Note: This article applies to Citrix SD-WAN WANOP.

Symptoms

The following are some of the symptoms:

• Caches are not load balanced properly (traffic gets redirected only to one or more but not all caches).

  Access the GUI of each Web Cache Communication Protocol (WCCP) cluster member and verify if all connections are getting accelerated by only one WCCP Cluster member:
  Monitoring > Optimization > Connections > Accelerated Connections

  CloudBridge Appliance #1 Cache

  

  CloudBridge Appliance #2 Cache

  

• Latency

• Disconnects

• Crashes

• Frequent migration of connections between caches causing unaccelerated connections.

Troubleshooting Steps

The following troubleshooting steps will not apply if the issue seen is uneven load balancing of traffic within CloudBridge4000/5000.

Note: In WCCP cluster, router load balances between CloudBridge4000/5000.

1. Verify that all appliances within a cluster are same model and run the same software release. CloudBridge700 and VPX does not support WCCP clustering. For more information, see Citrix eDocs - Limitations.

2. Verify if all appliances and instances (caches) are UP.

3. Access one WCCP cluster member from Monitoring > Appliance Performance > WCCP and verify the following tabs:

   1. Cache Status:

      • Verify if all caches have an assignment, status should be displayed as "Has Assignment".

   2. Routers:

      1. Verify if there is only one designated cache. For more information, see Citrix eDocs - Testing and Troubleshooting.

      2. Verify if all cache members of the cluster are correctly listed.

      3. Verify if all router members of the cluster are correctly listed. If there are multiple routers, ensure that all CloudBridge appliances have all the router's IPs configured in the cluster Service Groups (SG).

      4. If all router members are not listed, then add them under Configuration > Appliance Settings > WCCP > Configure Service Group.

      5. The following are some troubleshooting commands for router configuration reference (it is strongly recommended to engage Cisco TAC to validate the router(s) configuration).

• show ip wccp

• show ip wccp <service group >

• show ip wccp <service group > detail

• show ip wccp capabilities

4. If all caches have assignment but load is not equally distributed, then verify the following:

   1. LAN/WAN SG definition is defined equally on CloudBridge and the router:

      • WAN SG should be the SG that sees the largest pool of IP address. 
      • LAN SG should be the SG that sees the smallest pool of IP address. 

   2. Verify the value of the​ current mask used and confirm if this is the appropriate mask according to the cluster needs. For more information, see Citrix eDocs - Load-Balancing in the WCCP Cluster.

​Suggested Sanity Check: Count the number of caches and verify the range of mask elements that are assigned for the cluster. For example,

• You have five caches.

• Mask used is 0x1 (1 bit set), the number of mask elements of this mask is 2 (2 or less caches):

• 2^N= mask elements (N is the number of bits set).
• 2^1= 2 mask elements = 2 or less caches. Because there are five caches, this mask 0x1 is too small.

5. If the issue still remains, collect the following data:

   1. Screen shots.

      1. Monitoring page of WCCP:
         Monitoring > Appliance Performance > WCCP > Cache Status.
         Monitoring > Appliance Performance > WCCP > Routers.

      2. Configuration page of SG:
         Configuration > Appliance Settings > WCCP (select the SG and click Modify).

   2. Technical support file. For more information, see CTX133765 - How to Collect Diagnostic Data by using the Command Line Interface on a Branch Repeater Appliance or Branch Repeater VPX Instance and CTX135546 - How to Collect Diagnostic Data by using the Graphical User Interface on a Branch Repeater Appliance or Branch Repeater VPX Instance.

   3. Router(s) configuration.

   4. Detailed network topology from all sites. Clearly depict the (switches, routers, firewalls, IPs, servers) connectivity points between the devices.

   5. IP range/subnets that you are trying to optimize.