Skip to content

@taylorbot taylorbot released this Feb 12, 2019 · 6 commits to master since this release

This release of gsctl focuses on support new API features introduced to enable cluster autoscaling on AWS for releases starting at 6.3.0. (#325)

For more details on autoscaling, please consult our documentation.

Changes

  • gsctl create cluster
    • Two new (optional) flags --workers-min and --workers-max are introduced to set the initial scaling settings for the number of worker nodes in a cluster. The autoscaler will control the number of worker nodes within this range. For AWS releases prior to 6.3.0, and on Azure and KVM, both flags must be set to the same value. As before, once a cluster is created, the worker node count (or the range) cna be influenced using gsctl scale cluster.
    • The flag --num-workers can still be used to set the initial cluster size, as an alternative to --workers-min and --workers-max. The value given will be applied to both the minimum and maximum cluster size, effectively disabling autoscaling.
  • gsctl scale cluster
    • The new flags --workers-min and --workers-max are introduced to set the scaling limits for a cluster. On releases supporting autoscaling (AWS <= 6.3.0), they can be used independent of each other to adjust only the upper or lower limit.
    • The --num-workers/-w flag can still be used. In releases supporting autoscaling, tis will result in setting the minimum and maximum worker node count to the same value, effectively disabling autoscaling.
    • The flags --memory-gb, --num-cpus, and --storage-gb got removed as they were no longer functional
  • gsctl show cluster
    • The new line Worker node scaling is added, informing about the current scaling settings.
    • The new line Worker nodes running replaces the former Workers line, informing about the number of worker nodes running.
    • The new line Desired worker node count is shown when the cluster supports autoscaling and the worker node count is not pinned to a specific number. It shows the number of worker nodes the autoscaler attempts to have running currently.
  • gsctl delete cluster now accepts the ID of the cluster to be deleted as a positional argument. The old -c/--cluster flag is deprecated and will be removed in a future gsctl version. (#344, #346)

Bugfixes

  • The attempt to delete a cluster that does not exist now yields a user-friendly error message. (#346)
Assets 6
You can’t perform that action at this time.