EMC FAST and VMware SDRS I/O Metric – Redundant Employee

In my earlier post I talked about how VMware SIOC and EMC FAST VP goes hand in hand.

Now while I have been working on that same large scale deployment we were proposing to the customer that while we should employ EMC FAST VP and VMware Storage DRS, we need to disable the IO Metric based calculation in SDRS. So, in this was we would not employ both of them doing the same job. Now let us see what is does in the back end.

SDRS triggers action on either capacity and/or latency. Capacity stats are constantly gathered by vCenter, default threshold is 80%. I/O load trend is evaluated (default) every 8 hours based on a past day history, default threshold is 15ms. With that meaning that the Storage DRS algorithm will be invoked when these thresholds are exceeded. Now in the case of “utilized space” this happens when vCenter collects the datastore statistics and notices the threshold has been exceeded in the case of I/O load balancing.

Every 8 hours, Storage DRS will evaluate the I/O imbalance and will make recommendations if and when the thresholds are exceeded. Note that these recommendations will only be made when the difference between the source and destination is at least 5% and the cost/risk/benefit analysis has a positive result.
I/O Latency is the datastore response time measured in miliseconds. The default value is 15ms, which means that the datastore will not consider relocations until the datastore exceeds 15ms response time, and the imbalance rules are also satisfied.

In some cases it might be best to avoid the latency assessment because of variability in the application workload or storage configuration. for example, auto-tiered storage can provide varied I/O response times depending on the class of storage used by the tiered device. Additionally, tiered storage is designed to migrate blocks at a sub-LUN level. Since both SDRS and FAST address the same problem, although at a slightly different level, use only one to perform the migrations to avoid duplicate relocation operations, Do not allow both SDRS and FAST to perform automated relocations.

When using EMC FAST VP, use SDRS, but disable I/O metric here. This combination gives you the simplicity benefits of SDRS for automated placement and capacity balancing.

So, thumb rule is DO NOT enable I/O metric for SDRS recommendations. Let that be done by EMC FAST at the storage layer.


About Prasenjit Sarkar

Prasenjit Sarkar is a Product Manager at Oracle for their Public Cloud with primary focus on Cloud Strategy, Oracle Openstack, PaaS, Cloud Native Applications and API Platform. His primary focus is driving Oracle’s Cloud Computing business with commercial and public sector customers; helping to shape and deliver on a strategy to build broad use of Oracle’s Infrastructure as a Service (IaaS) and Platform as a Service (PaaS) offerings such as Compute, Storage, Java as a Service, and Database as a Service. He is also responsible for developing public/private cloud integration strategies, customer’s Cloud Computing architecture vision, future state architectures, and implementable architecture roadmaps in the context of the public, private, and hybrid cloud computing solutions Oracle can offer.

3 thoughts on “EMC FAST and VMware SDRS I/O Metric – Redundant Employee

  1. Pingback: Virtualizing Business Critical Applications ? – Consider Designing EMC FAST Cache & FAST VP - Stretch Cloud - Technology Undressed

  2. Pingback: EMC FAST-VP & PowerPath/VE with Storage DRS & Storage I/O Control | vcdx133