slapd does not start when already configured

[muellermartin]

I just found a bug when starting the container with an already configured slapd and prepopulated database e.g. by running the container with this command on pre-existing etc and lib folder:

docker run -d -p 389:389 --name ldap -v $PWD/etc:/etc/ldap -v $PWD/lib:/var/lib/ldap dinkel/openldap

docker ps -a returns this line for the started container (exit status 1):

77b130e80a2c        dinkel/openldap            "/entrypoint.sh sl..."   18 seconds ago      Exited (1) 18 seconds ago                            ldap

docker logs ldap returns nothing.

By debugging the entrypoint.sh shell script does exit at line 100:

slapd_configs_in_env=`env | grep 'SLAPD_'`

Due to setting set -e in line 8 and due to the fact that grep exits with a non-zero exit status when it does not find the search pattern, the script exits prematurely and does not proceed to the call of exec.

As a fix I suggest appending true to the subshell expression in line 100 like this:

slapd_configs_in_env=`env | grep 'SLAPD_' | true`

I will prepare a PR shortly for this. Edit: I saw there is already PR #18 which addresses this issue.

[muellermartin]

As a temporary workaround -e SLAPD_= can be passed to the docker run command.