02-19-2011 11:26 AM
I know that the ServerIron XL is ancient and I plan on replacing it asap. =)
I am having an issue where the ServerIron XL increases latency and drops packets whenever the healthcks are running.
Here is my configuration:
healthck node-a tcp
protocol http url "GET /test/file.gif"
healthck node-a-ssl tcp
protocol ssl url "GET /test/file.gif"
protocol ssl use-complete
server remote-name node-a x.x.x.x
port default disable
weight 2 0
port http healthck node-a
port http keepalive
port http url "HEAD /"
port ssl healthck node-a-ssl
port ssl keepalive
I have 20 remote servers and 2x healthcks each (for a total of 40)
I noticed that when I ping the ServerIron with the above configuration in place there was consistent latency and some packet loss.
When I add 'interval 30' to the healthck statements I notice that the pings are only timing out about once every 30 seconds or so, which leads me to believe that when the healthcks are running it is causing high cpu or memory utilization (possibly both).
Are healthcks pretty resource intensive?
Is there a better less resource intensive way to handle this?
02-20-2011 01:54 PM
Long time since I have used an XL
Yes L7 checks are resource heavy. And no there is not much you can do othen then use L4-check or drop the SSL check (this is the one that is taking the most recources).