Configuration of periodic tasks executed with a distributed lock. A dictionary where the keys are task names. If hosts are specified in run_on_hosts, the task will switch to any of them at the first opportunity and will keep running there until a new preferred host is specified or a failover occurs due to an error. By adjusting the timing settings, you can override the lock acquisition policy.
# yaml
default: {}
description: |
Configuration of periodic tasks executed with a distributed lock.
A dictionary where the keys are task names.
If hosts are specified in run_on_hosts, the task will switch to
any of them at the first opportunity and will keep running
there until a new preferred host is specified or a failover occurs due to
an error. By adjusting the timing settings,
you can override the lock acquisition policy.
schema:
type: object
additionalProperties:
$ref: '#/definitions/TaskSettings'
definitions:
TaskSettings:
type: object
description: Periodic task policy settings
additionalProperties: false
properties:
run_on_hosts:
description: |
USE IsCancelAdvised() OR IT DOESNT WORK
List of hosts on which the task
is allowed to run. The task will start on
any of the specified hosts—without
prioritization.
type: array
x-usrv-cpp-type: std::unordered_set
items:
type: string
acquire_interval:
description: |
How often to try to acquire the lock.
type: integer
x-usrv-cpp-type: std::chrono::milliseconds
prolong_interval:
description: |
How often to attempt to renew
the lock while holding it.
type: integer
x-usrv-cpp-type: std::chrono::milliseconds
lock_ttl:
description: Duration of lock
type: integer
x-usrv-cpp-type: std::chrono::milliseconds
forced_stop_margin:
description: |
How long to allow the component to
finish work when it cannot acquire the lock.
type: integer
x-usrv-cpp-type: std::chrono::milliseconds
worker_func_restart_delay:
description: |
Delay before attempting to restart a
crashed component.
type: integer
x-usrv-cpp-type: std::chrono::milliseconds
