Apache mod_cluster is reporting load as -1 or 0 and "Can't find worker for #"

Issue

• We are getting an error with load balancing. Mod_cluster-manager's INFO output

  Node: [1],Name: node1,Balancer: TestCluster,Domain: ,Host: 127.0.0.1,Port: 8009,Type: ajp,Flushpackets: Off,Flushwait: 10000,Ping: 10000000,Smax: 1,Ttl: 60000000,Elected: 0,Read: 0,Transfered: 0,Connected: 0,Load: 76
  Node: [2],Name: node2,Balancer: TestCluster,Domain: ,Host: 127.0.0.1,Port: 8109,Type: ajp,Flushpackets: Off,Flushwait: 10000,Ping: 10000000,Smax: 1,Ttl: 60000000,Elected: 0,Read: 0,Transfered: 0,Connected: 0,Load: -1
  

• We have some problems with mod_cluster and Load balancing.  In the Apache debug error_log we see

  [debug] mod_manager.c(1957): manager_handler STATUS (/) processing: "JVMRoute=node2&Load=75"
  [debug] mod_proxy_cluster.c(1481): proxy_cluster_isup: Can't find worker for 5
  

• Sticky sessions are broken to this node and requests are being routed to other nodes.

• All over the place, frequently nodes advertised them self with load -1 in the mod-cluster-manager page. After a reload of the Apache web server, things were normal again (positive load), but all as: not for long.
  This is unexpected and we have seen this literally all-over-the-place. So random nodes became unavailable to the Apache web server at random time.