Configuration and Setup
Many considerations go into setting up an EVO, but the most important thing is to plan your configuration. At its heart, EVO provides a highly flexible environment for sharing storage with multiple users. Take your workflow into consideration and think through the requirements that your storage system will need.
Note
Keep in mind that we offer assisted install services. Contact us at any time to arrange one of these services.
Planning the Configuration
Does your workflow require every user to have a “personal” logical disk that they use to store all of their projects? Does your workflow have users pulling files from one logical disk, working on them, and saving the output to another? Does it contain a centralized repository that everyone uses to pull source media? Do some users require higher-speed access than others? Are there different applications in use that put different demands on storage?
All of these questions should be thought through before starting the configuration of Disk Pools, Logical Disks, and the options for sharing and connecting them with other networked devices.
Disk size and type
EVO uses enterprise-class SSD or HDD drives. At the time of release, 4TB or 6TB HDDs are the most common choice, but 8TB, 10TB, and 12TB disks are also available. EVO rack-mounted systems can also use SAS disks.
Note
Disks are provided by SNS; contact us if interested in adding/upgrading storage capacity.
Disk Pool Considerations
EVO ships with and is frequently deployed with all disks comprising a single pool, which is ideal for most standard workflows. If planning an alternative deployment, it’s worth considering the following things:
Bandwidth Protection All physical disks in a drive pool are accessed when any logical disk in that pool is accessed. The bandwidth of the drives in one pool is protected from the bandwidth of drives in another. If there are different applications that may use the disks differently, or if different users may be accessing the same disks at the same time, overall performance may be gained by dividing the storage into more, smaller pools with discrete resources.
Spindle Count The overall bandwidth of a disk pool is a function of how many drives are in that pool. Configuring EVO with fewer pools, and more drives in each pool, increases the potential bandwidth for any given logical disk in the pool. Applications with very high bandwidth requirements, such as uncompressed video, may benefit from fewer, larger pools.
Logical Disk Size No logical disk can be larger than the pool that contains it. If you need a few very large logical disks for your workflow, you will need fewer, larger pools.
Expansion Storage You may get more capability for concurrent work by adding one (or more, depending on EVO model) expansion chassis. The performance is dependent on current and future system configuration, and the kind of work you do. For example, if your performance bottleneck is not the number of disks in your current EVO, then performance will not be increased simply by adding more disks.
A typical configuration is a single pool per chassis (4-disk pool in a Prodigy, 8-disk pool in an 8 Bay, 16-disk pool in 16 Bay or Nearline, 16-disk pool in each expansion chassis).
Note
A pool may consist of up to 32 disks.
Warning
If a pool spans more than one chassis, the integrity of the pool depends on the external cable connection and the independent power supply for each unit. It’s strongly recommended in this case that each chassis share a common power source to ensure all member disks are online at the same time.
Logical Disk Considerations
EVO supports different combinations of RAID levels and logical disk sizes. Your configuration may benefit from several, smaller sized logical disks or fewer, larger ones. No one method is correct for all users.
Here are things to consider when planning your logical disk configuration:
Disk Pool Placement - As described in Disk Pool Considerations, the placement of logical disks within a disk pool can change that logical disk’s performance, its available bandwidth, and its potential size. If your needs are best met with multiple drive pools, consider which pools are the best fit for the required logical disks.
RAID levels - EVO supports logical disks that use a virtualized form of RAID aggregation and protection. The system is tuned for performance at any RAID level, so preconceptions about performance differences between different RAID types may not apply to EVO RAID logical disks.
Note
The main trade-off with the available RAID levels is the balance between data protection and usable space. RAID protection should never be considered as an alternative to a backup plan. A RAID may protect data against one or more disk drive failures but not other situations, such as a user’s accidental deletion of important files.
RAID 5 - This is the default RAID level for 8 Bay EVO and EVO Prodigy. Data is written across all disks in the pool, including enough redundancy to withstand the loss of one physical disk. If a physical disk fails, the array operates in a degraded mode until it’s been replaced and parity is restored.
The space required for the parity data equals the capacity of one physical disk. An 8 bay with a single RAID 5 logical disk affords the usable capacity of seven physical disks.
RAID 6 - This is the default RAID expectation for 16 Bay (and larger) systems. Twice the parity is distributed across all disks in the pool, similar to RAID 5. The benefit of this extra parity information is that any two disks in a pool can fail while still allowing access to the logical disk(s). The extra parity calculation adds writes compared to RAID 5, while reads are unaffected or improved according to the number of disks available. A RAID 6 deployment using a single pool also allows for a single namespace for all its available content.
The space required for the parity data equals the capacity of two physical disks. A 16-disk RAID 6 affords the usable capacity of 14 disks.
Warning
Although EVO provides many features for reliability and redundancy, including RAID, it is imperative that you maintain frequent backups of all your critical data on at least one separate device. The redundancy built into any one system is insufficient to prevent against all potential disasters. A comprehensive backup/restore plan, well-followed procedures, and secondary backups preserved in an offsite location, are all critical to business continuity for any company with important data assets!
Logical Disk Size
EVO logical disks must be at least one gigabyte, and may be as big as your drive pool. Consider your workflow and the overall amount of usable space in your drive pool(s) when choosing logical disk size. A workflow that requires each user to have their own space to read from and write to may benefit from more, smaller logical disks. Large shared logical disks can be a better fit for a workflow where multiple users are working with the same source content.
How Many Bytes in a Gigabyte?
There is some disagreement in the industry over the exact definition of words like terabyte, gigabyte, megabyte, and kilobyte. This confusion stems from the early computer industry. The use of binary in computer systems leads to a preference for numbers that are powers of two. Two to the tenth power is 1024.
Humans think in base-ten, so the prefix kilo- (meaning thousand) was adopted to refer to this value, since it was considered close enough to one thousand.
As capacities increased, the difference has compounded. Using powers of two, one gigabyte is 1073741824 bytes. In an effort to resolve the disparity, the IEC created terms like kibibyte, mebibyte, gibibyte and tebibyte, where the second part of the prefix name is replaced by “bi” to signify that they are binary prefixes rather than decimal (base-10) prefixes.
Disk drive manufacturers use metric prefixes for marketing purposes (a 1 “terabyte” drive need only contain 1000000000000 bytes instead of 1099511627776, a 10% difference).
As a result of this, a “1TB” drive actually contains 931.32 Gigabytes as reported by most operating systems.
Note
EVO uses binary reporting for accuracy, so slightly differing disk statistics may be expected when comparing server and workstation volume capacity reports.
Connectivity Considerations
EVO supports copper and optical Ethernet, and larger chassis can accommodate many client connections without the need for a physical switch. When a switch is available, interfaces can also be aggregated so the switch can efficiently manage traffic distribution between clients.
If greater-than-1GbE connections are used, overall throughput may be increased by using jumbo frames. A greater MTU (maximum transmission unit) value provides efficiency for transmission of larger (e.g. video) files, since fewer, larger frames are transmitted.
As workstations generally do, EVO’s network adapters default to the standard frame size (1500 MTU) and if changed to jumbo (9000 MTU), then all ports in the connection path must match (EVO, switch, workstation). Any port can use standard frame sizes, while jumbo frames are out of spec for GbE.
While jumbo frames are often beneficial for greater-than-1GbE connections in a local environment, they should not be used for a gateway (internet) connection. Since all ports in the connection path would need to support jumbo frames, higher MTU should never be set for a port accessed by a remote user over the internet (VPN).
Gigabit Ethernet (GbE) - Most client workstations will likely offer one or more built in 1 Gbps Ethernet ports, which is sufficient for many workflows. In other environments, clients often connect their workstations directly via 10GbE or higher, while a GbE connection is reserved for LAN use, network time synchronization, email notifications, and so on.
10 gigabit Ethernet (10GbE) - A 10 Gbps Ethernet connection can be used to connect workstations directly, or in conjunction with the proper Ethernet switch to offer connectivity to multiple clients, or as an aggregate connection for multiple clients. This may be copper or optical Ethernet.
25/50 gigabit Ethernet (25GbE, 50GbE) - These adapters are optical-only, use specific cables/transceivers, and may require additional switch configuration.
Warning
Whether frames are set to standard or jumbo is not as important as ensuring the frame size is supported by sender, receiver, and every device between the two. This is an especially important consideration for remote connections.
Future Growth
EVO is designed to be modular in order to fit any situation, but your requirements may change over time. Perhaps some more editors have joined your team, or perhaps you land a big project that will require more storage than your EVO currently has. As long as your EVO has more available PCIe slots, you can add another Ethernet card to add more client ports.
Similarly, as long as your EVO has empty drive bays, you can purchase an additional quad of disks to add to the system. If your 8 Bay or 16 Bay EVO has used all of its internal drive bays, it is possible to add one or more expansion chassis that will each provide 16 more drive bays.
It is also possible to expand the size of existing logical disks by adding one or more disks to an existing pool. This type of in-place grow operation has caveats to consider (the RAID level is unchanged, for example), so consult with your reseller and/or SNS to ensure understanding of the implications for future storage expansion in your case.
EVO is a flexible and powerful system, and there is no single configuration that is right for all EVO users. Studio Network Solutions Product Specialists and our informed resellers are available to help with any and all planning. After careful consideration, you should have a plan for the following items:
Disk Pools - How many pools, and number of physical disks in each pool
Logical Disk(s) - How many, the size of each, the RAID level of each, and which disk pool(s) will be used
Connectivity - Which users will connect via Ethernet? Will multiple zones or VLANs be used to manage network bandwidth?
Now that you have a plan, let’s proceed with the actual configuration tasks.