SPECsfs2008_nfs.v3 Result

Oracle Corporation : Oracle ZFS Storage ZS3-4
SPECsfs2008_nfs.v3 = 450702 Ops/Sec (Overall Response Time = 0.70 msec)


44560 0.4
89370 0.5
134129 0.5
179075 0.6
224055 0.7
269180 0.7
319852 0.8
366649 0.9
407969 1.2
450702 1.4
Performance Graph

Product and Test Information

Tested By Oracle Corporation
Product Name Oracle ZFS Storage ZS3-4
Hardware Available September 2013
Software Available September 2013
Date Tested May 2013
SFS License Number 6
Licensee Locations Broomfield, CO USA

The Oracle ZFS Storage ZS3-4 is a high-performance storage system that offers enterprise-class NAS and SAN capabilities with industry-leading Oracle Database integration, in a cost-effective high-availability configuration. The Oracle ZFS Storage ZS3-4 offers simplified set up and management combined with industry-leading storage analytics and a performance-optimized platform that uses specialized Read and Write Flash-biased SSD caching devices. The Oracle ZFS Storage ZS3-4 can scale to 2 TB Memory, 80 CPU cores, and 3.4 PB capacity, with up to 12.8 TB of Flash Cache in a high-availability configuration. Oracle ZFS Storage Appliances deliver additional economic value bundled data services such as file- and block-level protocols including connectivity over InfiniBand, Compression, Deduplication, Thin provisioning, DTrace Analytics, Virus Scan, Snapshots, Triple Mirror, Triple Parity RAID, Phone-home, NDMP, Clustering, etc.

Configuration Bill of Materials

Item No Qty Type Vendor Model/Name Description
1 2 Storage Controller Oracle Oracle ZFS Storage ZS3-4: controller part# 7105725 Oracle ZFS Storage ZS3-4: controller
2 64 Memory Oracle 16 GB DDR3-1066 DIMMs part# 7105053(Qty:2) 16 GB DDR3-1066 DIMMs (for factory installation).Ordered in Quantity of 2
3 4 10 Gigabit Ethernet Adapter Oracle Sun Dual 10GbE SFP+ PCIe 2.0 Low Profile adapter part# 1109A-Z Sun Dual 10GbE SFP+ PCIe 2.0 Low Profile adapter incorporating Intel 82599 10 Gigabit Ethernet controller and supporting pluggable SFP+ Transceivers. ROHS-5. ATO option
4 8 Short Wave Pluggable Transceiver Oracle 10Gbps Short Wave Pluggable Transceiver (SFP+) part# 2129A Dual rate transceiver: SFP+ SR. Support 1 Gb/sec and 10 Gb/sec dual rate (for factory installation)
5 20 Storage Drive Enclosure Oracle Sun Disk Shelf SAS-2 part# DS2-0BASE Sun disk shelf: base chassis with 2 SAS-2 I/O modules, 2 AC PSUs and 2 cooling fans (for factory installation)
6 472 Disk Drives Oracle Disk Drives 300GB 15K RPM SAS-2 HDD part# 7101274 SAS-2 Disk Drives 300GB 15K RPM HDD
7 8 SSD Drives Oracle SAS-2 73GB 3.5-inch SSD Write Flash Accelerator part# 7101197 SLC SAS-2 SSD 3.5-inch write flash accelerator with stingray bracket (for factory installation)
8 4 SAS-2 Host Bus Adapter Oracle SAS-2 16-Port 6Gbps HBA part# 7103790 SAS PCIE 6Gbs 16 port (for factory installation)

Server Software

OS Name and Version Oracle ZFS Storage OS8
Other Software None
Filesystem Software ZFS

Server Tuning

Name Value Description
arc_shrink_shift 5 Control of arc shrink.

Server Tuning Notes

Download My Oracle Support document # 1577276.1 and apply included workflow. (Sets arc_shrink_shift to 5.)

Disks and Filesystems

Description Number of Disks Usable Size
300GB SAS 15K RPM Disk Drives 472 62.3 TB
SSD 73GB SAS-2 Write Flash Accelerator 8 584.0 GB
500GB SATA 7.2K RPM Disk Drives Each controller contains 2 of these disk drives and are mirrored and are not used for cache data or data storage. 4 899.0 GB
Total 484 63.8 TB
Number of Filesystems 216
Total Exported Capacity 60.80 TB
Filesystem Type ZFS
Filesystem Creation Options default
Filesystem Config 216 ZFS file systems
Fileset Size 52155.8 GB

Both controllers are setup with 8 pools total (4 pools per controller). Each of the controller's pools are configured with 56 disk drives, 1 write flash accelerator and 3 spare disk drives. The pools are also set up to mirror the data (RAID1) across all 56 drives. The write flash accelerator in each pool is used for the ZFS Intent Log (ZIL). All pools are configured with 27 ZFS file systems each. Since each controller is configured with 4 pools and each pool contains 27 ZFS file system, in total each controller has 108 ZFS file systems. Both of the controllers together in total have 216 ZFS file systems. There are 2 internal mirrored system disk drives per controller and are used only for the controllers core operating system. These drives are not used for data cache or storing user data.

Network Configuration

Item No Network Type Number of Ports Used Notes
1 10 Gigabit Ethernet 4 Each controller has 2 Dual port 10 Gigabit Ethernet cards using only a single port on each and using Jumbo Frames

Network Configuration Notes

There are 8 ports total but 4 are active at a time for high availibility. The MTU size is set to 9000 on each of the 10 Gigabit ports.

Benchmark Network

Each controller was configured to use a single port of two dual port 10GbE network adapters for the benchmark network. All of the 10GbE network ports on all the load generators systems were connected to the Arista 7124SX switch that provided connectivity.

Processing Elements

Item No Qty Type Description Processing Function
1 8 CPU 2.4GHz 10-core Intel Xeon(tm) Processor E7-4870 NFS, ZFS, TCP/IP, RAID/Storage Drivers

Processing Element Notes

Each Oracle ZFS Storage ZS3-4 controller contains 4 physical processors, each with 10 processing cores.


Description Size in GB Number of Instances Total GB Nonvolatile
Oracle ZFS Storage ZS3-4 controller memory 1024 2 2048 V
Grand Total Memory Gigabytes     2048  

Memory Notes

The Oracle ZFS Storage ZS3-4 controllers' main memory is used for the Adaptive Replacement Cache (ARC) the data cache and operating system memory.

Stable Storage

The Stable Storage requirement is guaranteed by the ZFS Intent Log (ZIL) which logs writes and other file system changing transactions to either a write flash accelerator or a disk drive. Writes and other file system changing transactions are not acknowledged until the data is written to stable storage. Since this is an active-active cluster high availability system, in the event of a controller failing or power loss, the other active controller can take over for the failed controller. Since the write flash accelerators or disk drives are located in the disk shelves and can be accessed via the 8 backend SAS channels from both controllers, the remaining active controller can complete any outstanding transactions using the ZIL. In the event of power loss to both controllers, the ZIL is used after power is restored to reinstate any writes and other file system changes.

System Under Test Configuration Notes

The system under test is a Oracle ZFS Storage ZS3-4 cluster set up in an active-active configuration.

Other System Notes


Test Environment Bill of Materials

Item No Qty Vendor Model/Name Description
1 2 Oracle Sun Fire X4270 M2 Sun Fire X4270 M2 with 144GB RAM and Oracle Solaris 11.1
2 1 Arista 7124SX Arista 7124SX 24-port 10Gb Ethernet switch

Load Generators

LG Type Name LG1
BOM Item # 1
Processor Name Intel Xeon(tm) X5680
Processor Speed 3.3 GHz
Number of Processors (chips) 2
Number of Cores/Chip 6
Memory Size 144 GB
Operating System Oracle Solaris 11.1
Network Type 10Gb Ethernet

Load Generator (LG) Configuration

Benchmark Parameters

Network Attached Storage Type NFS V3
Number of Load Generators 2
Number of Processes per LG 432
Biod Max Read Setting 2
Biod Max Write Setting 2
Block Size 64

Testbed Configuration

LG No LG Type Network Target Filesystems Notes
1..2 LG1 1 /export/sfs-1../export/sfs-216 None

Load Generator Configuration Notes

File systems were mounted on all clients and all were connected to the same physical and logical network.

Uniform Access Rule Compliance

Every client used 432 processes. All 216 file systems are mounted and accessed by each client and evenly divided amongst all network paths to the Oracle ZFS Storage ZS3-4 controllers. The file systems data was evenly distributed on the backend.

Other Notes

Other Test Notes: None

Config Diagrams

Generated on Mon Sep 09 18:45:30 2013 by SPECsfs2008 HTML Formatter
Copyright © 1997-2008 Standard Performance Evaluation Corporation

First published at SPEC.org on 09-Sep-2013