MemFS v2 û A Memory-based File System on HP-UX 11i v2

ManualsBrandsHP ManualsSoftwareHP-UX 11i v2 Foundation OE LTU

MemFS v2 – A Memory-based File System on HP-UX 11i v2

Technical white paper

Introduction......................................................................................................................................... 2

Why a Memory-based File System? ....................................................................................................... 2

The HP-UX Memory-based File System Design ......................................................................................... 3

MemFS user process backing store ........................................................................................................ 4

Mount options ..................................................................................................................................... 4

Performance........................................................................................................................................ 5

Postmark Benchmark ........................................................................................................................ 5

Connectathon Tests .......................................................................................................................... 7

SDET Benchmark.............................................................................................................................. 8

MemFS mount/umount performance................................................................................................... 9

Configuration Guidelines...................................................................................................................... 9

Recommended Use ............................................................................................................................ 11

Limitations and Future Work................................................................................................................ 11

Summary .......................................................................................................................................... 11

Summary of content (11 pages)

PAGE 1
MemFS v2 – A Memory-based File System on HP-UX 11i v2 Technical white paper Introduction......................................................................................................................................... 2 Why a Memory-based File System? ....................................................................................................... 2 The HP-UX Memory-based File System Design .........................................................................................
PAGE 2
Introduction A Memory-based File System (MemFS) is a file system that resides in memory. It does not normally write data out to stable storage. A MemFS is created from a mount operation, and ceases to exist when it is un-mounted. The purpose of such a file system is to provide fast access for temporary files that do not need to be kept for an indeterminate time. Because it does not normally have to do I/O to stable storage, the MemFS is able to provide extremely high throughput.
PAGE 3
Since the system sees a RAM disk as a device rather than a file system, any access to a file stored on it results in a second copy of the data being kept in the file system buffer. Some of the implementations of a RAM disk may provide an improved interface and some of these shortcomings may not be present. As attractive as the performance aspects sound, a memory-based file system only works efficiently under certain circumstances. The biggest disadvantage is that it does not preserve data across mounts.
PAGE 4
memory. The mount results in a combination of an mkfs(1M) and a mount(1M) without the disk operations. Subsequent to a mount, the MemFS can be used just like any other file system, with the same commands and system calls. From an application or user perspective there should be nothing that differentiates it from any other file system, except a better response time. Internally, MemFS behaves differently from other file systems.
PAGE 5
Also, note that the –r option is not supported. Since a read-only MemFS will always remain empty, this option, applied to a MemFS, serves no purpose. Other HFS options such as behind, delayed, fs_async, no_fs_async are not supported since they do not have any relevance in a Memory-based File System context. MemFS does not support quotas. The file-system specific option that MemFS supports is: size – Specifies the maximum size to which this MemFS instance is allowed to grow.
PAGE 6
quit RUN_LOAD={1000, 10000, 25000, 75000, 125000, 300000, 500000} The filesystems under test were: • tmpfs on Red Hat Linux 2.4.21-4 • MemFS on HP-UX 11iv2 of size=2GB. dbc_max_pct was set to 50% (4GB) • RAMdisk on HP-UX 11iv2: Basefile system: VxFS 3.5 using the delaylog mount option. • VxFS 3.
PAGE 7
directory lookups in HFS code that MemFS uses. tmpfs is not affected by the number of directory entries.
PAGE 8
The system under test consisted of a HP rx5670 server comprising of 2 x 900MHz Itanium CPUs, 8GB RAM. Figure 4: Connectathon benchmark results Hfs VxFS Ramdisk MemFS (sec) Linux ext3fs (sec) Linux tmpfs (sec) Default (sec) Log (sec) Delaylog (sec) VxFS (sec) Create 2501.46 699.41 23.48 6.44 4.91 197.51 0.82 Remove 989.77 677.42 8.54 5.38 3.82 26.88 0.91 Lookup across mounts 0.0 0.0 0.0 0.3 0.0 0.0 0.0 Setattr 0.0 6.0 0.1 0.1 0.0 0.0 0.0 Write 0.23 0.54 0.52 0.
PAGE 9
Figure 5: SDET Benchmark results SDET Results 10000 Transactions/sec 9000 8000 MemFS 7000 VxFS 6000 HFS 5000 RAMdisk/VxFS 4000 Linux/tmpfs 3000 Linux/ext2fs 2000 1000 0 50 0 10 80 60 40 20 1 0 Concurrent workload The following conclusions can be drawn: • MemFS performs better on the SDET benchmark compared to disk based filesystems and RAMdisk • MemFS performs better on the SDET benchmark compared to Linux’s tmpfs on the same test machine MemFS mount/umount performance Since MemFS
PAGE 10
A MemFS instance can be created either through a mount(1) command, or automatically through an entry in the /etc/fstab file. Creation of a MemFS instance through the mount(2) system call is not supported.
PAGE 11
Recommended Use Applications that will benefit most from MemFS are those that perform mostly meta-data intensive operations like creates and deletes of small files and directories. For example: compilers, editors and sorting applications. MemFS never writes out meta-data for files, so there is always a performance gain for directory and file manipulation. Applications can also interact with MemFS files using the mmap interface.