Least connection loadbalancing question.

Question

Hi,&nbsp;
I am testing ltm 11.5.2HF1 least connection loadbalancing method, anyways in my test I disable a pool member in a pool that is consisted of 16 members and after few seconds when all other members get connections I go and re-enable the disabled member and the test keeps running till it's time over which is about 5min, when the test is done the members don't have equal connections.&nbsp;
by the way CMP is disabled on the VIP.&nbsp;
my question: isn't the ltm supposed to send all the connections to the disabled member after re-enabling it till it becomes equal with other members and then continue rounrobin till the test is done ?&nbsp;
your thoughts please ?&nbsp;
Thanks,
Sahir&nbsp;

samir_jha_52506 · Accepted Answer

Hi Sahir,
This is not bug or issue. Hope you are aware of SRT(Slow Ramp Time). Whenever you enable any pool member, f5 doesn't give all new connection to that member. Slowly f5 share load to new member. By the way difference session is just 3 now, its not bad.&nbsp;

nitass · Answer

&gt;when the test is done the members don't have equal connections.

what connection do you mean? is it current connections (not maximum or total connections)?

sahir_180434 · Answer

both, the current connections for the disabled member is 3 connections less that all other members which reflects on the total connections beside the missed connections when the member was disabled.

nitass · Answer

i do not think total connection is right stats because least connections mode is based on open connections at the time new connection is coming but current connections should be. how much difference in current connections between pool members?

sahir_180434 · Answer

3 connections less, for all the members is 23, for the member that was disabled and moved to enable state is 20

nitass_89166 · Answer

i have not yet had an idea how to test with http traffic (short-lived connection) but it seems okay with ssh. i am running 11.6.0 hf4.
 configuration

root@(ve11a)(cfg-sync In Sync)(Active)(/Common)(tmos) list ltm virtual bar
ltm virtual bar {
    cmp-enabled no
    destination 172.28.24.10:22
    ip-protocol tcp
    mask 255.255.255.255
    pool foo
    profiles {
        tcp { }
    }
    rules {
        qux
    }
    source 0.0.0.0/0
    source-address-translation {
        type automap
    }
    vs-index 19
}
root@(ve11a)(cfg-sync In Sync)(Active)(/Common)(tmos) list ltm pool foo
ltm pool foo {
    load-balancing-mode least-connections-member
    members {
        200.200.200.101:22 {
            address 200.200.200.101
        }
        200.200.200.111:22 {
            address 200.200.200.111
        }
    }
    slow-ramp-time 0
}
root@(ve11a)(cfg-sync In Sync)(Active)(/Common)(tmos) list ltm rule qux
ltm rule qux {
    when RULE_INIT {
  set static::pool_name "foo"
  set static::members_cmd "members -list $static::pool_name"
  set static::pool_list [eval $static::members_cmd]
  unset -nocomplain static::pool_conn
  foreach static::pool_mbr $static::pool_list {
    set static::pool_conn([lindex $static::pool_mbr 0]) 0
  }
  unset static::pool_name static::members_cmd static::pool_mbr
}
when LB_SELECTED {
   increase counter
  set server_ip [LB::server addr]
  incr static::pool_conn($server_ip)

log
  set log_conn "client:[IP::client_addr]:[TCP::client_port] "
  foreach elm $static::pool_list {
    append log_conn "[lindex $elm 0]:$static::pool_conn([lindex $elm 0]) "
  }
  log local0. $log_conn
}
when CLIENT_CLOSED {
   decrease counter
  incr static::pool_conn($server_ip) -1
}

/var/log/ltm

[root@ve11a:Active:In Sync] config  tail -f /var/log/ltm
Jun  5 13:16:02 ve11a notice mcpd[8104]: 01070639:5: Pool /Common/foo member /Common/200.200.200.111:22 session status forced disabled.
Jun  5 13:16:24 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59163 200.200.200.111:0 200.200.200.101:1
Jun  5 13:16:37 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59167 200.200.200.111:0 200.200.200.101:2
Jun  5 13:16:46 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59169 200.200.200.111:0 200.200.200.101:3
Jun  5 13:16:55 ve11a notice mcpd[8104]: 01070639:5: Pool /Common/foo member /Common/200.200.200.111:22 session status enabled.
Jun  5 13:17:02 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59174 200.200.200.111:1 200.200.200.101:3
Jun  5 13:17:10 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59177 200.200.200.111:2 200.200.200.101:3
Jun  5 13:17:17 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59179 200.200.200.111:3 200.200.200.101:3
Jun  5 13:17:24 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59182 200.200.200.111:3 200.200.200.101:4
Jun  5 13:17:31 ve11a info tmm[30284]: Rule /Common/qux : client:192.168.206.117:59185 200.200.200.111:4 200.200.200.101:4

Forum Discussion

Least connection loadbalancing question.

10 Replies

Recent Discussions

Create a CSR and Key using the BigIP LTM GUI when renewing a certificate

F5 Rseries HA

Error when running bigip_command Playbook against LTM : Syntax Error: unexpected argument /bin/sh\n

Can iRule be used to perform exception of IPI category based on Geolocation

Can iRule mask the payload content on event logs of security

Related Content

F5 Connection Mirroring question

Health monitor question

Securely connecting Kubernetes Microservices with F5 Distributed Cloud

Troubeshooting website connection

Novice Question - irules