IT departments must ensure their customers, both internal and external, experience high availability and rapid responses from applications. When storage bottlenecks lead to slowdowns, however, customers end up dissatisfied. So let’s hone in on what can go wrong and what to monitor to get operations back on track as quickly as possible. Here are the top three storage issues to think about when planning storage and troubleshooting issues.
Virtual Storage Gone Wild
Storage virtualization, which offers the ability to pool storage devices and treat them as one, is offered to reduce complexity and costs. However, it is not without challenges.
Without a storage monitoring solution, you have no visibility to the volume of traffic at different times. That makes it difficult to slice up your storage array or subsystem among a variety of virtual machines (VMs). Like a highway where traffic flows smoothly most of the day, but slows to a crawl during rush hour, you might have a slowdown when everyone logs in when they get to work. If you’re not monitoring your storage, however, you won’t know about it until users complain. And if you believe in good service, that’s too late.
You cannot simply look at the number of VMs and allocate the same number to each LUN because it tells you nothing about the amount of activity the VMs generate. Some are always moving data around while others just keep the data safe and mostly in one place.
To ensure quick response at all times, you need data to guide you when assigning your VMs to a LUN. With an IT infrastructure performance management solution that monitors activity levels, you can balance the load on storage devices by marrying active and inactive VMs. You also need your solution to monitor your systems around the clock so that you can identify spikes in demand. Also, it must provide information on all your storage devices regardless of brand, allowing you to see all the data in one place when you divide up LUNs. Finally, you want a holistic view of your storage, including subsystems, controllers, storage arrays, and logical drives.
Storage Starved Applications
A similar issue can occur with applications. Some applications work harder than others. For instance, with 30 million monthly users, Amazon’s mobile app is on the go at all times. Not only that, but it’s also mission critical. If an application like this is placed incorrectly within a storage array, it’ll slow down, and users will abandon it.
Again, you need to monitor the activity level associated with applications so that you can plan your storage in a way that ensures response times meet your organization’s needs. Your monitoring solution should give you information on your system configuration and total resource capacity available to your applications, including disk space. Also, other parts of your infrastructure affect storage, so you want to understand server CPU and memory usage levels.
Storage is not a one-speed fits all solution. We now have tiering. Tier zero is for transactional data that needs to move without delay. Tier two is for data that can move at a more leisurely pace, such as backup and archived data. Finally, tier one is for data that falls somewhere in between.
While tier zero is the fastest, it’s also the most expensive. So to get the most for your money, you need to plan your storage according to your requirements for speed in accessing the data. If the foundational planning does not occur, an e-commerce website might end up on tier one, which is generally too slow to satisfy customers who expect immediate gratification. Plus, after implementing your storage plan, you need to continue to keep an eye on it because it will not remain static.
For tiering, you’ll need a monitoring solution that looks at IOPs and latency. Also, because server CPU and memory issues can interfere with performance, make sure it also tracks and measures them.
In summary, to avoid storage problems, it’s essential to understand your needs. For instance, make sure you know the activity levels of your VMs and applications before you allocate storage. Decide how quickly data has to move from storage to user. After you’ve implemented your storage plan, continue to monitor your IT infrastructure to make sure storage devices and the systems that support them remain healthy and continue to meet user needs.