Veritas File System 4.1 Release Notes (5900-0592, March 2010)

• Intent Log Resize
The VxFS intent log can be resized using the log option of the fsadm command. The fsadm
command supports online administration functions on VxFS file systems, Storage Checkpoints
and individual files and directories. You can use the fsadm command for file-system resizing,
extent reorganization, directory reorganization, and querying or changing the largefiles
flag. Changing the intent log size can improve system performance. The maximum intent
log size is now 2 gigabytes. See Veritas File System Administrator’s Guide and mkfs_vxfs (1M)
and fsadm_vxfs (1M) manpages for more information on intent log size.
• Default mount option is largefiles
The largefiles option is now the default file size option. See mkfs_vxfs (1M) and mount_vxfs
(1M) for more information. If largefiles is specified and the mount succeeds, then the
file system can contain files whose size is two gigabytes or larger, and also large files can
be created.
• Default Option for Intent Logging is delaylog
The delaylog option is the default intent logging mount option. In delaylog mode, the
effects of most system calls other than write (2), writev (2), and pwrite (2) are guaranteed to
be persistent approximately 15 to 20 seconds after the system call returns to the application.
The change of the default mount option from log to delaylog does not increase the risk
of data loss, but allows VxFS to cache data to improve performance. See Veritas File System
4.1 Administrator's Guide, for more information on the delaylog option.
• Recording Structural Changes to the File System
The histlog function was implemented in the fsdb_vxfs command. The history log
records structural changes to the file system to aid in product support.
• Renamed vxfsu_get_ioffsets Library
The vxfsu_get_ioffsets library call was renamed vxfs_get_ioffsets. The
vxfs_get_ioffsets function returns the offset and size, in bytes, of certain public fields
in the vx_inode inode structure, for any VxFS file.
• Caching of Inodes in a VxFS File System to an Inode Table
The VxFS file system caches the inodes in an inode table. The kernel tunable vx_ninode
determines the number of inodes in the inode table to help VxFS in caching.
The vx_ninode and vxfs_bc_bufhwm can be dynamically tuned without rebooting the
machine. The changes take effect immediately, although they can be optionally made to take
effect only after a reboot. See “Veritas File System 4.1 Administrator’s Guide? for more
information.
Limitations of VxFS 4.1
• Disk layout versions prior to version 4 cannot be mounted. You must use the vxfsconvert
command to convert them to disk layout version 4.
• The vxupgrade command upgrades only Disk Layout Version 4 and Disk Layout Version
5. Disk layout versions prior to version 4 cannot be mounted.
• VxFS 4.1 does not post EVM events when size of the file system expands during runtime.
• File system with 8 million inodes can be created only using the largefiles option of mkfs
command (the fsadm utility can also be used to set the largefiles flag on the file system).
See the mkfs_vxfs (1M) and fsadm_vxfs (1M) manpages for details.
Product Licensing (Currently not available)
The following table lists the Veritas File System features that require an additional license over
HP OnlineJFS.
8 Veritas File System 4.1 Release Notes