brooklyn-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <>
Subject [jira] [Commented] (BROOKLYN-212) Auto-scaling in finite BYON cluster keeps adding on-fire entities
Date Mon, 11 Jan 2016 11:59:39 GMT


ASF GitHub Bot commented on BROOKLYN-212:

GitHub user aledsage opened a pull request:

    BROOKLYN-212 fix


You can merge this pull request into a Git repository by running:

    $ git pull BROOKLYN-212

Alternatively you can review and apply these changes as the patch at:

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #1136
commit 810dd8d15078ebd197645e75ce32a64cef6de782
Author: Aled Sage <>
Date:   2016-01-08T21:27:23Z

    Asserts: expectedFailureOfType checks caused-by as well

commit f437e2ea962ea0228697ea41e8339202c67cd19d
Author: Aled Sage <>
Date:   2016-01-08T21:33:25Z

    BROOKLYN-212: AutoScaling doesn’t retry if InsufficientCapacity
    - Adds Resizable.InsufficientCapacityException, thrown by
      Resizable.resize() if could not grow at all.
    - DynamicCluster catches NoMachinesAvailableException, and rethrows
      as InsufficientCapacityException in resize().
    - AutoScalerPolicy catches InsufficientCapacityException, and sets
      insufficientCapacityHighWaterMark to record the max size it can get
      to. Does not try again to resize above that, unless the highWaterMark
      is explicitly cleared by reconfiguring that config value.
    - Tests:
      - Changes TestCluster to include history of sizes and desiredSizes
      - Changes TestCluster, so can throw InsufficientCapacityException
        when gets to a particular size.
      - Test for DynamicCluster throwing InsufficientCapacityException
      - Test for AutoScalerPolicyMetricTest, to not resize above the failure
        level again.
      - Test for AutoScalerPolicyNoMoreMachinesTest, for when BYON location
        has run out of machines in a DynamicCluster.

commit 230edaaab9d9bab8df8cc0a760d79d329071a93d
Author: Aled Sage <>
Date:   2016-01-11T11:39:51Z

    AutoScaler test: remove unnecessary timeout config
    succeedsEventually calls can just use the default timeout (unless it’s 
    a performance test where it must complete in a faster time!)


> Auto-scaling in finite BYON cluster keeps adding on-fire entities
> -----------------------------------------------------------------
>                 Key: BROOKLYN-212
>                 URL:
>             Project: Brooklyn
>          Issue Type: Bug
>            Reporter: Aled Sage
> When using a BYON location of a limited size (e.g. four vagrant VMs), if an auto-scaler
policy tries to resize above this number of VMs then it goes badly wrong!
> Auto-scaling causes a new entity to be added to the cluster, but it fails to start because
the {{location.obtain()}} call fails. That entity is left in quarantine, on-fire.
> The auto-scaler will then try again: another new entity is added, but this also fails.
This repeats, adding more and more entities.

This message was sent by Atlassian JIRA

View raw message