关于weblogic的Persistence

看了metoo的最常用的50个irules，其中有关于BEA Weblogic Persistence的一个irules。内容如下：


#  BEA Weblogic Persistence
#  CPU impact:   Medium
#  Requirement:  HTTP profile, Weblogic v8+
#
#  Weblogic servers have the ability to replicate session information
#  to a backup server.  Should the first primary server that a user is
#  using become unavailable, then the LTM should send a user to a specific
#  secondary server, rather than simply reloadbalancing to a new server.
#  For typical deployments, it is recommend to start with the LTM's built
#  in cookie persistence or the JSessionID persistence iRule.


when RULE_INIT {
  # set the name of the persistence token (cookie) here
  set ::SessionIDName "WebDomain_PORTALSESSION"
  set ::X_WebLogic_Cluster_List ""
}

when HTTP_REQUEST {
    if { [HTTP::cookie exists $::SessionIDName] } {
      set session_id [ substr [HTTP::cookie $::SessionIDName] 0 "!" ]
      set session_id_len [expr {[string length $session_id] + 1}]
      set primary_server_jvmid  [ substr [HTTP::cookie $::SessionIDName] $session_id_len "!" ]
      set primary_server_jvmid_len [expr {[string length $primary_server_jvmid] + 1}]
      set str_len [expr {$session_id_len + $primary_server_jvmid_len}]
      set secondary_server_jvmid  [ substr [HTTP::cookie $::SessionIDName] $str_len "!" ]
      persist uie $session_id

      # get primary server information from persistence record
      set LBPool [ persist lookup uie $session_id pool ]
      set LBMember [ persist lookup uie $session_id node ]
      set LBPort [ persist lookup uie $session_id port ]

      # start delete this section once tested and server addr is decoded
      set secondary_id [findstr $::X_WebLogic_Cluster_List $secondary_server_jvmid 0 "!-1|"]
      set secondary_server_jvmid_len [expr {[string length $secondary_server_jvmid] + 1}]
      set secondary_server_addr_str  [ substr $secondary_id $secondary_server_jvmid_len "!" ]
      set secondary_server_port  [ substr $secondary_id [expr {$secondary_server_jvmid_len + [string length $secondary_server_addr_str] + 1}] "!" ]

      # convert secondary string address to ipaddr
      set hex_ip [format %x $secondary_server_addr_str]

      set a [string range $hex_ip 0 1]
      set b [string range $hex_ip 2 3]
      set c [string range $hex_ip 4 5]
      set d [string range $hex_ip 6 7]
      set secondary_server_addr "[format %i 0x$a].[format %i 0x$b].[format %i 0x$c].[format %i 0x$d]"

      # heath check primary member
      if { $LBPool != "" } {   
        if {[LB::status pool $LBPool member $LBMember $LBPort] ne "up"} {
          # if primary failed, search header
          set secondary_id [findstr $::X_WebLogic_Cluster_List $secondary_server_jvmid 0 "!-1|"]
          set secondary_server_jvmid_len [expr {[string length $secondary_server_jvmid] + 1}]
          set secondary_server_addr_str  [ substr $secondary_id $secondary_server_jvmid_len "!" ]
          set secondary_server_port  [ substr $secondary_id [expr {$secondary_server_jvmid_len + [string length $secondary_server_addr_str] + 1}] "!" ]
          # binary format i3 {3 -3 65536 1}   
          # convert secondary string address to ipaddr
          set hex_ip [format %x $secondary_server_addr_str]

          set a [string range $hex_ip 0 1]
          set b [string range $hex_ip 2 3]
          set c [string range $hex_ip 4 5]
          set d [string range $hex_ip 6 7]
         
          set secondary_server_addr "[format %i 0x$a].[format %i 0x$b].[format %i 0x$c].[format %i 0x$d]"
          # health check secondary
          if {[LB::status pool $LBPool member $secondary_server_addr $secondary_server_port] eq "up"} {
            persist delete uie $session_id
            pool $LBPool member $secondary_server_addr $secondary_server_port
          } else {   
            # if secondary failed then re-load balance
            HTTP::redirect http://[getfield [HTTP::host] ":" 1][HTTP::uri]
          }
        }
      }
    }
    # This header instructs Weblogic to return cluster information
    HTTP::header insert X-Weblogic-Request-ClusterInfo true
}

when HTTP_RESPONSE {
    HTTP::header remove X-Powered-by
    if { [HTTP::cookie exists $::SessionIDName] } {   
      persist add uie [ substr [HTTP::cookie $::SessionIDName] 0 "!" ]   
    }
    if { [HTTP::header exists X-WebLogic-Cluster-List] }{
      set ::X_WebLogic_Cluster_List [HTTP::header X-WebLogic-Cluster-List]
      HTTP::header remove X-WebLogic-Cluster-List
    }
}

大概的意思能猜出来，但是不知道怎么用，需要weblogic进行额外配置或应用进行改造？其中WebDomain_PORTALSESSION从哪里来？用http watch看了交互过程没有发现这个head，只有：
Set-Cookie        JSESSIONID=K1zcZNfhzDKrtRQPg9Fv0TLQdznd1pbpRTZQvRMDw4DP7JqM1T1n!-1003613857!67506566; path=/
改为JSESSIONID之后应用倒是能访问，X-WebLogic-Cluster-List的值也能取到，但是当前连接的server关闭后却不能访问应用了，即使群集中的另一台server状态正常也不行。去掉irules之后却能够达到失败转移的效果，当然，这只是群集中只存在2个server的场景， 不存在选择问题。