Starting and stopping DSEE 6.3 directory and admin servers

Put this under the heading, “basic directory service operations for DSEE 6”.

While there are a lot of new things to learn about managing Sun’s Directory Server Enterprise Edition 6.3, actually knowing how to start and stop the various component parts is a good place to start. Particularly after your first host system reboot.

Let’s assume the environment in my previous post. Directory system user is sunds, who shares its home with the directory server itself under /opt/sun/dirsrv6. The environment variable $DH has been set to this location. The initial directory server has been created as $DH/var/ds-user1 and runs as sunds, but has been configured to start as root so that it can listen on standard LDAP and LDAPS ports 389 and 636, respectively.

Here’s the procedure for starting everything up:

1. Start tomcat app server. $DH/tomcat5/bin/startup.sh;

2. Start cacao server. $DH/dsee6/cacao_2/cacao/bin/cacaoadm start;

3. Start DSCC registry. $DH/ds6/bin/dsadm start $DH/var/dscc6/dcc/ads;

4. Start directory instance. $DH/ds6/bin/dsadm start $DH/var/ds-user1 (remember to execute this step as root if your server listens on a port below 1000).

To stop things gracefully, you’d go through the following steps:

1. Stop directory instance. $DH/ds6/bin/dsadm stop $DH/var/ds-user1 (remember to execute this step as root if your server listens on a port below 1000);

2. Stop the DSCC. $DH/ds6/bin/dsadm stop $DH/var/dscc6/dcc/ads;

3. Stop cacao. $DH/dsee6/cacao_2/cacao/bin/cacaoadm stop;

4. Stop tomcat. $DH/tomcat5/bin/shutdown.sh.

Of course when it comes shutdown, the only step I’d say would be really essential is #1. While the Sun Directory products have always been pretty robust when it comes to “surprise” reboots, that first experience of a corrupted directory database is one too many for most of us. You may really start or stop the directory instance itself at any time in the process, since it isn’t dependent on any of the other pieces (and none of them are dependent on it). In fact the only really necessary sequence would be steps 1 – 3 to get the administration server going properly. Even there, I’ve actually had things work even when that gets jumbled up, but you didn’t hear it from me.