complete newb... need sync script without rsync

[kevin1162]

hi,

let me start by saying my programming and linux knowledge/skill is VERY limited. I don't have much experience with either because it's not a large part of my job. I tinker here and there but nothing to write home about.

But due to recent events at my job i've been told i need to come up with a solution to sync a folder on one NAS to another NAS.

I started off simple enough with my limited knowledge. I just made a BASH script that contains this then runs as a cron job

cp -rup /nas1/dir1/* /nas2/dir2/

BUT then it dawned on me... If i delete a file or folder in /nas1/dir1 the old file will remain in /nas2/dir2. This is not acceptable, and this process has to be completley automated.

I can't for the life of me figure out how to do a comparison of the two directories and delete old files... I tried looking at using diff but can't wrap my head around figuring that out.

Something like rsync isn't an option because the servers this script will be executed on 1) don't have it installed and 2) i'm not allowed to install any other software on them. So i'm stuck using a bash script or perl, and i know Zero about perl.

Normally i wouldn't ask for help in this manner but i'm under the gun and have a VERY short deadline. If someone could guide me i'd appreciate it.

[proprietarysucks]

I haven't testing / won't test any of this stuff first, so deal with it.


computer a --> computer b
you could quickly remove all files from b which aren't on a by checking them, like

for filename in /nas2/dir2/*; do
[[ -d /nas1/dir1/$filename ]] || rm "/nas2/dir2/$filename"
done

then you could copy all the ones which are on a but not b, something like

for filename in /nas1/dir1/*; do
[[ -f /nas2/dir2/$filename ]] || cp -a /nas1/dir1/$filename /nas2/dir2
done

then you could only copy over the ones which have changed by doing something like

for filename in /nas1/dir1/*; do
[[ $(md5sum /nas1/dir1/$filename|cut -d' ' -f1) = $(md5sum /nas1/dir1/$filename|cut -d' ' -f1) ]] || cp -a /nas1/dir1/$filename /nas2/dir2
done


of course these could all be combined up and improved surely. the point is to get you in a working direction

[kevin1162]

Thank you so much for your reply...

Since I posted i was fiddling quite a bit and came up with this

#!/bin/bash

dir1="/directory_1/"
dir2="/directory_2/"

if [ -d $dir2 ]; then
 cd "$dir1"
 find . -print0 | while read -d $'\0' file; do
   [ -e "$dir2/$file" ] || echo "$file"
 done | xargs rm -rfv

 cp -rupv $dir2* $dir1
else
 echo Path Not found.. Check network status
fi

It's similar to your methodology, i still haven't tested it full scale in production only in my test setup and it seems to work.

Thanks for your input