Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have an LVS setup with NAT routing and two real servers (as an FYI, we're only routing http on tcp port 80). Currently requests to real server A respond just fine, with page load times of 3-4 seconds. However, requests to real server B are horribly slow with page load times upward of 5 minutes.

Now, the other day, it was just the opposite where requests to real server B were loading in an acceptable time frame but requests to server A were dog slow.

In either case, requests directly to the real servers bypassing the load blancer respond just fine. It's only requests that pass through the load balancer that have problems, and only requests to one of the two real servers at that.

As another FYI, the LVS server is running CentOS 6.3 with the default distro release of lvs.

Could anyone give me some suggestions as to where to look for what might be causing this? I would be happy to provide any details about my configuration if needed.

share|improve this question
Does the problem occurs under heavy load or all the time? – user130370 Dec 31 '12 at 15:42
All the time. In fact, there is rarely any load on either real server. – Alex Dec 31 '12 at 15:44
I guess you already looked at ? – user130370 Dec 31 '12 at 15:46
Yea, I've already looked at that. generic-receive-offload is already set to off. – Alex Dec 31 '12 at 17:27
anyway, centos6 ships recent kernel so it couldn't be that :/ real servers are running centos too? – user130370 Dec 31 '12 at 17:37

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.