How to fix the repetition of "org.jgroups.protocols.pbcast.NAKACK" issue?

Question

Ours is a java application that is using jgroups (version 2.6) to send and receive multicast messages. We got a heap dump in one of our environments. And the heap file got accumulated so quickly that is, within a month. But other setups are working normally. Please find the below error details,

One instance of org.jgroups.protocols.pbcast.NAKACK loaded by 
org.jboss.mx.loading.UnifiedClassLoader3 @ 0x804e98c0 occupies 1,966,616,080 (92.43%) bytes. 
The memory is accumulated in one instance of java.util.concurrent.ConcurrentHashMap$Node[], 
loaded by <system class loader>, which occupies 1,966,608,296 (92.43%) bytes.

Keywords :
org.jgroups.protocols.pbcast.NAKACK
org.jboss.mx.loading.UnifiedClassLoader3 @ 0x804e98c0
java.util.concurrent.ConcurrentHashMap$Node[]

But in NakReceiverWindow, both the "highest_received" and "highest_delivered" attributes have the same value. Please find the attached Image.

And in server.log we got the following details,

[org.jgroups.protocols.pbcast.NAKACK] sender XXX.XXX.XXX.XXX:61277 not found in xmit_table
ERROR [org.jgroups.protocols.pbcast.NAKACK] range is null

Could you kindly answer the following questions?

1. Whether the above error indicates a communication error between the clusters?
2. Whether there will be any network issue that causes the above issue?
3. And why the other setups didn't get the same error?
4. How to fix this issue permanently?

Thanks in Advance.