topic Is the ACE load Balancing correctly or is there an issue? in Application Networking

Is the ACE load Balancing correctly or is there an issue?

haidar_alm — Thu, 02 Jul 2015 08:09:06 GMT

Hi guys,

I had an e-mail from our Web team admin with a graph stating that there is an issue with the load balancing configuration.

We've got two front end servers SDHSPWFE01, and SDHSPWFE02.
According to the graph below, they can see http and https traffic hitting SDHSPWFE01 which is the green line. However, SDHSPWFE02 is the black line which is constantly on zero.

I've had a look at the Datacentre context and ran the following commands:

ACE-CR1-LOADBALANCER/Datacentre# sh stats loadbalance

+------------------------------------------+
+------- Loadbalance statistics -----------+
+------------------------------------------+
Total version mismatch : 0
Total Layer4 decisions : 2540
Total Layer4 rejections : 0
Total Layer7 decisions : 40621
Total Layer7 rejections : 0
Total Layer4 LB policy misses : 0
Total Layer7 LB policy misses : 0
Total times rserver was unavailable : 0
Total ACL denied : 0
Total IDMap Lookup Failures : 0
Total Misc Errors : 0
Total Cipher Lookup Failures : 0
Total Msg sent to Optimization : 0
Total Direct Msg received from Optimization : 0
Total Indirect Msg received from Optimization: 0
Total Optimization Msg sent to Real Servers : 0

I've ran many show commands and have a great deal of output, but was wondering if there is a specific one I can use to pin point if this is an ACE/Configuration issue or a server issue?

We are using Leastcount Predictor and sticky when configuring hosts.

Hi Haider,If it hasn't been

Kanwaljeet Singh — Sat, 08 Aug 2015 02:41:48 GMT

Hi Haider,

If it hasn't been taken care of yet, then do "show serverfarm <serverfarm name> detail" and see if there are any connections being forwarded to SDHSPWFE02. You can also see what is the status of this server, inservice or probe failed etc. May be that this server is out of service and hence no requests are being sent to it. ACE will take the server out of LB rotation if it is in PROBE_FAILED state or OUT OF SERVICE.

I would start checking from there.

Regards,

Kanwal

Note: Please mark answers if they are helpful.

If you're using sticky based

TIM JUDGE — Sun, 09 Aug 2015 06:28:03 GMT

If you're using sticky based on source, is the source address the same for multiple clients (proxy maybe or some other server)? Perhaps you have stickiness working against you in some way.

You could also try "show conn serverfarm <serverfarm name>" and see if there is some kind of pattern