|
|
This appendix provides information about troubleshooting Campus Manager applications. Each application section provides FAQs and a troubleshooting table. Information about these applications is listed in this appendix:
Use the information in the following topics to help you troubleshoot Topology Services:
Use the information in these sections to answer some of your common questions:
Topology Services requires Virtual Trunk Protocol (VTP) to be enabled in order to create VLANs. For most predictable results, Cisco recommends having at least one switch configured as a VTP server and the remaining switches configured as VTP clients.
If the two VTP servers are in the same domain and are connected by VTP trunks, it does not matter which switch the changes are made on. VTP ensures that the information on all VTP servers and clients in a single VTP domain are coordinated and share the same configuration.
If the servers are in different VTP domains, then they do not share VLAN states, and they are both known to Topology Services. You must select the VTP domain in which you want to make the VLAN changes, and the corresponding VTP servers will reflect those changes.
If there are two servers with the same VTP domain that are not connected by trunks the configurations managed by the two servers may diverge. This configuration is not supported by Topology Services, and it creates a discrepancy.
You can have VLANs with the same name provided that other characteristics, such as VLAN index and SAID value are also identical. Discrepancies occur when there are identically named VLANs with other attributes that are different (such as index, and so on). If the two VLANs share identical definitions, no discrepancy is detected.
It is usually best to use the default names that Topology Services creates for each VLAN.
Yes, but you must first create the VLAN (with identical attributes) on the transparent switch. To create a VLAN that exists on both the VTP server and all the transparent switches in a VTP domain, you can select the Create VLAN on all Transparent Switches check box when creating the VLAN.
Topology Services will attempt to correlate information from VTP transparent switches with the VTP server in the same domain. It will check the ISL index, VLAN name, and VLAN type for each VLAN, if the VLANs are identical, then they will be displayed
The Aggregate Link Attributes option is available when you right-click on a link in the LAN Edge network view. The Aggregate Link Attributes window shows the list of links between a switch cloud and a router. This cannot be indicated in the map itself because individual links are not shown, there can be more than one link between a router and the devices in a switch cloud.
LE Config Servers do not belong to a specific ATM-VLAN. Different LANE components could be going to different LE Config Servers (based on neighboring LS1010 registry info) to join the same ATM-VLAN.
Topology Services displays the Cisco CallManager servers discovered by the ANI Server in the Layer 2 view and all other applicable views. For example, if the Cisco CallManager server is connected to a device that belongs to a particular VTP domain, the Cisco CallManager server is displayed in that VTP domain view.
If a Windows NT server is used as a seed device for ANI Server discovery, it will appear in Topology Services as a Cisco CallManager (CCM) server. Because a CCM server is installed on Windows NT, Topology Services discovers both of these device types as one device.
 |
Note It is not recommended to use an end-station as a seed device. Refer to the ANI Online Help for information about choosing a seed device. |
Use the information in the Troubleshooting Topology Services table to troubleshoot the Topology Services application.
| Symptom | Probable Cause | Possible Solution |
|---|---|---|
Links that appeared correctly earlier, now show as dashed lines. | If you move a port, ANI cannot determine if a link has been removed or the link is down by looking at the CDP cache. It marks the link notInNetwork instead of removing it. | If you have changed the port for this link, and the link should no longer appear, delete the link. At the next discovery, it should no longer appear. |
After initial discovery, not all of the devices are discovered, and many devices are unreachable. | If devices are not discovered by Topology Services, it is likely that community strings are not entered correctly in the ANI Server. | Check the community strings on the devices and in the ANI Server Admin application. Then verify that the devices appear in the network view. |
Topology Services will not launch, when attempts are made to launch Topology Services an error message states that Topology Services is already running. | If you close the windows that appear when the application is starting up, and then try to relaunch the application, you will receive this error. | Wait for a reasonable amount of time, and then attempt to launch the application. Do not close the message windows that appear when the application is launching. |
Use the information in the following topics to help you troubleshoot User Tracking:
Use the information in these sections to answer some of your common questions:
User Tracking is an ANI client application. ANI provides several types of global discoveries, including:
The ANI Server stores this information in the ANI database. User Tracking discovers the host and user information in the ANI database, correlates this information, and displays it in the User Tracking table.
Before collecting user and host information, ANI first must complete a global discovery. During global discovery, ANI generates a device list to determine which switches and routers it should look at to obtain Media Access Control (MAC) and Internet Protocol (IP) address information.
With these device lists in place, the User Tracking service module of ANI performs the following steps:
Process | Description |
|
|
|
|
|
|
|
|
|
|
|
|
Refer to the ANI online help for further information about ANI Server discovery.
The criterion for whether or not User Tracking performs ping sweeps on a subnet is the number of hosts in the subnet.
If ping sweeps are not performed, User Tracking still obtains information from the router and switch mapping tables during a discovery.
Outdated entries result when:
User Tracking does not automatically delete outdated end user host entries. To delete these entries:
The IP phone display is a dialog box that displays noninteractive information. The menu functionality of the User Tracking display table does not apply to dialog boxes. For example:
User Tracking regenerates IP phone entries after a major aquisition or whenever you select Action > Discover IP Phones.
Select Query > Show IP Phones to view the latest information.
Indefinitely, until you delete the information.
User Tracking discovers all users and hosts in the network from the list of devices known to ANI. Refer to the Getting Started with the CiscoWorks2000 Server for more information about ANI discovery.User Tracking does not discover users and hosts that are connected to non-CDP devices. However, User Tracking may discover the non-CDP device itself as it would any other host machine.
|
Note Note: ANI discovers only those devices that are CDP-enabled. |
User Tracking uses the ANI Server error log. Refer to the Getting Started with the CiscoWorks2000 Server for more information about ANI discovery.
Use the information in Table A-2 to troubleshoot the User Tracking application.
| Symptom | Probable Cause | Possible Solution |
|---|---|---|
User Tracking cannot discover any users or hosts. | There may be no information in the ANI database. You must have valid ANI seed device(s) and run ANI discovery prior to running a User Tracking discovery. | See the ANI online help for further information. |
User Tracking cannot discover some users or hosts. | The ANI Server might not have discovered one or more devices to which users and hosts are connected. | Check the CiscoWorks2000 topology for the missing devices, make sure that CDP and SNMP are enabled on the devices, rediscover these devices, and verify that they appear on the topology map. |
User Tracking table is missing hostname, IP address, and subnet information for some hosts. | User Tracking is not finding the most recent network information. Network changes are not currently reflected in ARP information (routers) or bridge tables (switches). | Enable ping sweeps when User Tracking performs discovery. Ping sweeps are enabled by default. |
You made changes to the network and then ran User Tracking discovery, but the changes are not showing in the User Tracking display. | A complete ANI discovery has not finished since you added your changes. | User Tracking discovery is not a full network discovery; it discovers only the user and host data in your network. Changes that you make to your network might not appear after a User Tracking discovery. Try running a complete ANI discovery. |
Use the information in the following topics to help you troubleshoot Path Analysis:
Use the information in these sections to answer some of your common questions:
These are two of the more common errors:
The following are informational status messages:
All other status bar or alert box messages indicate that an error has occurred. See Table A-3 for definitions to the status bar messages and for definitions to the alert box messages.
| Message | Probable Cause | Severity | Possible Solution |
|---|---|---|---|
`From' field contains bad source endpoint | Invalid string entered for path trace destination. | Low | Confirm From field is a valid IP address or DNS name. |
`To' field contains bad source endpoint | Invalid string entered for path trace source. | Low | Confirm To field is a valid IP address or DNS name. |
Source and Destination Endpoints are on same device | Not possible to trace path between two IP addresses on one device. | Low | Make sure the To and From entries are not for the same device. |
Trace Running | Trace has not yet completed. | N/A | -- |
Trace Aborted | Error or anomalous situation prevented successful path trace. | Variable | -- |
Trace Timed Out | Trace attempt exceeded the timeout value (default 4 minutes). Path traces can take up to several minutes to complete. | Variable | Select Edit > Options... and increase the default trace timeout value. |
Trace Stopped | User stopped trace. | N/A | -- |
Trace Completed | Trace completed successfully. | N/A | -- |
| Message | Definition | Severity | Possible Solution |
|---|---|---|---|
Could not complete trace | One or more Layer 3 hops determined, but trace did not reach destination. Typically results from firewall rules, access lists, or blocked source-routed packets. | Variable | 1. Choose a destination on the same side of the firewall as the source. 2. Verify access lists along the path and correct any problems. 3. Confirm that no routers along the path block source-routed IP packets. |
Could not determine first hop | No known hops. Information to determine first hop is incomplete or inconsistent. | High | 1. Confirm source is in managed domain and SNMP-accessible. 2. Verify community strings. 3. Confirm end user stations are not multi-homed (containing multiple NICs). Path Analysis does not support this configuration. |
Could not reach source | No IP connectivity between Path Analysis server and source. | High | Run a path trace between the CiscoWorks2000 server and the source to analyze IP connectivity. |
The different kinds of lines and icons in Map View are visual clues, and each has a specific meaning. See the Path Analysis online help for detailed information about interpreting these lines and icons.
Yes. Path Analysis supports multiple, concurrent path traces from a single computer.
There are strict selection criteria for Path Analysis end-points. See the "Valid Data Path Trace End Points" section in "Tracking Packet Flow Using Path Analysis," for more details.
Yes, but only for LAN emulation (LANE).
If you encounter security errors from Microsoft Internet Explorer 5 while launching or running Path Analysis, you might consider making your CiscoWorks2000 server a trusted host. To do this:
1. In Internet Explorer, select Tools > Internet Options....
2. Select the Security tab.
3. Select the Trusted sites icon.
4. Click the Sites... button.
5. Enter the IP address of your CiscoWorks2000 server in the "Add this Web site to the zone" field.
6. Click Add.
7. Click OK.
Two common reasons a Layer 2 path exists with some Layer 2 devices missing are:
Do the following:
View the User Tracking entries for the unreachable end user station.
Alternatively, you can run a path trace from anywhere else on the network to this end user station. The last hop appears in Map view as a dotted line (best guess).
Path Analysis only uses User Tracking entries that have been discovered in the last 48 hours. If you want to use this end user station as a source or destination for a path trace, you must perform a User and Node Acquisition in User Tracking or run Discover All in Path Analysis.
Yes, you need to:
1. Enable Call Detail Records (CDRs) on Cisco CallManager (CCM).
2. Enable SNMP on CCM.
3. Provide community strings for all CCM devices to the ANI Server.
4. Provide IP addresses for discovery of CCM devices.
In nearly all cases, you must select the Voice Trace radio button in the Path Analysis main window to trace completed calls. However, there are two exceptions.
Do not select the Voice Trace radio button if you are looking at the signaling path by either:
In these two cases, select the Data Trace radio button instead.
Any parentheses, hyphens, spaces, or other non-numeric characters that you enter when specifying a telephone number for voice tracing are automatically deleted from your query.
The Calling Number value might not provide valid Call Detail Record (CDR) matches to your query, so in most cases you should not specify a calling number.
However, if you do not specify a calling number, and if your query results in too many CDRs found, then you can narrow your search criteria by specifying a calling number.
Yes, it is the best to specify the called number.
Yes. It is best to specify the start time and time range for voice tracing.
In many cases, a 15-minute range before and after the specified time is sufficient to locate a call without exceeding the limit of 100 Call Detail Records (CDRs).
If your query matches more than 100 CDRs, no records are shown. Instead, an error message appears.
If you specify a time range that is too brief, the search might not find your desired CDR.
IP phone calls routed over a Public Switch Telephone Network (PSTN) follow a uniform path:
Calls following this path produce one Call Detail Record (CDR) for the first leg of the call and another CDR for the final leg of the call. Each CDR contains a start and end point for its leg of the call. No CDR exists for the PSTN path.
When you search for this kind of call, both CDRs should appear in the Voice Trace Query Results window. You must then select each CDR independently to perform a trace on that leg of the call.
You cannot perform a voice trace on the PSTN leg of the call.
Yes, however you must perform a data trace using the IP addresses of the IP telephones. You cannot perform a voice trace using the calling and called numbers because busy or unanswered calls do not generate CDR records, which reference the calling and called numbers.
Use the information in Table A-5 to troubleshoot the Path Analysis application.
| Symptom | Probable Cause | Possible Solution |
|---|---|---|
| Data Traces | ||
Path Analysis client does not launch. | Path Analysis could take up to several minutes to launch because there are many JAR files that must be loaded and processed from the server. There is a problem with the server. | 1. Make sure you allow enough time for Path Analysis to load and process the JAR files from the server. You might want to use Internet Explorer, because typically, it performs this operation faster than other browsers. 2. Make sure you have specified the correct URL. 3. Make sure the server is up and running. 4. Select CiscoWorks2000 Server > Administration > Process Management > Process Status to confirm that all required services on the server are up and running. If any required services are not running, start them. Refer to the Getting Started with CiscoWorks2000 Server guide. 5. Reboot the server. |
Path Analysis client launches, but error message displays. | Server processes are not running normally. | Select CiscoWorks2000 Server > Administration > Process Management > Process Status to confirm that the ANI Server, ANI DB engine, Gatekeeper, EDS-TR, and EDS services are running on the server. If any processes are not running, start them. |
Start a path trace and get "undefined seed" error message. | Seed device was not specified in ANI. | Select CiscoWorks2000 Server > Setup > ANI Server Admin > Discovery Settings to define a seed device. |
Start a path trace and get "Initial discovery in progress" error. | After the ANI Server process has begun network discovery, it takes several minutes (or hours depending on the size of the network) to complete the discovery. When ANI discovery concludes, then user and host acquisition begins. No path trace is possible until both of these processes are complete. | Wait until ANI discovery, User Tracking ping sweeps, and user and host acquisition have completed before starting a path trace. You can monitor the progress of the discovery process in Path Analysis, which gives an indication of all three processes. |
Intermittent or recurring Path Analysis performance lag. | You have a misconfigured or non-functional DNS server. Your network is congested. One or more of your devices is too busy to respond to SNMP queries. | Confirm that your DNS servers are operational and properly configured. |
Path is discovered, but does not seem to function as shown. | An ACL (access control list) is allowing traffic from your CiscoWorks2000 server to the destination end-point, but is blocking traffic between your source and destination end-points. | None. |
No Layer 2 path for a given Layer 3 hop. | Layer 2 analysis is possible only when both ends of the Layer 3 hop are managed Cisco devices known to Topology Services or end user stations known to User Tracking. A question mark icon on either end of the Layer 3 hop indicates that these prerequisites have not been met. Refer to Path Analysis Concepts for more information about prerequisites for performing Layer 2 path analysis. | 1. Enable CDP for all Cisco devices. 2. In Topology Services, verify that all devices on this subnet are discovered and SNMP-accessible. If they are not, then verify their community strings and run ANI Discovery again. 3. Verify that end user stations are listed by User Tracking. If they are not, then run user and node acquisition again, preferably with ping sweeps enabled. 4. If there are VLANs or ELANs associated with this subnet, verify that the subnet mappings are correct for each Layer 3 hop and confirm that Topology Services has complete and accurate VLAN/ELAN information for all interfaces. |
Not all Layer 2 devices in the physical path are shown. | Path Analysis displays only those Layer 2 Cisco devices known to Topology Services and end user stations known to User Tracking. It does not include intervening hubs or non-Cisco or non-CDP devices. | No action required. |
No Layer 2 shortcutting (multilayer switching) is shown on the Map view. | Layer 2 shortcuts are supported on Catalyst 5000 switches (including the RSM module). Shortcut information only appears if you have specified the correct write community string. The Catalyst 5000 creates shortcuts on the first few packets in a flow, but they can be aged out. It is possible that a shortcut might not be present when you run a path trace. You can create a shortcut by having the source generate a few packets towards the destination before running a path trace. | If you do not see shortcut information on a Catalyst 5000 device, it might be because a shortcut does not exist or you do not have the correct write community string specified. Select CiscoWorks2000 > Setup > ANI Server Admin > SNMP Settings to confirm that you have the correct write community string specified. (The write community string is used to query the MLS MIB; not to configure the device.) |
A LANE segment appears in the Layer 2 trace, and shows connectivity between two LANE clients on either side of an ATM cloud, but does not show the intervening ATM switches. | There was no data-direct virtual channel between the two LANE clients during the trace. There are unsuppored software revisions on ATM switches and LANE cards in the ATM cloud. | See the Cisco documentation available on CCO. Select Cisco Product Documentation > Network Management > CiscoWorks2000 > Campus Manager/CWSI Campus > |
Path Analysis did not find a complete Layer 3 path between the source and the destination devices. | There could be many possible causes. | Check the status bar for messages and check for alert box messages. See Table A-3 and Table A-4. |
Table view does not contain the full set of information on Layer 3 interfaces. | Path Analysis collects this information about Cisco devices that are SNMP-reachable and on non-Cisco routers as long as they are SNMP-reachable and are within a subnet that is known to be within the organizational domain. The organizational domain comprises the subnets listed in the Subnet to VLAN/ELAN mapping table. | 1. Check the community strings and correct them if necessary. 2. Select Topology Services > View > Display View and confirm Cisco devices are SNMP-accessible (green icon). 3. Check the Subnet to VLAN/ELAN mapping table and make sure that the subnet of the device is listed. Refer to the "VLAN/ELAN Subnet Mapping Table and Layer 2 Tracing" section. |
| Voice Traces | ||
No records found in query. | Records might not be found because:
| 1. Select Topology Services > Layer 2 View and check whether the CCM server was discovered. The CCM server icon should be green. 2. Enter a less restrictive time range, exclude the Calling Number value, and/or enter a less specific called number. 3. Enable CDR logging on CCM. |
Too many matches found. | The search criteria are too broad. | Enter a more restrictive time range and/or a more specific calling/called number. |
Call Detail Record (CDR) query takes too long. | CDR query might take too long because:
| 1. Narrow your search criteria by reducing the time range or providing a calling number. 2. Make sure that all CCM servers that have been discovered in Topology Services are running and accessible. |
Posted: Thu Aug 10 13:22:00 PDT 2000
Copyright 1989-2000©Cisco Systems Inc.