Uploaded image for project: 'Lustre Documentation'
  1. Lustre Documentation
  2. LUDOC-485

document llite.*.read_ahead_range_kb parameter

Details

    • Improvement
    • Resolution: Unresolved
    • Minor
    • None
    • None
    • None
    • 9223372036854775807

    Description

      Add a description of the llite.*.read_ahead_range_kb parameter to LustreProc.xml. From patch https://review.whamcloud.com/38916 "LU-13669 llite: try to improve mmap performance":

          So the pattern could be something like this:
          
          [1M data] [hole] [0.5M data] [hole] [0.7M data] [1M data]
          
          Every time an application reads mmap data, it may not only
          read a single 4KB page, but also a cluster of nearby pages in
          a range(e.g. 1MB) of the first page after a cache miss.
          
          The readahead engine is modified to track the range size of
          a cluster of mmap reads, so that after a seek and/or cache miss,
          the range size is used to efficiently prefetch multiple pages
          in a single RPC rather than many small RPCs.
      

      Attachments

        Issue Links

          Activity

            [LUDOC-485] document llite.*.read_ahead_range_kb parameter
            adilger Andreas Dilger made changes -
            Link New: This issue is related to LU-13669 [ LU-13669 ]
            adilger Andreas Dilger made changes -
            Description Original: Add a description of the {{llite.\*.read_ahead_range_kb}} parameter to LustreProc.xml. New: Add a description of the {{llite.\*.read_ahead_range_kb}} parameter to LustreProc.xml. From patch https://review.whamcloud.com/38916 "{{LU-13669 llite: try to improve mmap performance}}":
            {quote}
            {noformat}
                So the pattern could be something like this:
                
                [1M data] [hole] [0.5M data] [hole] [0.7M data] [1M data]
                
                Every time an application reads mmap data, it may not only
                read a single 4KB page, but also a cluster of nearby pages in
                a range(e.g. 1MB) of the first page after a cache miss.
                
                The readahead engine is modified to track the range size of
                a cluster of mmap reads, so that after a seek and/or cache miss,
                the range size is used to efficiently prefetch multiple pages
                in a single RPC rather than many small RPCs.
            {noformat}
            {quote}
            adilger Andreas Dilger made changes -
            Link New: This issue is related to DDN-1769 [ DDN-1769 ]
            adilger Andreas Dilger created issue -

            People

              LM-Triage Lustre Manual Triage
              adilger Andreas Dilger
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated: