ow.mun.heng@xxxxxxx wrote:
-----Original Message-----In the ideal world - one would have access to a test (sometimes called staging) environment that has identical hardware and software to the production environment. One could then run a number of load and fault simulations over a set period of time to assess the suitability of the desired technology. Back in the real world - find a non-production system, set it up as you would for live - and thrash the living daylights out of it. chuck everything you can at it - taking note of load, processes, memory usage, disk access times, performace et al.
From: neil [mailto:neilcuk@xxxxxxx]
Sent: Wednesday, April 28, 2004 5:34 PM
To: For users of Fedora Core releases
I haven't used ReiserFS in production. In practice it's supposed to do
intelligent block management to maximise the capacity of your storage. For
example - if you specify blocks of 4k and your files end up being 1k or
smaller the ReiserFS system will still used the remaining space within each
block - very cool indeed! whereas in the extended filesystem you would
potentially loose 3/4 of your storage capacity if your file ended up being
1k or less. But read the links that have been posted and make sure you
test/test/test :-)
QUESTION :
So.. how do we test/test/test???
If it breaks: Can you recover? Is there a notable performance increase or overhead? was the transition troublesome on the test environment?
Really, if you have a think about it, one can envisage a mirriad of procedures to test a system.
<snip>
neil
