Plotting HHD wear and tear?

Now that quick plotting is achievable for most everyone with GPU plotting,

What are people’s thoughts about the wear and tear that is put on a HHD every time it is fully re-plotted?

Is it better to plot a HHD as little as possible?

Or,

Don’t worry about it and re-plot as many times as you need too, to keep up with the latest plots?

What is the desirable number of times you want to fully re-plot a HHD?

Or is fully re-plotting a HHD not to significant in the life of that HHD?

Just wondering in what others thoughts are on this issue?

Thanks!

:point_up_2: this, unless you plan on replotting every week :sweat_smile:

writing data to a hdd is really not a considering factor for life-span in our use case.

Temperature between 30-40C (no higher than 45C and no lower than 25C) and low vibration/shocks are important factors. Also number of power-on cycles, but with chia farming this is super low anyway.

2 Likes

Thanks!

So this helps with planning my overall strategy.

I’m going to plot to max compression I’m able too and stick with K32 size plots. When the plot filter reduction comes into effect or if I get better equipment, I’ll re-plot to adjust to stay efficient and keep my HHD I/O’s low.

Anyone simplifying this answer is misguiding you. First, there are new and used drives. Second, there are internal and external. Third, there are consumer, NAS and enterprise grade. Fourth, there are SAS and SATA. Fifth, is your temperatures of the drives and whether you monitor them or not.

If you have no idea what your starting point is, then to just say it’s all good? Like, oh yes, I can easily make my ROI on the drives? Really? You sure about that? All you will hear is conjecture. There is no exact Chia farming scenario that is tried, tested and true. No track record. Have you bothered to check the health of your existing drives? If not, then you have no clue where you are at on this. It’s not a simple yes/no although many Chia apologists want you to believe your HDD is going to last to ROI and last long enough past ROI to say it was worth running them for 24/7 for YEARS.

Aside from that have a great day!

It is as @Voodoo wrote.

If you are running a data center, where your hard drives are being hammered 24/7 with reads and writes, then a consumer grade vs an enterprise grade drive will be a prime factor. Such is not the case with Chia.

For writing contiguous, large files several times, such as Chia does when plotting or re-plotting… it is a walk in the park for any drives.

Improper cooling, dirty power, and constant power-ons will kill any class of drive sooner than the grade of the drive.

So keep your drives cool (a light breeze will do the job) and use a UPS (uninterruptible power supply).

And if you purchase a drive that vibrates more than average, I would return it.

Cough cough, FF reduces I/O by like 40%. Though FF Gigahorse may not do so as much yet as I think we switched back to the old C+ PoST from our Golang PoST to get Gigahorse working. Will need to test it.

As said, drive writes are are a non-issue…in fact the opposite is true - stale writes will eventually cause read errors…so expect to replot every ten years or so lolz

dont worry about you HDD
I just leave it here