EC2 best practices
Wed, 08/10/2011 - 23:34
We're looking to roll out a cluster with nodes on the order of 15 gigs. There are 2 aws options that look compelling:
* c1.xlarge - 15G and high IO. 4-way raid ephemeral drives and backup nightly by copying db files out to an ebs volume and snapshotting
* m1.xlarge - 17G and medium IO. 4-way raid EBS volumes and backup nightly by copying db files out to a single ebs (dont want to rely on a re-assembling a 4-way raid that may have been snapshotted at slightly different times).
Is it even standard practice to raid the underlying datastore? If this is not necessary, we may just forget about it and go with an m1.xlarge.
There's a bit of "your mileage may vary" and so your best bet is probably to discuss this directly with AWS.
We would recommend getting the most memory that your money can buy and monitoring the system appropriately.
Perry
Forum support is great for free but sometimes you need a guaranteed response time and dedicated resources for your questions or issues.
Consider purchasing enterprise-level support from Couchbase: http://www.couchbase.com/products-and-services/overview
Call or email "sales -at- couchbase-dot- com" today!