[mythtv-users] All encoder MPG stream buffers are full. Dropping data.

[George Mari]

Kelly Shutt wrote:
> Here's output from /proc/interrupts.  The storage for my recordings is 
> actually on the network device, and not local.  I noticed that IVTV and 
> my network interface are on the same interrupt, so maybe that's 
> contributing to the problem.  However, as I said before; I haven't had 
> any problems like this to date and it just started doing it after the 
> upgrade.  I've always had it on network storage without any problem. I
> suppose I could try moving the ethernet adapter to another PCI port, and 
> get it on a different interrupt.  That might free it up for things to 
> transfer properly.

The playing field changed when you upgraded - the goalposts moved.  Just 
because it worked before doesn't mean it should work now, unfortunately.
> 
> Yes, I have verified that my backup and optimization scripts are not the 
> problem.  In fact they're scheduled and run well outside my recording 
> times.  The issue is showing up at all times, no matter what is running 
> on the system.  It's not just a once in a while thing, it's dropping 
> like 25% of the data on every single recording.
> 
>            CPU0      
>   0:  373485936   IO-APIC-edge      timer
>   1:         12   IO-APIC-edge      i8042
>   6:          3   IO-APIC-edge      floppy
>   9:          0   IO-APIC-fasteoi   acpi
>  12:        114   IO-APIC-edge      i8042
>  14:   29776090   IO-APIC-edge      ide0
>  16:  116015732   IO-APIC-fasteoi   ivtv0, eth0
>  17:      32881   IO-APIC-fasteoi   sata_sil
>  18:      12037   IO-APIC-fasteoi   CMI8738-MC6
>  19:     872089   IO-APIC-fasteoi   nvidia
> NMI:          0   Non-maskable interrupts
> LOC:  373494563   Local timer interrupts
> TRM:          0   Thermal event interrupts
> SPU:          0   Spurious interrupts
> ERR:          0
> MIS:          0
> 
> Thanks,
> Kelly

Another thing I thought of - try temporarily turning off all logging 
from mythbackend.  I have a particularly slow master backend (Dual 
P3-600) and When I first upgraded it from .18 to .20 a while back, I had 
a performance problem that took me many days to track down.  Somehow, 
the default level of logging on mythbackend had changed during the 
upgrade, and it was writing all sorts of things to the mythbackend log 
file.  I tried turning off all logging, and my performance issues went 
away.  So I went turned on some logging, I think just the important 
ones, and it's been fine ever since.

Solving this type of thing is a process of elimination, one by one. 
Start with the easy-to-implement things first, then the harder ones.

You didn't mention anything about your PCI latency?