Hi all
I am having a little bit of confussing in raid(Redundant array of Inexpensive disks). The basic idea of RAID was to combine multiple small, inexpensive disk drives into an array of disk drives which yields performance exceeding that of a Single Large Expensive Drive.In Linux how internally it will do?. Can I call this as hardware cluster.Can any one plz help me regarding this.
Thanx in advance.
Regards, Srinivas Rao NatureSoft Pvt. Ltd. 11-20, 2nd Lane, Appu street, Mylapore, Chennai-600004.
Phones:044-24617193/94/95 (o)
Hi Srinivas, I believe the purpose of RAID is to create a failsafe system. Data is written to multiple disks simultanously. If one disk fails, the system continous to be usable with the other disk. I'm not sure if RAID is meant for performance enhancement. In fact when using RAID there is a slight reduction in performance. I could be wrong about this... guys please correct if so.
regards Parag
Hi all
I am having a little bit of confussing in raid(Redundant array of Inexpensive disks). The basic idea of RAID was to combine multiple small, inexpensive disk drives into an array of disk drives which yields performance exceeding that of a Single Large Expensive Drive.In Linux how internally it will do?. Can I call this as hardware cluster.Can any one plz help me regarding this.
Thanx in advance.
Regards, Srinivas Rao NatureSoft Pvt. Ltd. 11-20, 2nd Lane, Appu street, Mylapore, Chennai-600004.
Phones:044-24617193/94/95 (o)
Parag Shah writes:
Hi Srinivas, I believe the purpose of RAID is to create a failsafe system. Data is written to multiple disks simultanously. If one disk fails, the system continous to be usable with the other disk. I'm not sure if RAID is meant for performance enhancement. In fact when using RAID there is a slight reduction in performance. I could be wrong about this... guys please correct if so.
RAID is indeed used to create a failsafe system. If one of the array of disks fails the information stored on the erring disk can be reconstructed using the other disks. RAID also leads to performance enhancement in some cases(RAID has different level 0-5), this is particularly true if you have a hardware RAID controller. Some features of RAID can be implemented using software. Examples being mirroring of disk using LVM. Also the capacity of two disks can be concatenated for use as a single disk using LVM. I have written an article for LinuxGazette on LVM. Maybe U will find it useful.
<url> http://www.linuxgazette.com/issue84/vinayak.html </url>
Mirroring can lead to significant improvements in reads (parallel reads) thus reducing disk latency times.
Vinayak Hegde APGDST Student NCST-JUHU
RAID was designed to be a Redundant Array of INEXPENSIVE Disks - but the originators realised that the goals they were trying to achieve - of data protection and performance, werent acheivable with inexpensive disks.Hence the 'I' in the acronym now stands for Independent!! RAID is basically a data protection scheme. RAID-0, which is actually striping of data across n disks for performance, was not originally part of the RAID scheme. To acheive protection, you could go from high cost RAID 1, which duplicates all data, to doing parity calculations so as to reduce your costs. 'Software' or "hardware' RAID tells you whether this parity calculation is being done by software(that uses host cpu cycles) or a dedicated processor in the disk controller(called a RAID controller). It is this parity calculation that causes the drop in performance for software RAID. For all server environments, it is recommended to go for hardware RAID controllers. As for the performance part of it - to reduce the number of disk reads and writes - a cache is used in the disk controller. But then that is another story......
regds, kishor
__________________________________________________ Do you Yahoo!? Yahoo! Tax Center - forms, calculators, tips, more http://taxes.yahoo.com/
----- Original Message ----- From: "kishor bhagwat" aaaaarrrgghhh@yahoo.com To: linuxers@mm.ilug-bom.org.in Sent: Friday, February 21, 2003 8:46 PM Subject: Re: [ILUG-BOM] Re: About RAID
RAID was designed to be a Redundant Array of INEXPENSIVE Disks - but the originators realised that the goals they were trying to achieve - of data protection and performance, werent acheivable with inexpensive disks.Hence the 'I' in the acronym now stands for Independent!!
When they talk of Inexpensive Disks, they are not taking of a comparison between ide drives and scsi drives. They are talking about the cost differential between small capacity scsi drives against using a large capacity scsi drive. For example, a 100GB scsi drive may cost (hypothical cost --- I have no idea how much they actually cost) Rs. 50000, but using 5 scsi drives of 20GB each would cost (again hypothical) Rs. 20,000, giving the same result and better safety. That is why it was called Redundant Array of Inexpensive Disks.
Please remember that at the time when this term came into being, the smaller capacity hdds were much cheaper than the proportionate cost of a larger hdd. There was a premium on higher capacity drives because, first few companies could make it and because the tools available were not good enough to work with multiple drives (you didnt have vertual server and consolidation / transparent clustering accross the network or concepts like Network Area Storage). Since this pricing does not hold true today, the description of raid has changed to Redundent Arrey of Independent Disks
Regards Saswata
-
kishor bhagwat -
Parag Shah -
Saswata Banerjee & Associates -
srinivas -
vinayak_hegde＠softhome.net