[IMC-Docs] update: report about imc-brasil-tech's work on docs [2009-01-17]

[nah]

Report
======

Yesterday gdm turned on the humus vserver backup from 2008-09-28 in strummer,
kompost's colocation neighbour, and gave us access to it. We created an
indymedia user to be used to work on the docs by the members of the
collective. This user is going to expire on 2009-01-27.

We thought we should concentrate our work on 2 things:

  1. put the docs backup up, read-only; and 
  2. work towards backup update with non-corrupted data from humus latest
  image, using the analisys of the generated lists and data.


1. Docs backup
==============

After trying to understand the deepness of the TWiki config, we thought that
instead of modifying the existing TWiki config by trying to disable logins and
user creation, it would be better to create an empty instalation of foswiki[1]
with no user info, just the data.

So foswiki is installed and running with the docs backup in the usual address:

  http://docs.indymedia.org/

We did not migrate the users, so protected pages are not publicly available.
One idea would be to find out how to prevent updates to the content in foswiki
(that should be easier than with TWiki, as the commercial version is the only
one they provide support) and copy the user's passwords in the image .htpasswd
file, as it's not corrupted. 


2. Docs backup update
=====================

We ran scripts to generate a list of files which differ between the backup
and the image. The scripts and generated data are attached. Some numbers:

  45473 inspected_files
  832 inconsistent_files
  800 inexistent_files
  276 files with file system errors

These contain all the files under docs.indymedia.org/ folder. The files under
pub/ and data/ sum about 43.000 entries.

Inconsistent files are those that exist in the backup but whose contents are
different. They can either be updates or corrupted files. Inexistent files are
those which exist in the image but not in the backup. They can also be either
updates or corrupted files. File system errors prevented 276 files from being
read. These should be the ones we would have to recover by hand from web or
user's caches.

We are working now on scripts to test the consistency of the files and copy
newer data to the backup in order to obtain the most complete version
possible, regarding the time, knowledge and data we have.

We believe that in the next working session (which should happen tonight)
we'll be able to create lists of definitely corrupted files and newer files
which can be updated to the backup. That would allow writing a simple script
which updates the image based on this list.


Proposal for next Steps
=======================

  X. read-only backup: disable wiki updates and migrate users,
  Y. backup update: finish the scripts to generate the lists,
  Z. backup update: update the backup with newer files.


[1] http://www.foswiki.org/ - why this fork? http://foswiki.org/Home/WhyThisFork
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ellen-docs.tar.bz2
Type: application/octet-stream
Size: 812218 bytes
Desc: not available
Url : http://lists.indymedia.org/pipermail/imc-docs/attachments/20090119/4552a738/attachment-0001.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
Url : http://lists.indymedia.org/pipermail/imc-docs/attachments/20090119/4552a738/attachment-0001.pgp