Implement the first phase of the restore algorithm where the storage engine prepares the backup

XMLWordPrintableJSON

    • Type: Task
    • Resolution: Won't Do
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • None
    • Storage Execution
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      • Retrieve a list of the files that exist in the _restore directory.
      • Remove any entries from the WiredTiger.backup file for files not present in the _restore directory.
      • Start an independent instance of WiredTiger on the _restore directory.
        • WiredTiger will reconstruct the WiredTiger.wt file from the WiredTiger.backup file.
        • The journal files will be replayed. Replaying journal files can result in new files being created.
      • Using the list of files retrieved before starting WiredTiger, drop any newly created collection and index files.
      • Iterate over the records in the _mdb_catalog.wt file and do the following:
        • Records mapping to files not present in the _restore directory will be removed.
        • Check that all of the index files are present for a collection being restored.
        • Modify the namespace fields where a database rename was requested.
        • Modify the namespace fields of the record representing the oplog to local.oplog.restore.
      • If a collection to restore is specified in config.system.restore but is not present in the _mdb_catalog.wt and vice-versa, abort the procedure.
      • Take a checkpoint.
      • For each collection and its indexes to restore, and the oplog:
        • Export the file and table metadata from WiredTiger (see here).
        • Export the catalog entry (_mdb_catalog.wt) and fast count (sizeStorer.wt).
      • Shutdown the WiredTiger instance.

            Assignee:
            [DO NOT USE] Backlog - Storage Execution Team
            Reporter:
            Gregory Wlodarek
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated:
              Resolved: