On 2/20/2013 1:12 AM, Toke Eskildsen wrote:
On Tue, 2013-02-19 at 18:39 +0100, chamara wrote:
Hi Thanks Shawn for the Input, Yes i am using SolrCloud to replicate the
index to another server running with the same spec with 32cores and 72GB RAM
on each machine. I have to test the performance of RAID 10? Have you ever
done a deployment with RAID 10? The indexing will be NRT as far as i can see
, so it is going to be lots of writes to the DISK.
As you are replicating the index across servers, why not use a simple
RAID 0?
I can think of one simple reason to avoid raid0 - the amount of sysadmin
time it takes to set up a server, even if you're cloning it from another
one and changing unique IDs like MAC addresses, IP addresses, machine
name, and other OS-level identifiers, is significant. In my opinion,
the work to rebuild a server is enough that it's a good idea to ensure
that a single drive failure doesn't take out the OS. Drives *WILL*
fail. It's unavoidable. If you're OK with spending your time
rebuilding an OS that you have already set up once, then raid0 is a
viable solution for you. It's not for me.
If you have enough drives that you can set up the OS and all programs on
RAID1 and the data on raid0, then you can have both speed and capacity
for your data ... but there's a tradeoff. A single drive failure will
still take the server out of commission until you can replace the drive
and rebuild the data partition, but you won't have to rebuild the OS.
Thanks,
Shawn
