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

dubdub's avatar
dubdub
Icon for Nimbostratus rankNimbostratus
Feb 09, 2015

HTTP monitor and target web server down

Environment: 11.4.0 HF3

 

I have a reverse HTTP monitor that is assigned to two pool members. The HTTP server for this monitor is hosted on a third server, which is not a member of the pool. When this third server is up and serving requests, I can correctly execute the monitor for the pool members. However, when the HTTP server is down, my reverse monitors fail and the pool members get disabled. Is there a way to configure a reverse monitor to survive an outage on an alias address/port?

 

Thanks, Jen

 

application delivery
availability
design
LTM

3 Replies

  • Michael_Jenkins's avatar
    Michael_Jenkins
    Icon for Cirrostratus rankCirrostratus
    Feb 09, 2015
    Can you post the config for your monitor (sanitized if you need)? Just to help clarify how your monitor is working? Should be able to get it from the command line "tmsh list ltm monitor ".
  • dubdub's avatar
    dubdub
    Icon for Nimbostratus rankNimbostratus
    Feb 09, 2015
    Hi Michael, Sure - in this example, server1 is one of the two pool members. Server3 (which is running the HTTP server) is at 10.1.1.1. My question is, when server3 is down (reboot/patch/etc.), how can I ensure that the monitor doesn't fail so that server1 and server2 remain enabled in the pool? ltm monitor http Ultra_rev_server1 { defaults-from http destination 10.1.1.1:http interval 5 recv DOWN reverse enabled send "GET /down/server1.html\\r\\n" time-until-up 0 timeout 16 } Thanks, Jen
  • StephanManthey's avatar
    StephanManthey
    Icon for Nacreous rankNacreous
    Feb 09, 2015

    Hi dubdub,

     

    I´m not aware of a build-in solution for your requirement and I guess F5´s intention was to map the reverse monitor to the pool members directly. In case the monitoring condition is met or the pool member does not respond at all it will be marked as down until a following monitoring sequence is successful.

     

    Before this feature was build in I used a so called external monitor (based on a shell script) to poll a status page on a dedicated server. Based on the result the monitor modified the pool member state via tmsh.

     

    That´s the only approach coming close to what you have currently tried to implement.

     

    But how about using the API to control poolmember state?

     

    Thanks, Stephan