Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

HPCC-33137 Add supports for executing the cost optimizers in Thor #19423

Open
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

shamser
Copy link
Contributor

@shamser shamser commented Jan 16, 2025

Cost optimizer will execute in Thor for containerized deployments. It will continue to execute in EclAgent for non-containerized deployment.

The reason for this change is that in containerized deployments EclAgent cannot calculate the cost of the Thor cluster and so cannot not assigned the cost of issues reported by the cost optimizer. It cannot calculate the cost of the Thor cluster because Thor's cost parameters and resource configuration is not available to EclAgent - this information is only available in the Thor's configuration.

Note, that as the Thor manager executes on a per job basis, it executes the analyzer after every graph executes. The concept of "end of workunit" does not exist at present in Thor so the analyzeWhenComplete==true has not been implemented for Thor.

The changes:

  • Support for executing cost optimizer in Thor. This the default in containerized.
  • New 'analyzeInEclAgent' option - if true, the analyzer executes in EclAgent, otherwise it executes in Thor Manager.
  • By default 'analyzeInEclAgent' is true in bare-metal and false in containerized
  • The defaults for 'analyzeWhenComplete' has changed for containerized. It is now false by default in containerized. It remains true in bare-metal.
  • New 'disabled' option is available within analyzerOptions to disable cost optimizer. 'disabled' is false by default.

Type of change:

  • This change is a bug fix (non-breaking change which fixes an issue).
  • This change is a new feature (non-breaking change which adds functionality).
  • This change improves the code (refactor or other change that does not change the functionality)
  • This change fixes warnings (the fix does not alter the functionality or the generated code)
  • This change is a breaking change (fix or feature that will cause existing behavior to change).
  • This change alters the query API (existing queries will have to be recompiled)

Checklist:

  • My code follows the code style of this project.
    • My code does not create any new warnings from compiler, build system, or lint.
  • The commit message is properly formatted and free of typos.
    • The commit message title makes sense in a changelog, by itself.
    • The commit is signed.
  • My change requires a change to the documentation.
    • I have updated the documentation accordingly, or...
    • I have created a JIRA ticket to update the documentation.
    • Any new interfaces or exported functions are appropriately commented.
  • I have read the CONTRIBUTORS document.
  • The change has been fully tested:
    • I have added tests to cover my changes.
    • All new and existing tests passed.
    • I have checked that this change does not introduce memory leaks.
    • I have used Valgrind or similar tools to check for potential issues.
  • I have given due consideration to all of the following potential concerns:
    • Scalability
    • Performance
    • Security
    • Thread-safety
    • Cloud-compatibility
    • Premature optimization
    • Existing deployed queries will not be broken
    • This change fixes the problem, not just the symptom
    • The target branch of this pull request is appropriate for such a change.
  • There are no similar instances of the same problem that should be addressed
    • I have addressed them here
    • I have raised JIRA issues to address them separately
  • This is a user interface / front-end modification
    • I have tested my changes in multiple modern browsers
    • The component(s) render as expected

Smoketest:

  • Send notifications about my Pull Request position in Smoketest queue.
  • Test my draft Pull Request.

Testing:

@shamser shamser requested a review from ghalliday January 16, 2025 16:37
@shamser shamser changed the title HPCC-33137 Add supports for executing the cost optimizers execution in Thor HPCC-33137 Add supports for executing the cost optimizers in Thor Jan 16, 2025
Cost optimizer will execute in Thor for containerized deployments.  It will continue to execute
in EclAgent for non-containerized deployment.

The reason for this change is that in containerized deployments EclAgent cannot calculate the
cost of the Thor cluster and so cannot not assigned the cost of issues reported by the cost
optimizer.  It cannot calculate the cost of the Thor cluster because Thor's cost parameters
and resource configuration is not available to EclAgent - this information is only available
in the Thor's configuration.

Note, that as the Thor manager executes on a per job basis, it executes the analyzer after
every graph executes.  The concept of "end of workunit" does not exist at present in Thor so
the analyzeWhenComplete==true has not been implemented for Thor.

The changes:

* Support for executing cost optimizer in Thor.  This the default in containerized.
* New 'analyzeInEclAgent' option - if true, the analyzer executes in EclAgent, otherwise
it executes in Thor Manager.
* By default 'analyzeInEclAgent' is true in bare-metal and false in containerized
* The defaults for 'analyzeWhenComplete' has changed for containerized.  It is now false by
default in containerized.  It remains true in bare-metal.
* New 'disabled' option is available within analyzerOptions to disable cost optimizer.  'disabled'
is false by default.

Signed-off-by: Shamser Ahmed <[email protected]>
Copy link

Jira Issue: https://hpccsystems.atlassian.net//browse/HPCC-33137

Jirabot Action Result:
Workflow Transition To: Merge Pending
Updated PR

Copy link
Member

@ghalliday ghalliday left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@shamser Looks ok. How long does it take to run the analyser on a 2,000 activity graph?

if (w->getDebugValueBool("analyzeWorkunit", agentTopology->getPropBool("@analyzeWorkunit", true)))
if (w->hasDebugValue("analyzeWorkunit") && !w->getDebugValueBool("analyzeWorkunit", true))
return;
if (!getBoolWUOption(nullptr, nullptr, "analyzerOptions/@disabled", false))
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why not pass w and "analyzeWorkunit" to the function?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Because if these are passed into the function and 'analyzeWorkunit' was true, then the getBoolWUOption would also return true and that would mean the workunit would not be analyzed, which is not what it should do:

if analyzerOptions/@disabled == true, then it shouldn't execute the analysis. However, if analyzeWorkunit == true, it should analyze the workunit.

@jakesmith
Copy link
Member

How long does it take to run the analyser on a 2,000 activity graph?

I am not sure how to produce such a graph - probably utilizing ECL macros and #LOOP perhaps to produce long sequences of chained attributes(?), or a macro that produces a smaller subgraph variant that can be produced many times.
May need @ghalliday to produce this test, or reach out to ECL coders for some tips.

As discussed in my meeting with Shamser, if speed (and the incurred additional Thor cost) is an issue we could:

  1. Spin the analysis off onto an asynchronous thread, allowing the Thor to become available to run other jobs.
    Downside is that if the Thor hits a hard failure in a subsequent job before the analysis is complete, the analysis will be lost.
  2. Farm out the analysis to a service (e.g. a sasha service). i.e. notify (queue) a analysis job, and have a separate service perform the analysis asynchronously. The service could also be load balanced trivially this way, though that would probably be overkill.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants