Automate FTP process and autorestart on link failure

Applications IP Networking
1

Hi Guys,

i have this lil challenge;

i am to implement an automated script that searches/scans a directory for files then picks and sends this files to a very remote server via an ftp link.
the challenge here is that the ftp link fails due to netwrk issues maybe;
i therefore need to develop a proactive means of detecting the present ftp session is no more active,then quit and restart a new session;
as this files are not supposed to stay inthis location for up to an hour.
someone help pls;

2

an idea from me - if you have netcat installed use :
nc -z server.com 21
the response should be :

then echo $? will return 0, on failure will return 1. Check the error code and proceed based on the response.
-z flag just checks for a live listener / daemon on the remote port, without sending the data. This should ensure that the connection is OK.

3

Sounds like you should really be thinking about replacing that ftp link with something with sane error fallback and graceful degradation.

4

What would you suggest as an alternative to scripted FTP here? I guess there are tools better suited to failing links but have no idea what these would be?

5

The customary replies would be ssh/scp and/or rsync.

6

i wouldn't recommend using ftp as scp/sftp are much better easier in error handling and offer much better security.

Anyway, I made a while ago an auto-ftp script, find below some hints:

  • In fact you should make your script loop untill a success condition is reached, which in this case means all files are transferred successfully.
  • In your script, you should save the ftp output and have it checked to ensure last transfer is ok. this is what I came up with:

I just grabbed this code from the source scripts, so it might not be ready for use. it is just to get the idea.

In fact if you could grep ^226 $resultfile |grep -v "226 bytes" it should be a good indication the file is transmitted successfully.
The reason to grep -v "226 bytes" is in case the file is 226 bytes you'll grep
226 bytes transferred (or something like that).

#syntax:
#analyze_ftpresult <ftpresult file>

function analyze_ftpresult
{

\# check rfc 959 on http://www.ietf.org/rfc/rfc0959.txt
\#
resultfile=$1
is_error=0
is_noconnect=\`fgrep "Not connected" $\{resultfile\} 2&gt;/dev/null\`
    is_226=\`grep ^"226 " $\{resultfile\} 2&gt;/dev/null|grep -v ^"$\{errcode\} bytes" 2&gt;/dev/null\`
    is_250=\`grep ^"250 " $\{resultfile\} 2&gt;/dev/null|grep -v ^"$\{errcode\} bytes" 2&gt;/dev/null\`

if [ -z "$\{is_226\}" -a -z "$\{is_250\}" ] ;then
            return 1
    fi
if [ -n "$\{is_noconnect\}" ] ;then
	return 10
fi
	
$\{debug\} && set \+x
\# Too much debugging info
for errcode in \`echo 421 425 426 450 451 452 500 501 502 503 504 530 550 551 552 553\`
 do
	error=\`grep ^"$\{errcode\} " $\{resultfile\} 2&gt;/dev/null|grep -v ^"$\{errcode\} bytes" 2&gt;/dev/null\`
	case $\{error\} in
		421|425|426\)	return 10 ;;
	esac
	if [ -n "$\{error\}" ] ; then
		is_error=\`expr $\{is_error\} \+ 1\`
		echo "$\{stamp\} ERROR $\{_sfile\}: $\{error\}" &gt;&gt;$\{somelogfile\}
	fi
done
$\{debug\} && set -x
return $\{is_error\}
$\{debug\} && echo "&lt;&lt; end function analyze_ftpresult &gt;&gt;"

}