White Papers
Dell HPC NFS Storage Solution - High Availability (NSS-HA) Configuration with Dell PowerVault
MD3260/MD3060e Storage Arrays
11
4. Evaluation
The architecture proposed in this white paper was evaluated in the Dell HPC lab. This section describes
the test methodology and the test bed used for verification. It also contains details on the functionality
tests. Performance tests and results follow in Section 5.
4.1. Method
The NFS Storage Solution described in this solution guide was tested for HA functionality and
performance. A 360TB NSS4.5-HA configuration was used to test the HA functionality of the solution.
Different types of failures were introduced and the fault tolerance and robustness of the solution was
verified. Section 4.3 describes these HA functionality tests and their results. HA functionality testing
was similar to the work done in the previous versions of the solution
(3)
.
4.2. Test bed
The test bed used to evaluate the NSS4.5-HA functionality and performance is shown in Figure 3.
A 64 node HPC compute cluster was used to provide I/O traffic for the test bed.
A pair of Dell PowerEdge R620 servers were configured as an active-passive HA pair and function as
a NFS server for the HPC compute cluster (also called the clients).
Both NFS servers were connected to a shared Dell PowerVault MD3260 storage enclosure extended
with one Dell PowerVault MD3060e storage enclosure (Figure 3 shows a 360TB solution with two
PowerVault MD storage arrays) at the backend. The user data resided on an XFS file system created
on this storage. The XFS file system was exported using NFS to the clients.
The NFS servers were connected to the clients using the public network. This network was either
InfiniBand or 10GbE Ethernet.
For the HA functionality of the NFS servers, a private Gigabit Ethernet network was configured to
monitor server health and heartbeat, and to provide a route for the fencing operations using a
PowerConnect 5524 Gigabit Ethernet switch.
Power to the NFS servers was driven by two APC switched PDUs on two separate power buses.
Complete configuration details are provided in Table 3, Table 4, Table 5, and Table 6.