check_http services not working after 2014 upgrade

[vAJ]

This didn't come up in on my test server (very few checks), but upgraded to 2014R1 on my QA instance (many more checks) and all services utilizing check_http are timing out. Can curl the address/uri path just fine, but all I get is "CRITICAL - Socket timeout after 10 seconds"

I know plugins were updated with this release, anything in check_http I need to be worried about?

[vAJ]

No problems if I switch to headers only (using -N)

[abrist]

Can you give me an example of the command you are running?

[vAJ]

Code: Select all

check_http -H services-perf.dev.*****.com -f follow -I services-perf.dev.*****.com -u "/*****.AdminUser.WebApi/def" -p 80

Had to mask out company specific info... but it's about as vanilla of a check_http as you can get

[abrist]

Code: Select all

check_http -H services-perf.dev.*****.com -f follow -I services-perf.dev.*****.com -u "/*****.AdminUser.WebApi/def" -p 80 -vvv

Can you run the previous command with -vvv? I ran the command below and it works fine on my server. Can you try running this command as well?

Code: Select all

check_http -H support.nagios.com -u /forum/viewtopic.php?t=27000#97682 -vvv -f follow -I support.nagios.com -p 80

[vAJ]

Standard header call:

Code: Select all

User-Agent: check_http/v2.0.1 (nagios-plugins 2.0.1)
Connection: close
Host: services-perf.dev.*****.com
Accept: */*


CRITICAL - Socket timeout after 10 seconds

Check of your URI comes back fine:

Code: Select all

HTTP OK: HTTP/1.1 200 OK - 20817 bytes in 0.157 second response time |time=0.156747s;;;0.000000 size=20817B;;;0

Just realized that the hosts having problems are our VIPs served up by our A10 load balancers. Direct to the webservers behind those are fine. But the checks were working perfectly right before the upgrade... must be something in the call for page content that check_http is using that the A10 doesn't like. Very strange. Having a network engineer dump logs from this VIP.

[abrist]

Can you run the check with the verbose (-vvv) flag and post the output?

Code: Select all

 check_http -H services-perf.dev.*****.com -f follow -I services-perf.dev.*****.com -u "/*****.AdminUser.WebApi/def" -p 80 -vvv

[vAJ]

Code: Select all

 /usr/local/nagios/libexec/check_http -H services-perf.dev.*****.com  -f follow -u "/*****.AdminUser.WebApi/def" -p 80 -vvv
GET /*****.AdminUser.WebApi/def HTTP/1.1
User-Agent: check_http/v2.0.1 (nagios-plugins 2.0.1)
Connection: close
Host: services-perf.dev.*****.com
Accept: */*


CRITICAL - Socket timeout after 10 seconds

[vAJ]

Load-balancer shows it's returning an HTTP 200...

[vAJ]

I know check_http waits for TCP-FIN, which is why when I specify -N for headers only, it works. But something must have changed in plugins 2.0 or 2.0.1 with the session creation that is not working quite right...