Hi, Keith,
the data nodes in a cluster do not really need to be the same. The only
requirement is that they have to have the same endianness. They can have
different amounts of RAM, different CPU counts, hard drives, etc. But if
the data nodes are different, the performance will be pretty much
determined by the slowest node and the database size will be limited by
the node with the least amount or RAM and/or with the smallest hard
drive. It does not make much sense to use a powerful machine for a data
node if another data node is slow, because you will get practically the
same performance with two slow machines. Management node does not
monitor the "speed" of data nodes, it actually does not distribute the
load on data nodes at all. It is not how it is done in MySQL Cluster
architecture.
Regards,
Anatoly.
> Being a newbie to clusters this might be a dumb question. I haven't seen it directly answered anywhere so I am going to venture out on a limb and ask... > > I understand that each data node in a cluster needs the same amount of RAM . However, how will different cpu counts/hard drive setups and such affect the performance. Not talking celeron vs/quad core, but some systems might have raid and four cores while some might not have raid and only have two cpus. Does the management node monitor things and distribute the nodes or should I try as hard as I can to have equal hardware on the boxes? > > thanks, > > Keith >
--
MySQL Cluster Mailing List
For list archives:
http://lists.mysql.com/cluster
To unsubscribe:
http://lists.mysql.com/cluster?unsub=lists@pantek.com
Received on Thu Aug 16 17:29:40 2007
This archive was generated by hypermail 2.1.8
: Sun Oct 07 2007 - 10:15:00 EDT
|