Coping with increased capacity – preventing RAID failure
Posted on Friday, October 6, 2017
The race for increased capacity has seen magnetic disks breach the 12TB barrier for the first time. By replacing existing spinning disks with these larger drives, array capacity (and lifespan) can be increased for very little outlay.
There is a problem however. Every increase in drive capacity also increases the time required to rebuild an array after a failure. The longer this rebuild process takes, the greater the impact on system performance.
If your business is to take full advantage of these new, larger disk drives, you need to dramatically reduce the risk of array failure in the first place. Here are three key tips to assist:
1. Monitor drive health closely
Your RAID controllers all monitor drive health, providing warnings about potential failures. Your maintenance routines need to include a check on all drive alerts, providing early warning of potential failures.
2. Pull drives early
When the RAID controller detects a potential fault, pull the drive and replace it immediately. You may be tempted to ignore the warning, or to reinsert the same drive – don’t. Doing so simply delays the inevitable failure of the drive and a longer, more drawn out array rebuild.
3. Deploy disks strategically
Arrays are often built using a sequence of disks deployed in the same order they left the production line. Inserting these disks in serial number order, across a single shelf and RAID group, you risk magnifying the effect of a manufacturing defect on the array.
Instead, you should use disks from different production runs to minimize these risks.
Get help with your arrays
Although these tips are quite simple, they are extremely hard to manage in a massive data warehouse environment. Instead, you should look to partner with a vendor-agnostic maintenance and support provider like CDS who can perform the necessary reconfigurations allowing you to deploy large capacity drives in your existing arrays, and helps to reduce the risk of RAID failure.
To learn more about CDS can help you, please get in touch.
AI comes to enterprise storage management. Soon.
Hitachi is working to develop AI systems that will help CTOs better manage their enterprise storage arrays.
Increasing NAND flash costs affect disk array sales
Morgan Stanley estimates suggest businesses are delaying flash storage purchases as NAND costs increase.