Add larger drives to increase capacity

[BartBlaze]

My current cluster consists of four 4330 nodes with 450GB drives. To increase capacity I was considering replacing the 450GB drives with 900GB drives. This would double the capacity for about the same price as adding an additional node. 

I'm looking for best practices on how to accomplish this. I could replace one drive at a time and let the array rebuild but that seems like it could be quite time consuming. Would it be better to remove one of the nodes from the cluster and replace all the drives at once, then rejoin the node back to the cluster? 

[oikjn]

its a risk/reward question for you.  Safest way (assuming you have the capacity) would be to remove a node from the cluster, exchange all the disks and then join it.  This would ensure that you always remain at normal availability protection levels.

The alternatives are to remove one disk at a time while the node is in the cluster and have it rebuild each disk one at a time (which probably would actually take longer).  I wouldn't do this method unless you are tight for space already.  You could speed this process up slightly by trying to update two disks on two different nodes at the same time, but make sure that you pick two adjacent nodes in your cluster order so you don't risk restripes occuring on both copies of your data at the same time.

There are other ways that might be faster but would cause you would lose your node redundancy which probably isn't acceptible so I won't mention them unless those options don't work for you.