Don't setup selinux, and fixed ldap container setup

ldap reproducibly start the fail on my system with errors like this
(breaking the setup):

Mar 17 15:35:39 ldap ns-slapd[2955]: [17/Mar/2023:15:35:39.787214279 +0000] - NOTICE - ldbm_back_start - found 32755028k physical memory
Mar 17 15:35:39 ldap ns-slapd[2955]: [17/Mar/2023:15:35:39.790179451 +0000] - NOTICE - ldbm_back_start - found 21396840k available
Mar 17 15:35:39 ldap ns-slapd[2955]: [17/Mar/2023:15:35:39.793202804 +0000] - NOTICE - ldbm_back_start - cache autosizing: db cache: 1572864k
Mar 17 15:35:39 ldap ns-slapd[2955]: [17/Mar/2023:15:35:39.796339098 +0000] - NOTICE - ldbm_back_start - total cache size: 1610612736 B;
Mar 17 15:35:39 ldap ns-slapd[2955]: [17/Mar/2023:15:35:39.799636323 +0000] - ERR - no_diskspace - No enough space left on device (/dev/shm/slapd-kolab) (67104768 bytes); at least 1771674009 bytes space is needed for db region files
Mar 17 15:35:39 ldap ns-slapd[2955]: [17/Mar/2023:15:35:39.802385952 +0000] - ERR - ldbm_back_start - Failed to init database, err=28 No space left on device
Mar 17 15:35:39 ldap ns-slapd[2955]: [17/Mar/2023:15:35:39.805175036 +0000] - ERR - plugin_dependency_startall - Failed to start database plugin ldbm database

There doesn't seem to be a decent way to limit the size of cache in
dscreate (there's some magic in dscontainer).

I finally found ad020b0c3c82eec23f36a6759781cbc0a86865de in the ldap
repo, which probably explains what's going on (I assume we got bitten due to an ldap version
upgrade). See also https://pagure.io/389-ds-base/issue/51008
So we moved db/db_home back out of /dev/shm