We have an HP Blade Chassis that we recently moved from 4G San to 8G San. We have a particularly chatty & repetitive application. The application vendor has some delays with rewriting the code to be more efficient. We are on a much bigger install than what they are accustom to.
To alleviate some of their issues, our storage vendor recommended that we modify our round robin change rate from the default 1000 to 1. (EMC, we have the VMAX 40)
This isn't the only application on this cluster and it is in the process of growing.
VMware KB: Adjusting Round Robin IOPS limit from default 1000 to 1
While it worked great for this application, I want to know if there are any possible downsides to doing this?
I know that it is not recommended for Microsoft clustering - which we don't use.
But how about Oracle RAC?
Why would the default be at 1000 anyway?
Thanks - Nancy