Is RAID5 performance much worse than no RAID ?
I have a server with a RAID5 array. I understand that using RAID will slow down write speeds as it has to write to two different places. My question is, how much slower would it be compared to having no RAID, just one hard drive? Are we talking a few seconds or longer? Is it a noticeable difference?
Observing members:
0
Composing members:
0
2 Answers
Wikipedia sums it up well.
—
In an enterprise environment where we employ a high performance SAN – which is expressly engineered to improve the effective i/o throughput in any case – the performance differences between RAID0, RAID1, RAID5, RAID6 are practically so minimal that we almost always suggest the most cost-effective, simple, and relatively robust solution: RAID5. Some customers elect to go with RAID6 for the extra redundancy.
Point is, if you’re using a plain, vanilla minimal RAID5 setup, then yah you might see some performance differences as compared to other solutions. They’ll add up over the long-haul, but for day-to-day use of a transactional nature, it’d likely perform just fine as long as you weren’t whipping it non-stop all day.
RAID1 is probably the very best for performance and robustness. And also the most expensive since it is just a mirror, you need 1 drive for your data and a 2nd drive to mirror it. So your disk cost is double. And with twice as many drives, you fill up controllers fast so there’s additional cost there as well.
Finally, beware the siren call of best performance via RAID0. All that does is basically stripe a bunch of drives together so they act like a single unit. And so then, for each drive you add, you increase the chance of failure because if one drive fails, the entire unit is no good.
RAID is about redundancy. The performance hit is minimal.
Answer this question 
This question is in the General Section. Responses must be helpful and on-topic.
Composing members:
0
