News:

Tekforums.net - The improved home of Tekforums! :D

Main Menu

Whats the difference?

Started by dogbert, December 14, 2009, 18:03:56 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.


soopahfly

More features on the .12.

QuoteFeatures   Native Command Queuing (NCQ), Adaptive Fly Height, Clean Sweep, Enhanced G-Force Protection, Seagate SoftSonic, Perpendicular Magnetic Recording (PMR)  

vs

QuoteFeatures   3D Defence System, Native Command Queuing (NCQ)

bear

.9 is heavier,  slightly better latency

shock tollerance differ but I cant really tell which is better as it is measured differently !

andy[tek]

id also be surprised if that 7200.9 has the full warranty as they were discontinued a while ago. All the 7200.9 stock we get offered is all 1 yr RTB via the channel ?

Worth a check   -)

ps were cheaper !

hate to plug but ....

http://www.tekheads.co.uk/product/Seagate-Barracuda-7200.12-SATAII-NCQ-3Gbs-500GB-16MB-Cache-OEM_10435.html


Regards
Andrew
Tekheads.co.uk Purchasing

dogbert

OK, Looks like a bargain from TEK.

How safe is raid (appart from Raid5) for data, coss I still want the extra performance of stripping...would anyone use a pair of them stripped for a storage solution?

Clock'd 0Ne

RAID0 Striping offers no safety at all as it offers no redundancy. If anything I guess its worse on data safety as you are relying any one of the drives not packing in as if one drives failed the whole array collapses. Its bloody fast though :D

RAID1 is the safest (mirrored) offering complete redundancy and a read speed improvement I believe on some controllers (it can read the data asynchronous from both drives at once, but needs to write synchronously).

RAID3 trades one disk of the array space for data parity - you need at least three drives. RAID5 takes this a step further and distributes the parity data across all drive in the array, its the most widely used and probably the best of both worlds approach when looking at speed vs redundancy.

Serious

Raid 0 reduces safety proportional to the number of drives, so a 2 disk array has double the chance of a failure. Raid 1 does not offer any speed advantages, and halves the total storage area available but is far more secure, both drives have to fail at the same time to lose data.

Raid 2 stripped with hamming code parity. 3 disks minimum. Extremely fast data transfer but this is now never used because of requirements.

3, 4 and 5 offer better space and speed than raid 1 but need more drives and cpu cycles unless on a separate card. Data is stripped but recovery data is also stored. 3 has bit parity, which really needs synchronised data from the drives. 4 has block parity which is more tollerant and 5 has interleaved or distributed parity data. All 3 disks minimum.

Raid 6 has dual raid parity data stores, meaning it can recover from 2 drives failing simultaneously. 4 disks minimum.

Servers often have hot swappable drives, so you can pull a failed one and replace it while the system is running. Some of the above can also have hot spares additional drives that are used to rebuild the array immediately a failure happens.If possible get drives from several different batches, drives from the same batch seem to have an annoying tendency to go at the same time.