Printable View
This may be an odd request, or it may be a common thing-- no idea!
Anyone know of a way to do incremental backups of a single file?
In other words, back up changed parts of the file, rather than backing up the entire file every time.
Can it be done with rsync or fwbackup or anything else?
thanks
Rsync will perform incremental backups by default. This means that it will not recopy all of the files every single time a backup is performed. Instead, only the files that have been newly created or modified since the last backup will be copied
Another option is through SVN (Subversion) . Each time a change is made to the file it is committed to the SVN. If u want to test the fix that has been committed You can checkout the SVN on your local system in which latest changes will be noticed
Hi
Afraid you're misunderstanding. I'm not talking about multiple files. Just one, single, individual file. See?
Thanks for reply!
Hi.
K & P produced a suite of small tools to collect the history of a file. As with most versioning tools, they rely on diff to produce the differences.
I use the book: Amazon.com: Unix Programming Environment (Prentice-Hall Software Series) (9780139376818): Brian W. Kernighan, Rob Pike: Books , but it may be available elsewhere, possibly at Google for searching.
The code I mentioned is on page 167, put and get.
Personally, I find that the RCS commands ci and co are easy to script and the RCS repository (history directory) can be localized to the current directory. The bzr utility can also use a local repository.
Best wishes ... cheers, drl
This may be beyond me, but I'll explore it.
Thanks.
Hi.
The saving script about 30 lines long, and will require its companion get, as well as script overwrite. If nothing else, you'll be able to see some of the power of shell scripting, as well as a few techniques.Quote:
Originally Posted by johnywhy
The scripts can also be used as black boxes without knowing much about how they work.
I don't see how this task could be made much smaller, although perhaps a bare-bones RCS approach might work.
Good luck ... cheers, drl
Many thanks. Just ordered the book. I expect not to understand it without hours of struggle and research and kind help from others.
Hi.
While you are waiting for the arrival of the book, you can look at this demonstration of the RCS approach. This script initializes the environment by removing the test file, hi.txt, and the associated history (version) file, hi.txt,v. It steps through by displaying the file in several stages, and shows the files that remain after the steps, culminating in a display of the log of the actions to the file.
producing:Code:#!/usr/bin/env bash
# @(#) s1 Demonstrate RCS ci and co for keeping history of single file.
pe() { for _i;do printf "%s" "$_i";done; printf "\n"; }
pl() { pe;pe "-----" ;pe "$*"; }
db() { ( printf " db, ";for _i;do printf "%s" "$_i";done;printf "\n" ) >&2 ; }
db() { : ; }
C=$HOME/bin/context && [ -f $C ] && $C sed ci co
FILE=hi.txt
# Remove data file and history file.
rm -f $FILE ${FILE},v
pl " Current files before operations:"
ls -lgG
cat > $FILE <<EOF
one
two
three
EOF
pl " Data file $FILE:"
head $FILE
pl " Current files before checkin:"
ls -lgG
pl " Checking in file $FILE (simulated comment supply):"
ci $FILE <<<" First version."
pl " Current files after checkin:"
ls -lgG
pl " Now checkout the file, with lock:"
co -l $FILE
pl " Current files after checkout:"
ls -lgG
pl " Modify and checkin the file, effectively keeping the lock:"
sed -i -e '/two/d' -e '$a\
four' $FILE
head $FILE
ci -l $FILE <<<' Second edition.'
pl " Current files after modification and checkin:"
ls -lgG
pl " Display the modified file:"
head $FILE
pl " List the log and details of hi.txt:"
rlog hi.txt
exit 0
Comments:Code:% ./s1
Environment: LC_ALL = C, LANG = C
(Versions displayed with local utility "version")
OS, ker|rel, machine: Linux, 2.6.26-2-amd64, x86_64
Distribution : Debian GNU/Linux 5.0.8 (lenny)
bash GNU bash 3.2.39
sed GNU sed version 4.1.5
ci RCS version 5.7
co RCS version 5.7
-----
Current files before operations:
total 8
-rw-r--r-- 1 269 May 25 08:42 readme.txt
-rwxr--r-- 1 1114 May 25 09:39 s1
-----
Data file hi.txt:
one
two
three
-----
Current files before checkin:
total 12
-rw-r--r-- 1 14 May 25 09:46 hi.txt
-rw-r--r-- 1 269 May 25 08:42 readme.txt
-rwxr--r-- 1 1114 May 25 09:39 s1
-----
Checking in file hi.txt (simulated comment supply):
hi.txt,v <-- hi.txt
initial revision: 1.1
done
-----
Current files after checkin:
total 12
-r--r--r-- 1 210 May 25 09:46 hi.txt,v
-rw-r--r-- 1 269 May 25 08:42 readme.txt
-rwxr--r-- 1 1114 May 25 09:39 s1
-----
Now checkout the file, with lock:
hi.txt,v --> hi.txt
revision 1.1 (locked)
done
-----
Current files after checkout:
total 16
-rw-r--r-- 1 14 May 25 09:46 hi.txt
-r--r--r-- 1 223 May 25 09:46 hi.txt,v
-rw-r--r-- 1 269 May 25 08:42 readme.txt
-rwxr--r-- 1 1114 May 25 09:39 s1
-----
Modify and checkin the file, effectively keeping the lock:
one
three
four
hi.txt,v <-- hi.txt
new revision: 1.2; previous revision: 1.1
done
-----
Current files after modification and checkin:
total 16
-rw-r--r-- 1 15 May 25 09:46 hi.txt
-r--r--r-- 1 354 May 25 09:46 hi.txt,v
-rw-r--r-- 1 269 May 25 08:42 readme.txt
-rwxr--r-- 1 1114 May 25 09:39 s1
-----
Display the modified file:
one
three
four
-----
List the log and details of hi.txt:
RCS file: hi.txt,v
Working file: hi.txt
head: 1.2
branch:
locks: strict
drl: 1.2
access list:
symbolic names:
keyword substitution: kv
total revisions: 2; selected revisions: 2
description:
First version.
----------------------------
revision 1.2 locked by: drl;
date: 2012/05/25 14:46:51; author: drl; state: Exp; lines: +1 -1
Second edition.
----------------------------
revision 1.1
date: 2012/05/25 14:46:51; author: drl; state: Exp;
Initial revision
=============================================================================
1) You should be able to copy and paste this script into a file on your GNU/Linux system,
2) You didn't mention your platform, so I cannot advise you on how to install the rcs package if it is not already installed,
3) Your system will not be able to produce all of the environment or versions because that is a utility local to our organization.
See man pages for details.
Best wishes ... cheers, drl
( Edit 1: correct minor spelling error )
is this a text file? how often is the file updated? do you update it or is it updated automatically? Do you want back-ups of the old file? It could be a very simple script.
its a puppy linux save file. I believe it's binary ext3, with a lot of white space.Quote:
Originally Posted by mactruck
puppylinux.org/wikka/SaveFile
manually, not scheduled. Could be 3 in a day, then a month before the next one.Quote:
how often is the file updated?
i dont understand the question.Quote:
Do you want back-ups of the old file?
i'm now looking at these methods, which seem simpler than the suggestion above:
murga-linux.com/puppy/viewtopic.php?p=629862#629862Quote:
the 'diff' command and then you use the 'patch' command.... rdiff-backup would also work to keep one current copy of the puppy save file, and "binary diffs" that would allow recovery of a previous save file
If you have an alternative, would love to hear it.
thanks