[lvs-users] Load Balancer dropping requests

[Bruce Thompson]

Hi,

We are seeing a problem whereby the LB's for LDAPs in our performance test
environment drop connections sometimes to the SUN eXchange domains. This
doesn't appear to happen when going direct to an LDAP server instead of via
the LB's VIP.

The problem is that the symptoms only appear under load and long test runs.
I noticed by reading the how-to that the default timeout for a tcp
connection is 15 minutes, if it hasn't received tcpfin. Might I be better to
set this to 12 hours instead? A 10k test runs over 15 minutes, and we're
loosing 4000 - 7000 messages.

Here is our our ldirectord.cf:

b2b04util01:/etc/ha.d # cat ldirectord.cf |grep -v "^\#" |grep -v "^$"
checktimeout=10
checkinterval=5
autoreload=yes
logfile="/var/log/messages"
quiescent=no
virtual=10.32.24.48:1389
        real=10.32.24.186:1389 gate 10
        real=10.32.24.187:1389 gate 10
        real=10.32.24.10:1389 gate 10
        real=10.32.24.11:1389 gate 10
        real=10.32.24.12:1389 gate 10
        real=10.32.24.22:1389 gate 10
        service=ldap
        checktype=negotiate
        checkport=1389
        #negotiatetimeout=10
        request="dc=xxx,dc=xxx,dc=xxx"
        receive="dc=xxx,dc=xxx,dc=xxx"
        scheduler=rr
        protocol=tcp
b2b04util01:/etc/ha.d #

Any advice would be greatly appreciated.

Regards,
Bruce

-- 
m: +64 (0)21 962 085
h: +64 (0)4 384 6482
e: thompson.bj at gmail.com
skype: bjthompson