Uploaded image for project: 'Crowd Data Center'
  1. Crowd Data Center
  2. CWD-3940

AD Full Sync might leave the user map cache in a corrupted state if there is a failure when attempting to add a user to the internal directory

    XMLWordPrintable

Details

    • Bug
    • Resolution: Invalid
    • Medium
    • None
    • 2.8.0-OD-6, 2.8.0-OD-6-JIRA-01
    • Caching, Directory - LDAP
    • None

    Description

      This happens upon unsuccessfully adding the user to the internal directory.

      It does not matter why operation was unsuccessful, crowd will output a warning message and keep going with synchronization.

      Now the assumption that userMap contains the same set of users as the internal directory is not correct (user was already added to userMap in memory cache, but it was not added to the internal dir).

      Because synchronization will keep running in the next phase synchroniseGroupChanges is going to run and will try to create a membership for a non-existing user and we will see an error message in the logs.

      Crowd should be fixed such that we only add to the userMap the users that we were able to successfully add to the internal directory, and avoid running into problems in the later stages of the sync process.

      Attachments

        Issue Links

          Activity

            People

              aglowacki Arkadiusz Glowacki (Inactive)
              ohernandez@atlassian.com Oswaldo Hernandez (Inactive)
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: