Hi all,
we use a cisco content switching module (CSM) with software version 4.3(1) and we balance a ldap service on two ldap servers.
For probing the real servers we use the default script (see attachment).
The probing interval is set to 10 seconds, the retries to 2.
Monitoring the real servers with ldap probes does not work as expected in the special case when the real server needs very long to answer the probe requests.
So the CSM does not mark the real server down when the answer exceeds the two retries (e.g. after 40s the real server answers).
The config for the vserver is showed below:
===
vserver LDAPDIR
virtual <ip> tcp 389
vlan 999
replicate csrp connection
no persistent rebalance
slb-policy LDAPDIR-POL
inservice
policy LDAPDIR-POL
serverfarm LDAPDIR-F
serverfarm LDAPDIR-F
nat server
failaction purge
real name <server1>
inservice
real name <server2>
inservice
probe LDAP-HEALTH389
real <server1>
address <ip>
inservice
real <server2>
address <ip>
inservice
probe LDAP-HEALTH389 script
script LDAP_PROBE
interval 10
retries 2
port 389
===
Does anybody know what's the reaseon for this behaviour?
Many thanks in advance,
Thorsten Steffen
