"Spread Mode" question

[ACTCMS]

The concept of load-balancing is definitely a good one although predicting disk space required for the duration of a recording would never work accurately. The only way sub could implement that would be using a user configurable "fiddle-factor" or fuzzy logic, i.e., the user sets something which says 1 hour @ HD = X GB of storage requirement. I don't see that happening (or being reliable).

On the subject of load-balancing, however, perhaps using the logic used for allocating recordings to different tuners could also be used to provide a reliable spread. I do think it would be more complex to implement than the general theory appears to be though.

Cheers,
Brian

I thought I had Just said that Brian - in any case, no matter how compex the problem or how elegant the solution, it's only going to defer the inevitable by a couple of recordings...

Alex

[ttfitz]

Seriously though... if NPVR could do exactly what you asked for in your opening post, your two recordings might be fine, but what happens to any subsequent recordings?

A fine point, but what you don't understand is that when this was happening, it wasn't just two recordings, it was two every night. Since I couldn't tell where the space was, I thought I had cleared enough space, but I had done so on two different drives, and the same thing would happen repeatedly.

IMHO it's really not worth the coding effort just to save two extra recordings. Folks really need to take responsibility for their own recording space. After all, NPVR now has the Spread feature and it's had the series Keep so many episodes feature for some time.

Granted. I wish to point out that I brought this up not to have any complaint about what is done, just point out something that possibly hadn't been considered in the original formulation of this feature. No matter how good something is, there can always be room for improvement, right? I also think Mr. Johnson has a good point with the load balancing - while it might not matter a lot, it is certainly better to have two recordings at once being done on their own drives, wouldn't you think?

Also, while I think this is a good feature, I probably won't be using it, because like you said, I like to be responsible for, and in charge of, my recording space. In the same way, I tend not to use the recurring recording option that much, and I don't combine my QAM and Cablecard channels, preferring to select which show to record on which tuner. But I think all those features are great ones to have.

PS could I suggest that Tim gets a second NPVR system going so he can watch two recordings at the same time?

I was going to respond at more length, but was too busy watching two recordings at the same time - unfortunately now I have three tuners and I'm STILL falling behind!

[ACTCMS]

I was going to respond at more length, but was too busy watching two recordings at the same time - unfortunately now I have three tuners and I'm STILL falling behind!

So the feature you are really looking for is multiple PiP while watching recordings - let's call it the TV Wall feature...

[bgowland]

I thought I had Just said that Brian - in any case, no matter how compex the problem or how elegant the solution, it's only going to defer the inevitable by a couple of recordings...

Alex

Our post crossed - I was part way through my response when the puppy made a noise along the lines of "If you don't let me out the back door any time soon the carpet gets it".

I wish to point out that I brought this up not to have any complaint about what is done, just point out something that possibly hadn't been considered in the original formulation of this feature.

I doubt anyone sees your post as a complaint and it's often discussions like this which lead to changes in npvr functionality.

I was going to respond at more length, but was too busy watching two recordings at the same time - unfortunately now I have three tuners and I'm STILL falling behind!

Why am I suddenly reminded of "The Man Who Fell To Earth"?

Cheers,
Brian

[johnsonx42]

The concept of load-balancing is definitely a good one although predicting disk space required for the duration of a recording would never work accurately. The only way sub could implement that would be using a user configurable "fiddle-factor" or fuzzy logic, i.e., the user sets something which says 1 hour @ HD = X GB of storage requirement. I don't see that happening (or being reliable).

It doesn't really need to be accurate or reliable. The amount just has to be somewhat reasonable, within an order of magnitude is close enough. The logic is simply "ok, just started a recording on drive #1; now we're about to start another recording, so let's assume an average sized chunk of space will be used and recalculate the free space based on that assumption". I don't think there'd be any need to make a more accurate estimate than "average sized chunk of space", somewhere in the 2 to 5gb range. The goal of the logic is "fill all recording destinations roughly equally, but if two or more are pretty close, then load balance."