On 02/05/14 09:18, Craig Cook wrote:
>I did at $JOB[-1]
Opinions?
Good, bad, ugly? Great for X? Bad for Y?
Sorry for the long delay...
It is a bit of a lock-in to an appliance vendor. Though they do have an
soft-appliance in the form of an AMI and an RPM for development.
You can think of it as an SQL Hadoop cluster. They split tables among
cluster members, with duplication. The SQL is broken down and parts are
distributed to where the data is, in the cluster. Then summarize to the
requesting node.
They are good for single instance large databases. So, it depends on
what you are looking to run on the database. Really, all it does is
raise the threshold of when you are required to shard. Then again, if
you are going to be "cloud" based and have multiple geographies, then
you will need to shard anyway.
For their appliance, you are required a minimum of three boxes plus
provide an Infiniband switch. Yes, they are an Infiniband based cluster.
They claim wire level MySQL compatibility. While this is true, there
are higher level differences. First, table names are case insensitive,
but case preserving. And there are some SQL statements that might have
to be re-engineered a bit to make work. I think there might be a
limitation on foreign keys, also.
We had issues trying to make our product work in the lab, using MySQL in
place of Clustrix. Our coders were lazy in reference to table names.
They would create tables in CamelCase, but reference them in all lower.
We had to turn on the flag in MySQL to be case insensitive, but we
found that this flag was not 100%. Some migrations still failed because
of case issues.
--
Mr. Flibble
King of the Potato People
http://www.linkedin.com/in/RobertLanning
_______________________________________________
Tech mailing list
Tech@lists.lopsa.org
https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
http://lopsa.org/
