
  tbackup -- general backup system for Linux.

  V0.9

  Documentation for the experimental batch mode


Introduction

 If you have a tape drive, you may want to do backups
 noninteractively, e.g. at night.  Tbackup has an experimental `batch
 mode' feature for scheduling non-interactive (batch) backups as at(1)
 or crontab(1) jobs.

 The feature is experimental in the sense that it is not as
 thoroughly tested as other features.  It may not work on all linux
 system configurations.  Before you start relying on this feature, you
 should at least once test 

  1) a batch mode run in which you cause a fatal error on purpose, to
     see if the fatal errors is properly reported (by e-mail).

  2) a restore a backup made in batch mode.


1. Using the batch mode feature

 Scheduling a non-interactive backup run is a two-step process:

  1. Create a tbackup `argument file', which records the tbackup
     parameters to be used in the non-interactive run
 
  2. use at(1) or crontab(1) to schedule the run

 At the start of the non-interactive run, the `argument file' created
 in step 1 will be read by tbackup.  Independent runs must therefore
 have independent argument file names.

 If a fatal error occurs during the non-interactive run, tbackup uses
 a special error handler to ensure a clean abort.

1.1 Create an argument file

 The tbackup batch mode expects all parameters for the backup to be
 set via a command line argument to tbackup.
  
 You can make this argument by running tbackup -batcharg.
  
    | # tbackup -batcharg
    | 
    | The batcharg option creates an argument file in /etc/tbackup holding
    | the parameters supplied below.
    | 
    | Name for the argument to be created ---> [batch] mybatch
                                                       ^^^^^^^

    | The file /etc/tbackup/mybatch.arg already exists!!
    | Overwrite it? ---> (y/n) [n] y
                                   ^ 
    | 
    | The argument file /etc/tbackup/mybatch.arg will be created.
    | Tbackup can then be used in batch mode as 'tbackup -mybatch'.
    | 
    | Available choose methods:
    |   1 set   -- back up a directory set
    |   2 setrm -- back up a directory set, then remove its contents
    |   3 one   -- one user specified file or directory and below
    | Enter method name or number ---> [set] 
    [etc]

1.2 Schedule the batch run
  
 After this you can do something like
  
    | # at 12:30
    | tbackup -mybatch
      ^^^^^^^^^^^^^^^^
      ^D
      ^^(control-D to signify eof to at command)
    | Job a00c61d76.00 will be executed using /bin/sh
    | # 
  
 to have tbackup run at 12:30 with the options you just supplied.
 Don't forget the -mybatch argument (or whatever you call it).

 Note that the argument file (/etc/tbackup/mybatch.arg in the example
 above) is read by tbackup when the non-interactive job is started.
 By making different argument files with tbackup -batcharg, you can
 have multiple at(1) or crontab(1) jobs pending.

1.3 Error reporting

 If any error is encountered during the non-interactive job in batch
 mode, tbackup will try to abort cleanly.  The at(1) command, if used,
 will send you mail with the error message.
  
 As a side effect of the abort, if you have any commands after tbackup
 in your at(1) script they will not be run.

1.4 Other commands

 The trestore, tlist, tguess, and tverify commands in the tbackup also
 support batch mode operation.  Use goes along similar lines: run
 `tverify -batcharg', etc.

1.5 Limitations of the batch mode

 You can't use the floppy write method in batch mode, only the tape,
 filedev, and null methods.  When using tape, make sure there is a
 tape ready for writing in the tape drive.
    
 If the filesystem becomes full in batch mode, tbackup will pause till
 there is enough room again, even if this takes 5 days.

1.6 See also

 See also: the manual pages for at(1) and crontab(2), sections 9 and
 22 of README.txt, and the comments in the argument files produced by
 `tbackup -batcharg'.
