INTEL stuff

[sino_darenai]

I am looking at building 2 servers using the new dual PIII motherboard from Intel (Dodson SDS2) with their (Intel's) raid controller in their (intel's) server chasiss (SC5100)

I am looking to use one as a database server for an ecommerce application, and the other as the webserver/eccomerce software. The DB server will have either Raid 0+1 or Raid 5. The Webserver will just have Raid 1.

Does anyone have any experience with these products. I would appreciate any thoughts and/or suggestions. Do you think that I am going overboard with Intel?!?!

 

[FatBurger]

Go with RAID 5 for the database server, that'd be your best bet. I don't have any experience with the other stuff you mention, though. What kind of hard drives are you planning on using?

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft

 

[Acclrator]

Raid 5 isn't actually that reliable.
What it does, or is supposed to, is take a small portion of each of the 4 disks and copy files from the each other, so in a sense if one crashes all the files are saved on the other 3 and the one that crashed can be taken out, and put in a new one and the other 3 will just copy the old files back to the new HDD and it's like it never crashed. But from my experience.. If one crashes.. they all crash.

IMO mirroring would be the best.

 

[Crashman]

RAID5 offers protection against data corruption.

What's the frequency, Kenneth?

 

[Crashman]

Wouldn't RAID 50 be better? Doesn't RAID 50 offer all the data integrety of RAID 5 with the 100% security from loss of RAID1?

What's the frequency, Kenneth?

 

[FatBurger]

I'm not exactly sure how RAID 50 differs from RAID 5 in terms of data security, I'd be interested to know.

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft

 

[upec]

You must be very lucky to have 4 hard drives crash at once.

 

[Unolocogringo]

Actually raid 5/0 works a little differently. It writes the data to 2 or more drives in 4 bit chunks to each drive. then it adds the data written to the drives together. it then writes the PARITY chunk to the parity drive. This way if one(1) drive fails. the data can be recovered without loss .the other drives will also stop functioning because one drive has failed.then you replace the bad drive . once the bad drive is replaced the program will take over and rewrite the new drive with the info from the old drive, Even though it is not connected .
1011 first drive
1000 second drive
0110 third drive
1111 fourth drive
------------
3232 this is where the data is added together. Now if it is an even number the parity drive is written with a 0,if it is an odd number the parity is written with a 1.

1010 would be written to the parity drive. So now when a drive crashes, all can be recovered with simple math. And yes the whole system will be down while the new disk is written but nothing will be lost. hope this makes since to most of you it is hard to do diagrams on a typewriter.


I aint signing nothing!!!<P ID="edit"><FONT SIZE=-1><EM>Edited by Rick_Criswell on 04/18/02 08:07 PM.</EM></FONT></P>

 

[Acclrator]

I'm in the Air Force, and that's just what happened on the base. I wouldn't call it "lucky". It's just that the AF runs somewhat older machines, rarely is it top of the line, and at that decent.

It's just at that time, the HDD and system they were running on was older and I suppose all the hard drives said I give up at the same time, or around the same time.