Re: Gnutella specs wiki down?

El Thursday, 20 de March de 2008 11:29:59 Philippe Verdy escribió:
> So I think that all what is interesting: the various documents that define
> each extension developed by LimeWire and BearShare and the history of this
> discussion list, will help collect back some information.

archive.org has readable backups most of the sites from the two wikis - and 
for that they deserve a blessing, I think. 

> One main error was the multiplication of specification sites, and then
> their merging to a single place without verifying that they were safe for
> the long term: we thought we had enough backups.

> I don't understand why, given the experience of spammers trying to break
> all wikis, you did not perform any database dump for offsite backup...
> Notablyt when you can see that these spammers are generating lot of traffic
> that is inflating the volume of the history database, causing it the wiki
> to become slower.

I don't have backups, because I once tried to take them and gave up after 
several hours. 

Then I decided to trust my hoster... 

Annoyingly I now managed to get the backup stuff working in under 2 hours, 
after I got burned. 

> incident has been signaled!). A 3-days only backup serves absolutely
> nothing (notably when your ISP takes weeks to react after your signal the
> incident, and cannot keep at least one backup for giving you enough time to
> restore most missing data!), 

Problem is: They do longer backups, but not for databases :-( 

And they did react quite swiftly and backupped one of the four databases I 
have on my server. I then needed some time, because some things didn't work 
from the backup database they had created (typos and copy and paste can 
really kill - the database name ended in 34, not in 43...), so I decided to 
tackle them when my tests at university were done, which was a week later. 

Afterwards I found out, that they only backupped one of the four databases, 
and gnet-specs was on one of the other three. 

Since then I tried several things to get any data from the databases and just 
didn't get at it. 

> if the data is needed for very long term and 
> you cannot administrate the website at least every day to make sure it is
> fully functional; it may just be useful if the site is administered daily
> by people that will react a few hours after the incident, 365 days per year
> (spammers and hackers trying to pollute and/or break your site are active
> 24/24 during all year).

I couldn't have paid those people... 

But I know what I'd do the next time: Simply use a distributed Version control 
system like Mercurial ( http://www.selenic.com/mercurial/ ) as backend store 
for either static html sites or a wiki like ikiwiki ( http://ikiwiki.info/ ). 

I'm trying to pull the data from archive.org at the moment - via wget... I 
hope I'll get most of the content. 

Best wishes, 
Arne
-- 
Unpolitisch sein
Heißt politisch sein
Ohne es zu merken. 
- Arne Babenhauserheide ( http://draketo.de )
-- Weblog: http://blog.draketo.de

-- Mein öffentlicher Schlüssel (PGP/GnuPG): 
http://draketo.de/inhalt/ich/pubkey.txt


[Non-text portions of this message have been removed]