@Regnav,

The following is an answer from Dell’s product and support team regarding the issue you’re asking about.

Unlike some vendors that use commodity hardware for storage, redundant hardware components are used in EqualLogic arrays to help prevent such an occurrence of complete array failure. They are designed for 5-9s availability. Redundant component failures mean that you are having a really really bad day, and it is probably caused by much larger factors like an earthquake, natural disaster, power outage, etc. that are likely to affect your entire datacenter or campus. In the highly unlikely event that an array does go down, the data volumes that are contained on it go off-line (become inaccessible) until the array is back on line. Data is NOT corrupted or lost, you just can’t get to it. (It doesn’t matter what RAID type you had on which array). The same thing occurs if any other dual controller, highly redundant storage array (Dell/EMC Clariion, HP EVA…) goes off-line, the LUNS or volumes on that array become inaccessible.

The bottom line is that customers wishing to protect against this remote possibility should have a DR site established (either their own or a hosted site) and replicate their data to it safe keeping and recovery operations...
[Continued on 2nd reply]

-Dylan Locsin, Dell Storage

[Continued from previous reply @ Regnav]

I don’t know of any customers that have just “lost” an array due to inexplicable failure of redundant hardware for no reason. Rather, I have heard customers’ stories about the reliability of our products throughout disasters that blew my mind. One literally had water pouring through the roof of their data center into their racks and the EqualLogic arrays still ran. Another customer had an explosion at their datacenter which knocked almost all of their IT equipment offline – except the EqualLogic arrays. These extremely bad situations show that life’s common disasters, such as hurricanes, faulty facilities or power problems are likely to impact your entire data center, site or business, not just one piece of equipment, and you should protect your business from these threats with Disaster Recovery/Business Continuity measures, probably involving remote replication.

(Note, we do NOT guarantee our products will work through such disasters, and you definitely should NOT try to test any IT gear with such conditions, but I bet the customers that lived through them believe there’s value in purpose-built storage hardware and value in investing in off-site data protection.)

To your final question on removing disk drives – if there is a catastrophic failure of a chassis, but the drives are fine, yes, you can place those drives in a replacement chassis with controllers running a compatible firmware version, and the array will boot up and it will serve the data, as it was before. All the data and the configuration for an array is kept on the drives. That said, you should not move disk drives between EqualLogic arrays – unless support tells you to. If you’re in any sort of a recovery situation or want to prepare for one, contact Dell’s support team for EqualLogic products. They are experts and can walk you through the best practices.

Thanks,
Dylan Locsin, Dell Storage

Is it possible to prevent the group or volume from automatically allocating across arrays? Strange things happen in Iraq and we can't afford the downtime waiting for a replacement chassis.

Administrators can guide the EqualLogic system’s load balancing by making decisions on pool composition, what pool a volume is placed in, and by setting Volume RAID preferences to adjust the placement of data.

That said, EqualLogic arrays have purpose-built, highly available hardware, designed for 5-9s reliability, with hot-swappable components, but if you’re talking about high-risk environments, I’d suggest that the only practical way to assure your data is safe is to have copies available for recovery, whether off-site or locally. Consider replication. There are much higher probabilities that you’re going to have something happen to the environment around your EqualLogic array and your data center at large – building on fire, etc. than having an external event impact just 1 array in your group.