[Clipart] Help with the script - was: Re: A couple of server problems

[Alan]

Is there storage space to copy the svg and svgz files into a separate 
location say every 24 hours?

Automate the copy to happen once every 24 hours and then tarball from 
the copied files.

Nicu Buculei wrote:
> Frederik Elwert wrote:
>   
>> I haven't tried it, but I think xargs might be the wrong choice. It may
>> call tar multiple times, each time with some of the arguments. This
>> results in a final archive containing only the last files.
>>
>> You could use tar's -T/--files-from parameter to read the files from
>> filelist.txt
>>     
>
> Thanks Frederik, it worked!
>
> Now on http://download.openclipart.org/downloads/ is a tarball named 
> daily_SVG_snapshot.tar.bz2 which can be downloaded to get a snapshot of 
> the current submissions. It is quite large, 156 MB, will extracting 
> about 600MB and contains over 8000 images (some submissions have more 
> than one single SVG).
> I submitted a support request and it will be set to be updated daily.
>
> If someone can help wit a better version of the script, it will be 
> appreciated.
>
> The problem is: we have a directory  named "people" containing a lot of 
> files and subdirectories.
> The files in the root of the directory are not interesting, they are 
> failed uploads and temporary files. The subdirectories contain clipart 
> files (svg and svgz) and other not needed files (png, jpg, whatever - 
> avatars, screenshots, thumbnails).
> We need a tarball containing *only* the SVG and SVGZ files from 
> *subdirectories* and nothing else (the size is already large).
>
> I did it by running a find for *.svg and another for *.svgz, dump the 
> result in a text file and run tar with this text file as an input list.
>
> The weakness is that creating this tarball take a noticeable time (a few 
> minutes) and if some user deletes or renames an existing file (included 
> in that dumped find result) the tar operation will fail (it happened to 
> me the other day). Any idea?
>
>