Im having an issue with 1 specific VIP on my alteon 180e. The problem
is as follows:
Under periods of high(ish) load, one specific VIP seems to have trouble
loadbalancing. Every few seconds HTTP requests to that VIP take about
3.2 seconds to reply instead of the usual <200ms. When we graph the
HTTP request times, we can see the spikes occuring but the srange thing
is that the spikes are exactly the same height. ie. when the request
time slows down, it slows to the exact same time as the last spike
(about 3.2 seconds.) The web servers themselves are barely under any
load at all during this time. If we check HTP response times to the
Real IPs of the webservers, there is no slow down at all. The servers
are plugged directly into a port each on the Alteon.
We experience no other issues on any of the other VIPs.
The setup looks like this:
There are 2 webservers. Each webserver has 10 ip addresses configured
for 1 NIC. There are 10 websites running on each Webserver. We
therefore have 10 VIPS on the alteon. Each VIP contains 2 Real Ip
addresses in its group (One address from each webserver.)
Any help will be greatly apreciated - and dont be scared to point out
the obvious as this is the 1st time I'm dealing with Alteons.
The following is a copy of my config file with IPs and Mac`s changed to
garbage. The problem is occuring with the VIP at group 9.
script start "Alteon 180e" 4 /**** DO NOT EDIT THIS LINE!
/* Configuration dump taken 15:00:05 Tue Aug 23, 2005
/* Version 10.0.30, Base MAC address nn:yy:cf:xx:tt:c0
/c/slb/virt 1/service 7001