Skip to content

Latest commit

 

History

History
44 lines (41 loc) · 3.42 KB

csv-docs.md

File metadata and controls

44 lines (41 loc) · 3.42 KB

elbencho CSV Results File

elbencho supports operations against both directories (for example --mkdirs) and files (for example --write). Each elbencho operation however only deals with one of these so the generic name entry is used in the CSV fields. See the operation column for the row to determine if the operation was working with directory or file entries.

As mentioned in the main README file, elbencho presents two result columns: One showing the aggregate status at the point in time when the fastest I/O thread finished its work ("First Done") and one for the aggregate end result at the point in time when the slowest thread finished its work ("Last Done"). Results for each of these are stored separately in the CSV with columns appended with [first] and [last] respectively.

Fields

  • ISO date - timestamp when the row was written to the CSV file when the test completed
  • label - the label, if any, specified for this run on the command line (see --label)
  • path type - type of path passed into elbencho (dir or file)
  • paths - number of paths passed into elbencho
  • hosts - number of hosts generating load
  • threads - number of threads per host generating load
  • dirs - number of directories per thread
  • files - number of files per thread
  • file size - size of the files in bytes
  • block size - block size in bytes
  • direct IO - 1 if direct IO is used, else 0
  • random - 1 if random operations were used, else 0 (the default)
  • random aligned - 1 if random operations were block size aligned (the default), 0 if they were not (see --norandalign) or blank if no random operations were used
  • IO depth - IO depth used
  • shared paths - 1 if the paths were shared by all hosts (the default), else 0 (see --nosvcshare)
  • truncate - 1 if files were truncated before writing (see --trunc), else 0 (the default)
  • operation - file operation being performed (READ, WRITE, etc)
  • time ms [first] & [last] - time in milliseconds since the test started and the first and last thread finished
  • CPU% [first] & [last] - CPU percentage of the run as seen by the first and last thread finished
  • Operation results
    • entries/s [first] & [last] - number of entries per second for the first and last thread finished
    • IOPS [first] & [last] - number of IOPS for the first and last thread finished
    • MiB/s [first] & [last] - throughput in MiB/s for the first and last thread finished
    • entries [first] & [last] - total entries created by the first and last thread finished
    • MiB [first] & [last] - total MiB read/written by the first and last thread finished
    • Ent lat us [min] & [avg] & [max] - time in microseconds to complete an entry (file or directory)
    • IO lat us [min] & [avg] & [max] - time in microseconds to complete an IOP, that is: IO latency
  • When using --rwmixpct to do mixed reads and writes concurrently, the entries/s and related fields above report the data for the WRITE operations. The results of the READ operations are stored in the rwmix read fields.
    • rwmix read entries/s [first] & [last]
    • rwmix read IOPS [first] & [last]
    • rwmix read MiB/s [first] & [last]
    • rwmix read entries [first] & [last]
    • rwmix read MiB [first] & [last]
    • rwmix read Ent lat us [min] & [avg] & [max]
    • rwmix read IO lat us [min] & [avg] & [max]
  • version - version of elbencho
  • command - elbencho command line used