Expanding TiVo HD

gwmichel · 11-14-2011, 09:54 PM

So I bought a Fantom 1TB external drive for my TiVo HD. Of course it doesn't work. I read the FAQ on how to install it, and I am wondering:

1. How hard is it to do? I have a pc running Windows 7. Nothing fancy. Would I be better off removing the drive from the external drive and replacing, or marrying the 2 drives and using it externally? I don't even know if I have an Esata port on my pc.

2. would I be better off purchasing a plug and play drive? If so could someone make a recommendation? TiVo is not selling the drive on their websites- out of stock. I can't find it on Amazon either.

unitron · 11-14-2011, 10:26 PM

Quote:

Originally Posted by gwmichel

So I bought a Fantom 1TB external drive for my TiVo HD. Of course it doesn't work. I read the FAQ on how to install it, and I am wondering:

1. How hard is it to do? I have a pc running Windows 7. Nothing fancy. Would I be better off removing the drive from the external drive and replacing, or marrying the 2 drives and using it externally? I don't even know if I have an Esata port on my pc.

2. would I be better off purchasing a plug and play drive? If so could someone make a recommendation? TiVo is not selling the drive on their websites- out of stock. I can't find it on Amazon either.

The original S3, it was discovered, had some code in the software that let it use an external drive for additional space. Some drives worked, some didn't, but it wasn't an official TiVo feature.

Since the S3 HD, it has been an official TiVo feature, but only for certain Western Digital models. The TiVo checks some identifying string in the drive's firmware and only co-operates if it finds what it expects.

That way TiVo only has to provide customer service support for those models on which it's collaborated with WD, instead of whatever's on sale from Fly-By-Night, Inc. this week.

You can always format it as one big NTFS partition, hook it to your computer, and use the free version of TiVo Desktop to copy over anything the cable company didn't set the no-copy bit on.

ThAbtO · 11-15-2011, 01:40 AM

If you want to expand your storage, replacing the original drive with a bigger one is your best bet. Using a 2 drive system would only get you trouble in the future since recordings are stored across both drives and if 1 should fail, you can lose them.

dlfl · 11-15-2011, 08:33 AM

Quote:

Originally Posted by ThAbtO

If you want to expand your storage, replacing the original drive with a bigger one is your best bet. Using a 2 drive system would only get you trouble in the future since recordings are stored across both drives and if 1 should fail, you can lose them.

+1

Two HDD's means twice the probability of failure and makes it harder to figure out which drive failed. Check out DVR_DUDE on eBay and weaknees.com for plug-in drives. Both have good reputations.

Before doing anything else, divorce your expander (if not already done) and get the TiVo running stably on the internal drive. Then when you install the new internal drive, put the old one on the shelf. It can be subbed in later as a diagnostic technique when you suspect the new drive is failing -- and it's not worth much for any other purpose anyway.

jrtroo · 11-15-2011, 09:16 AM

Are you sure about "twice the probability of failure"? There are two points of failure, but I don't believe that their overall probability is additive, though it has been a long time since I have thought about such things.

Probability aside, two drives have many more opportunities to fail, not just including the discs themselves-- as you mention the enclosure, cabling, risk of getting knocked about, power supplies, ect.

I see two benefits of the upgrade- one is the capacity and the other is the drive that is on the shelf, ready to put into a tivo if the replacement drive fails.

lrhorer · 11-15-2011, 11:56 AM

Quote:

Originally Posted by jrtroo

Are you sure about "twice the probability of failure"? There are two points of failure, but I don't believe that their overall probability is additive, though it has been a long time since I have thought about such things.

You are correct. For example, if an event has a probability of .75 during an action, performing the action twice does not result in a probability of 1.5 the event will occur. The probability is .875. For events with a very small probability, however, doubing the probability space does roughly double the probability of the event. This is because for a very small value of x, 1 / (1 - x) is approximately equal to 1 + x.

Quote:

Originally Posted by jrtroo

I see two benefits of the upgrade- one is the capacity and the other is the drive that is on the shelf, ready to put into a tivo if the replacement drive fails.

That's right.

dlfl · 11-15-2011, 02:19 PM

Quote:

Originally Posted by jrtroo

Are you sure about "twice the probability of failure"? There are two points of failure, but I don't believe that their overall probability is additive, .......

As lrhorer said, you are correct -- they are not additive. And he's also correct that for small probabilities they are approximately additive.

Quote:

Originally Posted by lrhorer

....... For example, if an event has a probability of .75 during an action, performing the action twice does not result in a probability of 1.5 the event will occur. The probability is .875. .......

I'm not sure what problem you were solving here, so I won't say your result is wrong.

But it's not the right result for the problem at hand, which is: what is the probability that one or both HDD's will fail in a certain period of time given the probability for each drive independently is P. If each drive has a 75% probability of failing during that period, the probability that one (or both) will fail during that period is 93.75% (i.e., 15/16) -- thus 6.25% probability that both will make it through the period without failing.

I think a realistic case, based on my hazy memory of recent Google statistics on HDD's in their server farms, is a 50% failure probability over five years. Thus the probability for failure in a single-drive Tivo is 50% while the probabilty that at least one drive in a 2-drive TiVo will fail is 75%. Thus for a five year period, the probabilities are nowhere close to additive -- rather you get only 1.5 times the failure probability instead of double. If you rerun this case for one- or two-year periods the probabilities are much smaller and the results are much closer to additive. You could state accurately that with two HDD's you cut the probability of making it through five years without a failure in half, from 50% to 25%.

lrhorer · 11-15-2011, 03:11 PM

Quote:

Originally Posted by dlfl

I'm not sure what problem you were solving here, so I won't say your result is wrong.

I was just giving a very broad example of how probabilities of this sort are not additive.

Quote:

Originally Posted by dlfl

But it's not the right result for the problem at hand, which is: what is the probability that one or both HDD's will fail in a certain period of time given the probability for each drive independently is P. If each drive has a 75% probability of failing during that period, the probability that one (or both) will fail during that period is 93.75% (i.e., 15/16) -- thus 6.25% probability that both will make it through the period without failing.

Oops! 'Sorry about that. I whipped out that answer from my head, and instead of squaring 0.25, I divided by two. Silly me. You are right, of course:

1 - ( 1 - 0.75 ) ( 1 - 0.75 ) = 1 - 0.0625 = .9375 = 93.75%

Quote:

Originally Posted by dlfl

I think a realistic case, based on my hazy memory of recent Google statistics on HDD's in their server farms, is a 50% failure probability over five years.

That high? Really? That's surprising. Certainly I have not seen that level of failure in my systems. I don't have the number of drives online that Google does, of course, but still the number is statistically significant - about 100 drives, many of them spinning since 1992. I've lost 8 or 10.

Quote:

Originally Posted by dlfl

Thus the probability for failure in a single-drive Tivo is 50% while the probabilty that at least one drive in a 2-drive TiVo will fail is 75%. Thus for a five year period, the probabilities are nowhere close to additive

Yeah. If the number is 50%, then the probability is not small compared to 1.0. For double digit accuracy, one has to be closer to 10%, not 50%.

Stuxnet · 11-15-2011, 03:34 PM

Case A: You have one coin... Tails = Failure. You flip it and there are 2 possible outcomes equally obtainable... H & T —> You have a 50% likelihood of Failure (T)

Case B: You have two coins... any Tail = Failure. You flip both and there are 4 possible outcomes... HH, HT, TH, TT —> You have a 75% likelihood of Failure (HT, TH, TT)

The results here are exaggerated due to the high fail percentage. If your HDD had a 0.001% chance of failure per year (99.999% chance of no failure) and you had 2 units, your overall expected failure rate would be:

1-(.99999)^2 = 0.0000199999 or 0.00199999% (nearly double, but a really, really small likelihood).

Bottom line... Sure 2 HDDs will roughly double the failure rate... but in reality the likelihood of failure shouldn't materially sway your decision either way.

Put another way, it's like deciding between two $40,000 comparably equipped cars, where one costs 80 cents more than the other. Would you sweat that?

shwru980r · 11-15-2011, 04:11 PM

If you're not using a plug and play external drive, then you have to crack the case, remove the internal drive, marry the two drives and re-install the internal drive. For that much trouble, you might as well replace the internal drive with a single larger drive.

I'm not sure if it's possible, but it might be preferable to just use an external drive and remove the internal drive entirely. Then you don't have to crack the Tivo case anymore for drive failures.

dlfl · 11-15-2011, 09:30 PM

Quote:

Originally Posted by lrhorer

........That high? Really? That's surprising. Certainly I have not seen that level of failure in my systems. I don't have the number of drives online that Google does, of course, but still the number is statistically significant - about 100 drives, many of them spinning since 1992. I've lost 8 or 10.
...........

Sorry, you're right -- 50% failure probability over 5 years is too high. Here is the link to the study I mentioned:
http://static.googleusercontent.com/...k_failures.pdf
The relevant data is plotted in a bar chart on the 4th page. Adding up the bars for the first 5 years gets you to around 32% over 5 years. Note this data was based on consumer grade HDDs!

Quote:

Originally Posted by Stuxnet

.........Bottom line... Sure 2 HDDs will roughly double the failure rate... but in reality the likelihood of failure shouldn't materially sway your decision either way.

Put another way, it's like deciding between two $40,000 comparably equipped cars, where one costs 80 cents more than the other. Would you sweat that?

Assuming 32% failure probability over five years the failure probabilities for 1 and 2 HDD's are 32% and 54%. That seems more significant than 80 cents on a $40K car, to me at least.

In the absence of other factors that might favor an expander drive (e.g., ease of installation, or cost) this seems like a good argument for going single drive.

Stuxnet · 11-16-2011, 08:43 PM

Quote:

Originally Posted by dlfl

Assuming 32% failure probability over five years the failure probabilities for 1 and 2 HDD's are 32% and 54%. ... In the absence of other factors that might favor an expander drive (e.g., ease of installation, or cost) this seems like a good argument for going single drive.

If you're going to use a 32% 5-year fail rate, then your math is impeccable

I've been fortunate... never having a failed HDD... (there, I said it, now I'm jinxed).

Soapm · 11-16-2011, 09:15 PM

What are you guys, math professors? My head hurts reading that crap...

I say if it fails deal with it... If it don't fail, fail to deal with it...

unitron · 11-16-2011, 09:17 PM

[quote=Stuxnet;8813512...I've been fortunate... never having a failed HDD..[/QUOTE]

That's unpossible! Do you only buy new ones and replace them weekly?

Don't know why the quote tags failed.

In preview the first "quote" is in all caps.

ThAbtO · 11-16-2011, 09:20 PM

Quote:

Originally Posted by Soapm

What are you guys, math professors? My head hurts reading that crap...

I say if it fails deal with it... If it don't fail, fail to deal with it...

Here's a couple of Advil.

( O O )

dlfl · 11-16-2011, 09:28 PM

Quote:

Originally Posted by Soapm

What are you guys, math professors? My head hurts reading that crap...

I say if it fails deal with it... If it don't fail, fail to deal with it...

But they're so reliable we have all this spare time to post this crap.

I have to admit my personal experience suggests a probability of failure much less than 32% in five years. Google must be working their drives too hard.

But humans are very inaccurate at estimating probabilities too.

lillevig · 11-17-2011, 03:37 PM

In the end, theoretical failure rates can never replace empirical data and neither of those stands a chance against Murphy's Law.

WhiskeyTango · 11-17-2011, 08:05 PM

If you are willing to try, I'd recommend the DIY approach. The drives are much cheaper and there are step by step instructions on this site to copy the software to a blank drive that are very easy to do.

ThAbtO · 11-17-2011, 08:23 PM

Quote:

Originally Posted by WhiskeyTango

If you are willing to try, I'd recommend the DIY approach. The drives are much cheaper and there are step by step instructions on this site to copy the software to a blank drive that are very easy to do.

I suggest waiting until hard drive prices has come down, currently prices are 2-3 times the usual prices because of the Thailand flooding issues.

Soapm · 11-17-2011, 08:57 PM

Quote:

Originally Posted by ThAbtO

I suggest waiting until hard drive prices has come down, currently prices are 2-3 times the usual prices because of the Thailand flooding issues.

What's the odds of that???

ThAbtO · 11-17-2011, 09:02 PM

Its unknown when the disipate but right now a 1 tb green drive is about $120, a 2 tb is $170 at amazon.com. I recent got a 2 tb drive for only $70 last month.

WhiskeyTango · 11-18-2011, 08:35 AM

Quote:

Originally Posted by ThAbtO

I suggest waiting until hard drive prices has come down, currently prices are 2-3 times the usual prices because of the Thailand flooding issues.

True but DVR Dude is charging $150 and Weaknees $299 for 1 TB drives. You can get a Seagate 1 TB drive from Best Buy for $55 right now.

unitron · 11-18-2011, 09:28 AM

Quote:

Originally Posted by WhiskeyTango

True but DVR Dude is charging $150 and Weaknees $299 for 1 TB drives. You can get a Seagate 1 TB drive from Best Buy for $55 right now.

In fairness, we don't know if that drive has been tested in TiVos yet, and it certainly doesn't come pre-imaged, which is a convenience for those who don't waste all their time screwing around with computers like me and probably thee.