| #!/usr/bin/env bash | |
| # convert OpenLDAP schema file to LDIF file | |
| # | |
| # Copyright 2012 NDE Netzdesign und -entwicklung AG, Hamburg | |
| # Written by Jens-U. Mozdzen <[email protected]> | |
| # Copyright 2014 jaseg <[email protected]> | |
| # | |
| # Permission is granted to use, modify and redistribute this file as long as | |
| # - this copyright notice is left unmodified and included in the final code | |
| # - the original author is notified via email if this code is re-distributed as part of a paid-for deliverable | |
| # - the original author is not held liable for any damage, loss of profit, efforts or inconvenience of any sorts | |
| # that may result from using, modifying or redistributing this software. | |
| # | |
| # Use at your own risk - this code may not be suitable for your needs or even cause damage when used. | |
| # If you find any problems with this code, please let the author know so that it can be fixed or at least others | |
| # can be warned. | |
| # | |
| # Usage: schema2ldif.sh <fully-qualified schema file name> | |
| # | |
| # This program will try to convert the source file to an LDIF-style file, placing the resulting .ldif file | |
| # in the current directory. | |
| rc=0 | |
| if [ $# -gt 0 ] ; then | |
| slaptest=$(which slaptest 2>/dev/null ||ls /usr/sbin/slaptest||echo "") | |
| if [ -x "$slaptest" ] ; then | |
| schemaFile=$(readlink -f "$1") | |
| shift | |
| localdir=$(pwd) | |
| if [ -r "$schemaFile" ] ; then | |
| targetFile=$(basename "$schemaFile" .schema).ldif | |
| if [ ! -e "$localdir/$targetFile" ] ; then | |
| echo "$0: converting $schemaFile to LDIF $localdir/$targetFile" | |
| # create temp dir and config file | |
| tmpDir=$(mktemp -d) | |
| cd "$tmpDir" | |
| touch tmp.conf | |
| for dependency in "$@" | |
| do echo "include $dependency" >> tmp.conf | |
| done | |
| echo "include $schemaFile" >> tmp.conf | |
| # convert | |
| "$slaptest" -f tmp.conf -F "$tmpDir" | |
| # 3. rename and sanitize | |
| cd cn\=config/cn\=schema | |
| filenametmp=$(echo cn\=*"$targetFile") | |
| sed -r -e 's/^dn: cn=\{0\}(.*)$/dn: cn=\1,cn=schema,cn=config/' \ | |
| -e 's/cn: \{0\}(.*)$/cn: \1/' \ | |
| -e '/^structuralObjectClass: /d' \ | |
| -e '/^entryUUID: /d' \ | |
| -e '/^creatorsName: /d' \ | |
| -e '/^createTimestamp: /d' \ | |
| -e '/^entryCSN: /d' \ | |
| -e '/^modifiersName: /d' \ | |
| -e '/^modifyTimestamp: /d' \ | |
| -e '/^# AUTO-GENERATED FILE - DO NOT EDIT!! Use ldapmodify./d' \ | |
| -e '/^# CRC32 [0-9a-f]+/d' \ | |
| -e 's/^cn: \{[0-9]*\}(.*)$/cn: \1/' \ | |
| -e 's/^dn: cn=\{[0-9]*\}(.*)$/dn: cn=\1,cn=schema,cn=config/' < "$filenametmp" > "$localdir/$targetFile" | |
| # clean up | |
| echo "$0: LDIF file successfully created as $localdir/$targetFile" | |
| rc=0 | |
| rm -rf "$tmpDir" | |
| else | |
| echo "$0: target file $localdir/$targetFile already exists, aborting." >&2 | |
| rc=3 | |
| fi | |
| else | |
| echo "$0: source file $schemaFile could not be read, aborting." >&2 | |
| rc=2 | |
| fi | |
| else | |
| echo "$0: could not locate slaptest binary, exiting." >&2 | |
| rc=1 | |
| fi | |
| else | |
| echo "$0: usage: $0 <schemafile>" >&2 | |
| rc=99 | |
| fi | |
| exit $rc |
Wow this worked really well. Thanks
Thanks for the script, Jased. It was very helpful.
I suggest you to change line 52 to filenametmp=$(echo cn=*"${targetFile,,}") because if the initial schema filename contains uppercase characters, the script will fail.
Other suggestion is to change line 43 to echo "include $(realpath --relative-to=$localdir $dependency)", in order to allow the user to enter relative dependency paths.
This sadly won't work if multiple schemas are needed to be included, due to dependencies. For that you have to create your own conf file with includes, and adapt a part of this script to rewrite it.
This "tool" shows a complete lack of comprehension on how to use OpenLDAP and cn=config, I'd generally advise avoiding it. OpenLDAP ships with all standard schema in both a ".schema" file (for slapd.conf) and ".ldif" file for cn=config. ".ldif" file can be loaded into a cn=config database using the "ldapadd" utility (if it is online) or if one is doing a first time load of a cn=config database, one can simply use an include statement
This "tool" shows a complete lack of comprehension on how to use OpenLDAP and cn=config, I'd generally advise avoiding it. OpenLDAP ships with all standard schema in both a ".schema" file (for slapd.conf) and ".ldif" file for cn=config. ".ldif" file can be loaded into a cn=config database using the "ldapadd" utility (if it is online) or if one is doing a first time load of a cn=config database, one can simply use an include statement
So, just recently, I had a .schema file that I wanted to load into a fresh OpenLDAP 2.6.2 which is using cn=config. I wanted to add it to my list of includes (after core.ldif and cosine.ldif and so on. But I didn't have an .ldif version. What should I have done, if not what this script does? I'm new to OpenLDAP, and I'm eager to do things the right way....
Here's an example of how I converted the eduperson.schema file to a LDIF formatted schema file that can be used with ldapadd, etc:
mkdir -p /tmp/ldap/slapd.d
cp eduperson.schema /tmp/ldap
cd /tmp
I then created a slapd.conf file in /tmp/ldap with the following contents:
include /etc/ldap/schema/core.schema
include /etc/ldap/schema/cosine.schema
include /tmp/ldap/eduperson.schema
Then I converted this configuration over to cn=config format:
slaptest -f /tmp/ldap/slapd.conf -F /tmp/ldap/slapd.d
Then I exported just that schema file:
slapcat -n 0 -F /tmp/ldap/slapd.d -H "ldap:///???(cn={2}eduperson)" | grep -v ^structuralObjectClass: | grep -v ^entryUUID: | grep -v ^creatorsName: | grep -v ^createTimestamp: | grep -v ^entryCSN: | grep -v ^modifiersName: | grep -v ^modifyTimestamp: | sed -e "s/{2}eduperson/eduperson/" > eduperson.ldif
In the above example, the eduperson schema gets a weight of 2 because it was the 3rd schema file listed (core, cosine, eduperson). But now the end result is, I permanently have a converted form of the schema file I can use correctly with cn=config. We strip out the operational attributes as we would want those created fresh at import time (whether slapadd or ldapadd)
For an example of using an include statement with a cn=config database to use with slapadd, see my MMR blog post.
Here's an example of how I converted the eduperson.schema file to a LDIF formatted schema file that can be used with ldapadd, etc:
Excellent, thanks!
It looks like the big difference between your procedure and the script in this repo, is the use of slapcat to pull the data out of the temporary "cn=config" area....have I got that right? The script just hacks on the file "cn=config" file directly. I can see how it would be safer to trust slapcat to get the data...sort of follows the principle of using the "published API" rather than going to the underlying data.
@dukerobillard Generally, yes. You could also tweak it further to convert multiple schema at once and dump it directly to a file by using the -l option, and then pull out all of the schema you want converted in one go.
Struggeled with the same task to convert schema to ldif files wondering why there is no straight forward conversion tool available. Well there is:
https://github.com/fusiondirectory/schema2ldif
It is also available in the ubuntu repository.
slaptest literally is a straight forward conversion tool that ensures correctness of the result
slaptest literally is a straight forward conversion tool that ensures correctness of the result
Not quite, since you have to grep+sed it.
Indeed, sad that ldap does not come with an utility like
fuck-openldap.sh. Not sure why but it seems that everyone is having fun showing off their.schemabut when it is time to install it, there is no easy way. Withfuck-openldap.sh, it is easy and convenient.Thanks for this tool.