[mythtv-users] Dual-FireWire troubles in 0.20 solution found

Michael T. Dean mtdean at thirdcontact.com
Fri Sep 15 20:30:08 UTC 2006


On 09/15/06 16:03, Yeechang Lee wrote:

>Michael T. Dean <mtdean at thirdcontact.com> says:
>  
>
>>Or, so you tried to tell Daniel.  He responded back with
>>http://www.gossamer-threads.com/lists/mythtv/users/183967#183967 .
>>Remember the part about, "(To big a buffer can give you thrashing
>>which destroys performance.)"?
>>    
>>
>May I ask what your problem is?
>
Sorry.  I don't have time to type up all my problems--far too many to 
list.  But, I'm trying to work them out...

However, I can say, I really don't have a problem with you.

> This is the second time you've snapped
>at me out of the blue
>(<URL:http://www.gossamer-threads.com/lists/mythtv/users/185920#185920>
>and
><URL:http://www.gossamer-threads.com/lists/mythtv/users/186001#186001>)
>after I offer correct and, hopefully, useful information to the list.
>  
>
I apologize.

>For the record:
>
>* I have 1.5GB of RAM. One guess on what my swap usage is. (A clue:
>  It's the same as the level of your knowledge of basic social
>  etiquette.)
>* I had HD ringbuffer set to 96000 (or whatever that maximum value is)
>  in 0.18, 0.19, and 0.19-fixes with no problems recording from dual
>  FireWire boxes. 0.20 broke this. After Jim kindly helped me out I
>  decided to pass the word on for others' benefit. Acting as such is
>  generally considered a good thing.
>  
>
And I didn't argue with the fact that setting HDRingbufferSize to too 
large a value will cause problems or say that you shouldn't have told 
the list.  I just suggested that your "assigning" Jim a new development 
task ;) may be a bit pre-mature.

In other words, you said the max (96256) worked on previous versions.  
You know that you had issues with the default value (which is now 
4700--and is also the min).  So, basically, we have datapoints on the 
min and the max (for a 2-firewire system), but you haven't been running 
with your new setting long enough to know if you'll have issues with it 
(and at this point have no indication that you will as it seems to be 
working correctly--which was the point of your post).  And, if it ain't 
broke...

>>(Not to mention the part about, "It is very easy to break MythTV
>>playing with the TFW params.")
>>    
>>
>The current issue has nothing to do with the TFW parameters (although,
>for the record, boosting them caused me no problems I'm aware of and
>considerable good, as I've discussed on the list several times), as I
>haven't tweaked them at all in 0.20 as I can't get Axel's SRPMs to
>rebuild.
>
And until recently, you would have said that boosting the 
HDRingbufferSize to 96256 "caused me no problems I'm aware of and 
considerable good"...

>>IMHO, a "longer-term solution" isn't necessary just so you can use
>>your overly-large HDRingbufferSize.
>>    
>>
>Define "overly-large." 96MB is what mythfrontend permits it to go up
>to.
>
But your system requires 23.5MB, so 96MB is overly large.

>>If you want to use 96MB, get another backend and use 96MB in each or
>>get more RAM or put your MySQL on another system or ...
>>    
>>
>Let me repeat: I have 1.5GB of RAM.
>
Which doesn't seem to be enough for 2 firewire captures running with 
96MB ringbuffers each.

> And you, sir, need some
>counseling. I don't care if you are a MythTV developer;
>
Actually, I'm not a MythTV developer.

> that may give
>you some license to be rude, but not to be rude *and* wrong.
>
Nope.  I'm completely unlicensed.

Mike


More information about the mythtv-users mailing list