[mythtv-users] Re: Bleb.org down??

[David]

Simon Kenyon wrote:

>On Tuesday 20 April 2004 13:01, David wrote:
>  
>
>>I think the first problem here is copyright issues.
>>    
>>
>so what does bleb do about this?
>  
>
Dunno - good question. Seems he's not said...

>>How many users do you think we have in the UK? 10, 20? 1000?
>>If there are <100 (my guess) then it's not likely to draw attention and
>>it's worth the minimal risk - we're just avoiding hammering the RT site
>>and acting as a collective proxy.
>>    
>>
>
>the number will grow.
>  
>
to... ?
Seriously I can't see it getting to be a problem. And if it does the
copyright issue is likely to have ripped huge chunks out of us first ;)
(and if not - we'd have the numbers for a decent business case so we
could startup on our own!)
We are just talking about servicing the MythTV UK community aren't we?

>>If we make this a publicised / 'official' xmltv feed then we're asking
>>for trouble :)
>>
>>2nd - your design sounds complex.
>>    
>>
>
>yes. i'm a technical architect by trade. complex is what i do :-)
>
>  
>
Me too
Simple, effective and reliable is what I do ;-)

>>I think the simple thing is to set up 2 systems (you don't need slew of
>>them, simple redundancy is fine) running the grabbers and have a simple
>>http pull.
>>
>>bittorrent is great if you're a BB user - not so cool if you want to
>>pull a file ;)
>>
>>Databases are great - but... did you know that some people just use text
>>files and directories ;)
>>I think modern filesystems tend not to barf at lots of files - old ones
>>limit you to a mere 65000 or so ;)
>>(even that should see us through the next 20 yrs - in just 1 dir!)
>>
>>Why not write a nice easy xml file? (gonna have to handle XML anyway so
>>no overhead here)
>>
>>run grab_tv_uk_rt in daily mode.
>>Use an XML parser to split out the data on a 1 file per channel/day basis.
>>Then provide a rebuild XML parser (simple concatenation?) and let people
>>issue 1 wget per channel/day
>>
>>Almost simple ;)
>>    
>>
>
>fine. i like what you're saying. can i ask questions?
>
>  
>
yes ;)

>does each server run the grabbers in their entirety, so they are independent?
>  
>
yes - want to exchange xml and run a diff to validate things - cool.
(don't know how error prone the grabber is as I've never had the courage
- or time - to run it twice)

>how do we spread the load?
>  
>
what load? see sizing above.We're providing redundancy, not load
balancing. (IMHO)
On the servers? Simple http file serving? my 486 could probably handle it.
No database overhead.
Distribute a perl script that downloads the up to date list of servers
and tries something like apt-ping  to use least loaded.
It can also handle errors for the user - and maybe run mythfill - and
handle channel selection...

>how does the list of servers get distributed/updated?
>  
>
a mirrored web page saying "these are the servers - if you get a 500
blah, try another"
(and see above)


David