[Aside] NetScaler tracing, telnet, etc.

It is not possible to do a telnet from the NetScaler to any server to troubleshoot connectivity issues. The telnet may or may not succeed, but it doesn’t mean anything as the telnet is initiated from the NSIP where all NetScaler communications to its services happen from the SNIP. Only option in such cases is to create a service bound to that port & protocol, and monitor that.

At work, for instance, we had STA issues. So I created an HTTP service, bound to port 80, for each Delivery Controller. Then I created a new HTTP monitor that checks for /Scripts/CtxSta.dll and expects return code 406. This also lets me create an nstrace against this service itself to see what’s happening.

  • nstrace reference
    • Set the packet size to 0 and file size to 0.
    • Expression will be something like CONNECTION.SVCNAME.EQ("mySTA_svc_name")
  • There’s also nstcpdump.sh, which is a lighter version of nstrace. Less details, but quicker to get up and running. I prefer nstrace. :)
  • A blog post with examples of both.