27 replies to this topic

[fredp1]

Abosolutly awesome!!!!!!

[axoid]

I found this the most informative show yet. Looking forward the follow up.

[timekills]

Wow. Good stuff. Even some controversy on the use of certain "garbage" drives

[jmwills]

A little deep for a podcast but I did pickup links on a few good tools that I did not previously know about.

[mattd390]

Anyone want to buy some used green drives?? jk

[jmwills]

_{How much?}

[pcdoc]

Thanks guys for the feedback. Look forward to the second part as well.

[Joe_Miner]

Where do I get my CPE certificate signed for this podcast

Great show -- very instructive!

[Dude]

I've been listening to your show since podcast one and I must say I did enjoy this show a lot.
I do have a question for when Scott comes back on and it refers to the Green drive statement, I was a bit surprised by that comment.
I would have thought the technology in a green drive was superior to get the power savings, so would like to hear a little more on his experience with green drives.
I have Six at home three in my server and three I use for offsite back-up and whilst I've probably only had them a year I have never experienced any problems.

[pcdoc]

I've been listening to your show since podcast one and I must say I did enjoy this show a lot.
I do have a question for when Scott comes back on and it refers to the Green drive statement, I was a bit surprised by that comment.
I would have thought the technology in a green drive was superior to get the power savings, so would like to hear a little more on his experience with green drives.
I have Six at home three in my server and three I use for offsite back-up and whilst I've probably only had them a year I have never experienced any problems.

Thanks very much for the comments and for the support. I share your sentiments and I do not want to speak for Scott, however we have to remember what he is used to seeing as well as the fact that most of his customers are businesses. He also stated critical data which is different for many of us. I have no doubt that RE or other enterprise drives are better built but at twice the price they may not be the best match for the home user who is backing up or using it for backup. I think if I was going to run a business web server, I would probable use RE drives. I think everyone has to decide what works for them. For me, I have had excellent luck with the green drives and I have several levels of redundancy including cloud backup so I am not concerned. Never the less, he was a wealth of information and even though I am continue to buy green drives, I learned a heck of allot. Thanks again for your inputs and lets see what happens on the next part.

[no-control]

Personally I would challenge his statement to provide more clarity(facts) and less opinion. While I've known green drives are the low end as far as performance/reliability vs more expensive business class drives. I would have to interject context into his statement. At what point does a green drive provide a higher failure rate over RE drives? What is the cost:failure ratio? Etc...

His statements regarding how to set up a server are completely in line with what BYOB has preached (from myself anyway) since RAID became a major topic in the forums about a year ago. And my servers follow suit. Critical data on a RAID 1. <- theres a big period right there for emphasis. RAID 5 is a performance RAID not backup. I don't even agree with the backing up of a RAID 5 with another RAID 5 even on another server. Which brand/model of drives doesn't matter.

Now for business use there was one very short statement Scott made it was super short and I you may not realized the significance. For business critical applications use enterprise class SAS drives. Any enterprise grade server you buy from an OEM will contain SAS drives not SATA drives. I only use SAS in my business applications, the only exception is SOHO clients. Where its possible to use a cheaper drives in a RAID 10 for less money than a RAID1 of SAS. Even then I would say 50% spring for the more expensive SAS drives.

While Scott may think Green drives like the WD GP are "junk" I personally think they have a place in the market. Home servers are that market.

[timekills]

While we are on the debate train, I'd like some further clarification on the effects of heat on hard drives. This has been discussed on this site, but the summary of what Google found from their 100,000 hard drive failure study was that heat - mostly - had no effect on failure rate and in fact the higher average operating temperature drives had a lower than average failure rate. I'm curious why he (albeit briefly) correlated heat with failure.

http://static.google...sk_failures.pdf

Regarding SAS vs SATA failures, trend analysis shows a 1.9% SCSI failure rate, a 1.9% SAS and just over 2% for SATA. Not sure that is enough to warrant the extreme cost difference in drives. As above in Google's study, amount of use was not a direct correlation to failure rate, so although in all likelihood SAS drives see more use than SATA drives (given their usage scenarios) that does not appear to effect affect the failure rates.

(edited for Grammar school failure.)

[pcdoc]

Sound like we are going to have some QA on the next round.

[ikon]

OMG, OMG, OMG, what a fantastic podcast. The best BYOB so far. In fact, the best single podcast I've ever heard! Talk about nutrient dense. Scott must use a Tardis to package his presentations, because there is way more information on the inside than it might appear from the outside.

I was particularly interested to hear that DBAN is not truly the best way to nuke a drive before sending it to Computers For Schools or some other organisation. I will be informing my colleagues at work. Scott also reinforced my decision to no longer buy green drives, and that using SpinRite for maintenance like I do is valid.

I am soooo looking forward to the next chapter.

I have a suggestion: how about asking Scott to participate in a series of podcasts over the next few months? I think there is a ton more info to be derived.

Thanks again for having Scott on the show.

[Dude]

I've kept the podcast I'm going to listen to it again in a couple of weeks to see if there are other pieces I missed.
I was interested in his comments on the SSD drives and how the life was reduced subject to the technology being used, wasn't aware of that.
Like most of you I couldn't have a PC without a SSD anymore, I'm using a OCZ Vertex 3 for my C drive and have 2 Vertex 2 in Raid 0 for video editing, never going back to Spindles, except for back-up.

[Greg Welch]

Till they get him back on, 15 hours of scott multon

http://www.podnutz.com/myharddrivedied

[fblittle]

Scott Moulton is an awsome Guest. I can't wait for his next time on BYOB. I want to ask if the hosts can ask him what the drawbacks of using USB3 instead eSATA on an external drive box, like DROBO or Synology etc. . . for RAID or just JBOD etc. . .

In the past he has discouraged the use of USB to connect drives because of Command Queing and SMART.

"My Hard Drive Died" podcast was my favorite podcast (next to BYOB of course) in the past, so it is very good to hear him again on this podcast. If you could make him a semi regular guest he can enlighten us on many storage topics. He is very good at all aspects of RAID systems. One more thing, Ask him about RAID Cards, Brands, types etc . . .

Edited by fblittle, 17 January 2012 - 04:52 AM.

[N_Nescio]

TrueCrypt is open source. The source is freely available and has been examined by the community. KeePass is also open source, and has also been examined. The same goes for Linux and dm-crypt/LUKS. While nothing is 100%, and he is correct in that, Scott's comments came across as "well, the gov can probably do it". I don't buy it. We know of two, national level attacks against encryption, that failed. One was the Nicodemo Scarfo case, where the FBI had to resort to a keylogger called 'Magic Lantern' to get his PGP passphrase. (Interestingly, in that case, Symantec went on record as stating that they *wouldn't* detect Magic Lantern...nice huh?) The other is a recent case of Brazilian banker Daniel Dantas. The FBI had his hard drives for close to two years and could not get in. So I really wish Scott would have said something like "Unless you've examined the source yourself, you're relying on the community to indicate that there is no back door with open source software. In the two cases we know about, there wasn't". Great show though, thanks.

N

Edited by N_Nescio, 17 January 2012 - 01:54 PM.

[jmwills]

DOD has some amazing tools..............that's all I can say.

[N_Nescio]

DoD, yeah, I was with them for a while (SOCOM unit, not crypto related), and contracted with them in Iraq in '03-'04. Doesn't change the math. Assuming no backdoors or side channel attacks, they have to go after the algorithim (and which one? AES, Twofish, Serpent, MARS? They've cracked them all? Cascades too?), the key space (hashed password/phrase), or the password/phrase itself. Scott actually talked about that in his first few minutes, saying don't bother. His reference to TLAs was in regards to back doors. I'm up for the very remote possibility of a backdoor in TC, more so in PGP, and very likely in BitLocker. But GPG? Linux? LOL. No way a backdoor in dm-crypt/LUKS survives undetected. So these 'amazing tools' will have to do something that every single cryptographer on the planet says is, for all practical purposes, based on every piece of scientific data, impossible: brute force a 256bit key in less time than the life of the sun. Remember too, DoD had unencrypted Predator feeds, let the UAV fleet get infected, and handed the keys to the castle to Manning. I also believe DoD said there were WMD in Iraq...but they did get OBL...after 10 years and letting him escape from Tora Bora, but I digress. Going by facts, the Dantes case shows national level LEOs couldn't break it. Reading the Antisec dump of the LE forensics mailing list shows local and state level definitely can't. But differences of opinion are good, thanks for the reply.

N