Jira Project Import fails with "There was an unexpected problem processing the backup XML"

Summary

Fix Jira Data Center Project Import failure due to the error "There was an unexpected problem processing the backup XML"

Environment

Jira Data Center.

Diagnosis

When analyzing the logs, we can see that Jira failed to parse the XML backup with a NullPointer exception (NPE):

You might also see entries such as the one below, however, these are non-blocker errors. Jira will just skip that specific custom field value for a specific issue and proceed with the import accordingly:

One might believe that this is actually the custom field with problems when in fact it's not.

Cause

Verifying the product's source code, we can see that the NPE happened because Jira tried to find any of the following entries in the CustomFieldValue section of the XML file but failed:

• textvalue

• stringvalue

• numbervalue

This indicates an inconsistency in the source instance database. There are issues with reference to the customfieldvalue table in the database as if they had an entry for this custom field, however, they have a null value.

Solution

We can identify the problematic entries by using grep, which is available by default in Mac and Linux operating systems.

1. Extract the backup file using unzip, for example: unzip JiraBackup.xml

2. Run the following command to identify the entries causing the NPE: 

   1 cat entities.xml|egrep '<CustomFieldValue id="[0-9]*" issue="[0-9]*" customfield="[0-9]*"/>'

3. You should receive an output similar to: 

   1 2 3 <CustomFieldValue id="11077" issue="12269" customfield="10600"/> <CustomFieldValue id="11078" issue="13269" customfield="10600"/> <CustomFieldValue id="11079" issue="15187" customfield="10603"/>

This indicates that we have three issues with empty custom field values in the database. There are two ways to resolve this, the first of which is ideal as we will fix the data on the source and prevent this from impacting any future imports. 

Solution #1 (recommended)

1. On the source instance database, identify the custom field name that has a null value in the database:

   1 SELECT * FROM customfield WHERE id in (10600,10603);

2. Identify the problematic issues, e.g: issue="12269" (do this for all three issues with different ID):

   1 2 3 4 5 SELECT project.pkey, jiraissue.issuenum FROM jiraissue INNER JOIN project ON (jiraissue.project = project.id) WHERE jiraissue.id='12269'

3. Populate the custom field in the issue and re-generate the XML backup for project import in the destination instance

Solution #2 

Update the problematic entries in the existing XML backup and re-do the project import.

1. Open the entities.xml backup file, search for the problematic entries identified earlier in the KB

2. Add textvalue="" to these entries, so their outcome should be similar to:

   1 2 3 <CustomFieldValue id="11077" issue="12269" customfield="10600" textvalue=""/> <CustomFieldValue id="11078" issue="13269" customfield="10600" textvalue=""/> <CustomFieldValue id="11079" issue="15187" customfield="10603" textvalue=""/>

3. Zip the contents in .ZIP format and retry the project import.

If the problem persists after the steps above, please get in touch with Atlassian Support.