Torture Test Failure Follow-up

[jugbugs]

I have posted a thread noting that I was failing the torture test with my newly purchased hardware. The Hardware is as follows:

P4 2.8GHz 800MHz FSB
PC Chips M952 MB
512 Samsung DDR3200
Dual 80 G SATA Raid 0 HDDs

I sent the hardware back assuming that that was the problem and wanted to get that replaced. I sent it back to Compubuzz.com and they told me there was nothing wrong with my board. I mentioned to them to make sure it passes the prime 95 torture test. They told me that it fails the test along with all their other PC Chips/ ECS "testing" boards. Thus, thoroughly frustrated I purchased a ASUS p4p800SE MoBo from Newegg. Pluged all the same hardware from the old setup into this one and booted up. Needless to say ASUS MoBos kick ass and prime 95 works fine. I dont know what the deal is with pc chips but it seems prime wont run on those boards.

[Xyzzy]

This is not surprising at all... There is a big difference in quality between the bottom and top tier motherboard manufacturers...

You hear a lot of people using cheap motherboards for distributed computing, but they are usually running projects that don't check the results...

[smh]

Dual 80 G SATA Raid 0 HDDs

You know what RAID 0 means, don't you?

[jugbugs]

Originally Posted by smh

You know what RAID 0 means, don't you?

Of Course, here's a Googled version cause im lazy...

RAID 0...Also known as data striping, RAID 0 functionality divides data into blocks and distributes the blocks across multiple disks in a array. Distributing the disk I/O load across disks and controllers improves disk I/O performance. However, striping decreases availability because one disk failure makes the entire disk array unavailable.

[smh]

Originally Posted by jugbugs

Of Course, here's a Googled version cause im lazy...

RAID 0...Also known as data striping, RAID 0 functionality divides data into blocks and distributes the blocks across multiple disks in a array. Distributing the disk I/O load across disks and controllers improves disk I/O performance. However, striping decreases availability because one disk failure makes the entire disk array unavailable.

Just wanted to warn you what will happen in case of a disk failure.....

[jugbugs]

The same thing that happens to people using just 1 HDD.. Data Loss

[Moloch]

Originally Posted by jugbugs

The same thing that happens to people using just 1 HDD.. Data Loss

I dont believe he means data loss.. there is a critical problem with raid 0 where if one drive "breaks" you lose ALL your data due to the way the data is striped on the disks.

this is not a huge problem for a small array, say 2 disks. If you use decent hard drives you should be able to run this for many years with no problems. However, if you made a 16 drive raid 0 array, there is a pretty good chance one of those drives will crash due to poor handling (you should see the way these shipping companies throw around packages) or manufacturing defects. As far as I know there is no way to recover your data from the other 15 disks if one of them were to "crash".

[Reboot It]

Originally Posted by jugbugs

The same thing that happens to people using just 1 HDD.. Data Loss

True, but by having two (or more) separate disk drives you increase the overall probability of having at least one hardware failure. If you have that failure, of course as has been said you lose all of the data.

The increased overall probability of a failure is the important thing here. You might wish to think of two HDDs configured as RAID0 as equivalent to one much less reliable HDD.

[jugbugs]

Agreed. For me, what you get for performance increase, it is well worth it. Not to mention with regular backups, there wont be any data loss.