[mythtv-users] Why not MythArchive on Frontend only?
ryunokokoro at gmail.com
Mon Sep 10 18:34:18 UTC 2007
Brad DerManouelian wrote:
> On Sep 10, 2007, at 10:50 AM, Eric Robinson wrote:
>> Why not MythArchive on Frontend only?
> Because it hasn't been implemented yet.
>> Okay, I've thought about how many people may run minimum-CPU, diskless
>> Frontends and such but this feature would be really, really
>> Certainly a bit of transcoding has to occur. Why not have the backend
>> do the transcoding/ISO preparation and then ship it off to the
>> for burning? That would enable such scenarios:
> I believe the implementation would simply be that the frontend would
> do all the work and just be told where to find the files unless
> someone has other thoughts on how to delegate MythArchive jobs to
> available backend machines. Sounds like a lot of reworking of
> MythArchive to include its processes in the standard Job Queue.
Ahh, I confess knowing absolutely zilch about how Myth and it's plugins
interact underneath. T'was a scenario based on a few [perhaps intensely
incorrect] assumptions about what goes on behind the scenes and what
> <snips out illegal file sharing conversation>
Hahaha! True. Sorry about that. Just having a little fun.
>> [I assume that managing a Frontend only system is a lot easier than a
>> Frontend/Backend combination.]
> Not really. Combo machines are much easier to set up and maintain
> than separate machines. You don't have to worry about network access,
> bandwidth, mysql access, setting up the right IP addresses (which
> might change with DHCP networks), etc, etc.
By manage I meant "Keep up-to-date". Service, perhaps, would have been
a better term.
The setup that I'm planning has one very powerful computer (quadcore)
with two HD tuners [HDHomeRun, likely] and well over 3TB of storage
acting as the backend with three Mac Minis (mentioned in another mail to
the list) acting as Frontend machines (for an HD based implementation).
The plan includes access for pre-existing computers as well. The
network is being designed as a network of ever-on computers. That's why
we chose the Mac Mini's for frontend machines: low power profile.
Someone tested power consumption and came up with around 28 Watts when
watching a DVD from the drive! Idle is even less, of course.
As for network access and such, we'll be installing a special MythTV
router to dole out IPs to backend computers and allow frontends to float
(though we could change that to give everyone static IPs). The only
computers directly connected to the MythTV router will be the Mac
Minis. Everyone else would be accessing it from the standard
house-router. This keeps the Myth network from getting saturated with
normal internet traffic from laptops and such used throughout the
house. This would also be a fairly simple network to set up.
(Anyone see any glaring wholes in the current plan above?)
Installing Mythfrontend on Macs is a simple drag and drop process. I'm
extremely thankful for that, by the way (thanks snyperpad mans! I've
heard that someone's working on a Mythbackend.app as well. All power to
you, man!). The backend is a bit tougher. We'll be using Mythbuntu on
the backend and I have very little idea on what to expect there (aside
from the screenshots and comments in the lists which look/sound fantastic!).
While the 2Ghz Mac Mini's are likely fast enough to run as
frontend/backend machines that can handle MythArchive, the hope would be
that the Frontend could continue to watch [HD]TV while the backend
worked on the transcoding. Maybe it's powerful enough to do both? I've
Anyway, it's just a thought. It's *doable* but I understand when lofty
dreams don't really mesh with current implementations. Heh.
Would be a neat feature, at least.
More information about the mythtv-users