Hi. Does anyone know whether Netio uses local disk as part of its throughput tests? I have a customer claiming that he ran netio across 2 of his blade severs and is only seeing 30MB throughput over a 1GB internal connection.
We would expect that to be more, but it will be difficult to replicate their production environment/application set for test. I'm probably going to download netio and test on a base OS anyway, but I was wondering if anyone knows whether local disk performance can affect netio (I've never used it). These are older IBM blades that don't have disk cache, so local disk bottleneck could conceivably affect performance.
We've asked them to gather perfmon logs while running netio to verify whether the disk is being thrashed, but they have to schedule things as simple as perfom, and that could take days. Just trying to answer any questions I have in the meantime.
Thanks for any info.
We would expect that to be more, but it will be difficult to replicate their production environment/application set for test. I'm probably going to download netio and test on a base OS anyway, but I was wondering if anyone knows whether local disk performance can affect netio (I've never used it). These are older IBM blades that don't have disk cache, so local disk bottleneck could conceivably affect performance.
We've asked them to gather perfmon logs while running netio to verify whether the disk is being thrashed, but they have to schedule things as simple as perfom, and that could take days. Just trying to answer any questions I have in the meantime.
Thanks for any info.
