Software RAID configuration

We have configured software based RAID5 with LVM on our RHEL5 servers. Please let us know if its good to configure software RAID on live environment servers. What can be the disadvantages of software RAID against hardware RAID

As well, be advised that RAID 5 without sparing is potentially dangerous, depending on the need for availability.

If you lose 1 drive, you go into a degraded state, which will severely impact performance until another drive has been rebuilt. This is an issue with both software and hardware RAID.

Also, there are lots of things to investigate when using RAID 5, such as the chunk size in relation to your usage patterns as well as other tunables (read ahead, etc).

How many drives are you using in your array?

Thanks for the details, I am looking to the article provided.

Just on this 1 more question arises is that how can we know which Raid level is useful for the server E.g Mail, Web, Database server. Also things we can keep in mind before configuring RAID on live servers.

The choice of RAID level is based more on the number of available disks (read: budget) than the application using it. Typically, only 3 (4) types are used in practice:

  • RAID1 (mirroring) halves the available space, but the RAID is still available with either disk failing. A special case of that is RAID1+0 (RAID10), which combines the security of RAID1 with the speed of RAID0.
  • RAID5 (striping with parity) will cost you one disk out of the total, but it's fast and easy to set up. But as soon as more than 1 disk fails the whole array is gone.
  • RAID6 (striping with double parity) is like RAID5, only you'll loose 2 disks worth of space, and up to 2 drives can fail at the same time.

Setting up RAID on disks already in use is tedious and very very very dangerous. Don't attempt it unless you know what you're doing and have a proven good backup.