Bug 1695014

Description of problem:

Trying to change the NOFILE limits (limit of opened file descriptors that the kernel will allow), I opened the dirsrv@.service.
It redirects to /etc/sysconfig/dirsrv.system

```
# if you need to set other directives e.g. LimitNOFILE=8192
# set them in this file
.include /etc/sysconfig/dirsrv.system
```

So I opened the said file and changed the data there, as described :

```
# This controls the number of file handles avaliable. File handles
# correlate to sockets for the process, and our access to logs and
# databases.
#LimitNOFILE=16384
LimitNOFILE=65535
```

And restarted the service.

Upon restart, *only* the hard-limits was modified. Soft limit was still 1024 :

```
# cat /proc/$(cat /var/run/dirsrv/slapd-CEDRIC-COM.pid)/limits  | grep -e Limit -e "open files"
Limit                     Soft Limit           Hard Limit           Units     
Max open files            1024                 65535                files
```

Upon further search, it appears that ns-slapd overrides its NOFILE soft limits to `nsslapd-maxdescriptors` from cn=config, and defaults to 1024.

I would propose the following changes :

1) by default, if nsslapd-maxdescriptors isn't set in dse.ldif, ns-slapd should not override its corresponding soft limit value. (especially in systemd-enabled system, where there is already a configuration for this)

2) The comments in /etc/sysconfig/dirsrv.system should make it clear that there is a nsslapd-maxdescriptors cn=config entry that may/will override the soft limit