Auto retry

The auto retry function makes it easy to make sure that a job that runs on many machines really is done, without supervision. Examples: change workstation passwords, edit the registry etc.

Linux and Novell notice: auto retry depends on the [Machines] section and can thus not be used in Linux or Novell server scripts.

Background

Running a script on a large number of machines usually means trouble: not all machines are available all the time; some will be switched off when the job starts, some will be switched off during the job, and some will not reply due to network problems.

Normally SetupBatcher (1) ignores machines that didn't reply in the initial machine check and (2) asks questions during the job if a machine is unavailable or if there are errors, meaning that a job that runs on many machines usually must be looked over and restarted many times until finished.

The solution to this is the auto retry mode: the batch job runs without questions until all commands in fact has been executed on all machines in the [Machines] section.

To enable auto retry mode

Simply add the BatchSettings.AutoRetry setting to a script.

How auto retry mode differs from normal operation

·	A [Machines] section must be present in the script.

·	Listed machines are not contacted before the job starts.

·	During the job, when a machine is not available or if an error occurs, the machine is moved to the end of the "remaining" que. When the retry interval has passed it will be contacted again.

·	When a command fails, the script runs from the beginning the next time, it does not resume with the command where the error occurred.

·	The retry/ignore/abort error dialog is never displayed.

·	The [OnError] section can not be used.

Auto retry window from top to bottom

Item	Description
Job started	Time when the job started.
Time between retries	The BatchSettings.AutoRetry setting.
Current machine	The machine that SetupBatcher currently is working on.
Time to next attempt	Time until the next machine will be contacted. Visible when no current machine.
Remaining machines	Number of machines yet to finish.
Remaining machines list	This list is populated from the [Machines] section when the job starts with all machines not marked "OK-AllDone". · Last try - time for last try to contact the machine   · Machine - name of machine   · Attempts - number of attempts to contact the machine   · Errors - number of errors
Finished machines	Number of finished machines
Finished machines list	This list is populated from the [Machines] section when the job starts with all machines marked "OK-AllDone". Also, when a machine is finished it is moved from the "remaining" list to this list. · Time - time when machine was finished   · Machine - name of machine   · Attempt - finished on attempt   · Errors - number of errors
Errors	Total number of errors
Error list	A row is added to this list when an error occurs. · Time - time for the error   · Machine - machine where error occurred   · Attempt - attempt when error occurred   · Code - error code   · Line - line in [Batch] section where error occurred   · Command - command for which error occurred
Run now button	Click the "Run now" button to ignore the retry interval once: all remaining machines are contacted A.S.A.P one time.
Abort button	Click the "Abort" button to stop the job. Finished machines will be marked "OK-AllDone" in the [Machines] section and will not be contacted if the job is restarted.