ServeTheHome's RAID MTTDL Calculator Bug and Suggestion Box

PCBONEZ

New Member
Nov 1, 2017
16
0
1
61
I can not get the mentioned "other" box to work: "Enter # for MTBF (base 10)"
Been trying to enter 1.0M for MTBF.
No matter how I enter it doesn't calculate and what I entered blanks out.
Which syntax/style are you using for base 10 in that box?
Thanks.
 

PCBONEZ

New Member
Nov 1, 2017
16
0
1
61
Just a suggestion...
Might reduce computational load if you weren't doing numbers like 6,889,809,525,022.22 out to 2 decimal places.
From a practical standpoint 3 or 4 significant digits should be enough for comparisons - if that kind of rounding is supported in your program.
.
 

Patrick

Administrator
Staff member
Dec 21, 2010
12,181
5,257
113
I can not get the mentioned "other" box to work: "Enter # for MTBF (base 10)"
Been trying to enter 1.0M for MTBF.
No matter how I enter it doesn't calculate and what I entered blanks out.
Which syntax/style are you using for base 10 in that box?
Thanks.
1 works. 1.0 works. 1.2 works. Are you putting the M in there?

And the rounding - not an issue in a MTTDL model like this.
 

PCBONEZ

New Member
Nov 1, 2017
16
0
1
61
Firefox is blocking the page reload to the new numbers when I click "submit" - but only when I use the "other" box.
If I 'tell' Firefox to allow the page reload, it will, but it reverts back to the preset value (1.2M) in the process.
Once the blocking happens it locks-in and stays. I have to close the page and reopen to make it quit.

None of the other (other than "other" :confused:) boxes exhibit that behavior.
If I never touch the "other" box in a session everything works fine.
.
Blocked Reload.jpg
 
Last edited:

Patrick

Administrator
Staff member
Dec 21, 2010
12,181
5,257
113
We do not support folks having issues that are using ad blocking plug-ins. Too often they break things.
 

PCBONEZ

New Member
Nov 1, 2017
16
0
1
61
We do not support folks having issues that are using ad blocking plug-ins. Too often they break things.
First it makes no sense that an ad-blocker would -only- affect it when the 'other' box is used.
Second - I did a fresh install of Firefox with no ad-blocker. - Same exact behavior.

Maybe the problem is newer versions of Firefox don't like some of your older HTML.
.
 
Last edited:

nbritton

New Member
Nov 19, 2016
16
6
3
42
1 works. 1.0 works. 1.2 works. Are you putting the M in there?

And the rounding - not an issue in a MTTDL model like this.
No, the "other" box still does not work. Furthermore, it is not clear what the input should be, do you want it in base 10 scientific notation? Nothing I put in the box seems to do anything. It doesn't work in Safari, Chrome, or Firefox.
 

WarthogARJ

New Member
Feb 16, 2020
2
1
3
Hi,
Thanks for this calculator.
It's exactly what I was looking for, and I used it to help me figure some things out.

However, you have a bug in it.
Where you enter your own MTBF, by entering a Base 10 value.
It doesn't seem to react to that properly.

I wanted to use 1million, since I want to model a Seagate Ironwolf Drive, and that's the Manufacturer's suggested value.
But it gives me incorrect values.

Its' easy to see that without doing anything more, by using 1.2M and then 750k.
And what it gives for 1M is not in between that.

And then I got more complex, and put in a series of cases, and did a regression on the output.
And in fact it's a VERY nice fit.

I just did Raid 0, 1, 5 and 6.

If you plot MTBF on x-axis, and MTTDL on y-axis, then you get a perfect fit with: MTTDL = A_0 x (MTBF ^ A_1)
"Perfect" as in the R^2 is 1.00000

And there must be a relationship there, because the A-1's are related.
As in:
A_1 = 2.00 for Raid 0 and 1
A_1 = 1.00 for Raid 5
A_1 = 3.00 for Raid 6

With MTTDL in years.
And:
NRER = 10^14
Drive Capacity = 4TB
Sector Size = 512B
Disks = 5
Volumes = 1
Rebuild Speed = 100 MB/sec

The relationship for the A_0's is not so obvious:
Raid 0, A_1 = 2.13E-06
Raid 1, A_1 = 6.80E-08
Raid 5, A_1 = 1.18E-05
Raid 6, A_1 = 9.33E-10

And I suggest that you provide more space to see your inputs, it's too cramped.
At least on Chrome.

I'm now reading your Primer: I only saw it now....:-}

Thanks,
Alan
 

olee

New Member
Apr 6, 2021
2
0
1
I wanted to make a simulation for using SSDs in Raid 0, Raid 1, Raid 5, to see how much risk I take for losing drives.

I understand correctly this calculator is for HDDs only?

QNAP HS-453DX NAS, and I want to use it with SSD only.
  • 2x M.2 SATA slots, where I can put 2TB M.2 SATA SSDs, and
  • 2x 3.5 SATA bays, where I can put 4TB SATA SSDs.
Background:
I want to find a balance for myself between speed and capacity and risk of failure.
I hesitate between setting up a Raid 0 with 2x2TB and a Raid 0 2x4TB, for fast and big storage, with daily backup on an external USB drive, and the time to rebuild it if it fails. It would have the convenience to have everything on the NAS.

Or shall I content myself with less storage and speed with Raid 5 with 4x2TB, still with backup, and lose some comfort on continue keeping less used files on external USB hard drives.

Or go for 2TB Raid 1, and 4 TB Raid 1.

I have 10 GbE in the house.

Otherwise, thanks for the calculator, it is very useful. I found it via the YouTube video with Linus "Jellyfish Fryer".
 

Evan

Well-Known Member
Jan 6, 2016
3,280
566
113
The calculate is just as valid for SSD as HDD.
You enter all the relevant parameters in the same way and the output would be just as valid.

since it’s 4 SATA SSD even though 2 would be m.2 and 2 are 2.5” in a 3.5” slot I wound raid5 the 4 drives and enjoy. Anyway the limits become the sata interface so even if performance of the drives are not identical will be fine.

you can also get 3.84tb or 4tb SSD in both formats ;)
 

olee

New Member
Apr 6, 2021
2
0
1
I see, I was hoping SSDs would do a bit better in the calculator...

I didn't find any M.2 Sata SSD in 4TB capacity, only 2TB (I can find NVMe, but that doesn't fit in the NAS).

Can you suggest?

The simulation works better with Raid 5 with 4x4TB drives, that would be a good compromise for capacity, speed and risk.
 

Evan

Well-Known Member
Jan 6, 2016
3,280
566
113
Opps, right, alll the ones that I had in mind were 2TB in 2280 but 4TB only in 2.5”
yes NVMe can even get 8tb 2280 now. Sorry about that my mind had Samsung 860 Evo etc in mind.

the calculations between disks and ssd are the same, but ssd is in the end better due to;
- lower failures rates
- better non recoverable error rates usually
- longer MTBF mostly
- much much faster rebuild

biggest risk in ssd for raid is probably things like the 40k hour bugs and so on.