My itec professor is a genius

Page 10 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Sep 29, 2004
18,656
68
91
Originally posted by: theprodigalrebel
I've noticed that most non-gamers don't bother keeping up with what's new in the tech industry.
Gamers are always on the lookout for new GPU/CPUs, drivers, hacks, patches, tweaks etc.

Those working in IT obviously know what's new in computing (but not necessarily in audio/video)
Other people just...don't care much, I think.


best Buy doesn't stock those large capacity memory sticks ... atleast I didn't notice them.
 

TreyRandom

Diamond Member
Jun 29, 2001
3,346
0
76
Originally posted by: IHateMyJob2004
Originally posted by: theprodigalrebel
I've noticed that most non-gamers don't bother keeping up with what's new in the tech industry.
Gamers are always on the lookout for new GPU/CPUs, drivers, hacks, patches, tweaks etc.

Those working in IT obviously know what's new in computing (but not necessarily in audio/video)
Other people just...don't care much, I think.


best Buy doesn't stock those large capacity memory sticks ... atleast I didn't notice them.

Largest they've got is 4 GB. CompUSA's got larger ones, supposedly.
 

m1ldslide1

Platinum Member
Feb 20, 2006
2,321
0
0
Originally posted by: TreyRandom
Originally posted by: randay
Originally posted by: TreyRandom
Because RAID 0 contains neither redundancy nor fault-tolerance. An array of disks doesn't create redundancy unless the same information is on more than one disk. RAID 0 is for speed of data access, not for redundancy or fault tolerance.

So... I couldnt help but notice all those certs listed in your sig...
Yep. Been working in IT a while. Plus, I write certification training books and practice exams.

Maybe you could speed things up a bit for the new ONT exam (CCNP revision). Looks like I have to wait until June for Cisco Press to allow me to study... Impressive list of certs, btw.
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: randay
Originally posted by: TreyRandom
Originally posted by: randay
Originally posted by: TreyRandom
Originally posted by: randay
Originally posted by: TreyRandom
Originally posted by: randay
Originally posted by: TreyRandom
Your first paragraph is incorrect. If one disk fails in a RAID0 array, the remaining disk doesn't hold any usable data.

Are you trying to say that when a disk in a RAID array fails, that the data on the other disk automagically deletes itself or becomes corrupted? It is very possible to pull usable data off of said drive.

Not on a failed RAID 0 array, you cannot. The data is striped, meaning a little on drive A, a little on drive B, a little on drive C, a little on drive A, a little on drive B, a little on drive C. One drive fails, all data is unrecoverable. You must restore from backup. You can't recover the stripes of data - they are not able to be reassembled into anything meaningful. That's why you don't use RAID 0 unless you want to sacrifice reliability for sheer disk access speed.

look up block size, or stripe size. If the file is small enough to fit into a single block, then the entire file exists in that block which means a single drive.

I realize that. But how many files will you be able to recover? Very few, if any. Sure, you can get a whole bunch of <32K or <16K or <64K files... if they're not split... and you can reconstruct the allocation table so you can determine what those files are... but of what use is that, really?

Does that change the facts or are you just picking at straws? Can you or can you not retrieve data off of a functioning disk from a failed RAID0 array?

Oh, that's rich - the $40K sysadmin who makes up his own definition of redundancy is saying I'm grasping at straws. :D

Are you actually suggesting that that's a viable recovery strategy you implement in your networks? For all intents and purposes, the data is lost. Any data you recover is likely to be of very little use.

Answer the previous question posed by Spidey: Have you ever lost a drive in any kind of striped or RAID storage situation? Or is what you are spewing learned from books and the intarweb? I've lost drives in RAID arrays, but I've never lost drives in a RAID 0 array because only a complete moron would use them in a production environment.

Wow you come back for like 2 posts and you wanna go down that road now? Fine I'll oblige.

nah, forget it. I wont stoop to your levels. This thread has degenerated and the quality of your logic has also. There is no more point in trying to fend off the endless trolls and flames. Goodbye.

AND HE RUNS AWAY LIkE LITTLE GIRL..:laugh:


In this case, randay, although professing a deep intellectual understanding of the concept of RAID, does not indeed know the intricacies of the technology and the terminology associated with it. Randay fails to understand that a RAID "array" is defined by the management features of special disk controllers in junction with a series of inexpensive or interchangeable disks, and not one or the other. RAID controllers and RAID0/simple striping are synonymous due to the inclusion of simple striping functionality on most controllers. That however, is a product of customer demand and not a product of industry standards. He also professes to be intimately familiar with the term redundancy when he cannot even apply the general meaning to another setting (cars and trucks) correctly.

RAID, in an of itself, defines a method of data manipulation among a series of disks for some added benefit. The different RAID levels describe different methods of data manipulation that are used to provide INCREASED FAULT TOLERANCE to a system that uses disks that are not as reliable as enterprise-class disks (such as SAS), or to add additional fault tolerance to highly reliable systems. Redundancy is another characteristic of RAID arrays. By definition, THE RAID CONTROLLER MIRRORS THE SAME DATA TRANSFERS TO MULTIPLE DISKS/volumes. This serves no immediate purpose, and actually makes for an inefficient use of resources. That said, the redundancy in this system serves as a means to provide fault tolerance. Since the data is stored in a redundant fashion, faults can be tolerated since the raid array can use parity bits to reconstruct the array, or it's most basic form (RAID 1), simply switch is focus to the disk that is still operating normally after hardware failure and resume normal data transfers until the original disk is fixed. Depending on the RAID configuration used, the level of dedicated redundancy can be as much as 100% (RAID 1) allowing for not only simple fault tolerance but high-availability as well (systems stay up and don't skip a beat). In the end, RAID technology addresses the integrity and availability of data seen past the RAID controller, albeit in different ways depending on the configuration.


On a more serious note, I must admit that this does make for an interesting case study on how little the loudest people in a group actually know.

+5 BS points :cool:



<----by the way, he realizes that he too is one loud mofo and probably knows nothing but is ALWAYS ALWAYS ALWAYS willing to learn:D



 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: m1ldslide1
Originally posted by: TreyRandom
Originally posted by: randay
Originally posted by: TreyRandom
Because RAID 0 contains neither redundancy nor fault-tolerance. An array of disks doesn't create redundancy unless the same information is on more than one disk. RAID 0 is for speed of data access, not for redundancy or fault tolerance.

So... I couldnt help but notice all those certs listed in your sig...
Yep. Been working in IT a while. Plus, I write certification training books and practice exams.

Maybe you could speed things up a bit for the new ONT exam (CCNP revision). Looks like I have to wait until June for Cisco Press to allow me to study... Impressive list of certs, btw.

yeah...he has a crazy list.
 

m1ldslide1

Platinum Member
Feb 20, 2006
2,321
0
0
Originally posted by: randay
Originally posted by: TreyRandom
Originally posted by: randay
If one disk fails, the other disk still works.
If you have only one disk, if one disk fails, there are no other disks that still work.

from wikipedia
The term RAID was first defined by David A. Patterson, Garth A. Gibson and Randy Katz at the University of California, Berkeley in 1987.[2] They studied the possibility of using two or more disks to appear as a single device to the host system and published a paper: "A case for Redundant Arrays of Inexpensive Disks (RAID)" in June 1988 at the SIGMOD conference.

In RAID 0, if one disk fails, what good is it if you have one disk that works? The entire array is gone if one disk fails.


You can use it and a new drive to build a new array.

Listen, the original name was "Redundant Array of Inexpensive Disks". Think about that for a bit, and if you still feel the way you do about it then forget I said anything. Theres no point in trying to argue it any further.

LOL ;)

RAID0 = AID. No parity bits in raid0 man. I get that you're trying to say that since there's two disks, they're redundant, but that's not the concept of redundancy in the IT world, across ANY discipline. Redundancy means that one provides the same function as the other. That is NOT the case with raid0 - data is striped across both, and if you lose a disk you lose everything. Not redundant, even though there are two disks.

Edit: Like this post, redundant. If I had read the last 3 pages of the thread, I'd see that this post is technically unnecessary but in the event that the last 100 people explaining this topic were to have their posts deleted, my post would still explain it.
 

TreyRandom

Diamond Member
Jun 29, 2001
3,346
0
76
Originally posted by: m1ldslide1
Originally posted by: TreyRandom
Originally posted by: randay
Originally posted by: TreyRandom
Because RAID 0 contains neither redundancy nor fault-tolerance. An array of disks doesn't create redundancy unless the same information is on more than one disk. RAID 0 is for speed of data access, not for redundancy or fault tolerance.

So... I couldnt help but notice all those certs listed in your sig...
Yep. Been working in IT a while. Plus, I write certification training books and practice exams.

Maybe you could speed things up a bit for the new ONT exam (CCNP revision). Looks like I have to wait until June for Cisco Press to allow me to study... Impressive list of certs, btw.

I left the practice exam company I used to write for and started my own company with another ex-writer. Problem is, we've gotta start from scratch... meaning, we've gotta get the more popular exams done first (including the CCNA). :( Maybe we can get the CCNP stuff done by the end of this year or beginning of next, which will probably be way too late to help you.

As far as books are concerned, I recommend Cisco Press, which is what you're waiting on. And as you probably already know, Sybex is good for an alternate explanation of topics.

Thanks for the compliment. :) Gotta get the certifications in order to write about them, ya know? Plus, I know what you guys are looking for... I'm currently a network admin - just a tech like most of you guys... at least until the practice exam company takes off.
 

randay

Lifer
May 30, 2006
11,018
216
106
Originally posted by: Goosemaster

AND HE RUNS AWAY LIkE LITTLE GIRL..:laugh:


In this case, randay, although professing a deep intellectual understanding of the concept of RAID, does not indeed know the intricacies of the technology and the terminology associated with it. Randay fails to understand that a RAID "array" is defined by the management features of special disk controllers in junction with a series of inexpensive or interchangeable disks, and not one or the other. RAID controllers and RAID0/simple striping are synonymous due to the inclusion of simple striping functionality on most controllers. That however, is a product of customer demand and not a product of industry standards. He also professes to be intimately familiar with the term redundancy when he cannot even apply the general meaning to another setting (cars and trucks) correctly.

The point is, RAID, in an of itself, defines a method of data manipulation among a series of disks for some added benefit, and Randay is just not as knowledgeable as he professes to be. The different RAID levels describe different methods of data manipulation that are used to provide INCREASED FAULT TOLERANCE to a system that uses disks that are not as reliable as enterprise-class disks (such as SAS), or to add additional fault tolerance to highly reliable systems. Redundancy is another characteristic of RAID arrays. By definition, THE RAID CONTROLLER MIRRORS THE SAME DATA TRANSFERS TO MULTIPLE DISKS/volumes. This servers no immediate purpose, and actually makes for an inefficient use of resources. That said, this redundancy provides copies of data that can be used for data recovery should a physical disk fail. In a system that can tolerate physical or software faults, redundancy focuses on a method of data division and mirroring that provides


While all of these configurations use multiple drives for fault tolerance, asfor data integrIn the end, RAID technology addresses the integrity and availability of data seen past the RAID controller, albeit in different ways.


On a more serious note, I must admit that this does make for an interesting case study on how little the loudest people in a group actually know.


<by the way, he realizes that he too is one loud mofo and probably knows nothing but is ALWAYS ALWAYS ALWAYS willing to learn:D



+5 BS point :cool:

What happened to your little wiki reference? didnt work out anymore after I pointed out how you were wrong?
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: randay
Originally posted by: Goosemaster

AND HE RUNS AWAY LIkE LITTLE GIRL..:laugh:


In this case, randay, although professing a deep intellectual understanding of the concept of RAID, does not indeed know the intricacies of the technology and the terminology associated with it. Randay fails to understand that a RAID "array" is defined by the management features of special disk controllers in junction with a series of inexpensive or interchangeable disks, and not one or the other. RAID controllers and RAID0/simple striping are synonymous due to the inclusion of simple striping functionality on most controllers. That however, is a product of customer demand and not a product of industry standards. He also professes to be intimately familiar with the term redundancy when he cannot even apply the general meaning to another setting (cars and trucks) correctly.

The point is, RAID, in an of itself, defines a method of data manipulation among a series of disks for some added benefit, and Randay is just not as knowledgeable as he professes to be. The different RAID levels describe different methods of data manipulation that are used to provide INCREASED FAULT TOLERANCE to a system that uses disks that are not as reliable as enterprise-class disks (such as SAS), or to add additional fault tolerance to highly reliable systems. Redundancy is another characteristic of RAID arrays. By definition, THE RAID CONTROLLER MIRRORS THE SAME DATA TRANSFERS TO MULTIPLE DISKS/volumes. This servers no immediate purpose, and actually makes for an inefficient use of resources. That said, this redundancy provides copies of data that can be used for data recovery should a physical disk fail. In a system that can tolerate physical or software faults, redundancy focuses on a method of data division and mirroring that provides


While all of these configurations use multiple drives for fault tolerance, asfor data integrIn the end, RAID technology addresses the integrity and availability of data seen past the RAID controller, albeit in different ways.


On a more serious note, I must admit that this does make for an interesting case study on how little the loudest people in a group actually know.


<by the way, he realizes that he too is one loud mofo and probably knows nothing but is ALWAYS ALWAYS ALWAYS willing to learn:D



+5 BS point :cool:

What happened to your little wiki reference? didnt work out anymore after I pointed out how you were wrong?

actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)
 

m1ldslide1

Platinum Member
Feb 20, 2006
2,321
0
0
Originally posted by: randay

What happened to your little wiki reference? didnt work out anymore after I pointed out how you were wrong?

Here's the independent nomination for self-pwnage of 2007. Man, and it's only February.
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: m1ldslide1
Originally posted by: randay

What happened to your little wiki reference? didnt work out anymore after I pointed out how you were wrong?

Here's the independent nomination for self-pwnage of 2007. Man, and it's only February.

It's going to be a good year:beer:
 

randay

Lifer
May 30, 2006
11,018
216
106
Originally posted by: Goosemaster
actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)

No, really. whats wrong with

In computing, the acronym RAID (originally redundant array of inexpensive disks, also known as redundant array of independent disks) refers to a data storage scheme using multiple hard drives to share or replicate data among the drives. Depending on the configuration of the RAID (typically referred to as the RAID level), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In its original implementations, its key advantage was the ability to combine multiple low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than was affordably available in a single device using the newest technology.

Seems like a perfectly fine definition of RAID to me.
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: randay
Originally posted by: Goosemaster
actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)

No, really. whats wrong with

In computing, the acronym RAID (originally redundant array of inexpensive disks, also known as redundant array of independent disks) refers to a data storage scheme using multiple hard drives to share or replicate data among the drives. Depending on the configuration of the RAID (typically referred to as the RAID level), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In its original implementations, its key advantage was the ability to combine multiple low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than was affordably available in a single device using the newest technology.

Seems like a perfectly fine definition of RAID to me.

The keyword being "OR" which you completely ignored for a bunch of pages.
I truly wish you were being genuine, but from your attitude I gather that you still feel that you are "right" when you were arguing semantics, and incorrectly I might add.
 

randay

Lifer
May 30, 2006
11,018
216
106
Originally posted by: Goosemaster
Originally posted by: randay
Originally posted by: Goosemaster
actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)

No, really. whats wrong with

In computing, the acronym RAID (originally redundant array of inexpensive disks, also known as redundant array of independent disks) refers to a data storage scheme using multiple hard drives to share or replicate data among the drives. Depending on the configuration of the RAID (typically referred to as the RAID level), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In its original implementations, its key advantage was the ability to combine multiple low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than was affordably available in a single device using the newest technology.

Seems like a perfectly fine definition of RAID to me.

The keyword being "OR" which you completely ignored for a bunch of pages.
I truly wish you were being genuine, but from your attitude I gather that you still feel that you are "right" when you were arguing semantics, and incorrectly I might add.

Uh... you mean the bold part right? Like I said before, how does this disclude RAID0? Looks to me like it validates RAID0 as a legit RAID level.
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: randay
Originally posted by: Goosemaster
Originally posted by: randay
Originally posted by: Goosemaster
actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)

No, really. whats wrong with

In computing, the acronym RAID (originally redundant array of inexpensive disks, also known as redundant array of independent disks) refers to a data storage scheme using multiple hard drives to share or replicate data among the drives. Depending on the configuration of the RAID (typically referred to as the RAID level), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In its original implementations, its key advantage was the ability to combine multiple low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than was affordably available in a single device using the newest technology.

Seems like a perfectly fine definition of RAID to me.

The keyword being "OR" which you completely ignored for a bunch of pages.
I truly wish you were being genuine, but from your attitude I gather that you still feel that you are "right" when you were arguing semantics, and incorrectly I might add.

Uh... you mean the bold part right? Like I said before, how does this disclude RAID0? Looks to me like it validates RAID0 as a legit RAID level.


In regards to the wiki, fault tolerance is a product of certain disk and data manipulation methods, usually addressed as certain RAID levels. The original technology provides striping, which can be used on it's own, but that is regarding the original technology. The IMPLEMENTED TECHNOLOGY AND INDUSTRY STANDARDS DICTATE THAT FAULT TOLERANCE AND HIGH AVAILABILITY ARE THE MOST IMPORTANT CHARACTERISTICS OF AN ARRAY, and since simple striping is not REDUNDANT, AS THE DISK CONTROLLER SENDS DIFFERENT DATA IN AN UNRECOVERABLE FASHION TO EACH DISK OR VOLUME, it is not considered as a valid REDUNDANCY MECHANISM, and therefore the naming conventions applied to REDUNDANT ARRAYS OF INEXPENSIVE/INDEPENDENT DISKS ARE NOT VALID for simple striping scenarios expect for social reasons*.


*= sloth, easiness to remember, uniformity in marketing name conventions etc

edit: and by redundant I mean immediately INEFFICIENT.
 

TreyRandom

Diamond Member
Jun 29, 2001
3,346
0
76
Originally posted by: randay
Originally posted by: Goosemaster
Originally posted by: randay
Originally posted by: Goosemaster
actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)

No, really. whats wrong with

In computing, the acronym RAID (originally redundant array of inexpensive disks, also known as redundant array of independent disks) refers to a data storage scheme using multiple hard drives to share or replicate data among the drives. Depending on the configuration of the RAID (typically referred to as the RAID level), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In its original implementations, its key advantage was the ability to combine multiple low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than was affordably available in a single device using the newest technology.

Seems like a perfectly fine definition of RAID to me.

The keyword being "OR" which you completely ignored for a bunch of pages.
I truly wish you were being genuine, but from your attitude I gather that you still feel that you are "right" when you were arguing semantics, and incorrectly I might add.

Uh... you mean the bold part right? Like I said before, how does this disclude RAID0? Looks to me like it validates RAID0 as a legit RAID level.

Disclude? You mean exclude, right?

I'm not sure whether your four missing credit hours are in IT or English. :p
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: TreyRandom
Originally posted by: randay
Originally posted by: Goosemaster
Originally posted by: randay
Originally posted by: Goosemaster
actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)

No, really. whats wrong with

In computing, the acronym RAID (originally redundant array of inexpensive disks, also known as redundant array of independent disks) refers to a data storage scheme using multiple hard drives to share or replicate data among the drives. Depending on the configuration of the RAID (typically referred to as the RAID level), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In its original implementations, its key advantage was the ability to combine multiple low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than was affordably available in a single device using the newest technology.

Seems like a perfectly fine definition of RAID to me.

The keyword being "OR" which you completely ignored for a bunch of pages.
I truly wish you were being genuine, but from your attitude I gather that you still feel that you are "right" when you were arguing semantics, and incorrectly I might add.

Uh... you mean the bold part right? Like I said before, how does this disclude RAID0? Looks to me like it validates RAID0 as a legit RAID level.

Disclude? You mean exclude, right?

I'm not sure whether your four missing credit hours are in IT or English. :p

Now that sh!t was funny...
 

randay

Lifer
May 30, 2006
11,018
216
106
Originally posted by: Goosemaster
Originally posted by: randay
Originally posted by: Goosemaster
Originally posted by: randay
Originally posted by: Goosemaster
actually sicne you didn't even read it, I thought I would take the time to type out a nice description of RAID from my perspective:)

Just liek I do with everyone who isn't getting something.


<--math tutor:)

No, really. whats wrong with

In computing, the acronym RAID (originally redundant array of inexpensive disks, also known as redundant array of independent disks) refers to a data storage scheme using multiple hard drives to share or replicate data among the drives. Depending on the configuration of the RAID (typically referred to as the RAID level), the benefit of RAID is to increase data integrity, fault-tolerance, throughput or capacity, compared with single drives. In its original implementations, its key advantage was the ability to combine multiple low-cost devices using older technology into an array that offered greater capacity, reliability, speed, or a combination of these things, than was affordably available in a single device using the newest technology.

Seems like a perfectly fine definition of RAID to me.

The keyword being "OR" which you completely ignored for a bunch of pages.
I truly wish you were being genuine, but from your attitude I gather that you still feel that you are "right" when you were arguing semantics, and incorrectly I might add.

Uh... you mean the bold part right? Like I said before, how does this disclude RAID0? Looks to me like it validates RAID0 as a legit RAID level.


In regards to the wiki, fault tolerance is a product of certain disk and data manipulation methods, usually addressed as certain RAID levels. The original technology provides striping, which can be used on it's own, but that is regarding the original technology. The IMPLEMENTED TECHNOLOGY AND INDUSTRY STANDARDS DICTATE THAT FAULT TOLERANCE AND HIGH AVAILABILITY ARE THE MOST IMPORTANT CHARACTERISTICS OF AN ARRAY, and

Size and speed matters, or else everyone would use mirroring. Depending on your application, speed can be more important than fault tolerance. Processing large data sets for example, like images, video, audio, etc... Databases with huge amounts of transactions. Fault tolerance is probably the most common reason to use RAID though. RAID5 is probably the best RAID level even though it does not offer as high fault tolerance and availability as mirroring. Ive worked with more RAID3,5, and 0 arrays then RAID1s. Of course I always push for the RAID1 just to cover my ass. :p

since simple striping is not REDUNDANT, AS THE DISK CONTROLLER SENDS DIFFERENT DATA IN AN UNRECOVERABLE FASHION TO EACH DISK OR VOLUME, it is not considered as a valid REDUNDANCY MECHANISM.
noone is disputing this fact...
 

jakedeez

Golden Member
Jun 21, 2005
1,100
0
0
Originally posted by: JDMnAR1
Originally posted by: supafly
Also, all RAID setups have built in fault tolerance. Isn't that special?

If by "special" you mean "incorrect". ;) RAID 0 has no fault tolerance.

Why it really isn't raid...
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: randay

Size and speed matters, or else everyone would use mirroring. Depending on your application, speed can be more important than fault tolerance. Processing large data sets for example, like images, video, audio, etc... Databases with huge amounts of transactions. Fault tolerance is probably the most common reason to use RAID though. RAID5 is probably the best RAID level even though it does not offer as high fault tolerance and availability as mirroring. Ive worked with more RAID3,5, and 0 arrays then RAID1s. Of course I always push for the RAID1 just to cover my ass. :p

Okay mr. bigshot.

Web transactions or database queries. Give me numbers. Everyone including myself has spouted their share of BS but you are going over the line. Who in their right mind talks about VIDEO AND AUDIO when it comes to an enterprise environment. We are talking hardcore database and web transactions here...you know, things that go on in an enterprise enviroment.

That said, I can only talk about it so much more, as we are sort of getting to the bounds of my knowledge. Anything more that I can say is just numbers that I pull out of eweek or networkworld or books. I really have no real-world experience with databases in raid.

See...now you can trash me on the higher-level stuff which hopefully you know....right after we get done talking semantics...:p
 

TreyRandom

Diamond Member
Jun 29, 2001
3,346
0
76
Originally posted by: Goosemaster
Okay mr. bigshot.

Web transactions or database queries. Give me numbers. Everyone including myself has spouted their share of BS but you are going over the line. Who in their right mind talks about VIDEO AND AUDIO when it comes to an enterprise environment. We are talking hardcore database and web transactions here...you know, things that go on in an enterprise enviroment.

That said, I can only talk about it so much more, as we are sort of getting to the bounds of my knowledge. Anything more that I can say is just numbers that I pull out of eweek or networkworld or books. I really have no real-world experience with databases in raid.

See...now you can trash me on the higher-level stuff which hopefully you know....right after we get done talking semantics...:p

That's cool, Goose, I gotcha covered on databases in RAID. I work in a healthcare environment where we have loads of databases. I don't have access times on my specific boxen (like I would even consider tearing down my existing RAID arrays just to test access times for each database using different RAID levels...), but I can share my knowledge on what works best in different scenarios.

EDIT: That said, it's probably useless to discuss this stuff with Randay, because he'd debate me on the meaning of the word "is".
 

randay

Lifer
May 30, 2006
11,018
216
106
Originally posted by: Goosemaster
Originally posted by: randay

Size and speed matters, or else everyone would use mirroring. Depending on your application, speed can be more important than fault tolerance. Processing large data sets for example, like images, video, audio, etc... Databases with huge amounts of transactions. Fault tolerance is probably the most common reason to use RAID though. RAID5 is probably the best RAID level even though it does not offer as high fault tolerance and availability as mirroring. Ive worked with more RAID3,5, and 0 arrays then RAID1s. Of course I always push for the RAID1 just to cover my ass. :p

Okay mr. bigshot.

Web transactions or database queries. Give me numbers. Everyone including myself has spouted their share of BS but you are going over the line. Who in their right mind talks about VIDEO AND AUDIO when it comes to an enterprise environment. We are talking hardcore database and web transactions here...you know, things that go on in an enterprise enviroment.

That said, I can only talk about it so much more, as we are sort of getting to the bounds of my knowledge. Anything more that I can say is just numbers that I pull out of eweek or networkworld or books. I really have no real-world experience with databases in raid.

See...now you can trash me on the higher-level stuff which hopefully you know....right after we get done talking semantics...:p

Uh, you really dont need first hand experience to understand why speed would be more important or important at all. It also does not have to be an enterprise environment. I havent worked with any large databases myself but I do know a game called EVE online which is basically limited by the speed of its database drives. They use RAMSAN solid state disks so its no joke. Its not hard to imagine a database application needing really low access and read times above all else.

In my own personal experience, the data processing guys prefer RAID0 over RAID1. They process images, and I dont mean nekkid pics or anything like that, I mean sonar images, google earth type stuff. large images stitched together to make even larger images and then they have to make it all pretty and what not.

Also for some reason the programmers prefer really fast disks as well. They use a third party program for source control so I guess thats why they are not very concerned over fault tolerance and availability. In fact they've already had some failures and downtime but they still insist on RAID0 and speed speed speed. Something about wanting to compile faster. Hell who am I to complain.

RAID5 would be ideal but its just too expensive, most mobos only support 0 and 1 and you need more then just 2 drives :(
 

randay

Lifer
May 30, 2006
11,018
216
106
Oh and as for Video and Audio in an enterprise environment, think TV studio. HD broadcasts. etc... I used to work at one but it was back in the day and they were just starting to make the transition from tape to digital. We did have a 16 drive fibre array though. Running RAID3 or 5 at least.
 

TreyRandom

Diamond Member
Jun 29, 2001
3,346
0
76
Originally posted by: randay
Uh, you really dont need first hand experience to understand why speed would be more important or important at all. It also does not have to be an enterprise environment. I havent worked with any large databases myself but I do know a game called EVE online which is basically limited by the speed of its database drives. They use RAMSAN solid state disks so its no joke. Its not hard to imagine a database application needing really low access and read times above all else.

In my own personal experience, the data processing guys prefer RAID0 over RAID1. They process images, and I dont mean nekkid pics or anything like that, I mean sonar images, google earth type stuff. large images stitched together to make even larger images and then they have to make it all pretty and what not.

Also for some reason the programmers prefer really fast disks as well. They use a third party program for source control so I guess thats why they are not very concerned over fault tolerance and availability. In fact they've already had some failures and downtime but they still insist on RAID0 and speed speed speed. Something about wanting to compile faster. Hell who am I to complain.

RAID5 would be ideal but its just too expensive, most mobos only support 0 and 1 and you need more then just 2 drives :(

They'll prefer RAID 0... till they lose a disk. Then they'll wish they had RAID 1, 5, or a striped mirror (1+0, 0+1, 10, whichever they want to call it). We deal in large images as well - digital X-rays, MRIs, etc. Obviously, we don't want to lose that patient information, so we can't even consider RAID 0 on our servers.

I find it hard to believe that your programmers are OK with RAID 0, unless you mean they want it on their individual workstations. As long as they're saving everything to a server, then yeah, if they want RAID 0 on their workstations, go for it. What's there to lose? Just reimage the box if a drive dies.

Some workstation mobos have RAID 0 and 1, but you get into a whole new arena when you get to servers. You might have only had experience with RAID on workstations, so I'll cut you some slack. RAID 5 is no more expensive than RAID 0 or 1 on a server (with the obvious exception that you need 3 disks for RAID 5 and only 2 for RAID 0 or 1).