Trying to figure out an IO issue that has been bugging us for quite a while now and not sure where to start to troubleshoot the issue.
We are in the process of migrating an application from the AWS cloud to our physical datacenters.
What we are seeing is that a DB update single threaded process that takes about 8 minutes to complete in AWS takes well over 45 minutes to complete in our physical datacenter.
We have made sure the linux VM config matches up to what we have in AWS.
The only unknown here is what the optimized configs need to be for ESXi and the back end storage system which is presented thru NFS over a 10gb interconnect.
The storage system is Netapp running Data Ontap version 8.1.2 7-MODE
Anyone have any idea on how to go about troubleshooting this issue?
We have an open case with vmware and they have been online while we do the process and have confirmed that everything looks kosher on the vmware end.
Is there a tool we can use to test IO (latency/performance) between the ESXi host themselves and Netapp?
Or does anyone have a better way of trying to figure out where the bottleneck may be?
I would hate to have to tell the app owners that we can't figure out the issue and to keep the app in AWS.
Any help/direction will be much appreciated.
cheers