Forum Discussion

CuseHokie_18561's avatar
CuseHokie_18561
Icon for Nimbostratus rankNimbostratus
Jan 14, 2010

HTTP Monitor works on one cluster but not another?

I have a simple HTTP monitor that gets a file from the server and looks for a receive string.

 

 

I tested this out on my nonprod LTM and had it deployed to 2 prod clusters tonight.

 

 

For some reason, one prod cluster doesn't seem to be marking the same 10 servers up that the other cluster does...

 

 

 

Everything is the same...

 

 

 

I took a pcap and the bad cluster does not show protocol HTTP in wireshark for the send/receive, which I found is odd.

 

 

It's noting that the version of the bad cluster is 9.4.0 and the good is 9.4.8... The nonprod where I first tested is 9.4.7.

 

 

Any thoughts?

 

 

management
monitoring

2 Replies

Sort By
  • L4L7_53191's avatar
    L4L7_53191
    Icon for Nimbostratus rankNimbostratus
    Jan 14, 2010
    9.4.0 is *very* old code and you really should upgrade that environment. There are several known issues that could be at play here, and it's a decent bet that you've run into one of them. I'd definitely do a search on the support site and see if you can't find something that matches your monitor config.

     

     

    Barring that, could you post your monitor so we'll have that information?

     

     

    -Matt
  • hooleylist's avatar
    hooleylist
    Icon for Cirrostratus rankCirrostratus
    Jan 15, 2010
    As Matt suggested, seeing the monitor config (particularly the send and receive strings) and possibly a tcpdump from a working unit and one from a failing unit would help. I don't know of any significant changes in the monitoring daemon between 9.4.0 and 9.4.8 that would cause a monitor to fail.

     

     

    Aaron