Description Sönke Schwardt-Krummrich 2019-10-04 16:30:27 CEST

Scenario:
Users were created with the old user import (import_user) or manually via the UMC module "Benutzer (Schulen)" in LDAP. The LDAP attributes ucsschoolRecordUID and ucsschoolSourceUID are either not set or are set with ucsschoolRecordUID==$UID and ucsschoolSourceUID=="LegacyDB". This is usually impractical when switching to the new CLI import (ucs-school-import-user) or the UMC module "Benutzer-Import". For the two new import paths it is assumed that ucsschoolRecordUID and ucsschoolSourceUID are set with meaningful values.

A script is required which supports the migration of user objects from the old to the new import.

Solution approach:
A script supports the domain administrator in migrating the user objects by passing a CSV file to a script to be implemented. The CSV file consists of two columns with user name and RecordUID. The script then sets the corresponding RecordUID for the specified user objects. Optionally, a SourceUID can be specified via command line, which is then set homogeneously for all users specified in the file.
This step has to be done only once for the existing users. The user objects can then all be maintained via e.g. ucs-school-import-user.

Beispiel:
# cd /usr/share/ucs-school-import/scripts/
# cat mapping.csv
"username";"record_uid"
"anton123";"12345678"
"tom.teacher";"L887766"
"jeff.staff";"E-2345234-234-B"
# ./migrate_ucsschool_import_user --input-file=mapping.csv --modify-record-uid --source-uid="gsmitte"
User: 'anton123'   record_uid: [] ==> ['12345678']   source_uid: [] ==> ['gsmitte']
User: 'tom.teacher'   record_uid: [] ==> ['L887766']   source_uid: ['LegacyDB'] ==> ['gsmitte']
User: 'jeff.staff'   record_uid: [] ==> ['E-2345234-234-B']   source_uid: [] ==> ['gsmitte']
# cat /var/log/univention/ucs-school-migration-import-user.log
2019-09-13 15:49:20 INFO  migrate_ucsschool_import_user.__init__:142  Given arguments: ['/usr/share/ucs-school-import/scripts/migrate_ucsschool_import_user', '--modify-record-uid', '--input-file=mapping.csv', '--source-uid=gsmitte']
2019-09-13 15:49:20 INFO  migrate_ucsschool_import_user.modify_record_uid:294  User: 'anton123'   record_uid: [] ==> ['12345678']   source_uid: [] ==> ['gsmitte']
2019-09-13 15:49:20 INFO  migrate_ucsschool_import_user.modify_record_uid:294  User: 'tom.teacher'   record_uid: [] ==> ['L887766']   source_uid: ['LegacyDB'] ==> ['gsmitte']
2019-09-13 15:49:20 INFO  migrate_ucsschool_import_user.modify_record_uid:294  User: 'jeff.staff'   record_uid: [] ==> ['E-2345234-234-B']   source_uid: [] ==> ['gsmitte']
#

To simplify the creation of such a two-column CSV file, the future CSV file can be used for ucs-school-import-user if it contains the information "First Name", "Last Name" and "RecordUID" for each user. This file can be passed to the migrate_ucsschool_import_user script in a previous step using the --guess-usernames parameter. In addition to the parameter, the column in which the values for first name, last name and RecordUID are located must also be specified.

The script then attempts to identify the relevant user names using first and last names and generates the file "mapping.csv" required in the above example.
Since first and last names do not have to be unique, it is possible that either no user is found or even several user objects are found in the case of frequently occurring names. These problem cases are displayed in the output file at the very beginning and must be assigned manually. All cases that were uniquely assigned are listed below. In any case, the entire file should be checked again to avoid incorrect assignments.

Example:
# cd /usr/share/ucs-school-import/scripts/
# cat import.csv
"record_uid";"school";"lastname";"firstname";"classes";"comment"
"12345678";"gsmitte";"Meyer";"Anton";"gsmitte-1A";""
"L887766";"gsmitte";"Teacher";"Tom";"";"just some comments"
"E-2345234-234-B";"gsmitte";"Staff";"Jeff";"";"in-house technician"
# cat: import.csv: No such file or directory
# ./migrate_ucsschool_import_user --input-file=import.csv --guess-usernames --column-firstname=4 --column-lastname=3 --column-record-uid=1 --output-file=mapping.csv
# cat import.csv
"This CSV file consists of 2 sections:"
"The first section lists all problematic users for whom either *no* or *several* user names were found"
"in LDAP by first and last name. These entries must be checked and corrected manually. The entries of the"
"first section can be recognized by a note in the comment column and the column for the user name being empty."
"The second section contains all entries for which exactly one user was found in LDAP with the specified"
"first and last name. These should nevertheless be checked again for correctness."
"This file contains 2 ambiguous entries and 1 unambiguous entries."
"After completion of the corrections, the lines with this text paragraph must be removed!"
""
"username","record_uid","comment","input_data"
"","12345678","multiple users found: anton8,anton64,anton123","12345678 § gsmitte § Meyer § Anton § gsmitte-1A § "
"","E-2345234-234-B","no user found in LDAP","E-2345234-234-B § gsmitte § Staff § Jeff §  § in-house technician"
"tom.teacher","L887766","",""
#
<CORRECT THE FILE MANUALLY HERE ==> FILL COLUMN FOR USERNAME>
# cat mapping.csv
"username","record_uid","comment","input_data"
"anton123","12345678","multiple users found: anton8,anton64,anton123","12345678 § gsmitte § Meyer § Anton § gsmitte-1A § "
"jeff.staff","E-2345234-234-B","no user found in LDAP","E-2345234-234-B § gsmitte § Staff § Jeff §  § in-house technician"
"tom.teacher","L887766","",""
# ./migrate_ucsschool_import_user --input-file=mapping.csv --modify-record-uid --source-uid="gsmitte"
User: 'anton123'   record_uid: [] ==> ['12345678']   source_uid: [] ==> ['gsmitte']
User: 'jeff.staff'   record_uid: [] ==> ['E-2345234-234-B']   source_uid: [] ==> ['gsmitte']
User: 'tom.teacher'   record_uid: [] ==> ['L887766']   source_uid: ['LegacyDB'] ==> ['gsmitte']
#