v7
Applies to Campaign Classic v7 only
Applies to on-premise and hybrid deployments only
Switching to Unicode switching-to-unicode
For an existing prod instance in Linux/PostgreSQL, the steps for switching to unicode are as follows:
-
Stop the processes writing to the database:
code language-none su - neolane nlserver shutdown
-
Dump the database:
code language-none su - postgres pg_dump mydatabase > mydatabase.sql
-
Create a Unicode database:
code language-none createdb -E UNICODE mydatabase_unicode
-
Restore the database:
code language-none psql mydatabase_unicode < mydatabase.sql
-
Update the option indicating that the database is Unicode:
code language-none psql mydatabase_unicode update XtkOption set sStringValue = 'u'||sStringValue where sName='XtkDatabaseId' and sStringValue not like 'u%';
-
On the tracking servers:
code language-none su - neolane cd nl6/conf vi config-prod.xml
Add the u character in front of the value relating to the database identifier (databaseId):
code language-none <web> <redirection databaseId="u7F0000010554364C" trackingPassword="myPassword="/> </web>
-
On servers calling the database:
code language-none su - neolane cd nl6/conf vi config-prod.xml
Modify the database reference:
code language-none <dataSource name="default"> <dbcnx encrypted="1" login="<dbuser>:<base_unicode>" password="xxxx=" provider="postgresql" server="yyyy"/> </dataSource>
-
Reboot all the machines:
code language-none /etc/init.d/apache stop /etc/init.d/nlserver6 stop /etc/init.d/nlserver6 start /etc/init.d/apache start
-
Confirm the switch. To do this, connect via the Adobe Campaign console and:
- check that the data is displayed correctly, in particular the accentuated characters:
- launch a delivery and check that the tracking retrieval works.
recommendation-more-help
601d79c3-e613-4db3-889a-ae959cd9e3e1