[mythtv-users] mythtv frontend startup beautification

[Dan Wilga]

On 11/19/14 11:41 AM, Hika van den Hoven wrote:
> Hoi Dan,
>
> Wednesday, November 19, 2014, 5:26:21 PM, you wrote:
>
>> On 11/19/14 10:50 AM, Hika van den Hoven wrote:
>>> Hoi Dan,
>>>
>>> Wednesday, November 19, 2014, 4:03:23 PM, you wrote:
>>>
>>>> Personally, I use suspend to RAM rather than actually shutting down
>>>> my frontends. I still restart X and mythfrontend when resuming, but
>>>> that definitely lessens the ugliness. During suspend, I use this line
>>>> in irexec:
>>> No need to stop and restart X and mythfrontend. You might however need
>>> to save and restore your video memory. Older nvidia cards have problems
>>> with restoring from suspend. Also some kernel modules have problems
>>> especially some v4l and audio modules. You might have to reload them
>>> on resume. But that can all be automated.
>>>
>> I set this up quite some time ago, so I don't recall now why I had to
>> kill and restart X. I think it might have had something to do with
>> video, though these are ION2 boxes, so they are new-ish hardware. I may
>> try it again when I get a chance.
>> _______________________________________________
> I've done it with 0.26 and 0.27. Only as said with one slightly older
> nvidia card I have to save and restore the video memory or else I get
> a corrupted screen. I guess it reinitializes on resume. It's only
> slightly faster but the plus is you get your screen back as you left
> it. I must say I use the tuxonice kernel addition and associated
> scripts, which I find beter and easier to configure especially with
> suspend to disk as I use on my laptop.
>
I finally went back and tried this again. If I do nothing at all before 
or after sleep, it works fine for me--but only if I wake the computer 
soon after putting it to sleep. If some indeterminate amount of time 
passes (like a few hours), then when I wake the machine MFE won't 
respond to any remote control button presses.

What makes my frontends a bit unusual in this regard is that I'm using 
the "LIRC daemon socket" setting (LircSocket in the settings table) to 
talk to my backend server. I use RF remotes, which talk to their 
receiver, which then emits IR pulses that are then read by the IR 
receiver in my HDHomerun. The BE uses LIRC to route this data to the 
frontends.

I have compared an "lsof" listing of the open ports on a 
properly-working FE and the same after the failure, but can't see any 
difference. Most specifically, the LIRC port 8765 is still connected. 
Nonetheless, the FE acts as though this connection has been severed. 
What I need to do still is look at the server side of things. Maybe it's 
dropping the connection once it reaches a certain age.