---------------------------------------------
How to fix missing pictures:
_____________________________________________


How the picture uploads work:
-----------------------------
When a picture is uploaded a smaller preview version with 150
pixels width is automatically created. This will only be done
if the file is named picture.gif or picture.jpg. Other picture
file formats such as .tif, .psd or .bmp will not be resized
properly.
The picture thumbnails are set to appear in the newswire and
will link to the story where the picture can be seen in
original size. Both the thumbnail and the uploaded picture are
stored on stallman in:
/www/uk/local/webcast/uploads/metafiles
audiovisual stuff, pdf's, texts and older files are here:
/www/uk/local/webcast/uploads

In order to take as much as possible serving load off stallman,
all multimedia data and big text files are scripted to be
copied over to the media server images.indymedia.org which has
both a faster connection to the web and a bigger hard drive
than stallman. This media server is hosted on loudeye.com. The
copying over is done by a perl script called pushtoeye which
kicks in every 12 minutes and copies all audiovisual media,
photos and bigger text files to loudeye. There the files end up
in the following directory:
/metafiles/uk
Any file in here can be downloaded; here an example url:
 http://images.indymedia.org/imc/uk/anyfile.ext

When a file is copied over to loudeye, its URL is changed at
the same time. For exaample, a thumbnail appearing in the
newswire has for the first 12 minutes the following URL:
 http://uk.indymedia.org/local/webcast/uploads/metafiles/PICTURE-thumb.jpg
That means it's still linked to the file on stallman. But if
you check the URL after 12 minutes it will be
 http://images.indymedia.org/imc/uk/PICTURE-thumb.jpg
That means that it's now linked to the file on loudeye.

If something with the copying and linking over went wrong, the
picture will not appear anymore even though it was first
visible. This happens fairly often and has led a number of
people to believe that "the picture has been pulled" by an
editorial censor.

Copying over has been going wrong quite often, and in a variety
of ways. For example, all recent thumbnails are *linked* to
loudeye (even if they don't show in the wire), but only around
10 percent of the bigger media files are automatically linked
to loudeye. A number of files hasn't been sent to loudeye at
all, and others have ended up in wrong directories. There is a
catch-up script to trigger the copying over process manually,
but it doesn't work all the time either. So a lot of manual
fixing is needed.

If the script for copying and linking over fails it can have
several consequences: the hard drive can be getting full for
example if not enough manual fixing is done, as has happened
lately.

All the bigger media files should be linked to loudeye
and be deleted from stallman. Secondly, if not all big files
are linked to loudeye then the serving load on stallman
increases, causing the web serving performance to be more
sluggish than it need to be.

Still - remember that *no* software can be expected to work
faultlessly all of the time. Corporate software needs constant
maintenance, too, and they pay people to keep watching it.
Indymedia's non-corporate software is pretty good but can fail,
and as everybody in Indymedia gives their free time, there is
not always somebody around to correct things immediately.
This documentation is written to enable *you* to do it!

Here's how to fix it:
---------------------
First of all get the URL of the missing picture:
- On a PC, rightclick on the broken picture and select
properties.
- On a Mac, use iCab (download it from  http://www.icab.de).
Unlike netscrape/exploder, it has a an option to get URL's
of missing images. Click on the broken image, hold the
mouse button down and select "Image > Copy URL to
Clipboard" in the pop-up window.

If the URL is pointing to images.indymedia.org
----------------------------------------------
 http://images.indymedia.org/imc/uk/MISSING_PICTURE-thumb.jpg
this is the URL for loudeye's default directory
/metafiles/uk
connect with an ftp client to :

- the file is not there. the copying over has failed or erred
and you need to copy the file over manually:
1) Check first if it's not ended up one directory deeper
down, in /metafiles/uk/metafiles
2) If it's not in /metafiles/uk/metafiles then you need to
get it from stallman. it will be in
/www/uk/local/webcast/uploads/metafiles

- the file is there but the image is still broken:
1) Check first the file permissions: the file might not be
set as readable for everybody. correct it if necessary.
2) The file might be corrupt. Download it to your hard drive
and open it from there. If it's OK, delete the file from
loudeye and reupload it from your hard drive.
3) The file *is* corrupt, you cannot open it when you have it
on your hard drive: get the original from stallman.
4) The original on stallman is corrupt: post a comment on the
story saying you've checked the file but the upload was
corrupted, and ask to reupload it as a comment. If there's
an email of the author, send him/her a mail saying this.


If the URL is pointing to uk.indymedia.org:
-------------------------------------------

 http://uk.indymedia.org/local/webcast/uploads/metafiles/MISSING_PICTURE.jpg
this is the URL for stallman's default upload directory
/www/uk/local/webcast/uploads/metafiles
connect via ssh telnet to :
check if the file is in either of these directories:
/www/uk/local/webcast/uploads
/www/uk/local/webcast/uploads/metafiles/

What software to use to connect to stallman and loudeye:
--------------------------------------------------------
Using a Mac:
FTP client: Fetch, Transmit or whatever you prefer
SSH and SCP telnet: NiftyTelnet

Using a PC:
FTP client: CuteFTP, NWFTP or whatever you prefer
SSH telnet: Putty
SCP telnet: PSCP
WinSCP
iXplorer

For PC users:
Watch out, WinSCP or iXplorer *CANNOT* transfer files to and
from stallman's upload folders
/www/uk/local/webcast/uploads
/www/uk/local/webcast/uploads/metafiles/
If you want to download a file these folders *DO NOT* use
WinScp or iXplorer. Neither are able to display the amount of
files in these directories - they will simply hang and time
out. You DO need Putty and Pscp. Check with Putty if the file
is present and download it from a Command Prompt via PSCP.

Mac users:
NiftyTelnet does it all. You can telnet with the normal
'connect' button and SCP with a dedicated button. For more info
on how to download/upload heck
 http://uk.indymedia.org:8081/documentation/secure_copy_mac.txt
if you want to see screenshots with that documentation, go to
file  http://uk.indymedia.org:8081/documentation/ and download the
file secure_copy_mac.pdf. but make sure to keep your mouse held
down when you click on the link and download/save it to your disk.