Announcement

**walto** · Jun 07, 2011, 09:59 AM

Re: okay, now i need a little help ;-)

Are you sure that your files reside in the / directory of the ftp server?
Alternatively try getting the file with get and decompress them manually.

**wizard10000** · Jun 07, 2011, 10:05 AM

Re: okay, now i need a little help ;-)

Originally posted by walto

Are you sure that your files reside in the / directory of the ftp server?
Alternatively try getting the file with get and decompress them manually.

Yes, the file exists in that location. get works just fine and I don't need to decompress it - it's a daily website backup. The problem is that since the filename changes every day I have to use mget instead of get in my download script.

The script has worked for years - it quit working when I installed Natty and doesn't work in Squeeze either. mget doesn't work even in a manual ftp session. but does work with other distributions - the problem is client-side.

I'd be interested in whether anybody running Natty who also has access to a remote ftp server also sees this problem.

**dibl** · Jun 07, 2011, 10:17 AM

Re: okay, now i need a little help ;-)

@wiz, your procedure is way over my head. However, I looked for your mget command on my Natty system and found this:

Code:

don@natty-vm:~$ sudo apt-cache policy mget
N: Unable to locate package mget
don@natty-vm:~$ sudo apt-cache policy get
N: Unable to locate package get
don@natty-vm:~$ sudo apt-cache policy wget
wget:
 Installed: 1.12-2.1ubuntu2
 Candidate: 1.12-2.1ubuntu2
 Version table:
 *** 1.12-2.1ubuntu2 0
    500 [url]http://us.archive.ubuntu.com/ubuntu/[/url] natty/main i386 Packages
    100 /var/lib/dpkg/status
don@natty-vm:~$

:P

**wizard10000** · Jun 07, 2011, 10:28 AM

Re: okay, now i need a little help ;-)

Thanks, dibl - mget is internal to the command-line ftp client. It's not a separate package

mget is the same as get but will grab multiple files without prompting -

F'rinstance - if you've got four files

1.tgz
2.tgz
3.tgz
4.tgz

mget *.tgz will get them all without asking for confirmation. Since the filename changes every day that's the way my script downloads the daily website backup.

The download script looks like this -

#!/bin/bash

HOST=vps.com
USER=wizard
PASS=xxxxxxxx

cd /home/wizard/archive/website

ftp -inv $HOST << EOF
user $USER $PASS
binary
mget *.tgz
mdelete *.tgz
bye
EOF

find /home/wizard/archive/website/*.tgz -mtime +6 -exec rm {} \;

exit 0

It logs on to the ftp server, grabs the website archives, deletes them from the server then removes any archives more than a week old from my local PC.

**dibl** · Jun 07, 2011, 11:12 AM

Re: okay, now i need a little help ;-)

Ahhhhhhhhhh. OK, cool, I learned something today.

gftp is about as deep as I go with file transferring (obviously).

Thanks -- good luck with your challenge.

**tnorris** · Jun 07, 2011, 06:09 PM

Re: okay, now i need a little help ;-)

Assuming you are using bash (and some other scripting languages) you can use "HERE Documents" to automate FTP and other similar things. A simple example would be:

Code:

#!/bin/bash

site=ftp.example.com
username=user
passwd=password
backupdir=$HOME
filename="backup-$(date '+%F-%H%M').tar.gz"

echo "Creating a backup file $filename of $backupdir."

# Make a tar gzipped backup file
tar -cvzf "$filename" "$backupdir"

ftp -in <<EOF
open $site
user $username $passwd
bin
put $filename 
close 
bye
EOF

More details on how HERE Documents work in bash can be found here: http://tldp.org/LDP/abs/html/here-docs.html

The basic idea is the shell pipes the text of the HERE document into the command until it finds the "end marker" which was specified right after the original "<<". The magic of course is that you can use variables in the HERE document to control what gets piped into FTP.

You can also automate some stuff with FTP by using a .netrc file in your home directory. Just google for "ftp .netrc" for information on that.

Let me know if that's not clear and I'll provide more details.

**wizard10000** · Jun 08, 2011, 02:49 AM

Re: okay, now i need a little help ;-)

Originally posted by tnorris

Assuming you are using bash (and some other scripting languages) you can use "HERE Documents" to automate FTP and other similar things. A simple example would be...

That's actually pretty close to what I use now but there are a couple of issues. First, we're moving files in the wrong direction - it's mget that's not working, not put

My script works, it just doesn't work in Natty or Squeeze.

Second, since ftp isn't a shell (well, it is in the strictest sense) you can't pass variables to it. I found it kinda interesting that you can authenticate to an ftp daemon using shell variables but once you're logged on $filename doesn't work.

I read a couple of references that suggested putting the variable in parentheses or curly braces like this

${filename}

but no joy there either.

I do really appreciate the effort you put into this, though

thanks -

Announcement

okay, now i need a little help ;-)

okay, now i need a little help ;-)

Comment

Comment

Comment

Comment

Comment

Comment

Comment