Hi,

Can you help me a bit?

I have a Raspberry Pi and I run a script on it in every minute from cron.
The problem is that 90% of the day the ftp server is not available.
In the scipt I put an if which pings the server, and only when available jumps to this line:

wget -mnd --read-timeout=5 --tries=3 --user=$USER --password=$PASS ftp://$SERVER_IP:$SERVER_PORT/$SERVER_SD

But when i disconnect in the middle of downloading, wget just waits and waits, even if I told them the timeout to 5 minutes, and tries 3.
I would like wget to behave this way:
if during the downloading process the server disapears, than wait for instance 5 seconds, drop the file which was partially downloaded and let the script go forward.
(I download 10-50 MB mkv files from the server, so I don't care if 1 file will be redownloaded at the next time)

Try adding a --connect-timeout as well. You only have a timeout on the read. Your problem description is somewhat ambiguous. Sounds like wget hangs after the "disconnect" and then "just waits and waits" (on the next connect attempt).

1 members found this post helpful.

this rings a bell somewhere...
i think someone else was reporting some similar problem in connection with conky weather forecast scripts...
and the ultimate solution was to use curl instead.

Thanks, I've read this in the meantime in the manual.

Today I did --timeout=3 which will refer to all kind of timeout, read, dns etc. and this way it does what I wanted, wget drops the connection and the script can proceed ahead.

Ok, if I don't succeed with the --timeout parameter, I'll try with curl. I'm already tried lftp but I didn't like it

Use the -c flag as well so wget can restart a partial download where it left off.

1 members found this post helpful.

I know of this flag, thanks. In reality the video server creates 10-50MB files, and I wouldn't care if wget doesn't continue from where it left off. Let say 20MB of a file has already been downloaded, I don't know what wget does with a partially downloaded file without -c argument. In my case it's ok if wget deletes that partially downloaded file, and at the next time the server is available, downloads it again. But I'll be experimentng with the -c flag as well.
Here's my little script if you are interested: http://paste.debian.net/119687/
I just want wget not to wait forever if the server disapears. I run this script from cron in every 5 minutes.

Usually it adds a .1, .2, etc tag if you rerun it.

Yeah, but I don't want that numbering:-)

kzo81, you'll probably have to wrap that wget-command into a shell script or something that will do some cleanup before wget starts downloading.

what i remember, someone had problems with exactly that usage scenario in wget: wget just hangs there when the internet connection drops in the middle of a download. the ultimate solution was to use sth else - e.g. curl (very common on gnu/linux systems).

1 members found this post helpful.

If you use -c it simply resumes the original download. If you've already downloaded a file it'll state the file is complete and exit.

1 members found this post helpful.

Thank you for you guys, I managed to eliminate the problems with wget. I only used --timeout paramater and now it behaves the way I like, I also simplified my script.