For more information regarding the security incident at F5, the actions we are taking to address it, and our ongoing efforts to protect our customers, click here.

Forum Discussion

mfkk531_168091's avatar
mfkk531_168091
Icon for Nimbostratus rankNimbostratus
Dec 02, 2015

Pool Members Flapping Continuously

Hi Everyone, I have a pool that is continuously failing and passing health checks. I'm using http monitor

ltm pool t1b {
    members {
        10.1.14.135:any {
            address 10.1.14.135
            session monitor-enabled
            state down
        }
        10.1.14.77:any {
            address 10.1.14.77
            session monitor-enabled
            state down
        }
        10.1.14.142:any {
            address 10.1.14.142
            session monitor-enabled
            state up
        }
        10.1.14.143:any {
            address 10.1.14.143
            session monitor-enabled
            state down
        }
        10.1.14.144:any {
            address 10.1.14.144
            session monitor-enabled
            state up
        }
        10.1.14.145:any {
            address 10.1.14.145
            session monitor-enabled
            state up
        }
    }
    monitor http_80
}

  ltm monitor http http_80 {
            defaults-from http
            destination *.http
            interval 5
            send "GET /\\r\\n"
            time-until-up 0
            timeout 16
        }

Here's a tcpdump i took between my monitor local self-ip 10.1.14.22 and the server IPs tcpdump file - google drive

Appreciate your help!

2 Replies

  • JG's avatar
    JG
    Icon for Cumulonimbus rankCumulonimbus

    You probably want to create a customised monitor, e.g. with a send string:

     

    "GET / HTTP/1.0\r\n\r\n"

     

    just to see if your backend does not support HTTP/0.9, the default.

     

    Also some versions of BIG-IP transform "\r\n" to "\r\n" on upgrade, which gives me trouble.

     

  • Vik_K_236702's avatar
    Vik_K_236702
    Historic F5 Account

    Hi

     

    May I ask the time interval the pool members fail and is it a pattern?

     

    Moreover having a specific monitor will help your case even better and help you get an accurate application behavior.

     

    And any requirement to add a node with all ports as a part of a http pool ?

     

    Always best to add a member on the required port for our requirement.

     

    Regards, Vik