[X-Unix] Rsync, mysqldump and cron job help needed - part2

[Rad Craig]

5.) Am I on the right track?  I think it will be something like this:
a.  Setup a cron job for a mysql dump (whatever the correct command  
is), at say 10pm every Sunday night.
aa.  Repeat (make another cron job/command) for each different  
database.  Can I run them one minute apart, and they just run on  
separate threads?  Or do I need to guess at when they will finish and  
space them out?
b.  Setup a cron job to run rsync on most of my directory structure  
at 11pm every Sunday night.
c.  Setup a cron job to delete the mysql dumps at 1am, every Monday  
morning, at least after the first rsync has ran, because it will take  
longer to make the first full backup.

6.) Ok, here's one thing I'm confused about, rsync will make a full  
backup, then do incrementals.  Does it just update the entire  
directory structure I have downloaded, or does it make some kind of  
incremental files (like tape backups) that will just keep adding up  
(more files) each week, so that if I have to do a restore, I would  
have to restore the main files, then a zillion incremental versions?

6a.) How does that work, how do you keep it organized and not a total  
nightmare if you ever have to do a restore?  Or would I be better off  
doing a full backup each week as I have been doing, and just keeping  
6 or more weeks worth of backups?

6b.)  Do I run rsync from my webhost server via a cron job, to have  
it send the files to my shop machine, or do I run it from my machine  
at the shop that I'm actually downloading and storing everything on,  
where it fetches the files off the webhost?  This perspective eludes me.

6c.) I saw an example of:  rsync -azv foo:src/bar /data/tmp

Should I use the 'archive' flag?

6e.) If I want to copy a directory in my root of the webhost (not the  
machine/OS root), like /foo, including the foo directory itself, and  
all of its tree and contents, do I use /foo or /foo/?  In other words  
I want it to create a directory named /foo on my shop machine, then  
copy all of the contents and subdirectories in it, not just the  
contents and subdirectories.  What is the correct way to specify  
this?  If I understand the man page, I think I leave the trailing  
slash off and it will create the /foo directory, then copy all of its  
contents.  The wording in the man page about "copy the directory by  
name", is whats confusing me.  If it said either "copy the contents  
only" or "create the directory and copy its contents" it would be  
more clear, to me anyway.

6f.) Do I format it the command the same way for the destination?  I  
mean trailing slash or no?  I'm going to be copying from my root on  
the webhost, say a directory /foo, to some directory on my shop  
machine like "/website backups".  So would I use the trailing slash  
for the destination?

rsync -azv /foo "/website backups"
or
rsync -azv /foo "/website backups/"

I want it to create that foo directory inside the "website backups"  
directory, then copy all of its contents.

6g.) I have some directories with a space in the name, like the  
"website backups" example above.  Can I use a name like that or do I  
need to rename the directories that are named in that way?  If I can  
use them, do I use single or double quotes?

6h.) What are the following flags?  --backup? --update? --inplace?   
My man page lists about all of the flags and short descriptions, but  
when it gets down to the long descriptions, it doesn't show which  
flag the long description goes with, so it makes it very hard to get  
a detailed description of what a flag means.  It says something like  
"this flag specifies...", but doesn't tell me which flag.

6i.)  I'd like to verify the files after transfer, how do I do that?

6j.)  If rsync just makes a mirror, and only copies the changes after  
the initial download, then I have one mirror of the website.  How can  
I keep some type of 'roll-back' capable setup?  And I guess if it is  
making a mirror image of the files/directories I specify, I would  
also want it to delete the files on the destination that no longer  
exist on the source, if so, whats the correct flag for that?