Summary

An XML backup from coming from an Jira Data Center instance fails to be imported into another Jira Data Center instance, and the following symptoms are observed:

• the error Jira Software is currently unavailable. This might be because an upgrade task has failed to run or has not yet completed is thrown in the Jira UI, after the import progress reached 90%:
  

• the following error is thrown in the Jira application logs:

  2023-08-30 16:01:24,969+0000 JiraImportTaskExecutionThread-1 ERROR      [c.a.j.bc.dataimport.DefaultDataImportService] Error importing data: java.lang.IllegalArgumentException: Node must not be closed to be imported! </column>
  java.lang.IllegalArgumentException: Node must not be closed to be imported! </column>
  	at com.atlassian.dbexporter.importer.AbstractImporter.importNode(AbstractImporter.java:35)
  	at com.atlassian.dbexporter.DbImporter.importData(DbImporter.java:73)
  	at com.atlassian.activeobjects.backup.ActiveObjectsBackup.restore(ActiveObjectsBackup.java:170)
  	at com.atlassian.jira.bc.dataimport.DefaultDataImportService.restoreActiveObjects(DefaultDataImportService.java:561)
  	at com.atlassian.jira.bc.dataimport.DefaultDataImportService.performImport(DefaultDataImportService.java:736)
  	at com.atlassian.jira.bc.dataimport.DefaultDataImportService.doImport(DefaultDataImportService.java:323)
  	at com.atlassian.jira.web.action.setup.DataImportAsyncCommand.unsafeCall(DataImportAsyncCommand.java:82)
  	at com.atlassian.jira.web.action.setup.DataImportAsyncCommand.call(DataImportAsyncCommand.java:64)
  	at com.atlassian.jira.web.action.setup.DataImportAsyncCommand.call(DataImportAsyncCommand.java:30)
  	at com.atlassian.jira.task.ImportTaskManagerImpl$TaskCallableDecorator.call(ImportTaskManagerImpl.java:177)
  	at com.atlassian.jira.task.ImportTaskManagerImpl$TaskCallableDecorator.call(ImportTaskManagerImpl.java:149)
  	at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
  	at java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
  	at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
  	at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
  	at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
  	at java.base/java.lang.Thread.run(Thread.java:829)

The purpose of this article is to explain the meaning of this error, and to provide a way to address it.

Environment

Jira Server / Data Center on any version from 8.0.0.

Cause

The error found in the Jira application logs means that the activeobjects.xml file from the XML backup has an invalid format, meaning that it's missing at least one opening or missing tag.

For example, if we look at the extract shown below from an example of corrupted file, we can see that the tag <data tableName="AO_F1B27B_KEY_COMPONENT">  does not have a corresponding closing tag which should have been </data>:

...
  <data tableName="AO_F1B27B_KEY_COMPONENT">
    <column name="ID"/>
    <column name="KEY"/>
    <column name="TIMED_PROMISE_ID"/>
    <column name="VALUE"/>
    <row>
      <integer>31</integer>
      <string>issue.id</string>
      <integer>31</integer>
      <string>10708</string>
    </row>
  <data tableName="AO_F1B27B_KEY_COMP_HISTORY">
...

Even even though the error in the Jira logs is mentioning the tag </column>, it does not mean that the problem is with a tag called column. The word column is just a generic term thrown in the logs, and does not actually correspond to the tag that is missing an opening or closing tag. In the example shown above, even the closing tag </data> is missing from the XML file, the error in the Jira logs mentions the tag </column>.

The reason why this error is pointing to the activeobjects.xml file and not the entities.xml file is because of the method com.atlassian.activeobjects.backup.ActiveObjectsBackup.restore found in the stack trace. This method is responsible to extract and restore the data coming from the activeobjects.xml file.

Solution

Identifying the missing XML tag

This step can be a bit tricky, since there could be multiple missing tags, and since running various Linux commands on the file activeobjects.xml might take a lot of time depending on the size of this file.

For this section, we will take the example of a file activeobjects.xml that is missing one closing tag </data>.

Here are some steps you can follow to try to identify the missing XML tag:

• Unzip the XML backup

  • Replace xmlbackup.zip with the name of the zipped backup:

    unzip -a xmlbackup.zip

• Run the Linux command below, to check the validity of the activeobjects.xml file that came from the unzipped backup:

  xmllint --valid --stream activeobjects.xml

• Analyze the results:
  

  • The error that you are looking for is the error Opening and ending tag mismatch, as shown in the example of output below

  • In the example below, we can see that there is a tag mismatch between lines 46642 and 46761. In other words, there is an opening tag at line 46617, which is missing a corresponding closing tag

    activeobjects.xml:2: validity error : Validation failed: no DTD found !
    tp://www.atlassian.com/ao" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
                                                                                   ^
    activeobjects.xml:46761: parser error : Opening and ending tag mismatch: data line 46642 and backup
    </backup>
             ^
    Document activeobjects.xml does not validate
    activeobjects.xml : failed to parse

  • Note that you can ignore the error related to DTD, since this validation error is not relevant to the issue we are trying to fix

  • Note that the lines returned by the xmllint might not always be correct and might point to a non problematic tag rather than the problematic one, so please keep that in mind when analyzing the output of the command

• Let's check what is happening at the line 46642 along with the 20 lines after this line

  • Note that you might need to check more than 20 lines after the problematic line, depending on the content of the XML file

    sed -n '46642,46662p' activeobjects.xml

• Analyze the results:

  • We can see in the extract below from the XML file that the opening tag <data tableName="AO_F1B27B_KEY_COMPONENT"> does not have any closing tag </data>:

      <data tableName="AO_F1B27B_KEY_COMPONENT">
        <column name="ID"/>
        <column name="KEY"/>
        <column name="TIMED_PROMISE_ID"/>
        <column name="VALUE"/>
        <row>
          <integer>31</integer>
          <string>issue.id</string>
          <integer>31</integer>
          <string>10708</string>
        </row>
      <data tableName="AO_F1B27B_KEY_COMP_HISTORY">
        <column name="ID"/>
        <column name="KEY"/>
        <column name="TIMED_PROMISE_ID"/>
        <column name="VALUE"/>
        <row>
          <integer>3</integer>
          <string>issue.id</string>
          <integer>3</integer>
          <string>10400</string>

• One way to check if the <data> tag identified in the step above is the only <data> tag that is missing a closing tag is to run the 2 grep commands below and compare the results. The commands will result in exact counts of <data> and </data> tags:

  grep -c '<data ' activeobjects.xml
  grep -c '</data>' activeobjects.xml

• Analyse the results:

  • In the example below, we can see that there are 391 opening <data  ...> tags, while there are only 390 closing </data> tags. This is another indication that the XML file is missing 1 closing </data> tag:

    grep -c '<data ' activeobjects.xml
    391
    grep -c '</data>' activeobjects.xml
    390

Fixing the XML backup and performing the XML restore again

Now that we identified the missing XML tag in the previous step, let's add it to the activeobjects.xml file and re-zip the full XML backup:

• Add the missing tag to the activeobjects.xml file:
  

  • Let's assume that the line where the closing tag </data> is missing is the line number 46653.

  • In this case, the following command can be executed to insert the closing tag </data> at the line number 46653:

    sed -i '46653i</data>'activeobjects.xml

• Re-zip the full backup:

  • Replace archive.zip with the name of your choice

    zip archive.zip entities.xml activeobjects.xml		

• Try to restore the XML backup again on a working, vanilla Jira instance. In case the restoration finished properly, you will be seeing corresponding information message instead of the "Jira Software is currenaly unavailable" error.