[Server-devel] Backup scheme via Rsync and SSH

[Tony Pearson]

Team,
I have researched "backup methods" today.

There is a GPL tool that combines Rsync and SSH together called 
"rsnapshot".  I did a 
"yum install rsnapshot" and it is the latest version 1.3.0 release.

It is hosted on Sourceforge and here is their webpage: 
http://www.rsnapshot.org/

Rsnaphot looks like it has everything we need:

Can be used to backup local directories to a second disk on the local host
Can be used to backup remote directories to the local host
Backups are "pulled" (the backup server goes out and gets the updates from 
everyone)
Backups can be scheduled using cron
Handles rotation for hourly, daily, weekly and monthly versions
Handles backing up MySQL databases from Moodle
Can invoke a script before and/or after the backup on each client machine.

The rotation is standard, the default is:

        hourly backups taken every 4 hours 
                hourly.0  always represents the most recent
                hourly.1 the version before that
                etc.
        daily backups taken every day for 5 days
                daily.0 always represents the most recent
                daily.1 yesterday's
                etc.
        weekly backups taken once a week (say Monday) for 4 weeks
                weekly.0 represents the most recent
                weekly.1 represents the one before that
                etc.
        monthly backups taken once a month (say 1st of each month) for 3 
months
                monthly.0 represents the most recent
                monthly.1 the month before that
                etc.

The scheme is all done with hard-link pointers, so if a file hasn't 
changed
there will only be one copy of it, and the rest of the versions will 
merely point
to that.  Each time rsync is run, only the changed files since last time 
are sent.

If there are three machines (squid, school, and backup) here is the 
scheme.

Level 1 -- The "backup" server will schedule backups and collect updates 
from
the squid and school servers.  We can decide which particular directories 
we want
from each machine.  We may choose every four hours, for example.

Level 2 -- The "school" server will schedule backups and collect updates 
from
each XO laptop.  Here I suggest no more than once a day.  The school 
server
can also schedule a backup of the "backup server's" configuration.

Backups will be stored under /library/.snapshot/machine-id/ directory.

For test purposes.  I have only two machines, not three, so I will test 
the following:

Level 1 - The "squid" box will backup the "school" server every hour 
hours.
Level 2 - The "school" box will backup the "squid" server and my XO laptop 
once daily.

The two levels are uniquely different, so if you want me to focus on one 
over the other, please advise.

Otherwise, I will proceed working on each level, and try to determine the 
right directories, 
frequency, and configuration.

Thanks





Tony Pearson
Senior Storage Consultant, IBM System Storage?
Telephone: +1 520-799-4309 |  tie 321-4309 |  Cell: +1 520 990-8669
email: tpearson at us.ibm.com |  GSA: http://tucgsa.ibm.com/~tpearson
Blog: http://www.ibm.com/developerworks/blogs/page/InsideSystemStorage
AKA: 990tony Paravane, eightbar specialist 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/server-devel/attachments/20080224/1a52f153/attachment-0001.htm