Tips to Troubleshoot OpManager.

Following are a few tips which may be handy to get over your initial hiccups when using OpManager.

• Starting Trouble
• Discovery
• Mapping
• Monitoring
• Alerting and Notification

Starting Trouble!

• Failed to establish connection with Web Server. Gracefully shutting down.
• Error Code 500: Error in applying the OpManager 6.0 license over opmanager 5.6 or the version upgraded from 5.0
• Can't create tables or not all the tables are created properly' error is displayed during OpManager startup.
• Error downloading client files from BE

1. Failed to establish connection with Web Server. Gracefully shutting down.

Cause 1

While starting OpManager as 'root' user in Linux platform, the server goes down with the following message "Failed to establish connection  with web server. Gracefully shutting down ..". This is because  OpManager starts its Apache Web Server as 'nobody' user and 'nobody'  group. The Apache Server may not have read and execute permissions to  access the files under <OpManager Home> directory. Hence, the  connection to the Apache Server will not be established and the  OpManager server will gracefully shut down.

Solution

• Change the value of the parameter Group in httpd.conf file found  under <OpManager Home>/apache/conf/backup/ directory.
  Group #-1 to Group nobody
•  Provide executable permission to"httpd" file available under  <OpManager Home>/apache/bin/ by executing the following command:
   chmod 755 httpd

 OpManager server starts successfully after performing the above mentioned steps.

Cause 2

  If you are using Linux 8.0/9.0 :

 In Linux 8.0/9.0, a file named libdb.so is not bundled. In earlier  versions it was bundled. This file is needed by Apache. Without this,  apache does not start in Linux 8.0. This results in the issue you are  facing.

Solution

The file has been bundled with the product and is present in the  /lib/backup directory in the latest version of OpManager. Copy it to  the /lib directory and restart OpManager.

This solution has worked for those using Fedora and Madrake Linux too.

If you continue to face the problem, then execute the script  StartWebSvr (this will be a .bat file in Windows installation and .sh  file in Linux installation) in the /apache folder of OpManager installation and send us the output.

 If yours is a Debian Linux, then check if libgdbm.so.2 is available  under /usr/lib directory. If not, you can install the stable version  of libgdmg1. Download this package from the url  http://packages.debian.org/stable/libs/libgdbmg1


2. Error Code 500: Error in applying the OpManager 6.0 license over opmanager 5.6 or the version upgraded from 5.0

Cause

License issued for OpManager 6.0 fresh installation does not work when you upgrade 5.x version to 6.x

Solution

This issue is encountered when you evaluate OpManager 5.6 and subsequently apply the new license for OpManager 6.0. Follow these steps to move the database to the new fresh installation of OpManager 6.0 exe\bin

1. Shut down OpManager. (If OpManager is running as a service, stop the service from Control Panel > Services window)
2. Download upgrade pack from this link, https://www.manageengine.com/products/opmanager/service-packs.html
3. Run the script 'UpdateManager.bat (UpdateManager.sh for Linux) in the <opmanager-home>/bin folder. This opens the Update Manager tool.
4. Click "Install" and then click "Browse" to select the Upgrade Pack file (the .ppm file that you'd downloaded).
5. Follow the on-screen instructions to apply the Service/Upgrade Pack.
6. Once the upgrade is complete, start OpManager Server.
7. Take the backup of these folders
      /OpManager/mysql/data
      /OpManager/conf/
     /OpManager/users
8. Uninstall OpManager.
9. Download OpManager version 6.0 using this link https://www.manageengine.com/products/opmanager/download.html?pro
10. Move the  folders back to the new installation in the same location.
11. Start OpManager and apply the license.

If you feel this is cumbersome, the simplest solution would be to request OpManager 5.0 license from our license team :-)

3. Can't create tables or not all the tables are created properly' error is displayed during OpManager startup.

The data tables may be corrupted. You can repair the corrupt tables. Run the repairdb.bat under \bin directory. After this, run the ReInitializeOpManager.bat script in the same directory. This will remove all the tables created. Restart OpManager.

4. Error downloading client files from BE

Cause

This error occurs when the database tables are corrupted. The corruption can happen due to improper shutdown of OpManager such as during power outages.

Solution

The database must be repaired and OpManager needs a restart. Here are the detailed steps:

1. Stop OpManager Service
2. Open a command prompt and change directory to /opmanager/bin
3. Execute RepairDB.bat/sh. This repairs all the corrupt tables.
4. After it finishes executing, run it once again to ensure all corrupt tables are repaired
5. Restart OpManager.

Discovery

• Devices are not discovered
• Devices are identified by IP Address and not host names.

1. Devices are no8t discovered

Cause

This can happen if the ping requests to device get timed out.

Solution

To resolve this, increase the ping timeout in the file /conf/ping.properties and try again.

2. Devices are identified by IP addresses and not by host names

Cause

If DNS Server address is not set properly in the machine hosting OpManager, the DNS names of the managed devices cannot be obtained from the DNS server.

The other possible reasons could be:

• the DNS Server is not reachable
• the DNS Server is down during discovery
• the DNS Server does not exist

Solution

Ensure that the DNS Server is reachable and configure the DNS Server address properly.

Mapping

1. Some of my Routers are discovered as Desktops or Servers

Cause

The devices may not be SNMP enabled or the SNMP agent in the device is not responding to queries from OpManager.

Solution

Enable SNMP and rediscover the device. Despite this, if you face issues, troubleshoot as follows:

• Do you see a blue star in the device icon on the maps? This implies that the device responds to SNMP request from OpManager. The device is still not classified properly? Simply edit the category from the device snapshot page.
• If SNMP agent is not running on the router, it will be classified as a server or desktop.You can verify this by the blue star appearing on the top left corner of the device icon for the SNMP-enabled devices. To categorize the device properly, start the SNMP agent in the device. Refer to Configuring SNMP agents in Cisco Devices for details. Rediscover the device with correct SNMP parameters.
• If the SNMP agent is running on the router and you still do not see the blue star in the device icon, then check if the SNMP parameters are properly specified during discovery. If not, rediscover the device with correct SNMP parameters.
• The router is discovered as a server or desktop if the IP Forwarding parameter of the device is set to false. To set the value of this parameter to true,
  1. Invoke /opmanager/bin/MibBrowser.bat
  2. Expand RFC1213-MIB.
  3. In the ip table, click ipForwarding node.
  4. Type 1 in the Set Value box and click Set SNMP variable on the toolbar.
  5. Rediscover the device with correct SNMP parameters.

Similarly, for switches and printers too, enable SNMP in the device and rediscover.

2. How are Servers categorized in OpManager? Some servers are classified under desktops!

Following devices are automatically classified under servers based on response to SNMP/telnet request to the devices:

* Windows 2003 Server
* Windows 2000 Server
* Windows Terminal Server
* Windows NT Server
* Linux Servers
* Solaris Servers

Following devices are classified under desktops:

* Windows 2000 Professional
* Windows XP
* Windows NT Workstation.
* Windows Millinium Home Edition
* Devices not responding to SNMP and Telnet

If any of the servers are classified under desktops, simply import them into servers. Refer to the steps mentioned to check for SNMP.

Monitoring

• Despite SNMP being enabled on the device, the dial graphs for CPU, Memory, and Disk Utilization are not seen.
• Telnet/WMI-based resource monitor is not showing any data
• WMI Monitors are not working. It always says 'error # access denied'

1. Despite SNMP being enabled on the device, the dial graphs for CPU, Memory, and Disk Utilization are not seen.

Cause

SNMP may not be enabled, or the SNMP agent is not responding to requests.

Solution

Check the SNMP configurations, rediscover the device and re-add the monitors. Troubleshoot as follows:

The possible reasons for the graphs not appearing are:

• The Resource monitors may not have been associated to this device. Associate the monitors
• Check if SNMP is enabled properly on this device. If Yes, the Agent may not have responded to the SNMP request. Check if the Agent is responding using the Mib Browser.
• If the device has just been added, wait for the first poll to happen.

  Following are the steps to troubleshoot:

  1. In the device snapshot page, scroll down to the monitors list. Click the Edit icon against a monitor. For instance, let us try the CPU Utilization monitor. Click the Test Monitor link in the resulting screen. See if the monitor responds to the test request. If it does, you will see the dial graph.

  2. If there is an error message after step#1, it can be because of the snmp request to the cpu variable getting timed-out, or the oid may not be implemented in the MIB.

  3. To confirm the reasons mentioned above, invoke the tool MibBrowser.bat present in /bin directory. Load the Host Resource mib and query the oid .1.3.6.1.2.1.25.3.3.1.2 for the device that is not showing the cpu dial.

  4. If there is a response for the query in MibBrowser, it implies that the OID is implemented and the dial not appearing can be due to snmp timeout. So, you will need to configure the snmp timeout by including the parameter DATA_COLLECTION_SNMP_TIMEOUT 15 in the file NmsProcessesBE.conf for the process 'PROCESS com.adventnet.nms.poll.Collector'. Look for the following default entry in this file:

  PROCESS com.adventnet.nms.poll.Collector
  ARGS POLL_OBJECTS_IN_MEMORY 25 POLL_JDBC true MAX_OIDS_IN_ONE_POLL 15 AUTHORIZATION true DATA_COLLECTION_QUERY_INTERVAL 120000 PASS_THRO_ALL_POLLING_OBJECTS true CLEAN_DATA_INTERVAL 999999

  Include the mentioned additional parameter. Now the changed entry will be as shown below:

  PROCESS com.adventnet.nms.poll.Collector
  ARGS POLL_OBJECTS_IN_MEMORY 25 POLL_JDBC true MAX_OIDS_IN_ONE_POLL 15 AUTHORIZATION true DATA_COLLECTION_QUERY_INTERVAL 120000 PASS_THRO_ALL_POLLING_OBJECTS true CLEAN_DATA_INTERVAL 999999 DATA_COLLECTION_SNMP_TIMEOUT 15

  5. On the other hand, if there is no response in the Mib Browser, it implies that the OID is not implemented. The vendor must be requested to implement this variable for you. As an alternative, you can associate a telnet/wmi-based monitor for this device. Delete the existing SNMP-based monitor, Click the Add Monitor link again and select telnet/wmi-based monitors.

2. Telnet/WMI-based resource monitor is not showing any data

• If you have added a Telnet/SSH/WMI based Resource monitor, check if the UserName and Password specified are correct. Click the 'Password' link to configure the correct username and password to the device.
• Ensure that you have configured the domain administrator user name and password for WMI Monitors if the device is in a domain. Configure as <domain name>\<admin user name> in the User Name field. If the device is in a workgroup, it is sufficient to configure the device username and password.
• Despite the correct user name and password, if you are still unable to see the dial graphs in Windows devices, try the following steps:
  • Open a command prompt and change directory to /opmanager/conf/application/script
  • Type cscript cpu.vbs <device name> <domain name\admin username> <password>
    If this command returns a proper output, you should be able to see the dials. If you encounter an error such as Error # Access denied, very the login credentials once again.
  • If the monitored device is Windows XP, try the following option too:
    • Go to Administrative Tools -->Local Security Policy Select Security Options
    • From the options on the right, select Network access: Sharing and security model for local accounts
    • Right-click and select Properties
    • Change the privelege from Guest to Classic.
    • Remove and re-add the monitors.

Check to see if the monitors are up.

3. WMI Monitors are not working. It always says 'error # access denied'

Cause

Login credentials are incorrect.

Solution

Follow the steps below:

1.. Verify if you have provided the domain administrator username and password to connect to the device as mentioned in the above tip. If the device is in a domain the user name should be like "domain name\administrator name".

2. If the login credentials as specified in step 1 are correct, then try associating a WMI based monitor ( preferably, a Free/Used Space in MB/GB graph ) to the Exchange Server using the Resource Monitors -> Add Monitor -> WMI based monitor -> Free/Used Disk Space in MB/GB. You should get the list of drives available in the device.

3. If step #2 does not go through, then try enabling the WMI, RPC services on the Windows system and try the same again.

4. This can also happen if the DCOM settings are not configured properly.

You can check the exact error for this when you run a vbs script from the command prompt as in

cmd> cd [OpManagerHome]\conf\application\scripts\
cmd> cscript cpu.vbs [machinename] [domainname]\[username] [password]

5. You can also try configuring the dcom settings as mentioned below:

From the Run Prompt of your Windows 2k Server, type \"dcomcnfg\" and expand the tree under Component Services -> Computers. Click on the My Computer Icon from the Icon bar and select Default Properties. Check the following:

• Enable Distributed COM on this computer
• Enable COM Internet Services on this computer
• Select the Default Impersonation Level as \"Impersonate\".
• You can also edit the COM Security settings if needed.

6. If the above 4 steps do not help, try changing the service Log-on details as follows

• Go to Windows Service UI.
• Open "Properties" dialog of the "ManageEngine OpManager" service
• Go to "Log On" tab
• In the "Log on as" option select "This Account" and enter domain name\username and password, which has rights to access WMI data.
• Save and restart opmanager.
  
  Note: This will make the tray icon and splash screen disapper.

Alerting and Notification

• Email notifications are not received
• Error! page is displayed when a profile is selected

1. Email notifications are not received

Cause

Profile may not be associated to the device, or the mail-server settings may be incorrect

Solution

Check if the notification profile is associated to the device

Check if the correct criterion is selected in the profile configuration

Ensure the mail-server settings are configured correctly

2. Error! page is displayed when a profile is selected.

Cause

The profile name may contain special characters or a space

Solution

You will not be able to delete the profile from the client in such case. So, follow the steps below:

1. Stop OpManager
2. Open the file /conf/alert.filters
3. Remove the <FILTER>...</FILTER> element containing the profile configuration.
4. Restart OpManager.

Knowledgebase

For further tips to troubleshoot or find resolutions, dig into our online knowledgebase of write to us at http://support.opmanager.com.

Copyright © 2003-2006, AdventNet Inc. All Rights Reserved.