check_http services not working after 2014 upgrade

[sreinhardt]

The only real change, at a glace, seems to be the Accept: */* mimetype stuff, unless it was in one of the supporting files. Any chance you could PM me a url that has similar issues that I can test externally?

Edit: I should rephrase this as from 1.5(or what would have been) to 2.x there were minimal changes. There were MASSIVE changes from 1.4.xx to 1.5 to work with the redirect items. Also a side note, your use of redirect is a bit of an outdated edge case, that only http and browsers support, but we are still looking into handling this properly.

[abrist]

I did some digging, and it looks like you are using a scheme-relative (//path) redirection path instead of a standard relative path (/path or path). Not entirely best practice, but it is supported by most browsers and used by many sites, so we should patch this.

[vAJ]

This must be something in the A10 Load Balancer that's doing this. Checks directly to the IIS sites behind it work just fine...

I'll also have our network engineers look at the A10 config to see if this is something that can be tweaked there.

[slansing]

Let us know what you come up with on the Load Balancer's end, we have no reports of others using the same balancer with issues yet.

[vAJ]

Conclusion is that the 2014 upgrade forced the upgrade of the template "xiwizard_website_http_service" to use the "check_xi_service_http" (which uses check_http -H) instead of the straight "check_http" command (which uses only -I).

We're still trying to figure out why our A10 Load Balancers and the Nagios server don't correctly close the connection once the page is returned, but the fix is to force the template back to using "check_http" or update any broken service checks that were effected.

I think we can close this thread.

-Andrew