| Skip Navigation Links | |
| Exit Print View | |
|
Sun QFS File System 5.3 Configuration and Administration Guide Sun QFS and Sun Storage Archive Manager 5.3 Information Library |
| Skip Navigation Links | |
| Exit Print View | |
|
|
Sun QFS File System 5.3 Configuration and Administration Guide Sun QFS and Sun Storage Archive Manager 5.3 Information Library |
2. About the Master Configuration File
Configuration Examples for Local File Systems
Simple File System Configuration Example
How to Configure a Simple File System
Round-Robin Configuration Example
How to Configure the System for Round-Robin Allocation
Local Striping Configuration Example
How to Configure the System for Local Striping
Configuration Example for a Shared File System on an Oracle Solaris OS Platform
How to Configure the Shared File System
Configuration Examples for Highly Available File Systems
How to Create an mcf File for a Highly Available File System
Configuration Example for a Shared File System on an Oracle Solaris Cluster Platform
How to Create an mcf File for a Shared File System in an Oracle Solaris Cluster Environment
4. Configuring the File System
5. Configuring a Shared File System
6. Administering File System Quotas
7. Advanced File System Topics
9. Configuring WORM-FS File Systems
11. Using QFS File Systems with SANergy (SAN-QFS)
Use the configuration examples in this section for configuring the mcf file for a file system to be installed on a single Oracle Solaris OS host.
For examples that you can use in an Oracle Solaris Cluster environment, see Configuration Examples for Highly Available File Systems.
This example shows how to configure two file systems using a server that has a SCSI-attached StorageTek Multipack desktop array from Oracle.
You can use the format command to determine how the disks are partitioned. Example 3-1 shows the format command output.
Note - Only the last lines of format command output are shown.
Example 3-1 format Command Output for Configuration Example
# format < /dev/null
Searching for disks...done
AVAILABLE DISK SELECTIONS:
0. c0t10d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/sbus@3,0/SUNW,fas@3,8800000/sd@a,0
1. c0t11d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>
/sbus@3,0/SUNW,fas@3,8800000/sd@b,0
2. c6t2d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@7,4000/SUNW,isptwo@3/sd@2,0
3. c6t3d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@7,4000/SUNW,isptwo@3/sd@3,0
4. c6t4d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@7,4000/SUNW,isptwo@3/sd@4,0
5. c6t5d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@7,4000/SUNW,isptwo@3/sd@5,0
6. c8t2d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@b,4000/SUNW,isptwo@3/sd@2,0
7. c8t3d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@b,4000/SUNW,isptwo@3/sd@3,0
8. c8t4d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@b,4000/SUNW,isptwo@3/sd@4,0
9. c8t5d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@b,4000/SUNW,isptwo@3/sd@5,0
Specify disk (enter its number):
# format /dev/rdsk/c6t2d0s2
.
.
.
Part Tag Flag Cylinders Size Blocks
0 unassigned wm 0 0 (0/0/0) 0
1 unassigned wm 0 0 (0/0/0) 0
2 backup wu 0 - 4923 8.43GB (4924/0/0) 17682084
3 unassigned wm 0 0 (0/0/0) 0
4 unassigned wm 0 - 1229 2.11GB (1230/0/0) 4416930
5 unassigned wm 1230 - 2459 2.11GB (1230/0/0) 4416930
6 unassigned wm 2460 - 3689 2.11GB (1230/0/0) 4416930
7 unassigned wm 3690 - 4919 2.11GB (1230/0/0) 4416930
Begin writing the mcf file for this configuration example by defining the file system and its disk partitions, as described in this procedure.
The finished mcf file defines the following two file systems:
The qfs1 file system, which is created on slice 4 of the following disks: c8t2d0 (metadata), c6t2d0 (file data), and c6t3d0 (file data).
The qfs2 file system, which is created on slice 5 of the following disks: c8t2d0 (metadata), c6t2d0 (file data), and c6t3d0 (file data).
The following code example shows the resulting mcf file.
# cat /etc/opt/SUNWsamfs/mcf # # Equipment Eq Eq Family Device Additional # Identifier Ord Type Set State Parameters #----------- --- ---- ------ ------ ---------- # *qfs1 10 ma qfs1 on* /dev/dsk/c8t2d0s4 11 mm qfs1 on /dev/dsk/c6t2d0s4 12 mr qfs1 on /dev/dsk/c6t3d0s4 13 mr qfs1 on # *qfs2 20 ma qfs2 on* */dev/dsk/c8t2d0s5 21 mm qfs2 on* */dev/dsk/c6t2d0s5 22 mr qfs2 on* */dev/dsk/c6t3d0s5 23 mr qfs2 on*
Add entries in the /etc/vfstab file for the qfs1 and qfs2 file systems that you defined in the mcf file. The last two lines in the code example below show entries for these new file systems.
For a description of the fields in the /etc/vfstab file, see Table 3–2.
# cat /etc/vfstab # device device file mount # to to mount system fsck at mount # mount fsck point type pass boot params # ----- ---- ----- ---- ---- ---- ------ fd - /dev/fd fd - no - /proc - /proc proc - no - /dev/dsk/c0t10d0s1 - - swap - no - /dev/dsk/c0t10d0s0 /dev/rdsk/c0t10d0s0 / ufs 1 no logging swap - /tmp tmpfs - yes - *qfs1 - /qfs1 samfs - yes stripe=1* *qfs2 - /qfs2 samfs - yes stripe=1*
This example illustrates the configuration of a file system called qfs3 that uses round-robin allocation on four disk drives.
This example assumes the following:
The metadata device is a single partition (s1) used on controller 8, disk 4.
The data devices consist of four disks attached to controller 6. Each disk is on a separate target (1—4).
This example introduces the round-robin data layout.
The following code example shows the mcf file for this round-robin disk configuration.
# cat /etc/opt/SUNWsamfs/mcf # # Equipment Eq Eq Family Device Additional # Identifier Ord Type Set State Parameters #----------- --- ---- ------ ------ ---------- # *qfs3 10 ma qfs3 on* /dev/dsk/c8t4d0s4 11 mm qfs3 on /dev/dsk/c6t2d0s4 12 mr qfs3 on /dev/dsk/c6t3d0s4 13 mr qfs3 on /dev/dsk/c6t4d0s4 14 mr qfs3 on /dev/dsk/c6t5d0s4 15 mr qfs3 on
Edit the /etc/vfstab file to explicitly set round-robin allocation on the file system by specifying stripe=0 in the mount params field. The following code example shows stripe=0 for the qfs3 file system.
# cat /etc/vfstab #device device file mount #to to mount system fsck at mount #mount fsck point type pass boot params #----- ---- ----- ---- ---- ---- ------ fd - /dev/fd fd - no - /proc - /proc proc - no - /dev/dsk/c0t10d0s1 - - swap - no - /dev/dsk/c0t10d0s0 /dev/rdsk/c0t10d0s0 / ufs 1 no logging swap - /tmp tmpfs - yes - *qfs3 - /qfs3 samfs - yes stripe=0*
The default disk allocation unit (DAU) is 64 kilobytes, but the following example sets the DAU size to 128 kilobytes.
# sammkfs -a 128 qfs3
This example illustrates the configuration of a file system called qfs4 that stripes file data to four disk drives. This example assumes the following:
The metadata device is a single partition (s6) used on controller 0, logical unit number (LUN) 0.
The data devices consist of four disks attached to controller 6. Each disk is on a separate target (2—5).
The following code example shows the mcf file for this striped disk configuration.
# Equipment Eq Eq Family Device Additional # Identifier Ord Type Set State Parameters #----------- --- ---- ------ ------ ---------- # *qfs4 40 ma qfs4 on* /dev/dsk/c8t4d0s4 41 mm qfs4 on /dev/dsk/c6t2d0s4 42 mr qfs4 on /dev/dsk/c6t3d0s4 43 mr qfs4 on /dev/dsk/c6t4d0s4 44 mr qfs4 on /dev/dsk/c6t5d0s4 45 mr qfs4 on
Set the stripe width by using the stripe= option. The following code example shows the /etc/vfstab file with a mount parameter of stripe=1 set for the qfs4 file system.
# cat /etc/vfstab # #device device file mount #to to mount system fsck at mount #mount fsck point type pass boot params #----- ---- ----- ----- ---- ---- ------ fd - /dev/fd fd - no - /proc - /proc proc - no - /dev/dsk/c0t10d0s1 - - swap - no - /dev/dsk/c0t10d0s0 /dev/rdsk/c0t10d0s0 / ufs 1 no logging swap - /tmp tmpfs - yes - *qfs4 - /qfs4 samfs - yes stripe=1*
The stripe=1 specification stripes file data across all four of the mr data disks with a stripe width of one DAU. The DAU is the allocation unit you set when you use the sammkfs command to initialize the file system.
The following example sets the DAU size to 128 kilobytes.
# sammkfs -a 128 qfs4
With this striped disk configuration, any file written to this file system is striped across all of the devices in increments of 128 kilobytes. Files less than the aggregate stripe width times the number of devices still use 128 kilobytes of disk space. Files larger than 128 kilobytes have space allocated for them as needed in total space increments of 128 kilobytes.
Striped groups enable you to build RAID-0 devices of separate disk devices. Striped groups, however, have only one DAU per striped group. This method of writing large, effective DAUs across RAID devices saves system update time and supports high-speed sequential I/O. Striped groups are useful for writing very large files to groups of disk devices.
Note - A DAU is the minimum disk space allocated. The minimum disk space allocated in a striped group is as follows:
allocation-unit x number of disks in the group
Writing a single byte of data consumes a DAU on every member of the striped group. Make sure that you understand the effects of using striped groups with your file system.
The devices within a striped group must be the same size. You cannot increase the size of a striped group. You can add additional striped groups to the file system, however.
This example illustrates the configuration of a file system (called qfs5) that separates the metadata onto a low-latency disk. The mcf file defines two striped groups on four drives. This example assumes the following:
The metadata device is a single partition (s5) used on controller 8, disk 4.
The data devices consist of four disks (two groups of two identical disks) attached to controller 6. Each disk is on a separate target (2–5).
The following code example shows a sample mcf file for a striped group configuration.
# cat /etc/opt/SUNWsamfs/mcf # # Equipment Eq Eq Family Device Additional # Identifier Ord Type Set State Parameters #----------- --- ---- ------ ------ ---------- # *qfs5 50 ma qfs5 on* /dev/dsk/c8t4d0s5 51 mm qfs5 on /dev/dsk/c6t2d0s5 52 g0 qfs5 on /dev/dsk/c6t3d0s5 53 g0 qfs5 on /dev/dsk/c6t4d0s5 54 g1 qfs5 on /dev/dsk/c6t5d0s5 55 g1 qfs5 on
Set the stripe width by using the stripe= option. The following code example shows the /etc/vfstab file with a mount parameter of stripe=0, which specifies round-robin allocation between striped group g0 and striped group g1.
# cat /etc/vfstab #device device file mount #to to mount system fsck at mount #mount fsck point type pass boot params #----- ---- ----- ---- ---- ---- ------ fd - /dev/fd fd - no - /proc - /proc proc - no - /dev/dsk/c0t10d0s1 - - swap - no - /dev/dsk/c0t10d0s0 /dev/rdsk/c0t10d0s0 / ufs 1 no logging swap - /tmp tmpfs - yes - *qfs5 - /qfs5 samfs - yes stripe=0*
The -a option is not used with striped groups because the DAU is equal to the size of an allocation, or the size, of each group.
# sammkfs qfs5
This his example has two striped groups, g0 and g1. With stripe=0 in /etc/vfstab, devices 12 and 13 are striped, devices 14 and 15 are striped, and files are round-robined around the two striped groups. A striped group is treated as a bound entity. After you configure a stripe group, you cannot change it without issuing another sammkfs command.
|