github.com/iqoqo/nomad@v0.11.3-0.20200911112621-d7021c74d101/website/pages/docs/job-specification/restart.mdx (about)

     1  ---
     2  layout: docs
     3  page_title: restart Stanza - Job Specification
     4  sidebar_title: restart
     5  description: The "restart" stanza configures a group's behavior on task failure.
     6  ---
     7  
     8  # `restart` Stanza
     9  
    10  <Placement
    11    groups={[
    12      ['job', 'group', 'restart'],
    13      ['job', 'group', 'task', 'restart']
    14    ]}
    15  />
    16  
    17  The `restart` stanza configures a tasks's behavior on task failure. Restarts
    18  happen on the client that is running the task.
    19  
    20  ```hcl
    21  job "docs" {
    22    group "example" {
    23      restart {
    24        attempts = 3
    25        delay    = "30s"
    26      }
    27    }
    28  }
    29  ```
    30  
    31  If specified at the group level, the configuration is inherited by all
    32  tasks in the group. If present on the task, the policy is merged with
    33  the restart policy from the encapsulating task group.
    34  
    35  For example, assuming that the task group restart policy is: 
    36  ```hcl
    37  restart {
    38    interval = "30m"
    39    attempts = 2
    40    delay    = "15s"
    41    mode     = "fail"
    42  }
    43  ```
    44  and the individual task restart policy is: 
    45  ```hcl
    46  restart {
    47    attempts = 5
    48  }
    49  ```
    50  then the effective restart policy for the task will be: 
    51  ```hcl
    52  restart {
    53    interval = "30m"
    54    attempts = 5
    55    delay    = "15s"
    56    mode     = "fail"
    57  }
    58  ```
    59  
    60  ## `restart` Parameters
    61  
    62  - `attempts` `(int: <varies>)` - Specifies the number of restarts allowed in the
    63    configured interval. Defaults vary by job type, see below for more
    64    information.
    65  
    66  - `delay` `(string: "15s")` - Specifies the duration to wait before restarting a
    67    task. This is specified using a label suffix like "30s" or "1h". A random
    68    jitter of up to 25% is added to the delay.
    69  
    70  - `interval` `(string: <varies>)` - Specifies the duration which begins when the
    71    first task starts and ensures that only `attempts` number of restarts happens
    72    within it. If more than `attempts` number of failures happen, behavior is
    73    controlled by `mode`. This is specified using a label suffix like "30s" or
    74    "1h". Defaults vary by job type, see below for more information.
    75  
    76  - `mode` `(string: "fail")` - Controls the behavior when the task fails more
    77    than `attempts` times in an interval. For a detailed explanation of these
    78    values and their behavior, please see the [mode values section](#mode-values).
    79  
    80  ### `restart` Parameter Defaults
    81  
    82  The values for many of the `restart` parameters vary by job type. Here are the
    83  defaults by job type:
    84  
    85  - The default batch restart policy is:
    86  
    87    ```hcl
    88    restart {
    89      attempts = 3
    90      delay    = "15s"
    91      interval = "24h"
    92      mode     = "fail"
    93    }
    94    ```
    95  
    96  - The default service and system job restart policy is:
    97  
    98    ```hcl
    99    restart {
   100      interval = "30m"
   101      attempts = 2
   102      delay    = "15s"
   103      mode     = "fail"
   104    }
   105    ```
   106  
   107  ### `mode` Values
   108  
   109  This section details the specific values for the "mode" parameter in the Nomad
   110  job specification for constraints. The mode is always specified as a string
   111  
   112  ```hcl
   113  restart {
   114    mode = "..."
   115  }
   116  ```
   117  
   118  - `"delay"` - Instructs the scheduler to delay the next restart until the next
   119    `interval` is reached.
   120  
   121  - `"fail"` - Instructs the scheduler to not attempt to restart the task on
   122    failure. This is the default behavior. This mode is useful for non-idempotent jobs which are unlikely to
   123    succeed after a few failures. Failed jobs will be restarted according to
   124    the [`reschedule`](/docs/job-specification/reschedule) stanza.