BUG: restore from xml.zip failed

[Jack Wu]

When I try to restore the confluence site it gives me the error: null index column for collection: com.atlassian.crowd.model.application.ApplicationImpl.directoryMappings

I checked the contents of entities.xml and found that listIndex was handled incorrectly, which eventually led to an exception.

xml content:

<object class="DirectoryMapping" package="com.atlassian.crowd.model.application">
<id name="id">163841</id>
<property name="allowAllToAuthenticate">true</property>
<property name="listIndex" column="list_index">0</property>

After restoration, list_Index is set to null.

mysql> select * from cwd_app_dir_mapping;
+--------+----------------+--------------+-------- ---+----------------+
| id | application_id | directory_id | allow_all | list_index |
+--------+----------------+--------------+-------- ---+----------------+
| 163841 | 98305 | 131073 | T | NULL |
+--------+----------------+--------------+-------- ---+----------------+

Before restore, the value of list_index is 0, if I set it to 0 manually, confluence will work fine.

mysql> select * from cwd_app_dir_mapping;
+--------+----------------+--------------+-------- ---+----------------+
| id | application_id | directory_id | allow_all | list_index |
+--------+----------------+--------------+-------- ---+----------------+
| 163841 | 98305 | 131073 | T | 0 |
+--------+----------------+--------------+-------- ---+----------------+
1 row in set (0.01 sec)

I checked the hibernate configuration and found that it seems that the correct cloumn is not configured for listIndex?

<class name="com.atlassian.crowd.model.application.DirectoryMapping" table="cwd_app_dir_mapping">
        <cache usage="read-write"/>

        <id name="id" column="id" unsaved-value="null">
            <generator class="com.atlassian.confluence.impl.hibernate.ResettableTableHiLoGeneratorV5"/>
        </id>

        <many-to-one name="application" class="com.atlassian.crowd.model.application.ApplicationImpl"
                     update="false" insert="true" column="application_id" index="idx_app_dir_app"/>

        <many-to-one name="directory" class="com.atlassian.crowd.model.directory.DirectoryImpl"
                     column="directory_id" foreign-key="fk_app_dir_dir" index="idx_app_dir_dir" not-null="true"/>

        <property name="allowAllToAuthenticate" type="true_false" not-null="true" column="allow_all"/>

        <set name="authorisedGroups" inverse="true" cascade="all-delete-orphan" optimistic-lock="false">
            <cache usage="read-write"/>
            <key column="app_dir_mapping_id"/>
            <one-to-many class="com.atlassian.crowd.model.application.GroupMapping"/>
        </set>

        <set name="allowedOperations" table="cwd_app_dir_operation" cascade="all" optimistic-lock="false">
            <cache usage="read-write"/>
            <key column="app_dir_mapping_id" foreign-key="fk_app_dir_mapping"/>
            <element column="operation_type" length="32" not-null="true" type="com.atlassian.crowd.embedded.hibernate2.types.OperationPersistentType"/>
        </set>

    </class>

[Dan Breyen · Accepted solution]

Hi @Jack Wu welcome to the community.  I took at quick look at Atlassian's Jira site (jira.atlassian.com) and didn't find anything that resembles this issue, but I could have missed it. 

If the exporting isn't creating indexes correctly, you may want to contact support (support.atlassian.com/contact) and get some feedback from them.