unpacking across network behavior

[syth]

I recently set sabnzd to copy completed files in the "TV" category to a network mount, but the problem is the way that this is done is 1) the rars are unpacked to the network mount, then the file on the network mount is COPIED out of the unpack directory into the final directory, and finally the unpacked file is discarded.

The trouble is, this causes the unpack to take an absurdly long time, and it triples the bandwidth being used.

Is there any way to get sabnzbd to unpack the files in the download directory and only copy them to the final location after the complete file is extracted?

[markus101]

Easiest way is to have SAB unRAR them locally and then have a post processing script to move the files to their final destination.

The Bandwidth shouldn't be 3x, as the files from the unpack dir get moved to the final folder, not copied, from my experience at least. Unraring across the network is a major PITA, so the above is really the only way to handle it effectively, or run SAB on the Network server (if possible).

-Markus

[syth]

Yeah, I've watched the files getting copied. Running SAB on that server is not a good idea since almost all the files I download stay on this machine, only the TV shows get put on the network mount.

would the post-processing script be something as simple as

Code: Select all

#/bin/sh

mv $argv /path/to/network/mount

?? I've never looked into the post-processing scripts before.

and can I setup a post-process for just TV shows?

Suggestion: SABnzb needs a flag where the unrar takes place in the download location

OK, I looked at the data passed to post-processing scripts and no combination of data actually makes up the pathname to the file.  $1 is the folder but when i downlaoded a subpack to test, nothing contained the actual name(s) of the file(s):

"underbelly.a_tale_of_two_cities.2x03.brave_new_world.hdtv_xvid-fov.sub"
"underbelly.a_tale_of_two_cities.2x03.brave_new_world.hdtv_xvid-fov.idx"

So, feeling stupid here, how do I move the files out of $1 to another location? 

mv $1/* /path/to/mount/

??  That seems rather fugly.  What am i missing?

[markus101]

You're right with:

Code: Select all

mv $1/* /path/to/mount/

Not missing anything, that will move all the files, you could write logic to the script to only move avi files, but that seems overkill. Somethign like this will find only the AVI files and move them:

Code: Select all

FILEZ[1]="$1`find .  -regex '.*/.*\.avi' | sed 's/^\.//'`"
#Move the avi to the network share
mv "${FILEZ[1]}" "/path/to/mount/"  

You can assign post-processing scripts only to the Category TV, which will work wonders if you are using Newzbin, if you're not version 0.5 (Currently not released and available through SVN only) allows you to assign categories RSS feeds (if you're using RSS feeds).

-Markus

[syth]

You're right with:

Code: Select all

mv $1/* /path/to/mount/

Not missing anything, that will move all the files, you could write logic to the script to only move avi files, but that seems overkill. Somethign like this will find only the AVI files and move them:

Code: Select all

FILEZ[1]="$1`find .  -regex '.*/.*\.avi' | sed 's/^\.//'`"
#Move the avi to the network share
mv "${FILEZ[1]}" "/path/to/mount/"  

I ended up with a simple script

Code: Select all

#!/bin/sh

THEPATH=$1

cp -r "$THEPATH" /Volumes/Trinity/TV/C2D/

I tested this with a bunch of small files and it seems to be fine.

I like the find, but I don't understand the syntax "$1`find ... shouldn't there be a pipe or something?

[markus101]

No, you're no redirecting that output, but rather placing that before the actual AVI file name as is returned by the find statement. using the ` (below the Esc Key, not a single quote) you're running a command and appending its out put the {FILEZ[1]} which would make it /current/path/name.avi

Didn't make a whole bunch of sense to me at first, starting to get it now though.

-Markus

[syth]

... appending its out put the {FILEZ[1]} which would make it /current/path/name.avi

Oh, right.  That makes sense.

Wouldn't this all be easier if SABnzbd+ returned the path to the file, the filename (an array of filenames I guess), and the count of files?

Eh, whatever.  it works for my limited and lame needs, and that's ALL THAT MATTERS.  Thanks.

[markus101]

No problem, glad I could help.

-Markus