Forum Discussion

jerm1020_254086's avatar
jerm1020_254086
Icon for Nimbostratus rankNimbostratus
Jul 20, 2016

Troubleshooting monitor issue

Our QuickBuild Corp environment is fronted by an internal DNS entry (xxxxx.xxxxxxx.com) which points to the F5 virtual address 10.xx.xx.xx. This virtual server integrates with the pool xxxxx_xxxxx_p_https_pool whose members should respond to 443 health monitors and traffic. However, the current pool member MACHINE_NAME (10.xx.xx.xx) will only respond to port 80 and ICMP monitors. All local tests utilizing local hosts file modifications on MACHINE_NAME for xxxx.xxxxxx.com against HTTPS were successful, so there appears to be an issue routing HTTPS traffic between the NPCI VCMP and MACHINE_NAME. We have confirmed that telnet from the VCMP to the node/pool member via 443 is allowed, so we’re in need of remediation assistance:

 

a. Determine causes of HTTPS/443 health monitor down status for xxxxx_xxxxx_p_https_pool

 

b. Determine causes of inability for the F5 to pass traffic to a 443 pool members even when leveraging a high level health monitor such as icmp – To rule out health monitor issues we enable icmp as the only health monitor and received an UP status, but the 443 port designation for the pool member wouldn’t render site content.

 

2 Replies

  • it can be the webserver wants more, for example a host header or such.

     

    what if you do a curl -v -k https://MACHINE_NAME from the big-ip, do you get any response then?

     

  • Try these for some troubleshooting tips:

     

    SOL16165

     

    Thread on similar issue.

     

    If they don't work, provide the https monitor configuration. As you can ping and telnet, I don't think you have any routing issues.