On 2019 Mar 22 09:05:12, you wrote to me:

 ml>> interestng... rsync has/does the same... i use it to backup my
 ml>> systems here as well as backing up several multi-gigabyte git and svn
 ml>> repositories... there's only one copy, though, so no chance to go
 ml>> back further than the copy currently in place... plus it is a copy,
 ml>> not an archive of files backed up...

 PQ> It's the same outcome: a single generation.  OTOH the clensing
 PQ> function on the target is useful, to ensure an exact mirror of the
 PQ> source.  I don't recall rsync doing that[shrug]?

here's an example from one of our regularly used update scripts...

  rsync -av --delete $RTYPE.code.sf.net::p/$PROJECT/$REPO .

in this case:
  -a is archive
  -v is verbose
  --delete removes unknown files from the dest directory
  $RTYPE is either git or svn
  $PROJECT is the name of the project
  $REPO is the name of the repository being synced

 PQ> They do the same job: Left or Right hand sort of thing.  I found
 PQ> mirdir first, and have found the logfile(s) most informative.

those log files are handy, that's for sure... i have my scripts written from a 
template i hacked together over time... they log everything from stdout and 
stderr into a specified log file... the log files are rotated to a max of 10 
logs... the current one and the 9 previous ones... a different template doesn't 
do rotation of the logs but they are time stamped in their file names...

  eg: 20190322-foobar.log
      201903221800-fubar.log

)\/(ark

Always Mount a Scratch Monkey
Do you manage your own servers? If you are not running an IDS/IPS yer doin' it 
wrong...
... All ties go to the train!
---