I'm following a back up idea on this link http://esm2.imt-mrs.fr/~staffelb/gui...art1/cron.html

The only way I can get it to back a DIR of ours is to use the command line only ./backup

cron doesn't seem to be working

[root@localhost subs]# crontab -l
40 13 * * * ./usr/local/awstats/wwwroot/cgi-bin/./awstats.pl -update -config=www.rockinghamgateway.com
20 14 * * * ./srv/www/subs/backup
[root@localhost subs]#

I tried a few different ways by testing it by editing the crontab a few minutes forward to see if they are updateding or backing up.

hope that makes sence.

??? ??? ???

TT

tommytomato:
I believe the error is with your "tar" command.
Change from:
To:
The backup file's name must directly follow the "f" option.

Give that a whirl and let me know if it worked for you.

Nup nothing happing

10 * * * 0,1,2,3,4,5,6,7 /usr/local/awstats/tools/awstats_updateall.pl now >/dev/null 2>&1p
02 16 * * * ./srv/www/subs/backup

top one works and the 2nd one dont.

file back up has perm 0755 and has this in side the file

file was saved as backup, if I use the command line #./backup it works.

TT

Look at cron's logs to see what's happening.

Furthermore, cron uses it's own environment variables. $PWD, the present working directory, is one of them.
So I'm not sure "./", which is synonymous to $PWD, is set up correctly. In other words, I doubt if cron
can actually find your command, since it may not be running out of the directory that you expect it to.

It would be a good idea to change "./..." into an absolute path, or use ~root/ for the home directory of the
user root.
A alternative is to add the command
cd (or cd to your root's home dir)
to the top of your backup script, just to make sure it's running in the right directory.
A 3rd option is to specify absolute paths in the "rm" and "tar" commands.

If "./backup" works, assuming you're inside your home directory, then why do you ask cron to run it as
"./srv/www/subs/backup" (assuming for a moment that cron runs the commands out of your home directory)?
Shouldn't it simply be "./backup" instead? In other words, pay attention to the paths!
If you're worried that your path is wrong, try adding "ls" to the cron command:
02 16 * * * ls -l ./srv/www/subs/backup

This should send you an e-mail containing the output of the ls -l command. If it doesn't, add "> a_new_file" at the
end to make "ls" write it's output to a regular file (so you can read it afterwards).

How about specifying the full paths:

I've done what you guys said, and still no go, must be missing someting

my cron is

10 * * * 0,1,2,3,4,5,6,7 /usr/local/awstats/tools/awstats_updateall.pl now >/dev/null 2>&1p
27 17 * * * ./srv/www/subs/backup

I added that path into the backup file under subs
rm /srv/www/subs/backup.tar.gz
tar czf /srv/www/subs/backup.tar.gz .

must be the way I do my crons even the top cron is not work, I thought it was, OS is tinysofa odin

TT

Is crond running???
Try "pidof crond", or "ps aux | grep crond" to see.
If it's not, then fire it up and verify that at least one of your crons is running.

First of all, be carefull about the paths.
/srv/www/subs/backup.tar.gz
is NOT the same as
./srv/www/subs/backup.tar.gz

Secondly, to check if cron works, follow the following steps:
-Check if the cron daemon (crond) is running.
This can be done via the "ps" command (eg ps -efl|grep -i crond) or via
/sbin/chkconfig --list | grep -i cron
Via /sbin/service cron status (or something like that) may also work.

Seems WindowBreaker beat me to this one... - lol

-Secondly, your user must be allowed to run cron jobs. In other words, your username must appear in
the file /etc/cron.allow. If cron.allow doesn't exist, then your username may not appear in /etc/cron.deny.
If cron.deny and cron.allow do not exist, then you must create the cron.allow file and put "root" as well as
your username inside it. Be sure to create it as root.

For more info, see:
man cron
man crontab
man 8 crontab

ok guys thanks

I had to have a break so I went fishing, I'll double check every thing in the morning because I'm nacked.

I checked to see if cron is running here is what I get

Also I dont have /etc/cron.allow in my system

So which is the correct one, I have the top one /srv/www/subs/backup.tar.gz

TT

By the looks of it, your cron daemon is running alright.

Do you have /etc/cron.deny? If so, make sure that your user isn't listed in that file.
If you don't, then "su" to root, create the file /etc/cron.allow and put your username in there.
Something like:
su #switches to root user, you'll need the password...
echo "your_username_goes_here" > /etc/cron.allow
exit #exits su
(the text after the '#' sign are just comments, they are not executed by your system)

As to which is the right path, that depends on the directory you're working in.
Paths starting in "/" are called absolute paths. They are subdirectories located under
the "/" (top-level) directory of your disk.
The path that starts with "./" is different. "./" refers to the current directory (use the "pwd" command to see what it is). So, ./srv/www/subs/backup.tar.gz is located somewhere under the current directory, which may or may not be the "/" directory.

You can actually use any path you want. You just need to decide:
-where (in which directory) you want to store the backup.tar.gz file
-where the files you need to backup are stored.

Remove the . in the ./srv... should result in the script executing from cron. You only use a ./ in front of a command when your in a directory that has an executable that isn't in your present PATH. If your specifying the full path, you never need to put a . in the path name.

If you ask me, the entire tutorial is written badly.

Surely "tar -zcf backup.tar.gz ." should be in a bash script? So therefore:
Would make more sense?

I know that the #!/bin/bash isn't completely necessary, but there just doesn't seem to be any thought put into it.

Thats what I was thinking hey, while having a break and fishing I was thinking the same thing, I haven't had alot to do with scripts, but the ones I have use all start with #!/bin/bash.

So thanks again I'm going to give it another shot today

timmeke

I'm using root all the time, so do I still need to create that /etc/cron.allow folder ?

TT

ok here is what I did,

I renamed the backup file to backup.sh for starters then I edited the file like so.

then I edited my crontab like so, working only with backup.sh for now

Then I put the time foward a bit and waited for the backup to work, which it didn't to a point, but some how that crontab created a backup.sh.tar.gz file under the path /root/

I'm just wondering if the pathis wrong, the script jumps back to root dir cd ~is that correct if so should it be cd /srv/www/subs/

TT

Well guys it was the path

it now works, thanks for the help and tips, need to find out why the awstats one wont work.

TT