[Kea-users] Dropping the packets in load balancing

[Kraishak Mahtha]

Hi Darren/Peter , sorry for the delay in reply

I have added the logs for hooks and the allocation engine too and now I
observe the following case
Server A and Server B are my load-balancing servers
Case1)No active leases before and now I started service only on Server A
and kea-service on Server B is dead after some time
approx(6000) max-response-delay value it tried for every 10 seconds and
finally it set automatically to partner-down and worked as expected
Case 2) Started service on both servers A and B and I fed with some active
leases (total IP's-216 and at this instant we have 68 active leases) now I
stopped the service on Server B and even after 25 minutes the stage is
still in the communication-recovery stage, I don't see any much info in the
logs and I wanted to see whether now the server A can take control on the
scope range and provide leases but no after few active lease grants it
stopped responding even though we have free IP's because it is not in
partner down(to my understanding),

Please find the attached config and log files of the test flow:
 Out-put of the status get :
{
  "arguments": {
    "high-availability": [
      {
        "ha-mode": "load-balancing",
        "ha-servers": {
          "local": {
            "role": "primary",
            "scopes": [
              "dhcp1"
            ],
            "state": "communication-recovery"
          },
          "remote": {
            "age": 1005,
            "analyzed-packets": 175,
            "communication-interrupted": true,
            "connecting-clients": 128,
            "in-touch": true,
            "last-scopes": [
              "dhcp2"
            ],
            "last-state": "unavailable",
            "role": "secondary",
            "unacked-clients": 1,
            "unacked-clients-left": 13
          }
        }
      }
    ],
    "multi-threading-enabled": true,
    "packet-queue-size": 7,
    "packet-queue-statistics": [
      1.0704,
      2.60985,
      1.13931
    ],
    "pid": 32411,
    "reload": 1138,
    "sockets": {
      "status": "ready"
    },
    "thread-pool-size": 4,
    "uptime": 1138
  },
  "result": 0
}

The Server is not changing the status to partner-down if we already have
active leases or do we need any other approach or to add any config
parameter
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/kea-users/attachments/20230515/06f09399/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: kea-dhcp4.conf
Type: application/octet-stream
Size: 4698 bytes
Desc: not available
URL: <https://lists.isc.org/pipermail/kea-users/attachments/20230515/06f09399/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: kea_hooks.log
Type: application/octet-stream
Size: 50753 bytes
Desc: not available
URL: <https://lists.isc.org/pipermail/kea-users/attachments/20230515/06f09399/attachment-0003.obj>