Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-2904

Fix a bug where the reported checkpoint size could be many times data size

    • Storage 2016-10-31

      We've seen data source statistics where the reported checkpoint size of the table is many times (up to 15x) larger than the reported table size. They're accounted differently as described by keith.bostic:

      The two statistics are gathered differently: the checkpoint size is the last calculated checkpoint size based on the blocks allocated and discarded in the checkpoint (which could be from some time ago), where the file size is the current file size as the that instant. I don't expect that to make a difference though, because the principle way a file is truncated is by doing a checkpoint (but it's complicated enough code that I wouldn't want to swear it can't happen).

            keith.bostic@mongodb.com Keith Bostic (Inactive)
            sue.loverso@mongodb.com Susan LoVerso
            0 Vote for this issue
            5 Start watching this issue