Export/Import of Recipe Runs

Recipe runs can be exported to a file for for later analysis. For example, to gather data like CPU and iperf statistic when onboarding a new recipe.

The data that is exported is the instance of lnst.Controller.Recipe.RecipeRun that was run.

The RecipeRun object is “pickled” and compressed with LZMA/XZ compression using lzma.

By default the file will be contain a file extension .lrc which stands for “LNST Run, Compressed”.

Use export_recipe_run() to export and import_recipe_run() to import.

lnst.Controller.Recipe.export_recipe_run(run: lnst.Controller.Recipe.RecipeRun, export_dir: str = None, name: str = None) → str

Export a recipe run to a file. RecipeRun is pickled and compressed.

Parameters:

• run (RecipeRun) – RecipeRun object to export.
• export_dir (str) – Directory to export file to. Defaults to run.log_dir
• name (str) – Name of output (exclusive of directory). Defaults to <recipename>-run-<timestamp>.lrc.

Returns:

Path of output file.

Return type:

str

Example:

ctl = Controller(...)
recipe = BondRecipe(...)
ctl.run(recipe)
...
>>> from lnst.Controller.Recipe import export_recipe_run
>>> path = export_recipe_run(recipe.run[0])
2020-10-02 15:20:58 (localhost) - INFO: Exported BondRecipe run data to /tmp/lnst-logs/2020-10-02_15:20:18/BondRecipe_match_0/BondRecipe-run-2020-10-02_15:20:58.lrc
>>> print(path)
/tmp/lnst-logs/2020-10-02_15:20:18/BondRecipe_match_0/BondRecipe-run-2020-10-02_15:20:58.lrc

lnst.Controller.Recipe.import_recipe_run(path: str) → lnst.Controller.Recipe.RecipeRun

Import a recipe run that was exported using export_recipe_run()

Parameters:path (str) – Path to file to import Returns:object which contains the imported recipe run Return type:RecipeRun

Example:

>>> from lnst.Controller.Recipe import import_recipe_run
>>> run = import_recipe_run("/tmp/lnst-logs/2020-10-02_15:20:18/BondRecipe_match_0/BondRecipe-run-2020-10-02_15:20:58.lrc")
>>> type(run)
<class 'lnst.Controller.Recipe.RecipeRun'>
>>> run.recipe.__class__
<class 'lnst.Recipes.ENRT.BondRecipe.BondRecipe'>
>>> run.results[38]
<lnst.Controller.RecipeResults.Result object at 0x7f20727e1e20>
>>> run.results[38].data
{'cpu': [[[<lnst.RecipeCommon.Perf.Results.PerfInterval object at 0x7f20727e1e50>,...]]], ... }
>>> print(run.results[38].description)
CPU Utilization on host host1:
cpu 'cpu': 45.40 +-0.00 time units per second
cpu 'cpu0': 45.40 +-0.00 time units per second