The Week In Storage: Please Delete Your Data Properly; Storage Game Of Thrones Unfolds

Status
Not open for further replies.

jimmysmitty

Polypheme
Moderator
Dispose Of Your Data Correctly - Deleting And Quick Format Do NOT Count
I like to do what is called a Guttman which is 33 passes of writing 0s to the entire drive. Takes forever but you can't find anything after that.

It is better than even government standard, which is 7 passes.
 

Haravikk

Honorable
Sep 14, 2013
317
0
10,790
1
I like to do what is called a Guttman which is 33 passes of writing 0s to the entire drive. Takes forever but you can't find anything after that.
This seems wasteful; encryption is the best technique IMO, as it's something that's worth doing anyway, and all you have to do is destroy the password-protected key (if it's even on the same device, if you use Bitlocker it may not be). Of course you could use your 33 pass erase on the key if you like, and it'll be significantly faster than for the full drive.

Once the key is gone the data may as well be unrecoverable. If you require more protection than that then the next step is destroying the drive really, which will likewise be much faster.
 

wussupi83

Distinguished
Oct 29, 2009
41
0
18,530
0
Wow Tom's, Seagate sales are down and you avoid every possible way of saying it had anything to do with the fact that their drives were failing at rates much higher than their competitors in Backblazes environment. The fact that you avoided this MAJOR detail and the fact that just days after that Backblaze data was released you ran a huge article praising Seagate up and down after a visit to their factory is pretty good evidence, in my worthless opinion, that there was some advertising dollars thrown your way to sweep it under the rug (not to mention the background of your site right now is advertising a Seagate drive). I suppose you're just a business at the end of the day and one can't take it too personal but seriously, when it gets in the way of journalistic integrity it's hard to look the other way.
 

PaulAlcorn

Senior Editor
Editor
Feb 24, 2015
678
25
5,010
0



If we were beholden to Seagate we would not be mentioning the fact that Seagate is having severe issues, a fact that I have covered in many dedicated stand-alone articles as the news progressed. You will not find articles from a reputable source that draw a direct line between Seagate's current situation and the Backblaze tests, as the real reason is the declining PC market, as we cite in every piece (and as many other sites report).

I have criticized the Backblaze findings at another site that I worked for, long long before I ever came to Tom's. If you wish to find said articles, a simple Google will suffice. I am not influenced in any way, shape, or form by advertising dollars. The site where I published the previous article exposing the incorrect Backblaze test methodology has never received any advertising money from Seagate, to this day (to my knowledge).

Tom's Hardware has never attempted to shape/alter my coverage, and frankly, I do not appreciate the suggestion that it has.

I cannot speak to the article about the factory visit, but if it is the one I am thinking of, that was published years after the initial Backblaze "reports", which began several years ago.
 

jimmysmitty

Polypheme
Moderator


100% this. The Backblaze data is useless since there are so many unaccounted variables.
 

wussupi83

Distinguished
Oct 29, 2009
41
0
18,530
0



Paul,

I appreciate your response but if the avoiding the mentioning of the Backblaze data was not deliberate than it was very poor oversight on your part considering that it was major headlines on your site less than 6 months ago and also on many other major tech sites; Therefore, I advise you to perform more research before publishing such data while neglecting a major data point. Also the article was published right after the latest scandal headlines, you may have ran one years ago but that is not the one I am referring to.

Source:
Tom's Article On Seagate Class Action Published Feb 2, 2016: http://www.tomshardware.com/news/seagate-hdd-failure-lawsuit-3tb,31118.html

Tom's Article On "How Seagate Tests It's Hard Drives" Factory Tour
Feb 9, 2016: http://www.tomshardware.com/reviews/how-seagate-tests-hard-drives,4408.html
 

wussupi83

Distinguished
Oct 29, 2009
41
0
18,530
0


Anyone who actually has taken the time to read the data on Backblazes website about their storage pods and hard drive usage knows that answer is a complete cough out and you're plain wrong.

First of all, Backblaze ran enterprise drives in the exact same configuration as these consumer drives and there was absolutely not a shred of difference in failure rates for the 3 years the study ran vs their consumer counterparts. Seagate's drives experienced high failure rates within the same 3 years.

Second of all, Seagate's competitor's consumer drives were performing statistically significantly better. Therefore, even if you say it had to do with the environment, Seagate's drive quality was still way below their competitors and deserved the bad press.
 

PaulAlcorn

Senior Editor
Editor
Feb 24, 2015
678
25
5,010
0



Yes, I wrote the piece on the class action lawsuit that the firm was attempting to drum up. Note, there has been no news on their attempts since (ie., it failed).

It was not poor oversight to not mention it. The vast majority of Seagate's sales are in OEM contracts into the consumer notebook and PC market. These buyers are not concerned with the Backblaze findings, as they are all too aware that the "findings" are inherently flawed.

Perhaps you should go read some of the latest updates, which now have WD in the lead position as the worst for 2016 - in fact, the WD failure rate is double that of Seagate. Don't you think I would have jumped on that and written that for the world to see if I was attempting to glorify Seagate?

I cannot speak to the release of the factory tour article, as I was not involved in that project and had no knowledge of it. However, I know that piece was likely not done in 7 days (visit, writeup, edit). It would be hard for our edit team to schedule the piece around a time that does not coincide with a Backblaze update, if you keep track of them then you would be aware that they occur frequently.

The outline of the Seagate issues in this piece was brief, as it is merely an update to an ongoing issue that we are covering. If you click the link I provided in the text it will take you to an article, written by myself, that goes into the full outline of just what market forces are causing Seagate to plunge, and why. You will also note that I did not dive in to the Intel reasons either, as I provided a link to further analysis I had already written. I did break down Micron a bit, but merely because I had to, as I had not covered the issue before.

Backblaze certainly hurt Seagate's reputation badly, there is no doubt of that. However, if you go and read market breakdowns that show the decline in units shipped, as I do on a quarterly basis, then you will see that the WD and Seagate sales and units shipped are declining in lockstep (at nearly the same rate). If Backblaze had caused the Seagate plunge, dont you think it would have not affected WD, in fact, would WD not have TAKEN those sales? Well, neither happened, because it just is not a factor in the larger scheme of things.

As much as I am loathe to link another site, please go to the second chart on this article to see the progression (and first chart on page 2). This type of data, and cold hard facts, are what I base my analysis upon, not simply because someone wishes that I mention something. http://www.anandtech.com/show/10315/market-views-hdd-shipments-down-q1-2016

Please note that article does not mention Backblaze once. Also, note that if you read ANY analyst, or perhaps sourced paid analyst reports (some of which we have access to) they do not mention Backblaze.

You claim that my non-mention is due to either "oversight" or "poor research". In fact, it is due to actual reasearch that goes beyond reading articles on the flawed Backblaze test from writers that wouldn't know the difference between a SATA or SAS HDD if I smacked them upside the head with it.
 

PaulAlcorn

Senior Editor
Editor
Feb 24, 2015
678
25
5,010
0




Sorry, but every professional that I know that has examined the Backblaze methodology agrees it is flawed. If you read the current updates you will note that WD is failing at double the rate of Seagate in 2016.

I wouldn't put too much stock in that, either, as you have no idea which chassis they are using, the data is worthless. They had several prior revisions that are inherently flawed and resulted in excessive failures. In fact, they provide well-documented updates on that - YET they continue to use the old chassis.
 

jimmysmitty

Polypheme
Moderator


It is not. The pods have been shown to be flawed, i.e. they do not properly secure the HDDs via the screws like the HDDs are designed to.

Second they are using consumer drives that are not rated for 24x7 use in a high heat/extreme environment.

We have multiple servers at work with plenty of HDDs. Want to know how many of them are consumer rated HDDs? None. Want to know the top failure brand? All of them. We have had Toshibas, Seagates, WDs, Samsungs all of them fail. None fail more than the others and all are rated as enterprise 24x7 drives because they have to be.

https://www.backblaze.com/blog/hard-drive-reliability-stats-q1-2016/

Look at that. Seagate now (as of 2016) has the second lowest failure rate.

You know what is even better? The lowest failure rate HGST drive, HMS5C4040BLE640, is a enterprise grade HDD:

http://www.newegg.com/Product/Product.aspx?Item=N82E16822145981

The Lowest failing 4TB Seagate HDD, ST4000DM000, is a consumer grade desktop HDD not rated for 24x7 server/enterprise use:

http://www.newegg.com/Product/Product.aspx?Item=N82E16822178338

ALL of the Seagate HDDs used are consumer grade desktop drives, they are all Barracudas not Constellations which are the enterprise grade drives.

Now given that kind of information, do you think it makes a difference when these drives, one being 24x7 rated the other not, in the terms of failure rates? Do you think it is possible that the testing methodology of Backblaze is still fair when comparing enterprise grand HDDs to non enterprise grade HDDs?
 

targetdrone

Honorable
Mar 26, 2012
300
0
10,780
0


Are you SURE that overwrites works on SSDs?
 

jimmysmitty

Polypheme
Moderator


Yes it actually does however it will also lower the life of the drive more than is needed.
 

PaulAlcorn

Senior Editor
Editor
Feb 24, 2015
678
25
5,010
0


I think that you are referring to the FTL (Flash Translation Layer) not providing 100 percent assurance that data is deleted in its entirety, and there is some merit to that. The FTL creates difficulties assuring that files are erased, but this is usually when speaking in terms of single files or folders. As there is an FTL and there isn't a 1 to 1 mapping there is a chance that the actual surface of the NAND is not overwritten, and that the FTL has merely been modified while the data is still there. For government agencies and businesses this is a real concern, and there is a special interest group that has been formed to attempt to rectify this issue in a standardized method. They have actually briefed us, Micron is one of the founding members.

In the case of and just nuking the whole drive (instead of a few files) of consumer-level data that isn't going to be subject to NSA-like extraction techniques I feel that a full format followed by a 2X overwrite is sufficient. Of course, as noted, there is the option to go full-nutso with 35 passes. There is a small chance that there could be some data that could be extracted after a 2X pass, but it will likely just be a few interspersed 1's and 0's as opposed to an actual file. However, it would require a very professional, and likely state-funded entity, to extract that mostly-useless data. For normal users that isn't going to be a problem, if their data was worthy of that sort of scrutiny they are likely well aware of it (at least hopefully) and taking the necessary precautions, such as smashing it with a hammer.

 
Status
Not open for further replies.

ASK THE COMMUNITY

TRENDING THREADS