Originally posted by: Zstream
Do you know the performance difference between raid 1 & 5? I would imagine Raid 1 is just like a single HDD?
RAID1 uses two separate physical drives, each a mirror copy of the other. Every delete and / or write operation is done on both drives (one first, then the other VERY soon after). A read operation will be done from the first drive that has the data available; a sustained sequential read may be done by alternating between drives to speed it up. In the background the RAID system constantly checks the drives to be sure they both are working properly and that their data matches. In the event of a failure, the system automatically reverts to using ONLY the good drive, and alerts the user to the problem. The software also usually includes tools to help identify the faulty drive, and then (following replacement) make a fresh complete mirror copy of the good drive to the new one, restoring the full mirrored RAID1 operation. RAID1 uses two drives to give you one drive's storage capacity. OP, the RAID1 system I used has tools included that will let you take a partially-filled regular drive and add a new empty drive, then convert them both into a RAID1 array containing all the original data.
RAID5 uses several physical drives (minimum 3, the classic structure is 5) and splits every byte written to the array into four pieces, then calculates a fifth byte with XOR data. All five bytes are written to different places - in the classic structure, one of each of the four pieces to the first four drives, and the extra byte to the fifth. The mathematics involved means that if you lose any one of those five pieces due to a disk problem, the remaining four pieces are used to completely regenerate the original byte unambiguously. The system monitors itself and, for a minor error, it can give you the right data and then go back to fix the record on the disks. If that record can't be fixed then it can detect which disk has the faulty information, stop using it, and report the problem. The entire system can continue to operate with slightly longer data access times due to the extra CPU workload required, until the faulty drive is replaced. Once that is done, the RAID5 system can re-construct the data on the new replacement drive from the surviving data on the other four, and that completely restores the RAID5 array. Really good RAID5 systems use dedicated controller cards to reduce CPU workload, and these often also make it possible to "Hot Swap" a replacement drive in place of a faulty one, then rebuild the array as a background task, so the the system experiences no downtime at all. That feature is not usually available in RAID1 or RAID5 built into the mobo's chipset. RAID5 (classic form) uses five drives to give you four drives' storage capacity.
Both systems offer protection against data loss, ability to recover and restore themselves, and assurance of continued system function during the fault stages. However, the do their jobs quite differently.
Both systems, obviously, cannot recover from simultaneous failure of two drives (I've seen it happen - remember to BACKUP!) RAID6 does something like RAID5 but calculates and stores an additional relationship between the four original byte pieces, with the result that the byte can be regenerated unambiguously from any 4 of the 6 pieces of stored information. So, it can recover from simultaneous failures in two drives, a rare event! RAID6 uses six drives to delivery four drives' storage capacity. The level of data security in RAID5 or RAID6 usually is deployed only in large data servers that MUST ensure on-line performance with almost no downtime. They require some professional-quality maintenance and monitoring. In those settings the systems usually benefit from correct backup systems, also, unlike so many home / SOHO systems.