[mythtv-users] Myth problem or HDPVR probmem?

[Don Lewis]

On 14 Feb, Robert McNamara wrote:
> On Mon, Feb 14, 2011 at 12:09 PM, Dan Armbrust <
> daniel.armbrust.list at gmail.com> wrote:
> 
>> Oh, and a side issue is that even if it is a hardware problem / driver
>> problem... mythTV does a horrible job at handling the error.  It
>> doesn't recognize that the capture card has failed... it doesn't try
>> another one, doesn't stop attempting to use the non-responsive one.
>>
>> Just carries on, making 0 byte recordings of everything, until someone
>> notices it.
>>
>>
> Yes, MythTV unfortunately presumes that your hardware and drivers are in
> good working order.  I think I can speak for all of us that we would happily
> accept patches that more gracefully deal with those pre-requisites not being
> true, but ultimately we have to make the assumption that the devices the
> user configures do in fact work.

If I take MythTV out of the picture, then my HD-PVR and the drivers in
Fedora 14 appear to be working properly.  I just ran a test where I
looped doing channel changes and captures from the /dev/video device
using dd and I got 150 successful captures before I got tired of
testing.

When my HD-PVR is being controlled by mythbackend (0.24), I used to have
problems with the HD-PVR locking up a couple times a month when I was
running the i386 version of Fedora 12.  After upgrading to the x86_64
version of Fedora 14, my HD-PVR would lock up every day or so, with
maybe three successful recordings before it failed.

About a week ago I increased the delay at the end of my channel change
script from 4 seconds to 10 seconds, increased the HD-PVR tuning timeout
in mythtvsetup, covered the IR sensor on the HD-PVR, and waved a dead
chicken over it.  It works somewhat better now and only locks up about
every other day (with the blue halo off).  When it locks up, the cat
/dev/video test fails with an I/O error.  Power cycling the HD-PVR
brings it back to life for a couple of days.

Prior to this, I tried replacing the HD-PVR wall wart, which didn't help
a bit.  All the hardware is on an UPS.  We've had some power glitches in
the past few days that haven't triggered any problems.  The last lockup
was overnight when the power was clean.

The problem seems to be triggered by some sort of interaction between
mythbackend and the HD-PVR, since I don't have the same problem using dd
to capture with the same tuning script.