Upgrading with a XML backup restore will succeed even when an upgrade task is throwing an UpgradeError

While testing CONF-20489 i found that while an inplace upgrade will correctly fail to start up after an UpgradeError, an upgrade using a site backup will succeed.

Details:

• Upgrading a 3.0.2 instance to the latest 3.5 trunk
• Added the following lines to NetworkAndSiteNotificationTypesUpgradeTask:

  final UpgradeError upgradeError = new UpgradeError("Error while upgrading notifications table - FORCED", new Exception());
  errors.add(upgradeError);

• Performed an inplace upgrade - error was correctly thrown and Confluence will not start up:

  Upgrade failed. Please consult the system logs for details. You will need to fix these problems and restart.

• Repeated the steps, but upgrading by restoring an XML backup. The restore is successful and a 3.5 instance using the upgarded data is available. Error given in the logs is:

  2011-01-10 14:05:45,413 ERROR [Indexer: 1] [atlassian.bonnie.search.BaseDocument
  Builder] getDocument Error extracting search fields from comment: Erroneous orph
  aned comment v.1 (32790) using BackwardsCompatibleExtractor wrapping com.atlassi
  an.confluence.contributors.search.extractors.CommentContributionExtractor@9a9445
   (com.atlassian.confluence.contributors:commentContributionExtractor): null
   -- url: /confluence/admin/restore.action | userName: admin | referer: http://lo
  calhost:8080/confluence/admin/backup.action | action: restore
  java.lang.NullPointerException
          at com.atlassian.confluence.contributors.search.extractors.CommentContri
  butionExtractor.addFields(CommentContributionExtractor.java:30)
          at com.atlassian.confluence.plugin.descriptor.ExtractorModuleDescriptor$
  BackwardsCompatibleExtractor.addFields(ExtractorModuleDescriptor.java:45)
          at com.atlassian.bonnie.search.BaseDocumentBuilder.getDocument(BaseDocum
  entBuilder.java:104)
          at com.atlassian.confluence.search.lucene.ConfluenceDocumentBuilder.getD
  ocument(ConfluenceDocumentBuilder.java:102)
          at com.atlassian.confluence.search.lucene.tasks.AddDocumentIndexTask.per
  form(AddDocumentIndexTask.java:43)
          at com.atlassian.bonnie.index.TempIndexWriter.perform(TempIndexWriter.ja
  va:73)
          at com.atlassian.confluence.search.lucene.TempIndexWriterStrategy.perfor
  m(TempIndexWriterStrategy.java:43)
          at com.atlassian.confluence.search.lucene.tasks.TempIndexBackedIndexTask
  Performer.perform(TempIndexBackedIndexTaskPerformer.java:21)
          at com.atlassian.confluence.search.lucene.DefaultObjectQueueWorker.index
  Collection(DefaultObjectQueueWorker.java:78)
          at com.atlassian.confluence.search.lucene.DefaultObjectQueueWorker$1.doI
  nTransactionWithoutResult(DefaultObjectQueueWorker.java:62)
          at org.springframework.transaction.support.TransactionCallbackWithoutRes
  ult.doInTransaction(TransactionCallbackWithoutResult.java:33)
          at org.springframework.transaction.support.TransactionTemplate.execute(T
  ransactionTemplate.java:127)
          at com.atlassian.confluence.search.lucene.DefaultObjectQueueWorker.run(D
  efaultObjectQueueWorker.java:51)
          at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExec
  utor.java:886)
          at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor
  .java:908)
          at java.lang.Thread.run(Thread.java:619)
  2011-01-10 14:05:45,413 ERROR [Indexer: 1] [confluence.search.lucene.DefaultObje
  ctQueueWorker] indexCollection Error indexing document
   -- url: /confluence/admin/restore.action | userName: admin | referer: http://lo
  calhost:8080/confluence/admin/backup.action | action: restore
  java.lang.NullPointerException
          at com.atlassian.confluence.pages.Comment.getUrlPathForVersion(Comment.j
  ava:92)
          at com.atlassian.confluence.pages.Comment.getUrlPath(Comment.java:84)
          at com.atlassian.confluence.search.lucene.extractor.AddressableExtractor
  .addFields(AddressableExtractor.java:21)
          at com.atlassian.confluence.search.lucene.ChangeDocumentBuilder.getDocum
  ent(ChangeDocumentBuilder.java:97)
          at com.atlassian.confluence.search.lucene.tasks.AddChangeDocumentIndexTa
  sk.perform(AddChangeDocumentIndexTask.java:71)
          at com.atlassian.confluence.search.lucene.tasks.RebuildChangeDocumentsIn
  dexTask.perform(RebuildChangeDocumentsIndexTask.java:79)
          at com.atlassian.bonnie.index.TempIndexWriter.perform(TempIndexWriter.ja
  va:73)
          at com.atlassian.confluence.search.lucene.TempIndexWriterStrategy.perfor
  m(TempIndexWriterStrategy.java:43)
          at com.atlassian.confluence.search.lucene.tasks.TempIndexBackedIndexTask
  Performer.perform(TempIndexBackedIndexTaskPerformer.java:21)
          at com.atlassian.confluence.search.lucene.DefaultObjectQueueWorker.index
  Collection(DefaultObjectQueueWorker.java:83)
          at com.atlassian.confluence.search.lucene.DefaultObjectQueueWorker$1.doI
  nTransactionWithoutResult(DefaultObjectQueueWorker.java:62)
          at org.springframework.transaction.support.TransactionCallbackWithoutRes
  ult.doInTransaction(TransactionCallbackWithoutResult.java:33)
          at org.springframework.transaction.support.TransactionTemplate.execute(T
  ransactionTemplate.java:127)
          at com.atlassian.confluence.search.lucene.DefaultObjectQueueWorker.run(D
  efaultObjectQueueWorker.java:51)
          at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExec
  utor.java:886)
          at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor
  .java:908)
          at java.lang.Thread.run(Thread.java:619)