Backing up servers that participate in attachment consolidation (DAOS)

When databases on a server are set to participate in attachment consolidation, backing up the server requires an additional step of backing up all .NLO files stored in the Domino® Attachment and Object Service repository. Such files are never modified after creation, so after an initial backup, you need only perform incremental backups.

About this task

This procedure assumes that, in case of failure, you plan to restore the files to the originating Domino® server. If .NLO files have been encrypted with the originating server's key, they will be readable only when restored to the originating Domino® server. In addition, make sure that the file-access permissions and file ownership rights in the operating system to the DAOS directory remain unchanged after you run the backup utility; the Domino® and Attachment Object Service cannot continue working with an .NLO file if the original owner no longer has access to it.


  1. If you are using deferred deletion with DAOS, set the interval to longer than the interval between your backups. For example, if you back up weekly, specify 8 days for the setting "Defer deletion of DAOS objects n days" in the Server document.
  2. Back up .NSF files on the server using a backup utility that is compatible with NSF files. The utility must be able to use the backup and recovery methods of the Domino® C API Toolkit.
  3. Back up the DAOSCAT.NSF and DAOS.CFG files. These files are located in the data directory.
  4. Back up all .NLO files in your DAOS repository. You can use any flat file backup utility of your choice (such as Tivoli® Storage Manager). If DAOS has created subdirectories, maintain the directory hierarchy in your backup.
  5. After the first backup of the DAOS repository, perform incremental or full backups as desired of both .NSF and .NLO files.
  6. Optional: It is highly recommended that you archive any transaction logs so that changes that occurred after the last backup can be replayed for the most complete restoration of data.