2012-12-21, 22:35
(2012-12-21, 22:13)janbar Wrote: Hi,
Probably you have a network issue. But also since we have limited the buffer size request depending of the backend response to avoid any break of show, this probem could occur. For me HD use 35~45 mb/s. Also you use a remote storage, so you need at least twice of the bandwidth. That means you need 70~100 mb/s. This is the reason why i said in a previous post you should have a gigabit network to manage these cases.
The best thing should be to use two different networks (so two ethernet cards on the backend) , one for the storage and the other for the remote client.
I will make a branch with an algorithm less aggressive for the buffer size. You could try it.
br.
Just to clarify, I have a single server with all my disk space built in.
I do not have a separate NAS box for my video captures.
(I understand how a separate LAN segment would be a good idea if that were the case)
I have tested with 2 separate targets (Win/Linux laptop and RPi card) connected to server via GigE.
Since I (only) see this issue using RPi PVR/Myth, and not RPi Videos/NFS, I'm pretty sure it's not my network.
Since I do not see the issue on Laptop/Win/XBMC or Laptop/Linux/Myth-frontend, I assume my issue is specific to RPi.
I thought it might be resource scheduling on the server side (mythbackend process)
but not being able to duplicate on Laptop/Linux with mythfrontend kind of eliminated that track for me.
(I did play some with process scheduling for mythbackend on the server, to no avail)
I am at your disposal for testing any fixes. Thanks for the prompt response!
EDIT: Oh, BTW my RPi is the 256M version, if that makes a difference.