Skip to content
This repository was archived by the owner on Jan 1, 2024. It is now read-only.

1.9.0 - closer to reality

Choose a tag to compare

View all tags
@hackallcode hackallcode released this 30 Apr 15:35
· 66 commits to master since this release
1.9.0
3dd8a96
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Verified
Learn about vigilant mode.

Bug fixes

  • Running the role with python 2.7
  • Skipping instances restart when package was updated, but configuration wasn't
  • Missing default config for machine with stateboard
  • Specifying cartridge_app_name other than the TGZ package name
  • Creating unnamed replicasets with instances without replicaset_alias set
  • Getting control instance:
    • Now one not expelled instance should also be alive; it's checked
      by creating connection using instances advertise URIs
    • Control instance should be alive
    • If there are some joined instances, but no one of them isn't alive,
      getting control instance fails.
  • set_control_instance is improved to consider non-joined instance status
  • edit_topology step now considers roles dependencies, permanent and hidden roles
    and don't perform unnecessary calls if enabled roles list isn't changed

Features

  • failover_promote step to promote replicasets leaders
  • Allowed to skip user and group creation for tgz
  • Debug control instance and one not expelled instance
  • Timeouts for two-phase commits:
    • twophase_netbox_call_timeout
    • twophase_upload_config_timeout
    • twophase_apply_config_timeout
  • eval and eval_on_control_instance steps to eval code on instances
  • Step stop_instance to stop and disable instance systemd service
  • Step start_instance to start and enable instance systemd service
  • Step restart_instance_force to restart systemd service without any conditions
  • New cartridge_failover_params fields:
    • failover_timeout
    • fencing_enabled
    • fencing_timeout
    • fencing_pause
  • edit_topology_allow_missed_instances variable to allow replicasets containing
    the instances that are not started yet
  • upload_app_config step to load the file or directory config
    (Cartridge and TDG are supported)

Changes

  • Timeout to wait for cluster health after topology editing
    renamed from edit_topology_timeout to edit_topology_healthy_timeout
  • cartridge_cluster_cookie now is required only for configure_instance,
    restart_instance and upload_app_config steps
Assets 2
Loading