Show HN: BackBlaze Hard Drive Data Failure Viewer(drives.fail) |
Show HN: BackBlaze Hard Drive Data Failure Viewer(drives.fail) |
I'm a huge fan of the BackBlaze Drive Failure Reports[0] and use them exclusively when shopping for drives to use for my NAS. This project was an excuse to dive into F# and how it's used for Web Devlopment (SAFE Stack[1]) to provide me with a quick way to view the drive reports and search the drives on Amazon.
It's very much a MVP. I'd like to improve on it when I find more time. Ideally, the next steps would be for me to actually level up my front end game to have something much prettier to look at. As well as provide common functionality that you'd expect from something like this. (cough Column sorting cough)
Hope some of you find this useful or interesting.
Thanks!
And some fields need to have a "<" or ">" factor -- for example, I don't want to see any results that have less than 100 drives in the fleet. At those small sample sizes, they're still just testing them and not really buying them for real work.
I just updated the UI and now have a better filtering panel and also have column sorting.
Whoops!
Weird.
Now the only remaining weird thing I've noticed is that when I select a small handful of drive capacities in the filter selection (say, 14TB and larger), I get a nice small list in the main display. But when I try to sort that small list by any of the columns, it throws away the small list of drives and goes back to the full list which is then sorted by that column.
Not sure what's going on there.
Now I can pick the large capacities that I'm interested in (14TB and larger), sort by DriveDays, and then look at annualized failure rates, and easily notice that the Toshiba MG07ACA14TA has by far the most drive days of those models (over 13m) and it has a failure rate below 1%.
Or, sort by failure rates and see that the WD WUH721414ALE6L4 has over 1m drive days and a failure rate of less than 0.5%.
I think either of those drives might fit my requirements. I'm a big fan of the Ultrastar models, and I might even forgive WD for buying the company, so I might go that direction. OTOH, Fujitsu also has an excellent reputation.
So, now I have to make a decision between these two choices. It's good place to be!
Now, in doing this process, I noticed that you are just searching on the Amazon site for the product model number. Can I convince you to find the appropriate Amazon created ASIN for each model number and plug that into your database, and then search by the ASIN instead? An ASIN is like an ISBN for books, and for books I think the ISBN is the ASIN.
I think the model number search would be okay as a backup, but there's always a lot of hits when you search by something like model number, and you can never be sure that you've got the right one. But if you search by ASIN, you have a very high confidence that you're getting exactly what you were searching for and nothing else.
Of course, you also want to plug in your referral code as well, but I think you definitely want to search by ASIN instead of the product model number.
Thanks!
I honestly did the referral link to the model search because prior to doing so I was trying to price check, but don't have a programmatic way of doing that without access to the Associates API.
I did it manually at first, but then saw price drifts within a couple of days. On top of that, some models on the list were out of production and I was hesitant to link pricing to models that were likely to be refurbished, old stock, or listed as used.
I wasn't aware of ASIN, I do like the approach better, so I can work on getting that switched over when I get free time over the weekend.
Thanks again for your feedback!
In particular, I know nothing about the API, although I did learn a few years ago about what an ASIN is. That's about as far as I can go, with regards to any kind of automated processing related to the consumer side.
I look forward to future enhancements to this tool! And let me know if you need any further testing.