|
|
This chapter describes how to detect and correct problems with a Maps application. It includes the following major sections:
Table 11-1 lists the executable programs that run for each Maps application.
| Application | Polling | Motif Interface | Web Interface |
|---|---|---|---|
DLSw | cwbdlswpollerd | dlsw | |
RSRB | cwbrsrbpollerd | rsrb | cwbrsrbpollerd |
APPN | appnpoller | appn |
This section describes some general DLSw and RSRB troubleshooting procedures for initial problem resolution, and then lists specific symptoms and possible remedies.
If the RSRB or DLSw application fails to operate correctly, first perform the generalized diagnostic tasks, described in the following sections, as the root user.
In the runrsrb or rundlsw script, verify that the environment variables are set according to the values in Table 3-2 in the "Installing and Configuring CiscoWorks Blue Maps" chapter.
Use the showserver command in the $SYBASE/install directory to verify that the Sybase server, named during installation, is active. If the showserver command does not work properly, issue the following command to verify that the Sybase server, named during Sybase configuration, is active.
# ps -ef | grep sybase_server_name
You can also use the showserver command to find out which user started Sybase (usually it is username sybase, root, or cscworks).
Use the dbutil command in the $CWBROOT/etc directory (or use the Database Check option of the Maps web browser interface) to verify the integrity of the Maps database and to check for error messages. The dbutil command syntax is described in the appendix "CiscoWorks Blue Maps Daemons and Processes."
Use the cwbdaemons command in the $CWBROOT/etc directory to verify the status of the Maps daemons. The cwbdaemons command displays the running status of each Maps daemon.
% cd $CWBROOT/etc
% ./cwbdaemons
If you still have not resolved the problem, a Maps daemon might have failed. Go on to the list of symptoms in the section "Diagnosing DLSw and RSRB Symptoms." For more information about the cwbdaemons command, see the appendix "CiscoWorks Blue Maps Daemons and Processes."
If there is an SNMP error while requesting information from the router, log on to the router and use the following command to check the current configuration:
snmp-server community read_community_string RO
The read_community_string should match the read community string used by Maps. You can check and edit this value by clicking a router icon with the right mouse button, and then selecting Edit>Modify.
The default SNMP server packet size is 1500. If the packet size is too low, then some commands may fail. If commands fail, you can raise the packet size to 8192.
You can use the Maps application menu bar to rediscover, add, or delete the router, as shown below.
If a trap occurs, but the status change is not reflected in the map, check the router configuration, as described in the "Configuring the Cisco IOS Software for CiscoWorks Blue Maps" section of the "Preparing to Install CiscoWorks Blue Maps," and use the
ps -ef grep cwbtrapd command to ensure that the Maps trap daemon is running.
Table 11-2 lists symptoms that your DLSw or RSRB application might demonstrate, and it refers you to various chapter sections for diagnostic techniques to resolve the problems.
| For this symptom ... | See this section ... |
|---|---|
Cannot configure Sybase database | |
Core dump occurs | |
Device does not appear on the map | |
Global view is empty | |
Incorrect device information | |
Nodes not synchronized | |
Sybase numbers not synchronized | |
Sybase VDEVNOS error | |
Unexpected dialog box prompt appears | |
Wrong map view | |
"Could not open object class" message |
If the RSRB or DLSw map is blank, synchronization with the network management system or seed file may have failed. Ensure that network devices have been discovered by the network management system. From the RSRB or DLSw main window, select Admin> Process Control to reset the sync daemon.
If a DLSw key routers view is blank, select Edit>Key Devices from the menu bar to ensure that you have designated routers as key devices in the seed file. If the key routers view is still blank, select View>Global to display a global view.
SunNet Manager maintains a separate device database for each user. In this case, you must use the RSRB or DLSw application from the username whose database you want to synchronize.
You can also try logging in to the Sybase server to check whether the devices in the seed file are also in the database. The following procedure describes how to check the database with the isql command.
$SYBASE/bin/isql -Usa -Psybasesa
1> use SNA
2> go
1> select * from SNA.devices
2> go
You should see all devices that are in your seed file.
If a device fails to appear on the map, the discovery daemon may have failed. Perform the following steps:
Step 1 Use the isql command, as shown below, to verify the device's entry in the Sybase Devices table. This example is for a database named SNA.
%$SYBASE/bin/isql -Usa -Psybasesa
1> use SNA
2> go
1> select count(*) from snasuper.devices
2> go
If the reported count is zero, the Devices table is empty.
If the Devices table is not empty, use the following commands to see if the device's entry is in the Devices table.
1> select * from snasuper.devices where device_name=device_name
2> go
1> quit
Step 2 If the device named device_name is not in the Devices table, check for its existence in the network management system database or the seed file. If the device is in the Devices table, verify that the device is running the correct release of Cisco IOS software and that you can ping the device from the network management system. Select Edit>Modify and enter the device name and correct read community string.
Step 3 Use Edit>Add Device to add the device to the map. The RSRB or DLSw application attempts to discover a newly added device automatically.
Step 4 Use the ping command to verify that you can locate the device.
Step 5 Check whether SNMP is configured on the router and verify the read community string. SNMP must be configured on the router for DLSw and RSRB.
Step 6 Check the access list in the router configuration to verify that the IP address can be reached.
If a device appears on the map but the information about it appears to be out-of-date, use the following procedure.
Step 1 Select Edit>Rediscover Device(s) to rediscover the device.
Step 2 Select Admin>Process Control and reset the poller and monitor daemons.
Step 3 Use the dbutil command in the $CWBROOT/etc directory, or use the Database Check option of the Maps web browser, to verify that all Sybase triggers and stored procedures are in place.
If the problem still has not been resolved, ensure that both the network management system trap daemon and the Maps trap daemon are active. See the section "Running Daemons in Debug Mode."
If the nodes in your network are not synchronized with your database, try one of the following procedures.
Select all the devices on the map and rediscover them.
Step 1 Select Layout>See All.
Step 2 Select all the devices on the map, as described in the section "Selecting Multiple Map Icons" in the chapter "Using the Motif Applications."
Step 3 Select Edit>Rediscover Device(s).
Use the following commands:
% cd $CWBROOT/etc
% ./runprocess dbutil -c
The dbutil -c command removes all entries from the database tables. (For information about the dbutil command, see the appendix "CiscoWorks Blue Maps Daemons and Processes.") Rerun Maps and SNA View discovery processes to repopulate the database tables. In Maps, select Admin>Discover from the menu bar. In SNA View, select the Status/Discovery button.
During installation, you might receive this Sybase error message:
More vdevno's are needed. The CW_SYBASE server needs to be reconfigured.
Use following steps to reconfigure your dataserver:
Login to Sybase as 'sa' user
Then execute following commands:
>sp_configure
>go
>sp_configure "devices", 15
>go
>RECONFIGURE WITH OVERRIDE
>go
>shutdown
>go
Please contact your Sybase System Administrator
If you receive that message, use the following procedure:
Step 1 Use the showserver command in the $SYBASE/install directory to find out which user started Sybase (usually root, sybase, or cscworks).
Step 2 Use the su command to become that user:
#su - username
If you don't know the user's password, but you know root's password, you can become the root user:
#su root
Step 3 Log in to Sybase using the isql command:
#sybase_dir/bin/isql -Usa -Psybasesa
Step 4 Look at the current number of devices:
>sp_configure
>go
Step 5 Increase the number of devices by at least 3. Here, we increase it to 20.
>sp_configure devices, 20
>go
>reconfigure with override
>go
Step 6 Shut down and restart Sybase:
>shutdown
>go
#cd <sybase_dir>/install
#startserver -f ./RUN_CW_SYBASE
#startserver -f ./RUN_CW_BACKUP_SERVER.
Caution Shutting down Sybase may affect other running programs.
During deinstallation you may see the following error message in the SMIT log:
0518-305 odmdelete: Could not open object class sm_cmd_hdr
Check path name and permissions.
This message occurs when the links from /etc/objrepos to /usr/lib/objrepos, for the following objects, are missing.
To correct this problem, recreate links and reinstall Maps.
During installation, the following error message might appear if the Sybase numbers are out of synchronization:
Building the SNA data device...
An error has occurred while building the SNA data device. Check the file /tmp/cwb_sqll.log for more info. Exiting.
If this error message appears, do the following.
Step 1 Log in to Sybase using the following command:
#sybase_dir/bin/isql -Usa -Psybasesa
Step 2 Shut down and restart Sybase.
>shutdown
>go
#cd sybase_dir/install
#startserver -f ./RUN_CW_SYBASE
#startserver -f ./RUN_CW_SYBASE_BACKUP
You can restart any or all suspect daemons in the debug mode (using the -d option that is detailed in the appendix "CiscoWorks Blue Maps Daemons and Processes"). Read the output on your display (stdout), and report your observations to the TAC, if necessary. If you restart the monitor daemon with the -d option, you must restart any RSRB, DLSw, and APPN applications that were running.
On startup, every daemon checks for a running instance of itself. If you restart any daemon with the -f or -s option, an existing instance is automatically killed. Without -f or -s, the existing instance survives and a new instance does not start.
To see a list of active RSRB and DLSw daemons and their process identification numbers, use the cwbdaemons command in the $CWBROOT/etc directory. The cwbdaemons command is described in the appendix "CiscoWorks Blue Maps Daemons and Processes."
These error messages are used by both the RSRB and DLSw applications.
Explanation
RSRB or DLSw either cannot get to the data server or cannot log in.
Recommended Action
Check the environment variables; try to log in to the database server from a separate window and check the transaction log. Then try to restart the application.
view_name is the name of the view that you are trying to save to a file:
filename is the name of the file in which you are trying to save the view.
Explanation You tried to use File>Save As to save a map to a file, but the file prefix is one of the reserved Maps file prefixes. Maps reserves the following file prefixes:
Recommended Action Do not use File>Save As to save to a file with one of the reserved prefixes.
Explanation
A copy of the process called daemon_name is already running.
Recommended Action
Select Admin>Process Control and click the Reset button.
Explanation
RSRB or DLSw failed to connect to the network management system database before starting the discovery daemon.
Recommended Action
Verify the integrity of the network management processes. If you have HP OpenView or NetView installed, you can use the ovstatus command, which is typically located in the
/usr/OV/bin or /opt/OV/bin directory. The ovstatus command displays the status of daemons and processes. You can ensure that all network management daemons are running and restart those that are not running.
Explanation
When you selected Edit> Add, Maps detected that the device you specified already had an entry in the database.
Recommended Action
If you receive this message in error, delete the device by selecting Edit>Delete Device(s) from the Maps menu bar.
Explanation
The device_name is not known to the operating system or to the domain name server.
Recommended Action
Check to see whether the domain name server has been configured. Verify that the device name is in the /etc/hosts file.
Explanation
RSRB or DLSw failed to query network devices.
Recommended Action
If you have HP OpenView or NetView installed, you can use the ovstatus command, which is typically located in the /usr/OV/bin or /opt/OV/bin directory. The ovstatus command displays the status of daemons and processes. You can ensure that all network management daemons are running and restart those that are not running.
Explanation When switching views, for example from global view to focus view, a correct filename and format is read, but the view type in the file is wrong. For example, the map is changing to a focus view, but the view type says it is a virtual ring view.
Recommended Action Delete or move the particular map file.
Explanation
RSRB or DLSw does not know the community string for this router.
Recommended Action
Use the Edit>Modify menu item to set the correct community string.
Explanation
The monitor daemon is not running.
Recommended Action
Use Admin>Process Control to start or reset the monitor daemon.
Explanation
A copy of the monitor daemon is already running on the workstation, or some other program is using this port.
Recommended Action
If another copy of the monitor daemon is already running, this is an informational message. No action is necessary. If another program is using this port, you can edit the rundlsw or runrsrb script and change the setting of the -u monitor port option.
Explanation You are reading a map file with an incorrect format or version. This happens when a map file format changes between versions of maps and an attempt is made to try to read the incompatible format.
Recommended Action Delete or move the particular map file.
Explanation
The SNMP query to a device was not successful. Either the device is no longer reachable from the network management workstation or it is not responding to the SNMP query.
Step 1 Use the ping command to see whether the router is active.
Step 2 Use the telnet command to log in to the router to check the router configuration. Ensure that the router is configured for the protocol that you are using, that the router is configured for SNMP, and that the read community string is correct.
Explanation
You tried to start the cwbtrapd trap daemon, but the network management system daemons were not running. The network management system daemons must be running because cwbtrapd must register with the NMS daemons before it can receive trap messages.
Recommended Action
Start the NMS daemons and then restart the cwbtrapd trap daemon.
This section describes some general APPN troubleshooting procedures for initial problem resolution, and then lists specific symptoms and possible remedies.
If the APPN application fails to operate correctly, verify that the environment variables are set in the runappn script according to the values in Table 3-2 in the "Installing and Configuring CiscoWorks Blue Maps" chapter.
Go to another view, then return to this view. If the problem has not cleared up, exit the application, remove the map files again, and restart the application.
Table 11-3 lists symptoms that your APPN application might demonstrate and refers you to some diagnostic techniques to resolve the problems.
| For this symptom | See this section |
|---|---|
Cannot find LU | |
Global view is empty | |
Good connection displayed as degraded | |
Incorrect status | |
Link station is not displayed | |
Network node status is unknown | |
Wrong map view displayed |
If the global map is blank, you might have one of the following problems:
To resolve this problem, perform one or more of the following:
If the status of all network nodes is unknown to the application, the network topology agent could have failed or APPN could have been stopped in that device, or the SNMP request may have failed. Use the ping command or select Tools>PathTool from the menu bar to verify that the network topology agent is accessible. If necessary, use the telnet command to log in to the agent to start the APPN protocol.
If APPN cannot find a known LU, try the following procedure:
Step 1 Select each node on the map, one at a time.
Step 2 At each node, click the right mouse button.
Step 3 Select Get Directory from the menu.
Step 4 Repeat Step 1 through Step 3 for the next node.
If a known link station does not appear in the List TGs and Links window, use the following procedure:
Step 1 Close the List TGs and Links window.
Step 2 Select the node on which the missing link is defined.
Step 3 Click the right mouse button.
Step 4 Select Get Local Topology from the menu.
Step 5 Reopen the List TGs and Links window to see if the link is listed. Or, set the autolocaltopo variable in the cwbinit file to ALL or NN_ONLY and restart the application.
If a known good connection appears degraded (yellow), a defunct TG number may exist in the network topology agent's database. In a number of days, the network topology agent will delete the defunct number. When you start the APPN application after that deletion occurs, the connection does not appear degraded.
If the status of a port, link, or transmission group is reported incorrectly, refresh the information by using the following procedure.
Step 1 On the map, select the owning node.
Step 2 Click the right mouse button.
Step 3 Select Get Local Topology from the menu.
For more information, see the section "Displaying the APPN Popup Menu Items" in the "Using the APPN Application" chapter.
The following error messages are unique to the APPN application.
view_name is the name of the view that you are trying to save to a file:
filename is the name of the file in which you are trying to save the view.
Explanation You tried to use File>Save As to save a map to a file, but the file prefix is one of the reserved Maps file prefixes. Maps reserves the following file prefixes:
Recommended Action Do not use File>Save As to save to a file with one of the reserved prefixes.
Explanation
APPN failed to connect to the network management system database before discovery.
Recommended Action
If you have HP OpenView or NetView installed, you can use the ovstatus command, which is typically located in the /usr/OV/bin or /opt/OV/bin directory. The ovstatus command displays the status of daemons and processes. You can ensure that all network management daemons are running and restart those that are not running.
Explanation
APPN must send an SNMP MIB request to a node, but no device name is known for this node.
Recommended Action
If the displayed default device name is not correct, enter the correct device name and read community string. Your input also establishes the default name for future SNMP MIB requests. If you do not know the device name, or if the node cannot be an SNMP agent, click Cancel to terminate the request.
Explanation When switching views, for example from global view to focus view, a correct filename and format is read but the view type in the file is wrong. For example, the map is changing to a focus view, but the view type says it is a Virtual Ring view.
Recommended Action Delete or move the particular map file.
Explanation
Discovery was completed but a network topology agent was not found.
Recommended Action
Review the devices in the seed file or network management system database to ensure that at least one active APPN network node exists. Then use Admin>Discover to retry the request, or exit the application and restart it. In the Startup Query window, specify a topology agent that is known to be reliable.
Explanation
No devices were available to be discovered.
Recommended Action
Review the device list in the seed file or network management system database so that discovery has a chance of success. Then use Admin>Discover to retry the discovery, or exit the application and restart it. In the Startup Query window, specify a topology agent that is known to be reliable.
filename is the name of a map file.
Explanation You are reading a map file with an incorrect format or version. This error occurs when a map file format changes between versions of maps and you attempt to read the incompatible format.
Recommended Action Delete or move the particular map file.
Explanation
The Startup Query window is being displayed because the application has not yet accessed a network topology agent. If a valid CP name is displayed in the window, perhaps the device name is also needed.
Recommended Action
Enter a valid CP name or device name or discover a network topology agent. To prevent this dialog box from being shown in the future, specify a network topology agent in the cwbinit file or as an option with the runappn command.
Explanation
A map view requires data that is not available to the APPN application. This happens typically when you request the ports and links view but local topology is not available from the target node.
Recommended Action
If the target node can act as a network topology agent, and you know its IP address or its device name and read community string, select Get Local Topology from its popup menu. Get Local Topology collects the local topology, as explained in the section "Displaying the APPN Popup Menu Items" in the "Using the APPN Application" chapter.
Explanation
An Adjacent Nodes view is shown for a node that does not have local topology information available. Only adjacent nodes learned from the global view or from local topology collection from other nodes are displayed. Adjacent EN and LEN nodes are, most likely, not shown.
Recommended Action
If this node can act as an agent and you know its IP address or its device name and read community string, collect its local topology now.
Explanation
A network topology agent has been selected but is not returning network topology information. No global map is currently available. The application will retry this agent at the network topology polling rate.
Recommended Action
Perform the following steps in order. Stop when one is successful.
Step 1 Select Admin>Discovery to discover a new network topology agent.
Step 2 Ensure that the selected network topology agent is an APPN network node with the APPN MIB, and that the IP address or device name you specified is correct. If not, stop the APPN application and restart it with a new network topology agent.
Step 3 Ensure that the APPN protocol is active on the target agent.
Step 4 Use the ping command to see if the agent is alive and accessible.
This section presents troubleshooting techniques for the web browser.
If your web browser cannot access the web server, try the following.
If the web browser cannot access the web server, look at the web server's log files. The CiscoWorks Blue web server log files are $CWBROOT/apache/logs/access_log and $CWBROOT/apache/logs/error_log. The $CWBROOT/apache/logs/access_log file lists all accesses to the web server. The $CWBROOT/apache/logs/error_log file lists individual errors in accessing the web server.
If selecting Tools>Web Browser fails, look at the $CWBROOT/etc/runweb file and ensure that the environment variables are set correctly for your directory structure.
You can also try turning off proxy server use at the web browser. To turn off proxy use for the Netscape browser, for example, select Options>Network Preferences. Select the Proxies tab, and click No Proxies. Click OK and try to access the web server again.
If you are also running CiscoWorks CiscoConnect, we recommend that you use that web server instead of the CiscoWorks Blue Maps web server. For instructions, see the "Using the Web Browser" chapter.
If you cannot access the cwblue.html page, use the following command to ensure that the web server is running:
# ps -ef | grep httpd
To restart the web server, see the section "Starting and Stopping the http Server at the Maps Workstation" in the "Using the Web Browser" chapter.
If you are using a web server other than the web server provided with Maps, see the section "Before You Access Any Other http Server" in the "Using the Web Browser" chapter.
If there is no output from the HTTP server, or if only the "Last Update at" time appears, perform the following steps:
Step 1 Verify the execute permission of the files using the following command:
# ls -l /$CWBROOT/cgi-bin/
Step 2 Use the following command to verify the presence of the
$CWBROOT/cgi-bin/mib.txt file and to ensure that it is linked to the $CWBROOT/etc/mib.txt file.
# ls -l $CWBROOT/cgi-bin/mib.txt
Sometimes the Sybase database is inaccessible from the web browser because the permissions are not set correctly for the Sybase directory. When you connect to the maps web server, you might see the following error messages:
Database login failed
DSQUERY = CW_SYBASE
DBNAME = SNA
MASTER_DSQUERY = CW_SYBASE
MASTER_DBNAME = SNA
It may be that one of the listed variables is set incorrectly or that the processes started by the web server do not have execute permission for the Sybase directory. Try to add the permissions to the directory. For example, if you are logged in to a Solaris workstation as root, you would enter these commands:
# cd /opt
# ls -ald CSCO*
drwxrwxrwx 17 cwblue other 512 Sep 30 16:58 CSCOcb
drwxr-x--- 13 cscworks cscworks 512 Sep 5 11:13 CSCOcwh
drwxr-x--- 12 sybase cscworks 512 Sep 30 17:03 CSCOsyb
Notice that the permissions are not set correctly. You could enter these commands to set the permissions:
# chmod 751 CSCOcwh CSCOsyb
# ls -ald CSCO*
drwxrwxrwx 17 cwblue other 512 Sep 30 16:58 CSCOcb
drwxr-x--x 13 cscworks cscworks 512 Sep 5 11:13 CSCOcwh
drwxr-x--x 12 sybase cscworks 512 Sep 30 17:03 CSCOsyb
Now try the web browser again.
If you accessed the CiscoWorks Blue web server but saw some other page, perhaps another web server is already using port 80. For example, when you installed CiscoWorks Blue Maps, no other web server was running so the CiscoWorks Blue web server was configured for port 80. But later, when you started the CiscoWorks Blue web server or tried to display the CiscoWorks Blue web page, some other web server is already using port 80.
Reconfigure the CiscoWorks Blue web server to use the backup port (port 1742), and specify port 172 in the web browser's URL.
To reconfigure the CiscoWorks Blue web server, edit the file
$CWBROOT/apache/conf/httpd.conf. Find the port specification, which is on the following line in the file:
Port 80
Change the port specification on that line from port 80 to port 1742:
Port 1742
And save the $CWBROOT/apache/conf/httpd.conf file.
To specify port 1742 in the web browser, enter the following URL in the browser:
http://workstation_name/cw-blue:80/cwblue.html
Where workstation_name is the name of the workstation where the Maps web server is installed.
If you are using a web browser other than the CiscoWorks Blue web browser and you cannot display the help system, try this procedure:
Step 1 Locate the configuration file for the httpd server that you are using.
Step 2 Search for the following line:
AddType text/x-server-parsed-html.htm
If you do not find the line, add it anywhere in the configuration file.
Posted: Thu Aug 19 11:03:59 PDT 1999
Copyright 1989-1999©Cisco Systems Inc.