cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Robson <mar...@gmail.com>
Subject Re: Best Practics for Cassandra in Production?
Date Thu, 14 Jan 2010 23:35:24 GMT
2010/1/14 shiv shivaji <shivajisus@yahoo.com>

> I have looked at performance posts in the forum but was wondering if there
> are general suggestions for using cassandra in production.
>

I'd say pretty obvious stuff:

- Performance test as much as you can
- Choose the following carefully as they're difficult to change: key
selection strategy, partitioner, replication factor
- Get your ops team to pick the hardware which gives the best balance of
disc capacity, disc performance, CPU and RAM; you probably want to test some
configurations and plan this carefully. Work with your hardware vendor on
this one. As Cassandra can store data redundantly, you may want to go for
non-redundant discs.
- Work out the sizing of your cluster(s) correctly - it may be more sensible
to go with a larger number of lower spec machines (again, check with your
ops team and vendor as power consumption is usually the main cost issue with
clusters hosted in datacentres). The more machines you have, the less
capacity you lose when one of them is down (this is obvious but I thought
I'd say it anyway)
- Work out how your clients are going to connect to the servers, as they
should probably pick a server based on availability and load balancing but
if you run another tier on the same hardware they might just connect to the
local node (hopefully your load balancing / HA system for the next tier up
will do the trick).

Mark

Mime
View raw message