change the missing ID in the EnrichmentBusinessView with an existing one from the Topology_TQL definition and save the resource.
     • Add the missing ID in the Topology_TQL definition by replacing an existing one and then save the resource.

     Here is an example:

     <tql:node class="node" name="Node" id="19">
             <tql:where>
                 <tql:data-stores>
                     <tql:data-store>UCMDB</tql:data-store>
                 </tql:data-stores>
             </tql:where>
     </tql:node>

     If you do not have id 12 in the TQL query, you can change it in the following way:

     <tql:node class="node" name="Node" id="12">

Workaround Switch the UCMDB Server back to HTTP protocol by changing the following setting in the \UCMDBServer\conf\settings.override.properties file to true and then restart the UCMDB server:

jetty.connections.http.enabled=true

Workaround  

• Use the UCMDB Local Client available in 2018.05. For details, see Launch UCMDB UI as a Desktop Application from UCMDB Local Client.

• If you still want to resolve the issue with Chrome, you need to enable NPAPI manually and keep Chrome below version 45 to make UCMDB UI client applet work properly. For details, see KM01656540.

  To launch UCMDB UI from Chrome version 45 (or later), follow the instructions in How to Launch UCMDB UI from Chrome 43+, Firefox 48+, Microsoft Edge, or Safari 10+.

Workaround You can deploy larger packages using the deployPackages JMX method (go to the JMX Console > UCMDB:service=Packaging Services > deployPackages).

You can also control the maximum size of the packages to be uploaded by setting the value of the infrastructure setting File upload maximum file size for archive type to a larger value (default value: 200 MB).

Workaround Check and make sure that the files you import into UCMDB are using the same encoding as defined in the XML. For example, if you export the XML files in UTF-8 encoding, then you should import files encoded in UTF-8 as well. The problem would occur if you import files encoded in UTF-8 BOM and the XML content specifies it is UTF-8.

Workaround In case the JMX console Quick Search does not work on IE, do either of the following:

• On IE, go to Tools > Compatibility View settings. If the Display intranet sites in Compatibility View setting option is selected, clear the checkbox.
• Use Firefox as an alternative web browser.

Workaround If UCMDB server cannot start Solr for some reason after the server starts, just stop the UCMDB server, start Solr manually and start the UCMDB server again.

Workaround If the problem appears, modify the wrapper.conf file as follows:

1. Open the <UCMDB_Server_Home>\bin\wrapper.conf file using a text editor.

2. Uncomment the line with wrapper.java.additional.46, and specify a server name or simply use the localhost name.

   To find the localhost name in console, simply enter hostname.

Workaround In order to change the Linux Max Open files, you will have to increase the maximum number of open files system-wide and per process of a specific user:

• System-wide:

  Set system-wide file descriptors (FD) limits by editing the /etc/sysctl.conf file, so that after reboot the setting will remain as it is:

  1. Locate and open the /etc/sysctl.conf file.

  2. Append a config directive as follows:

     fs.file-max = 100000

  3. Save and close the file.

  4. Users need to log out and log back in again for the changes take effect, or just type the following command: # sysctl -p

• User Level FD Limits:

  After setting system-wide FD limits, you can still limit httpd user (or any other users) to specific limits by editing the /etc/security/limits.conf file.

  1. Set httpd user soft and hard limits as follows:

     httpd soft nofile 4096

     httpd hard nofile 10240

  2. Save and close the file.

To view limits, enter:

# su - httpd

$ ulimit -Hn

$ ulimit -Sn

To check the limits of a process:

1. Get the PID number by entering:

   - #ps aux | grep <process-name> #

2. Check the limits of a process:

   - cat /proc/<PID>/limits`

Workaround Manually set the network interface in the wrapper.conf file when you have more than 1 network interface. To do so,

1. On a UCMDB server instance in the HA cluster, open the c:\hp\UCMDB\UCMDBServer\bin\wrapper.conf file in a text editor.

2. Add a wrapper.java.additional.nn entry into the file, which looks similar to the following:

   #wrapper.java.additional.27=-Djava.compiler=NONE
   
   #wrapper.java.additional.28=-Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=5005
   wrapper.java.additional.30=-Djgroups.bind_addr=<ip-address>   

   where <ip-address> is the IP of the right interface to be used.

3. Save the file.
4. Repeat the above steps on each of the remaining UCMDB servers in the cluster, and use for each the right IP/Interface assigned for that server.
5. Restart all UCMDB servers.

Workaround The problem can be resolved by improving the performance of the Garbage Collector by using G1 algorithm for the Garbage Collector. For enterprise environments where you have more than 16 GB of RAM allocated just for UCMDB (see the wrapper.java.maxmemory property value in the ..\UCMDB\UCMDBServer\bin\wrapper-platform.conf file), check and make sure you add the following settings to the wrapper-custom.conf file if they are not present:

#Enable the following parameters for JVM G1 garbage collector 
in enterprise environments
wrapper.java.additional.54=-XX:+ParallelRefProcEnabled


wrapper.java.additional.55=-XX:G1HeapRegionSize=32
wrapper.java.additional.56=-XX:InitiatingHeapOccupancyPercent=70       

Workaround Create a joinf link in the TQL query between the concerning two CI Types by joining them on an attribute value. This way the enrichment manager will know how to perform correctly the enrichment actions.

Workaround The first time the enrichment is executed, it will have a full TQL query layout. If the enrichment rule is active and it is notified of a change, it will run on a partially TQL query layout. This is designed so for performance reasons.

To workaround the issue, you may take either of the following:

• Run enrichments on a scheduled notice. Nowadays it is not recommended to have many active enrichments.

• Split the enrichment into two parts, one does the CREATE of new CIs which will run periodically, one does the UPDATE which is active.

Workaround To resolve the issue,

1. On the probe machine, open the java.policy file (in the <DataFlowProbe>\bin\jre\lib\security directory) using a text editor.

2. Add the following permission:

   javax.management.MBeanTrustPermission "register"

3. Save the file.
4. Restart the probe.

Workaround For successful dispatch, make sure either of the following:

• When naming a data flow probe, do not use the same name as that of any domain; or

• Use the same domain name for a probe under the conditions that:

  • The probe is the only probe contained in the domain
  • The probe does not share the same name with any other domain

Workaround Right-click an adapter and select Go to Adapter, then go to the Adapter Configuration tab, in the Trigger Dispatch Options section, select the check box for Override default probe selection, and in the value field specify the target probe you want to use.

Workaround To resolve this issue, the easiest way to modify any IP range or credentials (such as a protocol) from UCMDB server, which will trigger the automatic upgrade process for the probes. As a result, the updated DSD/DRD files as well as the upgrade packages are downloaded to the probes, then the probes are upgraded to the latest version automatically.

Workaround To fix this issue, do the following:

1. Stop Data Flow Probe.
2. Add wrapper.java.additional.58=-XX:+DisableAttachMechanism into
   <DataFlowProbe_Home>\bin\WrapperGateway.conf and
   <DataFlowProbe_Home>\bin\WrapperManager.conf.
3. Restart Data Flow Probe.

Workaround Manually re-associate default opening programs for the three file types: Inventory Tools Viewer for .xsf files, and Inventory Tools Analysis Workbench for .aws files and .awcs files.

Workaround Do not configure too many IP ranges and always keep the total number of IP ranges below 1500 per probe.

Workaround Delete the auto.zsai file on all remote servers directly so that it will be automatically regenerated:

1. Delete the auto.zsai file on all of the remote servers.

2. Restart the XML Enricher.

Workaround The default connection timeout value of the Universal Discovery protocol is 20 seconds. If the command execution in your organization's discovery job takes some time, to ensure that the job runs successfully, increase the timeout value in both of the following places, for example, to 45 seconds:

• The Connection Timeout parameter value in the Universal Discovery Protocol Parameters dialog (go to UCMDB UI > Data Flow Management > Data Flow Probe Setup > Domains and Probes > DefaultDomain(Default) > Credentials > Universal Discovery Protocol)

• The shellGlobalCommandTimeout property value in the globalSettings.xml file (go to UCMDB UI > Data Flow Management > Adapter Management module, under Resources > Packages > AutoDiscoveryContent > Configuration Files)

Workaround To resolve the issue, perform the following:

• Always make sure that the PostgreSQL install directory is added into the anti-virus software exclusion list. The exclusion of data files will not introduce any potential security risk.

• If you need to run weekly-based scan, monitor the probeerror.log file, and if a database error shows up (for example, a database error related to the Discovery_result table), do the following:

  1. Clean the probe log folder.

  2. Run Clear Probe Results Cache from UCMDB UI > Data Flow Management > Universal Discovery > Discovery Modules/Jobs to clean the problematic table.

  This should resolve the issue.

Workaround To resolve the issue, append the missing ../lib/discovery-content-api.jar to the value for the remoteJVMClasspath adapter parameter.

Note: When upgrade a probe by reinstall, do not copy the old dataflowprobe.properties file from the old probe to the upgraded probe directly, because the values of the properties might have been changed during the upgrade.

Workaround Re-deploy the ASM_Enhanced.zip package after upgrade as follows:

1. Unzip the CP18 package and locate the ASM_Enhanced.zip package in the packages folder.

2. In UCMDB UI, navigate to Administration > Package Manager, and click the Deploy packages to server button.

3. Select the ASM_Enhanced.zip package and import it.

Workaround To workaround the issue, locate and open the hp\UCMDB\DataFlowProbe\pgsql\data\postgresql.conf file in a text editor, and then modify the default value of the statement_timeout setting from 0 to 3600000.

Workaround To modify the parameters in the service discovery activity template successfully, select a different row or click Enter to complete the change you made in the Value field of parameters. After that, click OKto save the changes. This ensures that the parameters values are saved successfully and the communication log reflects the change properly, then the new parameter value can be used in discovery jobs correctly.

Workaround The only way to update these files is to call the editResource JMX method (from the UCMDB:service=Packaging Services or UCMDB:service=URM Services category).

If you want to update IP Ranges and Credentials, you can also use the importCredentialsAndRangesInformation and exportCredentialsAndRangesInformation JMX methods. For details, see Hardening section of the UCMDB Help or JMX Reference section of the UCMDB Help.

Workaround To resolve the issue, do the following:

• If the value of initString is changed from UCMDB UI, regardless whether the probe is connected to or disconnected from the UCMDB server, you can use the setLWSSOInitString JMX method on the probe side to keep the same initString value as on the UCMDB server side, then credentials from CM can be successfully synchronized.

  To do so,

  1. On the Probe machine, launch the Web browser and enter the following address: https://localhost:8453.

     Note If the Probe Manager and the Probe Gateway are running as separate processes, the address should be entered on the machine that is running the Probe Manager as follows: https://localhost:8454.

  2. Click type=CMClient to open the JMX MBEAN View page.
  3. Locate the setLWSSOInitString method and provide the same init string that was provided for UCMDB's LW-SSO configuration.
  4. Click the setLWSSOInitString button.

• If the value of initString is changed from JMX on the UCMDB server side, and

  • the probe is disconnected from the UCMDB server, use the setLWSSOInitString JMX method on the probe side to resolve the issue. For detailed instructions, see steps above.

  • the probe is connected to the UCMDB server, invoke the setInitString JMX method on the UCMDB server side to enable automatic synchronization of the initString value.

    For instructions on how to invoke the setInitString JMX method, see JMX Reference section of the UCMDB Help.

Workaround Before re-imaging your Virtual Machine system, remove the CI instance of the old Windows/Unix system from UCMDB manually.

Workaround To work around this issue, manually set the IP range for the problematic Management Zone, and then move the Management Zone(s) to the target folder.

Workaround The probe requires at least 4G memory to run the WebSphere to Web Server Dependencies job. Therefore, allocate at least 4G memory for the probe.

Workaround Update the default credentials for basic authentication manually. To do so,

1. Open the JMX Console of the UCMDB Server side, enter changeBasicAuthenticationCredential in the quick search field and click the link that appears.

2. Specify values for userName and password.

   Note: Password must contain at least six characters.

3. Click Invoke.

4. Encrypt the password by using the JMX Console of the probe side as follows:

   1. Open the JMX Console of the probe side, enter getEncryptedKeyPassword in the quick search field and click the link that appears.

   2. Enter the password that you specified.

   3. Click Invoke and then the encrypted password is generated.
5. Copy the encrypted password.

6. Edit the DataFlowProbe.properties file as follows:

   appilog.agent.Probe.BasicAuth.User = <the user name 
   that you specified>

   appilog.agent.Probe.BasicAuth.Pwd = <the encrypted 
   password that you just copied

7. Save the change and then restart the probe.

Workaround To clear the data, run the following:

..\UCMDB\UCMDBServer\integrations\tools\clearProbeData.bat

Workaround To make the Check Network Availability permission available for the out-of-the-box role Discovery And Integrations Admin, you need to manually deploy the Basic_Authorization.zip package. To do so,

1. Copy the Basic_Authorization.zip package from the <UCMDB_Server_Home>\content\basic_packages directory to the machine from where you will deploy the package.

2. Launch UCMDB UI and navigate to Administration > Package Manager, and click the Deploy packages to server button.

   In the popup dialog, select Basic_Authorization.zip, and then click Deploy.

3. After the deployment is completed, go to Security > Roles Manager, select Discovery and Integration Admin > CIT Menu Items, then select Node CI type, the Check Network Availability action should show up in the Selected Actions pane.

Workaround To work around this issue, you can use a remote process of JVM 7 to run these jobs. For detailed instructions, see KM01746334.

Workaround Set up a DNS alias, such as UDCallHome, that in each network resolves to a local (for that region) UD probe to receive the agent call home event. This way the same initial agent call home configuration could be used throughout.

Workaround To resolve the issue, use the C:\hp\UCMDB\DataFlowProbe\pgsql\bin\pgAdmin3.exe GUI to perform DB dump. Just right click the Data Flow Probe DB, and choose backup.

For example, from the C:\hp\UCMDB\DataFlowProbe\pgsql\bin directory, execute the following command:

psql.exe --username=postgres --command 
"VACUUM  full verbose" --dbname=dataflowprobe

Workaround If some jobs run in the remote process mode, make sure that you set the value of the appilog.collectors.storeDomainScopeDocument property in the DataFlowProbe.properties file to true.

Workaround In order for the UCMDB_Probe_DB service to start properly, you may configure the Log On options for the UCMDB_Probe_DB service as follows:

1. Locate the UCMDB_Probe_DB service in the Services window.
2. Right-click UCMDB_Probe_DB and select Properties.

3. In the UCMDB_Probe_DB Properties window, go to the Log On tab, and do either of the following:

   • (Recommended) Select the Local System account option.

     Note: This is the recommended option, because the SYSTEM account has access to all necessary folders according to the default settings on Windows.

   • Select This account, and specify the account name and password.

     Note This option is NOT recommended. Even if you specify an administrator account, it will be treated as a common user account by PostgreSQL, because PostgreSQL is able to "irrevocably give up administrative rights at startup".

     If you DO need to specify a different account, make sure that the USERS group on your Windows platform has:

     • Read and Write access to the C:/UCMDB/DataFlowProbe/pgsql folder.

     • Read access to the files in system library (for example, the C:/Windows/System32 folder) that PostgreSQL needs to access.

       For the list of files that PostgreSQL needs to access, see PostgreSQL documentation.

4. Click OK.

Workaround If you want to remove one of the entries,

1. Go to Start > Run, enter regedit and click OK to open the Registry Editor.

2. Navigate to HKEY_LOCAL_MACHINE\Software\Microsoft\Windows\CurrentVersion\Uninstall, and locate the entry you want to remove.

   Note The registry path may vary with Windows operating systems, make sure you locate the entry for your OS.

3. Delete the selected entry.

Workaround To perform manual upgrade of probes on Linux, run the <DataFlowProbe Installation Folder>/tools/upgrade/extractUpgradePackage.sh script.

Workaround When you deploy a probe update manually via UCMDB UI, the maximum package size allowed is 200 MB. If the package size is greater than 200 MB, deploy it via JMX console using the deployProbePatch JMX method from the UCMDB:service=Discovery Manager category.

Also, note that the UCMDB UI client for applet requires 1280 MB memory at least.

Workaround Check and make sure that the sizes of attributes between UCMDB and RTSM/BSM are synchronized before pushing CIs with large values.

Workaround To make the connection work, add the certificate of the remote machine into the cacert file in the \DataFlowProbe\bin\jre\lib\security\cacert directory.

Workaround Whenever you encounter a similar timeout issue and see DataAccessAdapterQueryGetSupportedClassConfigs in the log, you can go to the JMX console and increase the values for the following two settings:

• task.DataAccess.Manager.getAdapterClassesConfig.timeOut. Default value for the setting is 20000.

• configuration.remote.action.timeout. Default value for the setting is 35000.

Workaround To create an integration point for both data push and population, do the following:

1. Update the value for Object Root from managed_object to root.

   1. Log in to UCMDB with an administrator account and go to Administration > Infrastructure Settings Manager.
   2. From the Filter by column drop-down list, click Name, type Object Root in the text box and double-click the Object Root entry that is displayed.
   3. In the Properties dialog box, go to the Current Value field, change managed_object to root, and then click Save.
   4. Log out of UCMDB and log in to UCMDB again for the change to take effect.

2. Add the discoverypattern_mdr_type attribute to Auto Discovery Pattern definition.

   1. Go to Modeling > CI Type Manager > CI Types pane > Root tree > Data > Object > Configuration, and click Auto Discovery Pattern definition.
   2. In the right pane, click the Attributes tab and click the Add button.
   3. In the Add Attribute dialog box, type discoverypattern_mdr_type in the Attribute Name and Display Name fields, and click OK.
3. If you already create an integration point, delete it and create it again.

4. Open the Adapter Source Editor dialog box and click Save.

   1. Go to Data Flow Management > Adapter Management > Resources pane, select the adapter that is related to your integration point.
   2. Right-click the adapter, select Edit adapter source.
   3. In the Adapter Source Editor dialog box, click Save.

Workaround When creating the new CI Type inherited from Person, change first the identification of Person by removing all attributes conditions, so the new identification of the newly created CI Type will be empty.

Workaround If you want to use the Integration Service to run a remote process, add the following jars to the basic_discovery_minimal_classpath setting in the <UCMDB_Server_Home>\integrations\conf\DataFlowProbe.properties file manually:

../lib/cryptojce.jar;../lib/cryptojcommon.jar;../lib/jcmFIPS.jar;

Workaround Use Assign Tenants functionality from the CI’s context menu.

Workaround If you want to switch back (to disable Tenant aware reconciliation), do the following:

1. Remove the ID_ATTRIBUTE qualifier for the TenantOwner attribute on the managed_object CIT.

2. Remove the value of the reconciliation.tenantaware.citypes setting.

3. Reload the class model from persistency (go to the JMX console > UCMDB:service=Class Model Services, and invoke the reloadClassModelFromPersistency method).

4. Go to JMX console > UCMDB:service=Model Services, invoke the recalculateID method with classname field empty.

5. Go to JMX console > UCMDB:service=Model Services, invoke the updateClasModel method.

Workaround To avoid removing the System Default Tenant from the Consumer Tenants list, make sure you set the System Default Tenant as consumer.

Only when the System Default Tenant is not set as consumer, the Owner Tenant will be overwritten with the System Default Tenant when trying to save.

Workaround If there are enrichments which are creating new CIs, after setting the environment as tenant aware, the attribute Owner tenant should be set for those CI Types which are being created through enrichments.