Troubleshoot > Troubleshooting Data Flow Management > Troubleshooting and Limitations

Troubleshooting and Limitations – Data Flow Probe Setup

Data Flow Probe Setup - Troubleshooting

ClosedProblem: You cannot transfer a Data Flow Probe from one domain to another.

Reason: Once you have defined the domain of a Probe, you can change its ranges, but not the domain.

Solution: Install the Probe again:

  1. (Optional) If you are going to use the same ranges for the Probe in the new domain, export the ranges before removing the Probe. For details, see Ranges Pane.

  2. Remove the existing Probe from UCMDB. For details, see the Remove Domain or Probe button in Data Flow Probe Setup Window.

  3. Install the Probe. For details, see the section about installing the Data Flow Probe in the interactive Universal CMDB Deployment Guide.

  4. During installation, make sure to give the new Probe a different name to the name given to the old Probe, or make sure you delete the reference to Probe from the original domain.

ClosedProblem: Discovery shows a disconnected status for a Probe.

Solution: Check the following on the Probe machine:

  • That the Probe is running

  • That there are no network problems

Solution: The probe status is Disconnected or Disconnected (being restarted).

  • Search for restart messages in the wrapperProbeGW logs.

  • If the probe does not restart, try to take probe thread dump from the disconnected time and search for the ProbeGW Tasks Downloader thread.

  • If there is no probe thread dump, investigate the problematic timeframe in the wrapperProbeGw log. In particular:

    • Check if the probe tasks confirmer has been running for more than 5 minutes.
    • Check if some of the resources are being downloaded for more than 5 minutes.
ClosedProblem: The connection between the Universal CMDB server and the Probe fails due to an HTTP exception.

Solution: Ensure that none of the Probe ports are in use by another process.

ClosedProblem: A Data Flow Probe node name cannot be resolved to its IP address. If this happens, the host cannot be discovered, and the Probe does not function correctly.

Solution: Add the host machine name to the Windows HOSTS file on the Data Flow Probe machine.

ClosedProblem: After uninstalling the Data Flow Probe, mysqld.exe and associated files are not deleted.

Solution: To delete all files, restart the machine on which the Data Flow Probe was installed.

ClosedProblem: After the UCMDB Server CUP is updated, the Probe fails to start or fails to connect to server

Solution: The Probe's CUP version must be the same as UCMDB Server's CUP version. If the CUP versions are not aligned, you must update the Probe's CUP version. To do this, see How to Deploy a Data Flow Probe CUP.

In some cases, the CUP may need to be deployed manually on a Probe. For details, see How to Deploy a Data Flow Probe CUP Manually.

ClosedProblem: I want to check if my integration probe is connected, but I can't see it listed in the Data Flow Probe Setup module tree.

Reason: The Data Flow Probe Setup module displays only Data Flow Probes for discovery. Integration Probes—that is, Probes on Linux machines, and Windows Probes configured for integration only— are not displayed in the Data Flow Probe Setup module.

Workaround: To see if an integration Probe is connected, create a dummy integration point and verify that the Probe is listed among the Probes that can be selected for the integration point (in the Data Flow Probe field). For details, see How to Set Up an Integration Point.

ClosedProblem: Troubleshooting PostgreSQL Issues

Solution:

The table below lists the Data Flow Probe database scripts. These scripts can be modified for administration purposes, both in Windows and Linux environments.

Note  

  • The scripts are located on the Data Flow Probe machine, in the following location:

  • Data Flow Probe database scripts should be changed for specific administration purposes only.
Script Description
exportPostgresql [PostgreSQL root account password] Exports all data from the DataFlowProbe database schema to data_flow_probe_export.bin in the current directory
importPostgresql [Export file name] [PostgreSQL root account password Imports data from a file created by the exportPostgresql script into the DataFlowProbe schema
enable_remote_user_access Configures the PostgreSQL Data Flow Probe account to be accessible from remote machines
remove_remote_user_access Configures the PostgreSQL Data Flow Probe account to be accessible only from the local machine (default)
set_db_user_password [new PostgreSQL Data Flow Probe account password] [PostgreSQL root account password] Modifies the PostgreSQL Data Flow Probe account password
set_root_password [new PostgreSQL root account password] [Current PostgreSQL root account password] Modifies the PostgreSQL root account password

 

ClosedProblem: The Data Flow Probe database service cannot start.

  • Reason: Hosts machine must not contain "localhost".

    Solution: On the Data Flow Probe machine, open

    • Windows: %systemroot%\system32\drivers\etc\hosts
    • Linux: /etc/hosts

    and ensure that all lines containing "localhost" are commented out.

  • Reason: Microsoft Visual C++ 2010 x64 Redistributable is installed during the installation of the Probe. If for some reason this redistributable is uninstalled, PostgreSQL stops working

    Solution: Check if Microsoft Visual C++ 2010 x64 Redistributable is installed. If not, reinstall it.

Data Flow Probe Setup - Limitations

  • Limitation: The Probe cluster range distribution purely handles the potential number of active or inactive IP addresses available in its network range, not the actual number of IP addresses populated or the number of devices available inside those ranges. For example, You have 100 IP addresses, and Probe cluster distributes them into two Probes: 50 IP addresses to one Probe and 50 IP addresses to the other Probe. If only one IP address is active on Probe A and 50 IP addresses are active on Probe B, Probe B would take a disproportional share of the load.

    Reason: The actual number of IP addresses are unavailable before the actual discovery.

    Workaround: If many IP ranges are unevenly populated, in order to have full control of the IP address distribution, it is best to avoid using the Probe cluster feature and simply use Probes separately.

  • When the Probe is running in separate mode on a machine where both the Gateways and the Manager share a same installation folder, the Data Flow Probe CUP must be installed manually. For details, see How to Deploy a Data Flow Probe CUP Manually.

  • Data Flow Probe CUPs that were deployed manually can be uninstalled using manual methods only. For details, see How to Uninstall Probe CUPs Manually.

  • Universal Discovery Agent may not callhome in, but not limited to, the following scenario:

    • The callhome IP address that is configured on the Universal Discovery Agent belongs to a client type range that is added to a cluster.

      Note The Universal Discovery Agent supports 1 primary and 1 secondary probe.

    • The range is a member of a probe cluster.
    • The cluster contains two or more probes.

    In this scenario, callhome may not work as expected. Contact Support for assistance in configuring callhome.

Send Help Center feedback