Monday, October 19, 2009

Fix, /opt/iplanet/ldap/slapd-hostname/start-slapd: not found

Eric reported me slow login on Solaris 10 machines, i started my investigation from name resolution & i found that the machine was unable to resolve hostname(s) from DNS. initially i was expecting some kind of misconfiguration of resolv.conf but lately i found that DNS server of that environment was down!

So i simply shoot

#svcadm -v restart network/iplanetldap

but found that service did'nt started :(

anyhow with further investigation i found following,


when you will run svcadm -v restart network/iplanetldap you will get following error in log file i.e.


# cat /var/svc/log/network-iplanetldap:default.log

[ Oct 18 23:02:34 Executing start method ("/opt/iplanet/ldap/slapd-hostname/start-slapd") ]
/sbin/sh: /opt/iplanet/ldap/slapd-hostname/start-slapd: not found


HOSTNAME in error was wrong! so i found the cause, that somehow ldap daemon startup script is not getting correct hostname of the machine.


Run following command to check store configuration for ldap daemon.

Step 01: #svccfg export network/iplanetldap > /tmp/ldap-old-config.xml

this will create a file “/tmp/ldap-old-config.xml” you should vi/edit this file & correct hostname inside.

Output of the file is:

bash-3.00# vi /tmp/ldap-old-config.xml

edit following lines:

name='start' type='method' exec='/opt/iplanet/ldap/slapd-hostname/start-slapd' timeout_seconds='60'
name='stop' type='method' exec='/opt/iplanet/ldap/slapd-hostname/stop-slapd' timeout_seconds='60'

Step 02: svccfg import /tmp/ldap-old.xml

Step 03: svcadm -v restart network/iplanetldap

Step 04: svcadm -v clear network/iplanetldap

Step 05: svcs -l network/iplanetldap


No comments: