ESG Lab Review: HP P4000: Affordable, Scalable, Reliable Storage Hyper-Redundant Clustered Storage


The HP P4000 SAN extends the fault tolerance provided by traditional modular disk arrays using a combination of clustering and Network RAID technologies. Network RAID slices data and distributes it over multiple nodes in a P4000 cluster. Up to four copies can be created and the cluster can be “stretched” within a campus or metropolitan network to survive a data center level outage. As a result, a P4000 SAN can be configured to survive the failure of one or more drives, one or more network interfaces, and one or more nodes in a P4000 cluster. HP refers to this capability as hyper- redundant clustered storage.


As shown in Figure 6, a number of errors were introduced to test the fault tolerance of a four node HP P4000 cluster. A node failure and a network failure were injected as host IO was running against a hyper-redundant Network RAID-10 volume and an unprotected Network RAID-0 volume. Multiple drive failures were tested during a previous ESG Lab validation.


Figure 6. Error Injection Testing


7


The hyper-redundant Network RAID-10 volume was accessed from a Windows 2008 server as the E: drive. The unprotected Network RAID-0 volume was accessed as the F: drive. Files were copied to both volumes as catastrophic errors were introduced. The power cables were unplugged from one of the P4000 nodes in the four node SAN cluster. As expected, the unprotected RAID-0 volume stopped working and was no longer accessible from Windows Explorer. The RAID-10 volume remained accessible. The failure was noted on the SANIQ management console approximately 10 seconds after the error had been introduced. The cluster automatically healed itself and the F: drive became usable again after the failed node was powered back on. The test was repeated using a pulled network interface card between the server and one of the P4000 nodes. As expected, the RAID-0 volume was inaccessible until after the network cable had been plugged back in and the RAID-10 volume was available for the duration of the test.


During a previous ESG Lab validation of the P4000 SAN, the system remained fully available as a series of errors was introduced on a stretched cluster configured within labs located on different floors within the same building. ESG Lab:


 Pulled a team bonded redundant Ethernet back-end interface;  Pulled two drives;  Powered off the entire upstairs half of the cluster;  Re-plugged Ethernet;  Powered up the entire upstairs half of the cluster;  Replaced the two failed drives.


Why This Matters


As virtual server and storage deployments grow in size and complexity, so too do the consequences of things going wrong. Regardless of the number and types of hardware failures that may occur during the lifecycle of electronically stored information assets, employees, managers, and customers expect that their data will always be available. ESG Lab as confirmed that the hyper-redundant cluster storage architecture of the HP P4000 SAN can be used to create a self-healing virtual storage solution that never needs to be turned off.


© 2010 Enterprise Strategy Group, Inc. All Rights Reserved.


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8