File System Enhancements

The Solaris 9 release includes the following file system enhancements.

Extended File Attributes

The UFS, NFS, and TMPFS file systems have been enhanced to include extended file attributes. Application developers can associate specific attributes to a file. For example, a developer of a file management application for a windowing system might choose to associate a display icon with a file.

Extended attributes are logically represented as files within a hidden directory that is associated with the target file.

You can use the extended file attribute API and a set of shell commands to add and manipulate file system attributes. See the fsattr(5), openat(2), and runat(1) man pages for more information.

Many file system commands in Solaris provide an attribute-aware option that you can use to query, copy, modify, or find file attributes. For more information, see the specific file system command in the man pages.

See also the System Administration Guide: Basic Administration for more information.

Improved UFS Direct I/O Concurrency

The performance of direct I/O is used by database applications to access unbuffered file-system data. Direct I/O improvements allow concurrent read access and write access to regular UFS files. Previously, an operation that updated file data would lock out all other read or write accesses until the update operation was completed.

See the System Administration Guide: Basic Administration and the man page, mount_ufs(1M), for more information.

DNLC Improvements

The directory name look-up cache (DNLC) is enhanced to provide improved performance when you access files in large directories with 1000 or more files.

The DNLC is a general file-system service. DNLC caches the most recently referenced directory names and their associated vnodes. UFS directory entries are stored linearly on disk. This means that locating an entry requires searching each entry for the name. Adding a new entry requires searching the entire directory to ensure the name does not exist. To solve this performance problem, entire directories are cached in memory by the DNLC.

Another feature in this release is that DNLC caches file objects that have been looked up but do not exist. This feature is known as negative caching, and is useful because some applications repeatedly test to check if a file exists.

New tunable parameters are associated with the DNLC improvements. These parameters are set optimally. Do not change the parameters casually.

See the Solaris Tunable Parameters Reference Manual for further information.

UFS Snapshots (fssnap)

You can use the fssnap command to create a snapshot of a file system. A snapshot is a file system's temporary image that is intended for backup operations.

When run, the fssnap command creates a virtual device and a backing-store file. You can back up the virtual device, which looks and acts like a real device, with any of the existing Solaris backup commands. The backing-store file is a bitmapped file that contains copies of pre-snapshot data that has been modified since the snapshot was taken.

See the System Administration Guide: Basic Administration and the man page, fssnap(1M), for more information.

Updated mkfs Command

The mkfs command has been updated to improve performance when you create file systems. Improved mkfs performance is often 10 times faster than in previous Solaris releases. Performance improvements are seen on systems when you create both large file systems and small file systems. However, the biggest mkfs performance improvements occur on systems with high-capacity or high-speed disks.

Installation Enhancements

The Solaris 9 release includes the following installation enhancements.

Solaris Live Upgrade 2.0

Note - Solaris Live Upgrade 2.0 is available in the Solaris 9 operating environment. This description covers version 2.0.

Solaris Live Update 2.1 is now available in the Solaris 9 8/03 Update release. For a description of Solaris Live Update 2.1, see Solaris Live Upgrade 2.1.

Solaris Live Upgrade provides a method of upgrading that substantially reduces the usual service outage that is associated with an operating system upgrade. You can duplicate your current running boot environment, then while the original boot environment continues to run, you can upgrade the duplicate. The duplicate boot environment is then activated to become the active boot environment when the system is rebooted. If a failure occurs, you can quickly fall back to the original boot environment with a simple reboot. This feature eliminates the downtime for the production environment that is associated with a normal test and evaluation process.

In addition to upgrading a boot environment, you can install a Web Start Flash archive on an inactive boot environment. When you reboot the system, the configuration that you installed on an inactive boot environment is active.

The Solaris 9 release includes several Live Upgrade enhancements that apply to the command-line interface only. These enhancements affect the following:

• Progress Reporting - When using Solaris Live Upgrade to upgrade or install a Web Start Flash archive, the percentage of the upgrade or install completed is reported.

• Changes to the lumount and luumount Commands - The lumount command mounts all the boot environment's file systems. If you do not explicitly specify a mount point, lumount creates a mount point. This mount point uses the boot environment name rather than a random set of numbers, thus preventing a proliferation of mount points. This enhancement makes the luumount command easier to use.

  The luumount command unmounts the boot environment's root file system. The luumount command now accepts a mount point as well as the boot environment name. And with the -f option, a boot environment's file system can be forcibly unmounted.

  See the man pages lumount(1M) and luumount(1M).

• Scheduling Priorities - Solaris Live Upgrade's main purpose is to minimize production environment downtime while migrating to a new operating system. Some Solaris Live Upgrade operations, such as upgrading and copying file systems, can cause significant load to a system. Solaris Live Upgrade now has the tools to control scheduling by priorities. This feature minimizes performance degradation on the production system. You can change the defaults in the /etc/default/lu file.

• Naming Boot Environments - Solaris Live Upgrade commands allow long names for boot environments. The commands can now associate a description of any length with a boot environment name.

  For further information, see the Solaris 9 4/04 Installation Guide or the man page, ludesc(1M).

For further information on Solaris Live Upgrade, see "Solaris Live Upgrade (Topics)" in Solaris 9 4/04 Installation Guide.

Note - In the Solaris 9 Update releases, note the following name change:

Solaris Flash (formerly Web Start Flash)

Web Start Flash Installation Feature

The Web Start Flash installation feature enables you to create a single reference installation of the Solaris operating environment on a machine. You can then replicate that installation on several machines.

For further information, see "Solaris Flash Installation Feature (Topics)" in Solaris 9 4/04 Installation Guide.

Note - In the Solaris 9 Update releases, note the following name change:

Solaris Flash (formerly Web Start Flash)

Web Start Flash Archive Retrieval Using FTP

The Web Start Flash program has been updated to allow you to retrieve a Web Start Flash archive by using FTP. When installing an archive, you can specify the location of an archive on an FTP server.

For details about how to retrieve an archive from an FTP server, see the Solaris 9 4/04 Installation Guide.

Note - In the Solaris 9 Update releases, note the following name change:

Solaris Flash (formerly Web Start Flash)

Minimal Installation

Files that constitute several features in the core software group, or metacluster, are now moved into separate, more logically organized packages. You can optionally exclude these packages from the Solaris operating environment when you install the Solaris software. You can also remove these packages by using the pkgrm command after installation. See the pkgrm(1M) man page.

Files that constitute the following features are moved into new packages or existing packages:

• Cache file system

• NFS

• Kerberos security

• Distributed file system

• NIS-related

• Network routing daemons

• Remote network r* commands

• telnet server

• tftp server

• Domain name server

• DARPA name server

• Remote procedure call services

• Boot or install server

• setuid and setgid