Veritas™ File System 5.0.1 Administrator's Guide HP-UX 11i v3 HP Part Number: 5900-0082 Published: November 2009 Edition: 1.
© Copyright 2009 Hewlett-Packard Development Company, L.P. Confidential computer software. Valid license from HP required for possession, use or copying. Consistent with FAR 12.211 and 12.212, Commercial Computer Software, Computer Software Documentation and Technical Data for Commercial Items are licensed to the U.S. Government under vendor’s standard commercial license. The information contained herein is subject to change without notice.
Contents Technical Support ............................................................................................... 4 Chapter 1 Introducing Veritas File System ....................................... 15 About Veritas File System .............................................................. Logging ................................................................................ Extents ................................................................................ File system disk layouts ........
8 Contents Chapter 2 VxFS performance: creating, mounting, and tuning file systems ..................................................................... 33 Creating a VxFS file system ............................................................ Block size ............................................................................. Intent log size ........................................................................ Mounting a VxFS file system ..........................................................
Contents Chapter 4 VxFS I/O Overview ............................................................... 69 About VxFS I/O ............................................................................ Buffered and Direct I/O ................................................................. Direct I/O ............................................................................. Unbuffered I/O ...................................................................... Data synchronous I/O ............................
10 Contents Chapter 7 Quotas .................................................................................. 109 About quota limits ...................................................................... About quota files on Veritas File System ......................................... About quota commands ............................................................... Using quotas ............................................................................. Turning on quotas ..........................
Contents Volume encapsulation ................................................................. Encapsulating a volume ......................................................... Deencapsulating a volume ..................................................... Reporting file extents .................................................................. Examples of reporting file extents ........................................... Load balancing .......................................................................
12 Contents File placement policy rule and statement ordering ............................ 176 File placement policies and extending files ...................................... 179 Appendix A Quick Reference ................................................................. 181 Command summary .................................................................... Online manual pages ................................................................... Creating a VxFS file system ............................
Contents About unique message identifiers .................................................. 250 Unique message identifiers .......................................................... 251 Appendix C Disk layout .......................................................................... 255 About disk layouts ...................................................................... Supported disk layouts and operating systems ................................. VxFS Version 4 disk layout ......................
14 Contents
Chapter 1 Introducing Veritas File System This chapter includes the following topics: ■ About Veritas File System ■ Veritas File System features ■ Veritas File System performance enhancements ■ Using Veritas File System About Veritas File System A file system is simply a method for storing and organizing computer files and the data they contain to make it easy to find and access them.
16 Introducing Veritas File System Veritas File System features ■ File system disk layouts Logging A key aspect of any file system is how to recover if a system crash occurs. Earlier methods required a time-consuming scan of the entire file system. A better solution is the method of logging (or journaling) the metadata of files. VxFS logs new attribute information into a reserved area of the file system, whenever file system changes occur.
Introducing Veritas File System Veritas File System features ■ Extent-based allocation Extents allow disk I/O to take place in units of multiple blocks if storage is allocated in consecutive blocks. ■ Extent attributes Extent attributes are the extent allocation policies associated with a file. ■ Fast file system recovery VxFS provides fast recovery of a file system from system failure.
18 Introducing Veritas File System Veritas File System features ■ Multi-volume support The multi-volume support feature allows several volumes to be represented by a single logical object. ■ Dynamic Storage Tiering The Dynamic Storage Tiering (DST) option allows you to configure policies that automatically relocate files from one volume to another, or relocate files by running file relocation commands, which can improve performance for applications that access specific types of files.
Introducing Veritas File System Veritas File System features first Used for single indirection. Each entry in the extent indicates the starting block number of an indirect data extent second Used for double indirection. Each entry in the extent indicates the starting block number of a single indirect address extent. Each indirect address extent is 8K long and contains 2048 entries.
20 Introducing Veritas File System Veritas File System features ■ While there are no limits on the levels of indirection, lower levels are expected in this format since data extents have variable lengths. ■ This format uses a type indicator that determines its record format and content and accommodates new requirements and functionality for future types. The current typed format is used on regular files and directories only when indirection is needed.
Introducing Veritas File System Veritas File System features a full structural check of the entire file system. Replaying the intent log may not completely recover the damaged file system structure if there was a disk hardware failure; hardware problems may require a complete system check using the fsck utility provided with VxFS. See “The log option and data integrity” on page 22. VxFS intent log resizing The VxFS intent log is allocated when the file system is first created.
22 Introducing Veritas File System Veritas File System features Enhanced data integrity modes For most UNIX file systems, including VxFS, the default mode for writing to a file is delayed, or buffered, meaning that the data to be written is copied to the file system cache and later flushed to disk. A delayed write provides much better performance than synchronously writing the data to disk.
Introducing Veritas File System Veritas File System features disk to guarantee the persistence of the file data before renaming it. The rename() call is also guaranteed to be persistent when the system call returns. The changes to file system data and metadata caused by the fsync(2) and fdatasync(2) system calls are guaranteed to be persistent once the calls return. Enhanced performance mode VxFS has a mount option that improves performance: delaylog.
24 Introducing Veritas File System Veritas File System features Warning: Some applications and utilities may not work on large files. Access Control Lists An Access Control List (ACL) stores a series of entries that identify specific users or groups and their access privileges for a directory or file. A file may have its own ACL or may share an ACL with other files. ACLs have the advantage of specifying detailed access permissions for multiple users and groups.
Introducing Veritas File System Veritas File System features Quotas VxFS supports quotas, which allocate per-user quotas and limit the use of two principal resources: files and data blocks. You can assign quotas for each of these resources. Each quota consists of two limits for each resource: hard limit and soft limit. The hard limit represents an absolute limit on data blocks or files. A user can never exceed the hard limit under any circumstances.
26 Introducing Veritas File System Veritas File System features log. File system operations, such as allocating or deleting files, can originate from any node in the cluster. Installing VxFS and enabling the cluster feature does not create a cluster file system configuration. HP Serviceguard Storage Management environments require HP Serviceguard for file system clustering. To be a cluster mount, a file system must be mounted using the mount -o cluster option.
Introducing Veritas File System Veritas File System performance enhancements You can then configure policies that automatically relocate files from one volume to another, or relocate files by running file relocation commands. Having multiple volumes lets you determine where files are located, which can improve performance for applications that access specific types of files. DST functionality is a separately licensed feature and is available with the VRTSfppm package.
28 Introducing Veritas File System Veritas File System performance enhancements ■ Tunable indirect data extent size ■ Integration with VxVM™ ■ Support for large directories Note: VxFS reduces the file lookup time in directories with an extremely large number of files. About enhanced I/O performance VxFS provides enhanced I/O performance by applying an aggressive I/O clustering policy, integrating with VxVM, and allowing application specific parameters to be set on a per-file system basis.
Introducing Veritas File System Using Veritas File System This value defines the maximum size of a single direct I/O. See the vxtunefs(1M) and tunefstab(4) manual pages.
30 Introducing Veritas File System Using Veritas File System Online system administration VxFS provides command line interface (CLI) operations that are described throughout this guide and in manual pages. VxFS allows you to run a number of administration tasks while the file system is online. Two of the more important tasks include: ■ Defragmentation ■ File system resizing About defragmentation Free resources are initially aligned and allocated to files in an order that provides optimal performance.
Introducing Veritas File System Using Veritas File System and the file system. The vxresize command guarantees that the file system shrinks or grows along with the volume. Do not use the vxassist and fsadm_vxfs commands for this purpose. See the vxresize(1M) manual page. See the Veritas Volume Manager Administrator's Guide.
32 Introducing Veritas File System Using Veritas File System
Chapter 2 VxFS performance: creating, mounting, and tuning file systems This chapter includes the following topics: ■ Creating a VxFS file system ■ Mounting a VxFS file system ■ Tuning the VxFS file system ■ Monitoring free space ■ Tuning I/O Creating a VxFS file system When you create a file system with the mkfs command, you can select the following characteristics: ■ Block size ■ Intent log size Block size The unit of allocation in VxFS is a block.
34 VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system You specify the block size when creating a file system by using the mkfs -o bsize option. The block size cannot be altered after the file system is created. The smallest available block size for VxFS is 1K. The default block size is 1024 bytes for file systems smaller than 1 TB, and 8192 bytes for file systems 1 TB or larger. Choose a block size based on the type of application being run.
VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system ■ tmplog ■ logsize ■ nodatainlog ■ blkclear ■ mincache ■ convosync ■ ioerror ■ largefiles|nolargefiles ■ cio ■ mntlock|mntunlock ■ tranflush Caching behavior can be altered with the mincache option, and the behavior of O_SYNC and D_SYNC writes can be altered with the convosync option. See the fcntl(2) manual page. The delaylog and tmplog modes can significantly improve performance.
36 VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system writing the new file contents to a temporary file and then renaming it on top of the target file. The delaylog mode The default logging mode is delaylog. 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.
VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system The behavior of NFS servers on a VxFS file system is unaffected by the log and tmplog mount options, but not delaylog. In all cases except for tmplog, VxFS complies with the persistency requirements of the NFS v2 and NFS v3 standard.
38 VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system The blkclear mode The blkclear mode is used in increased data security environments. The blkclear mode guarantees that uninitialized storage never appears in files. The increased integrity is provided by clearing extents on disk when they are allocated within a file. This mode does not affect extending writes.
VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system ■ The mincache=direct, mincache=unbuffered, and mincache=dsync modes also flush file data on close as mincache=closesync does. Because the mincache=direct, mincache=unbuffered, and mincache=dsync modes change non-synchronous I/O to synchronous I/O, throughput can substantially degrade for small to medium size files with most applications.
40 VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system See the open(2), fcntl(2), and vxfsio(7) manual pages. Warning: Be very careful when using the convosync=closesync or convosync=delay mode because they actually change synchronous I/O into non-synchronous I/O. Applications that use synchronous I/O for data reliability may fail if the system crashes and synchronously written data is lost.
VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system The disable policy If disable is selected, VxFS disables the file system after detecting any I/O error. You must then unmount the file system and correct the condition causing the I/O error. After the problem is repaired, run fsck and mount the file system again. In most cases, replay fsck to repair the file system. A full fsck is required only in cases of structural damage to the file system's metadata.
42 VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system confined to data extents. mdisable is the default ioerror mount option for cluster mounts. The largefiles|nolargefiles option The section includes the following topics : ■ Creating a file system with large files ■ Mounting a file system with large files ■ Managing a file system with large files VxFS supports files larger than 2 gigabytes. The maximum file size that can be created is 2 terabytes.
VxFS performance: creating, mounting, and tuning file systems Mounting a VxFS file system is not to specify either option. After a file system is mounted, you can use the fsadm utility to change the large files option.
44 VxFS performance: creating, mounting, and tuning file systems Tuning the VxFS file system The mntunlock option of the vxumount command reverses the mntlock option if you previously locked the file system. Combining mount command options Although mount options can be combined arbitrarily, some combinations do not make sense. The following examples provide some common and reasonable mount option combinations.
VxFS performance: creating, mounting, and tuning file systems Tuning the VxFS file system ■ VxFS buffer cache high water mark ■ Number of links to a file ■ VxFS inode free time lag Tuning inode table size VxFS caches inodes in an inode table. There is a dynamic tunable in VxFS called vx_ninode that determines the number of entries in the inode table. You can dynamically change the value of vx_ninode by using the sam or kctune commands. See the sam(1M) and kctune(1M) manual pages.
46 VxFS performance: creating, mounting, and tuning file systems Tuning the VxFS file system either the system memory size, which is the default, or the value of the tunable if explicitly set, whichever is larger. Thus, dynamically increasing the tunable to a value that is more than two times either the default value or the user-defined value, if larger, may cause performance degradation unless the system is rebooted.
VxFS performance: creating, mounting, and tuning file systems Tuning the VxFS file system VxFS buffer cache high water mark VxFS maintains its own buffer cache in the kernel for frequently accessed file system metadata. This cache is different from the HP-UX kernel buffer cache that caches file data. The vx_bc_bufhwm dynamic, global, tunable parameter lets you change the VxFS buffer cache high water mark, that is, the maximum amount of memory that can be used to cache VxFS metadata.
48 VxFS performance: creating, mounting, and tuning file systems Monitoring free space See the sam(1M) and kctune(1M) manual pages. You can also add an entry to the system configuration file as shown in the following example: vxfs_maxlink vx_maxlink 40000 This sets the value of vx_maxlink to 40,000 links. VxFS inode free time lag In VxFS, an inode is put on a freelist if it is not being used. The memory space for this unused inode can be freed if it stays on the freelist for a specified amount of time.
VxFS performance: creating, mounting, and tuning file systems Monitoring free space Full file systems may have an adverse effect on file system performance. Full file systems should therefore have some files removed, or should be expanded. See the fsadm_vxfs(1M) manual page. Monitoring fragmentation Fragmentation reduces performance and availability. Regular use of fsadm's fragmentation reporting and reorganization facilities is therefore advisable.
50 VxFS performance: creating, mounting, and tuning file systems Monitoring free space The “after” result is an indication of how well the reorganizer has performed. The degree of fragmentation should be close to the characteristics of an unfragmented file system. If not, it may be a good idea to resize the file system; full file systems tend to fragment and are difficult to defragment.
VxFS performance: creating, mounting, and tuning file systems Tuning I/O Note: Thin Reclamation is a slow process and may take several hours to complete, depending on the file system size. Thin Reclamation is not guaranteed to reclaim 100% of the free space. You can track the progress of the Thin Reclamation process by using the vxtask list command when using the Veritas Volume Manager (VxVM) command vxdisk reclaim. See the vxtask(1M) and vxdisk(1M) manual pages.
52 VxFS performance: creating, mounting, and tuning file systems Tuning I/O ■ The mount command queries VxVM when the file system is mounted and downloads the I/O parameters. If the default parameters are not acceptable or the file system is being used without VxVM, then the /etc/vx/tunefstab file can be used to set values for I/O parameters. The mount command reads the /etc/vx/tunefstab file and downloads any parameters specified for a file system.
VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description write_pref_io The preferred write request size. The file system uses this in conjunction with the write_nstream value to determine how to do flush behind on writes. The default value is 64K. read_nstream The number of parallel read requests of size read_pref_io to have outstanding at one time.
54 VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description fcl_keeptime Specifies the minimum amount of time, in seconds, that the VxFS File Change Log (FCL) keeps records in the log. When the oldest 8K block of FCL records have been kept longer than the value of fcl_keeptime, they are purged from the FCL and the extents nearest to the beginning of the FCL file are freed.
VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description fcl_winterval Specifies the time, in seconds, that must elapse before the VxFS File Change Log (FCL) records a data overwrite, data extending write, or data truncate for a file. The ability to limit the number of repetitive FCL records for continuous writes to the same file is important for file system performance and for applications processing the FCL.
56 VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description initial_extent_size Changes the default initial extent size. VxFS determines, based on the first write to a new file, the size of the first extent to be allocated to the file. Normally the first extent is the smallest power of 2 that is larger than the size of the first write. If that power of 2 is less than 8K, the first extent allocated is 8K.
VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description inode_aging_size Specifies the minimum size to qualify a deleted inode for inode aging. Inode aging is used in conjunction with file system Storage Checkpoints to allow quick restoration of large, recently deleted files. For best performance, it is advisable to age only a limited number of larger files before completion of the removal process.
58 VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description default_indir_ size On VxFS, files can have up to ten direct extents of variable size stored in the inode. After these extents are used up, the file must use indirect extents which are a fixed size that is set when the file first uses indirect extents. These indirect extents are 8K by default.
VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description read_ahead The default for all VxFS read operations is to perform sequential read ahead. You can specify the read_ahead cache advisory to implement the VxFS enhanced read ahead functionality.
60 VxFS performance: creating, mounting, and tuning file systems Tuning I/O Table 2-1 Tunable VxFS I/O parameters (continued) Parameter Description write_throttle The write_throttle parameter is useful in special situations where a computer system has a combination of a large amount of memory and slow storage devices. In this configuration, sync operations, such as fsync(), may take long enough to complete that a system appears to hang.
VxFS performance: creating, mounting, and tuning file systems Tuning I/O Note: VxFS does not query VxVM with multiple volume sets. To improve I/O performance when using multiple volume sets, use the vxtunefs command. If the file system is being used with a hardware disk array or volume manager other than VxVM, try to align the parameters to match the geometry of the logical disk.
62 VxFS performance: creating, mounting, and tuning file systems Tuning I/O
Chapter 3 Extent attributes This chapter includes the following topics: ■ About extent attributes ■ Commands related to extent attributes About extent attributes Veritas File System (VxFS) allocates disk space to files in groups of one or more adjacent blocks called extents. VxFS defines an application interface that allows programs to control various aspects of the extent allocation for a given file. The extent allocation policies associated with a file are referred to as extent attributes.
64 Extent attributes About extent attributes Some of the extent attributes are persistent and become part of the on-disk information about the file, while other attributes are temporary and are lost after the file is closed or the system is rebooted. The persistent attributes are similar to the file's permissions and are written in the inode for the file. When a file is copied, moved, or archived, only the persistent attributes of the source file are preserved in the new file.
Extent attributes About extent attributes smaller pieces. By erring on the side of minimizing fragmentation for the file system, files may become so non-contiguous that their I/O characteristics would degrade. Fixed extent sizes are particularly appropriate in the following situations: ■ If a file is large and contiguous, a large fixed extent size can minimize the number of extents in the file.
66 Extent attributes Commands related to extent attributes Write operations beyond reservation A reservation request can specify that no allocations can take place after a write operation fills the last available block in the reservation. This request can be used a way similar to the function of the ulimit command to prevent a file's uncontrolled growth. Reservation trimming A reservation request can specify that any unused reservation be released when the file is closed.
Extent attributes Commands related to extent attributes file system does not support extent attributes, has a different block size than the source file system, or lacks free extents appropriate to satisfy the extent attribute requirements.
68 Extent attributes Commands related to extent attributes
Chapter 4 VxFS I/O Overview This chapter includes the following topics: ■ About VxFS I/O ■ Buffered and Direct I/O ■ Concurrent I/O ■ Cache advisories ■ Freezing and thawing a file system ■ Getting the I/O size About VxFS I/O VxFS processes two basic types of file system I/O: ■ Sequential ■ Random or I/O that is not sequential For sequential I/O, VxFS employs a read-ahead policy by default when the application is reading data. For writing, it allocates contiguous blocks if possible.
70 VxFS I/O Overview Buffered and Direct I/O Buffered and Direct I/O VxFS responds with read-ahead for sequential read I/O. This results in buffered I/O. The data is prefetched and retained in buffers for the application. This is the default VxFS behavior. On the other hand, direct I/O does not buffer the data when the I/O to the underlying device is completed. This saves system resources like memory and CPU usage. Direct I/O is possible only when alignment and sizing criteria are satisfied.
VxFS I/O Overview Buffered and Direct I/O Direct I/O versus synchronous I/O Because direct I/O maintains the same data integrity as synchronous I/O, it can be used in many applications that currently use synchronous I/O. If a direct I/O request does not allocate storage or extend the file, the inode is not immediately written. Direct I/O CPU overhead The CPU cost of direct I/O is about the same as a raw disk transfer.
72 VxFS I/O Overview Concurrent I/O transferred to disk synchronously before the write returns to the user. If the file is not extended by the write, the times are updated in memory, and the call returns to the user. If the file is extended by the operation, the inode is written before the write returns. The direct I/O and VX_DSYNC advisories are maintained on a per-file-descriptor basis. Data synchronous I/O vs.
VxFS I/O Overview Cache advisories ■ By using the cio mount option. The read(2) and write(2) operations occurring on all of the files in this particular file system will use concurrent I/O. See “The cio option” on page 43. See the mount_vxfs(1M) manual page. Cache advisories VxFS allows an application to set cache advisories for use when accessing files.
74 VxFS I/O Overview Getting the I/O size When the file system is frozen, any attempt to use the frozen file system, except for a VX_THAW ioctl command, is blocked until a process executes the VX_THAW ioctl command or the time-out on the freeze expires. Getting the I/O size VxFS provides the VX_GET_IOPARAMETERS ioctl to get the recommended I/O sizes to use on a file system. This ioctl can be used by the application to make decisions about the I/O sizes issued to VxFS for a file or file device.
Chapter 5 Online backup using file system snapshots This chapter includes the following topics: ■ About snapshot file systems ■ Snapshot file system backups ■ Creating a snapshot file system ■ Backup examples ■ Snapshot file system performance ■ Differences Between Snapshots and Storage Checkpoints ■ About snapshot file system disk structure ■ How a snapshot file system works About snapshot file systems A snapshot file system is an exact image of a VxFS file system, referred to as the snap
76 Online backup using file system snapshots Snapshot file system backups its snapshots are unmounted. Although it is possible to have multiple snapshots of a file system made at different times, it is not possible to make a snapshot of a snapshot. Note: A snapshot file system ceases to exist when unmounted. If mounted again, it is actually a fresh snapshot of the snapped file system. A snapshot file system must be unmounted before its dependent snapped file system can be unmounted.
Online backup using file system snapshots Creating a snapshot file system Creating a snapshot file system You create a snapshot file system by using the -o snapof= option of the mount command. The -o snapsize= option may also be required if the device you are mounting does not identify the device size in its disk label, or if you want a size smaller than the entire device.
78 Online backup using file system snapshots Snapshot file system performance To create a backup using a snapshop file system 1 To back up files changed within the last week using cpio: # mount -F vxfs -o snapof=/home,snapsize=100000 \ /dev/vx/dsk/fsvol/vol1 /backup/home # cd /backup # find home -ctime -7 -depth -print | cpio -oc > /dev/rmt/0m # umount /backup/home 2 To do a level 3 backup of /dev/vx/dsk/fsvol/vol1 and collect those files that have changed in the current directory: # vxdump 3f - /dev/
Online backup using file system snapshots Differences Between Snapshots and Storage Checkpoints application running an online transaction processing (OLTP) workload on a snapped file system was measured at about 15 to 20 percent slower than a file system that was not snapped.
80 Online backup using file system snapshots How a snapshot file system works Figure 5-1 The Snapshot Disk Structure super-block bitmap blockmap data block The super-block is similar to the super-block of a standard VxFS file system, but the magic number is different and many of the fields are not applicable. The bitmap contains one bit for every block on the snapped file system. Initially, all bitmap entries are zero.
Online backup using file system snapshots How a snapshot file system works data for block n can be found on the snapshot file system. The blockmap entry for block n is changed from 0 to the block number on the snapshot file system containing the old data. A subsequent read request for block n on the snapshot file system will be satisfied by checking the bitmap entry for block n and reading the data from the indicated block on the snapshot file system, instead of from block n on the snapped file system.
82 Online backup using file system snapshots How a snapshot file system works
Chapter 6 Storage Checkpoints This chapter includes the following topics: ■ About Storage Checkpoints ■ How a Storage Checkpoint works ■ Types of Storage Checkpoints ■ Storage Checkpoint administration ■ Space management considerations ■ Restoring a file system from a Storage Checkpoint ■ Storage Checkpoint quotas About Storage Checkpoints Veritas File System provides a Storage Checkpoint feature that quickly creates a persistent image of a file system at an exact point in time.
84 Storage Checkpoints About Storage Checkpoints See “How a Storage Checkpoint works” on page 85. Unlike a disk-based mirroring technology that requires a separate storage space, Storage Checkpoints minimize the use of disk space by using a Storage Checkpoint within the same free space available to the file system.
Storage Checkpoints How a Storage Checkpoint works availability and data integrity by increasing the frequency of backup and replication solutions. Storage Checkpoints can be taken in environments with a large number of files, such as file servers with millions of files, with little adverse impact on performance. Because the file system does not remain frozen during Storage Checkpoint creation, applications can access the file system even while the Storage Checkpoint is taken.
86 Storage Checkpoints How a Storage Checkpoint works Primary fileset and its Storage Checkpoint Figure 6-1 Primary fileset Storage Checkpoint /database emp.dbf /database jun.dbf emp.dbf jun.dbf In Figure 6-2, a square represents each block of the file system. This figure shows a Storage Checkpoint containing pointers to the primary fileset at the time the Storage Checkpoint is taken, as in Figure 6-1.
Storage Checkpoints How a Storage Checkpoint works data is copied to the Storage Checkpoint before the new data is written. When a write operation changes a specific data block in the primary fileset, the old data is first read and copied to the Storage Checkpoint before the primary fileset is updated. Subsequent writes to the specified data block on the primary fileset do not result in additional updates to the Storage Checkpoint because the old data needs to be saved only once.
88 Storage Checkpoints Types of Storage Checkpoints Types of Storage Checkpoints You can create the following types of Storage Checkpoints: ■ Data Storage Checkpoints ■ nodata Storage Checkpoints ■ Removable Storage Checkpoints ■ Non-mountable Storage Checkpoints Data Storage Checkpoints A data Storage Checkpoint is a complete image of the file system at the time the Storage Checkpoint is created. This type of Storage Checkpoint contains the file system metadata and file data blocks.
Storage Checkpoints Types of Storage Checkpoints Figure 6-4 Updates to a nodata clone Primary fileset Storage Checkpoint A’ B C D E See “Showing the difference between a data and a nodata Storage Checkpoint” on page 95. Removable Storage Checkpoints A removable Storage Checkpoint can “self-destruct” under certain conditions when the file system runs out of space. See “Space management considerations” on page 101.
90 Storage Checkpoints Storage Checkpoint administration Use this type of Storage Checkpoint as a security feature which prevents other applications from accessing the Storage Checkpoint and modifying it. Storage Checkpoint administration Storage Checkpoint administrative operations require the fsckptadm utility. See the fsckptadm(1M) manual page. You can use the fsckptadm utility to create and remove Storage Checkpoints, change attributes, and ascertain statistical data.
Storage Checkpoints Storage Checkpoint administration For disk layout Version 6 or 7, multiply the number of inodes by 1 byte, and add 1 or 2 megabytes to get the approximate amount of space required. You can determine the number of inodes with the fsckptadm utility as above. Using the output from the example for disk layout Version 5, the approximate amount of space required by the metadata is just over one or two megabytes (23,872 x 1 byte, plus 1 or 2 megabytes).
92 Storage Checkpoints Storage Checkpoint administration Removing a Storage Checkpoint You can delete a Storage Checkpoint by specifying the remove keyword of the fsckptadm command. Specifically, you can use either the synchronous or asynchronous method of removing a Storage Checkpoint; the asynchronous method is the default method. The synchronous method entirely removes the Storage Checkpoint and returns all of the blocks to the file system before completing the fsckptadm operation.
Storage Checkpoints Storage Checkpoint administration ■ To mount a Storage Checkpoint of a file system, first mount the file system itself. ■ To unmount a file system, first unmount all of its Storage Checkpoints. 93 Warning: If you create a Storage Checkpoint for backup purposes, do not mount it as a writable Storage Checkpoint. You will lose the point-in-time image if you accidently write to the Storage Checkpoint. A Storage Checkpoint is mounted on a special pseudo device.
94 Storage Checkpoints Storage Checkpoint administration vol1 /dev/vx/dsk/fsvol/ vol1:may_23 ■ /fsvol_may_23 vxfs ckpt=may_23 0 To mount a Storage Checkpoint of a cluster file system, you must also use the -o cluster option: # mount -F vxfs -o cluster,ckpt=may_23 \ /dev/vx/dsk/fsvol/vol1:may_23 /fsvol_may_23 You can only mount a Storage Checkpoint cluster-wide if the file system that the Storage Checkpoint belongs to is also mounted cluster-wide.
Storage Checkpoints Storage Checkpoint administration file system, use the asynchronous method to mark the Storage Checkpoint you want to convert for a delayed conversion. In this case, the actual conversion will continue to be delayed until the Storage Checkpoint becomes the oldest Storage Checkpoint in the file system, or all of the older Storage Checkpoints have been converted to nodata Storage Checkpoints.
96 Storage Checkpoints Storage Checkpoint administration 4 Examine the content of the original file and the Storage Checkpoint file: # cat /mnt0/file hello, world # cat /mnt0@5_30pm/file hello, world 5 Change the content of the original file: # echo "goodbye" > /mnt0/file 6 Examine the content of the original file and the Storage Checkpoint file.
Storage Checkpoints Storage Checkpoint administration Converting multiple Storage Checkpoints You can convert Storage Checkpoints to nodata Storage Checkpoints, when dealing with older Storage Checkpoints on the same file system.
98 Storage Checkpoints Storage Checkpoint administration 3 Try to convert synchronously the latest Storage Checkpoint to a nodata Storage Checkpoint. The attempt will fail because the Storage Checkpoints older than the latest Storage Checkpoint are data Storage Checkpoints, namely the Storage Checkpoints old, older, and oldest: # fsckptadm -s set nodata latest /mnt0 UX:vxfs fsckptadm: ERROR: V-3-24632: Storage Checkpoint set failed on latest.
Storage Checkpoints Storage Checkpoint administration To create a delayed nodata Storage Checkpoint 1 Remove the latest Storage Checkpoint.
100 Storage Checkpoints Storage Checkpoint administration 3 Convert the oldest Storage Checkpoint to a nodata Storage Checkpoint because no older Storage Checkpoints exist that contain data in the file system. Note: This step can be done synchronously.
Storage Checkpoints Space management considerations 4 Remove the older and old Storage Checkpoints.
102 Storage Checkpoints Restoring a file system from a Storage Checkpoint ■ Remove the oldest Storage Checkpoint first. Restoring a file system from a Storage Checkpoint Mountable data Storage Checkpoints on a consistent and undamaged file system can be used by backup and restore applications to restore either individual files or an entire file system.
Storage Checkpoints Restoring a file system from a Storage Checkpoint To restore a file from a Storage Checkpoint 1 Create the Storage Checkpoint CKPT1 of /home. $ fckptadm create CKPT1 /home 2 Mount Storage Checkpoint CKPT1 on the directory /home/checkpoints/mar_4. $ mount -F vxfs -o ckpt=CKPT1 /dev/vx/dsk/dg1/vol- \ 01:CKPT1 /home/checkpoints/mar_4 3 Delete the file MyFile.txt from your home directory. $ cd /home/users/me $ rm MyFile.
104 Storage Checkpoints Restoring a file system from a Storage Checkpoint To restore a file system from a Storage Checkpoint 1 Run the fsckpt_restore command: # fsckpt_restore -l /dev/vx/dsk/dg1/vol2 /dev/vx/dsk/dg1/vol2: UNNAMED: ctime = Thu 08 May 2004 06:28:26 PM PST mtime = Thu 08 May 2004 06:28:26 PM PST flags = largefiles, file system root CKPT6: ctime = Thu 08 May 2004 06:28:35 PM PST mtime = Thu 08 May 2004 06:28:35 PM PST flags = largefiles CKPT5: ctime = Thu 08 May 2004 06:28:34 PM PST mtime =
Storage Checkpoints Restoring a file system from a Storage Checkpoint 2 In this example, select the Storage Checkpoint CKPT3 as the new root fileset: Select Storage Checkpoint for restore operation or (EOF) to exit or to list Storage Checkpoints: CKPT3 CKPT3: ctime = Thu 08 May 2004 06:28:31 PM PST mtime = Thu 08 May 2004 06:28:36 PM PST flags = largefiles UX:vxfs fsckpt_restore: WARNING: V-3-24640: Any file system changes or Storage Checkpoints made after Thu 08 May 2004 06:28:31 PM
106 Storage Checkpoints Restoring a file system from a Storage Checkpoint 3 Type y to restore the file system from CKPT3: Restore the file system from Storage Checkpoint CKPT3 ? (ynq) y (Yes) UX:vxfs fsckpt_restore: INFO: V-3-23760: File system restored from CKPT3 If the filesets are listed at this point, it shows that the former UNNAMED root fileset and CKPT6, CKPT5, and CKPT4 were removed, and that CKPT3 is now the primary fileset. CKPT3 is now the fileset that will be mounted by default.
Storage Checkpoints Storage Checkpoint quotas Storage Checkpoint quotas VxFS provides options to the fsckptadm command interface to administer Storage Checkpoint quotas. Storage Checkpoint quotas set the following limits on the number of blocks used by all Storage Checkpoints of a primary file set: hard limit An absolute limit that cannot be exceeded. If a hard limit is exceeded, all further allocations on any of the Storage Checkpoints fail, but existing Storage Checkpoints are preserved.
108 Storage Checkpoints Storage Checkpoint quotas
Chapter 7 Quotas This chapter includes the following topics: ■ About quota limits ■ About quota files on Veritas File System ■ About quota commands ■ Using quotas About quota limits Veritas File System (VxFS) supports user quotas. The quota system limits the use of two principal resources of a file system: files and data blocks. For each of these resources, you can assign quotas to individual users to limit their usage.
110 Quotas About quota files on Veritas File System See “About quota files on Veritas File System” on page 110. The quota soft limit can be exceeded when VxFS preallocates space to a file. Quota limits cannot exceed two terabytes on a Version 5 disk layout. See “About extent attributes” on page 63. About quota files on Veritas File System A quotas file (named quotas) must exist in the root directory of a file system for any of the quota commands to work.
Quotas Using quotas Turning on quotas To use the quota functionality on a file system, quotas must be turned on. You can turn quotas on at mount time or after a file system is mounted. To turn on quotas ◆ To turn on user quotas for a VxFS file system, enter: # quotaon /mount_point Turning on quotas at mount time Quotas can be turned on with the mount command when you mount a file system.
112 Quotas Using quotas To modify time limits ◆ Specify the -t option to modify time limits for any user: # edquota -t Viewing disk quotas and usage Use the quota command to view a user's disk quotas and usage on VxFS file systems.
Chapter 8 File Change Log This chapter includes the following topics: ■ About File Change Log ■ About the File Change Log file ■ File Change Log administrative interface ■ File Change Log programmatic interface ■ Summary of API functions ■ Reverse path name lookup About File Change Log The VxFS File Change Log (FCL) tracks changes to files and directories in a file system.
114 File Change Log About the File Change Log file About the File Change Log file File Change Log records file system changes such as creates, links, unlinks, renaming, data appended, data overwritten, data truncated, extended attribute modifications, holes punched, and miscellaneous file property updates. Note: FCL is supported only on disk layout Version 6 and later. FCL stores changes in a sparse file in the file system namespace. The FCL file is located in mount_point/lost+found/changelog.
File Change Log File Change Log administrative interface File Change Log administrative interface The FCL can be set up and tuned through the fcladm and vxtunefs VxFS administrative commands. See the fcladm(1M) and vxtunefs(1M) manual pages. The FCL keywords for fcladm are as follows: clear Disables the recording of the audit, open, close, and statistical events after it has been set. dump Creates a regular file image of the FCL file that can be downloaded too an off-host processing system.
116 File Change Log File Change Log administrative interface fcl_keeptime Specifies the duration in seconds that FCL records stay in the FCL file before they can be purged. The first records to be purged are the oldest ones, which are located at the beginning of the file. Additionally, records at the beginning of the file can be purged if allocation to the FCL file exceeds fcl_maxalloc bytes. The default value of fcl_keeptime is 0.
File Change Log File Change Log programmatic interface # fcladm off mount_point To remove the FCL file for a mounted file system, on which FCL must be turned off, type the following: # fcladm rm mount_point To obtain the current FCL state for a mounted file system, type the following: # fcladm state mount_point To enable tracking of the file opens along with access information with each event in the FCL, type the following: # fcladm set fileopen,accessinfo mount_point To stop tracking file I/O statisti
118 File Change Log File Change Log programmatic interface Backward compatibility Providing API access for the FCL feature allows backward compatibility for applications. The API allows applications to parse the FCL file independent of the FCL layout changes. Even if the hidden disk layout of the FCL changes, the API automatically translates the returned data to match the expected output record.
File Change Log Summary of API functions return EIO; } if (fclsb.
120 File Change Log Reverse path name lookup vxfs_fcl_seek() Extracts data from the specified cookie and then seeks to the specified offset. vxfs_fcl_seektime() Seeks to the first record in the FCL after the specified time. Reverse path name lookup The reverse path name lookup feature obtains the full path name of a file or directory from the inode number of that file or directory.
Chapter 9 Multi-volume file systems This chapter includes the following topics: ■ About multi-volume support ■ About volume types ■ Features implemented using multi-volume support ■ About volume sets ■ Creating multi-volume file systems ■ Converting a single volume file system to a multi-volume file system ■ Removing a volume from a multi-volume file system ■ About allocation policies ■ Assigning allocation policies ■ Querying allocation policies ■ Assigning pattern tables to director
122 Multi-volume file systems About multi-volume support About multi-volume support VxFS provides support for multi-volume file systems when used in conjunction with the Veritas Volume Manager. Using multi-volume support (MVS), a single file system can be created over multiple volumes, each volume having its own properties. For example, it is possible to place metadata on mirrored storage while placing file data on better-performing volume types such as RAID-1+0 (striped and mirrored).
Multi-volume file systems Features implemented using multi-volume support See “About Dynamic Storage Tiering” on page 143. ■ Placing the VxFS intent log on its own volume to minimize disk head movement and thereby increase performance. This functionality can be used to migrate from the Veritas QuickLog™ feature. ■ Separating Storage Checkpoints so that data allocated to a Storage Checkpoint is isolated from the rest of the file system. ■ Separating metadata from file data.
124 Multi-volume file systems About volume sets Volume availability is supported only on a file system with disk layout Version 7 or later. Note: Do not mount a multi-volume system with the ioerror=disable or ioerror=wdisable mount options if the volumes have different availability properties. Symantec recommends the ioerror=mdisable mount option both for cluster mounts and for local mounts. About volume sets Veritas Volume Manager exports a data object called a volume set.
Multi-volume file systems Creating multi-volume file systems 3 125 List the component volumes of the previously created volume set: # vxvset -g dg1 list myvset 4 VOLUME vol1 INDEX 0 LENGTH 20480 STATE ACTIVE CONTEXT - vol2 vol3 1 2 102400 102400 ACTIVE ACTIVE - Use the ls command to see that when a volume set is created, the volumes contained by the volume set are removed from the namespace and are instead accessed through the volume set name: # ls -l /dev/vx/rdsk/rootdg/myvset crw------- 1 r
126 Multi-volume file systems Creating multi-volume file systems Example of creating a multi-volume file system The following procedure is an example of creating a multi-volume file system.
Multi-volume file systems Converting a single volume file system to a multi-volume file system 4 List the volume availability flags using the fsvoladm command: # fsvoladm queryflags /mnt1 5 volname vol1 flags metadataok vol2 vol3 vol4 vol5 dataonly dataonly dataonly dataonly Increase the metadata space in the file system using the fsvoladm command: # fsvoladm clearflags dataonly /mnt1 vol2 # fsvoladm queryflags /mnt1 volname vol1 vol2 vol3 vol4 vol5 flags metadataok metadataok dataonly dataonly dat
128 Multi-volume file systems Removing a volume from a multi-volume file system 4 If the disk layout version is less than 6, upgrade to Version 7.
Multi-volume file systems About allocation policies Forcibly removing a volume If you must forcibly remove a volume from a file system, such as if a volume is permanently destroyed and you want to clean up the dangling pointers to the lost volume, use the fsck -o zapvol=volname command. The zapvol option performs a full file system check and zaps all inodes that refer to the specified volume. The fsck command prints the inode numbers of all files that the command destroys; the file names are not printed.
130 Multi-volume file systems Assigning allocation policies To assign allocation policies 1 List the volumes in the volume set: # vxvset -g rootdg list myvset VOLUME vol1 vol2 vol3 vol4 2 INDEX 0 1 2 3 LENGTH 102400 102400 102400 102400 STATE ACTIVE ACTIVE ACTIVE ACTIVE CONTEXT - Create a file system on the myvset volume set and mount the file system: # mkfs -F vxfs /dev/vx/rdsk/rootdg/myvset version 7 layout 204800 sectors, 102400 blocks of size 1024, log size 1024 blocks largefiles supported # m
Multi-volume file systems Querying allocation policies 3 131 Define three allocation policies, v1, bal_34, and rr_all, that allocate from the volumes using different methods: # fsapadm define /mnt1 v1 vol1 # fsapadm define -o balance -c 64k /mnt1 bal_34 vol3 vol4 # fsapadm define -o round-robin /mnt1 rr_all vol1 vol2 vol3 vol4 # fsapadm list /mnt1 name rr_all bal_34 v1 order round-robin balance as-given flags 0 0 0 chunk 0 64.
132 Multi-volume file systems Assigning pattern tables to directories Assigning pattern tables to directories A pattern table contains patterns against which a file's name and creating process' UID and GID are matched as a file is created in a specified directory. The first successful match is used to set the allocation policies of the file, taking precedence over inheriting per-file allocation policies. See the fsapadm(1M) manual page.
Multi-volume file systems Allocating data To assign pattern tables to directories 1 Define two allocation policies called mydata and mymeta to refer to the vol1 and vol2 volumes: # fsapadm define /mnt1 mydata vol1 # fsapadm define /mnt1 mymeta vol2 2 Assign the pattern table: # fsapadm assignfspat -F mypatternfile /mnt1 Allocating data The following script creates a large number of files to demonstrate the benefit of allocating data: i=1 while [ $i -lt 1000 ] do dd if=/dev/zero of=/mnt1/$i bs=65536 co
134 Multi-volume file systems Volume encapsulation Allocating data from vol1 to vol2 1 Define an allocation policy, lf_12, that allocates user data to the least full volume between vol1 and vol2: # fsapadm define -o least-full /mnt1 lf_12 vol1 vol2 2 Assign the allocation policy lf_12 as the data allocation policy to the file system mounted at /mnt1: # fsapadm assignfs /mnt1 lf_12 '' Metadata allocations use the default policy, as indicated by the empty string ('').
Multi-volume file systems Volume encapsulation To encapsulate a volume 1 List the volumes: # vxvset -g dg1 list myvset VOLUME vol1 vol2 INDEX 0 1 LENGTH 102400 102400 STATE ACTIVE ACTIVE CONTEXT - The volume set has two volumes.
136 Multi-volume file systems Reporting file extents 6 Encapsulate dbvol: # fsvoladm encapsulate /mnt1/dbfile dbvol 100m # ls -l /mnt1/dbfile -rw------- 1 root other 104857600 May 22 11:30 /mnt1/dbfile 7 Examine the contents of dbfile to see that it can be accessed as a file: # head -2 /mnt1/dbfile root:x:0:1:Super-User:/:/sbin/sh daemon:x:1:1::/: The passwd file that was written to the raw volume is now visible in the new file.
Multi-volume file systems Reporting file extents volume name, logical offset, and size of data extents, or the volume name and size of indirect extents associated with a file on a multi-volume file system. The fsvmap command maps volumes to the files that have extents on those volumes. See the fsmap(1M) and fsvmap(1M) manual pages. The fsmap command requires open() permission for each file or directory specified. Root permission is required to report the list of files with extents on a particular volume.
138 Multi-volume file systems Load balancing Load balancing An allocation policy with the balance allocation order can be defined and assigned to files that must have their allocations distributed at random between a set of specified volumes. Each extent associated with these files are limited to a maximum size that is defined as the required chunk size in the allocation policy. The distribution of the extents is mostly equal if none of the volumes are full or disabled.
Multi-volume file systems Converting a multi-volume file system to a single volume file system 139 extents on the volumes being removed are automatically relocated to other volumes within the policy. The following example redefines a policy that has four volumes by adding two new volumes, removing an existing volume, and enforcing the policy for rebalancing.
140 Multi-volume file systems Converting a multi-volume file system to a single volume file system Note: Steps 5, 6, 7, and 8 are optional, and can be performed if you prefer to remove the wrapper of the volume set object.
Multi-volume file systems Converting a multi-volume file system to a single volume file system 7 Edit the /etc/fstab file to replace the volume set name, vset1, with the volume device name, vol1.
142 Multi-volume file systems Converting a multi-volume file system to a single volume file system
Chapter 10 Dynamic Storage Tiering This chapter includes the following topics: ■ About Dynamic Storage Tiering ■ Placement classes ■ Administering placement policies ■ File placement policy grammar ■ File placement policy rules ■ Calculating I/O temperature and access temperature ■ Multiple criteria in file placement policy rule statements ■ File placement policy rule and statement ordering ■ File placement policies and extending files About Dynamic Storage Tiering VxFS uses multi-tier o
144 Dynamic Storage Tiering About Dynamic Storage Tiering Note: Some of the commands have changed or been removed between the 4.1 release and the 5.0 release to make placement policy management more user-friendly. The following commands have been removed: fsrpadm, fsmove, and fssweep. The output of the queryfile, queryfs, and list options of the fsapadm command now print the allocation order by name instead of number.
Dynamic Storage Tiering Placement classes Placement classes A placement class is a Dynamic Storage Tiering attribute of a given volume in a volume set of a multi-volume file system. This attribute is a character string, and is known as a volume tag. A volume can have different tags, one of which can be the placement class. The placement class tag makes a volume distinguishable by DST. Volume tags are organized as hierarchical name spaces in which periods separate the levels of the hierarchy .
146 Dynamic Storage Tiering Administering placement policies Tagging volumes as placement classes The following example tags the vsavola volume as placement class tier1, vsavolb as placement class tier2, vsavolc as placement class tier3, and vsavold as placement class tier4 using the vxadm command. To tag volumes ◆ Tag the volumes as placement classes: # vxvoladm -g cfsdg settag vsavola vxfs.placement_class.tier1 # vxvoladm -g cfsdg settag vsavolb vxfs.placement_class.
Dynamic Storage Tiering Administering placement policies for which each document is the current active policy. When a policy document is updated, SFMS can assign the updated document to all file systems whose current active policies are based on that document. By default, SFMS does not update file system active policies that have been created or modified locally, that is by the hosts that control the placement policies' file systems.
148 Dynamic Storage Tiering Administering placement policies Querying which files will be affected by enforcing a placement policy The following example uses the fsppadm query command to generate a list of files that will be affected by enforcing a placement policy. The command provides details about where the files currently reside, to where the files will be relocated, and which rule in the placement policy applies to the files.
Dynamic Storage Tiering Administering placement policies 149 You can specify the -T option to specify the placement classes that contain files for the fsppadm command to sweep and relocate selectively. You can specify the -T option only if the policy uses the Prefer criteria forIOTEMP. See the fsppadm(1M) manual page.
150 Dynamic Storage Tiering File placement policy grammar File placement policy grammar VxFS allocates and relocates files within a multi-volume file system based on properties in the file system metadata that pertains to the files. Placement decisions may be based on file name, directory of residence, time of last access, access frequency, file size, and ownership. An individual file system's criteria for allocating and relocating files are expressed in the file system's file placement policy.
Dynamic Storage Tiering File placement policy rules SELECT statement The VxFS placement policy rule SELECT statement designates the collection of files to which a rule applies.
152 Dynamic Storage Tiering File placement policy rules Either an exact file name or a pattern using a single wildcard character (*). For example, the pattern “abc*" denotes all files whose names begin with “abc". The pattern “abc.*" denotes all files whose names are exactly "abc" followed by a period and any extension. The pattern “*abc" denotes all files whose names end in “abc", even if the name is all or part of an extension. The pattern “*.
Dynamic Storage Tiering File placement policy rules In the following example, only files that reside in either the ora/db or the crash/dump directory, and whose owner is either user1 or user2 are selected for possible action: A rule may include multiple SELECT statements.
154 Dynamic Storage Tiering File placement policy rules the last rule in the policy document on which the file system's active placement policy is based should specify * as the only selection criterion in its SELECT statement, and a CREATE statement naming the desired placement class for files not selected by other rules.
Dynamic Storage Tiering File placement policy rules space for new files to which the rule applies on the specified placement classes. Failing that, VxFS resorts to its internal space allocation algorithms, so file allocation does not fail unless there is no available space any-where in the file system's volume set.
156 Dynamic Storage Tiering File placement policy rules tier2 1 The element with a value of one megabyte is specified for allocations on tier2 volumes. For files allocated on tier2 volumes, the first megabyte would be allocated on the first volume, the second on the second volume, and so forth.
Dynamic Storage Tiering File placement policy rules additional_placement_class_specifications relocation_conditions A RELOCATE statement contains the following clauses: An optional clause that contains a list of placement classes from whose volumes designated files should be relocated if the files meet the conditions specified in the clause.
158 Dynamic Storage Tiering File placement policy rules Indicates the placement classes to which qualifying files should be relocated. Unlike the source placement class list in a FROM clause, placement classes in a clause are specified in priority order. Files are relocated to volumes in the first specified placement class if possible, to the second if not, and so forth.
Dynamic Storage Tiering File placement policy rules This criterion is met when files are unmodified for a designated period or during a designated period relative to the time at which the fsppadm enforce command was issued. This criterion is met when files exceed or drop below a designated size or fall within a designated size range. This criterion is met when files exceed or drop below a designated I/O temperature, or fall within a designated I/O temperature range.
160 Dynamic Storage Tiering File placement policy rules max_access_age min_modification_age max_modification_age min_size max_size min_I/O_temperature
Dynamic Storage Tiering File placement policy rules Both the and elements require Flags attributes to direct their operation. For , the following Flags attributes values may be specified: gt The time of last access must be greater than the specified interval. eq The time of last access must be equal to the specified interval. gteq The time of last access must be greater than or equal to the specified interval. For , the following Flags attributes values may be specified.
162 Dynamic Storage Tiering File placement policy rules GB Gigabytes Specifying the I/O temperature relocation criterion The I/O temperature relocation criterion, , causes files to be relocated if their I/O temperatures rise above or drop below specified values over a specified period immediately prior to the time at which the fsppadm enforce command was issued. A file's I/O temperature is a measure of the read, write, or total I/O activity against it normalized to the file's size.
Dynamic Storage Tiering File placement policy rules I/O temperature is a softer measure of I/O activity than access age. With access age, a single access to a file resets the file's atime to the current time. In contrast, a file's I/O temperature decreases gradually as time passes without the file being accessed, and increases gradually as the file is accessed periodically.
164 Dynamic Storage Tiering File placement policy rules The files designated by the rule's SELECT statement that reside on volumes in placement class tier1 at the time the fsppadm enforce command executes would be unconditionally relocated to volumes in placement class tier2 as long as space permitted. This type of rule might be used, for example, with applications that create and access new files but seldom access existing files once they have been processed.
Dynamic Storage Tiering File placement policy rules The following example illustrates a possible companion rule that relocates files from tier2 volumes to tier1 ones based on their I/O temperatures. This rule might be used to return files that had been relocated to tier2 volumes due to inactivity to tier1 volumes when application activity against them increases.
166 Dynamic Storage Tiering File placement policy rules tier2 tier3 4 3 This rule relocates files whose 3-day I/O temperatures are less than 4 and which reside on tier1 volumes.
Dynamic Storage Tiering File placement policy rules tier2 10 100 tier3 100 This rule relocates files smaller than 10 megabytes to tier1 volumes, files between 10 and 100 megabytes to tier2 volumes, and files large
168 Dynamic Storage Tiering File placement policy rules placement_class_name additional_placement_class_specifications relocation_conditions A DELETE statement contains the following clauses: An optional clause that contains a list of placement classes from whose volumes designated files should be deleted if the files meet the conditions specified in the clause.
Dynamic Storage Tiering Calculating I/O temperature and access temperature 120 The first DELETE statement unconditionally deletes files designated by the rule's SELECT statement that reside on tier3 volumes when the fsppadm enforce command is issued. The absence of a clause in the DELETE statement indicates that deletion of designated files is unconditional.
170 Dynamic Storage Tiering Calculating I/O temperature and access temperature files that have experienced I/O activity in the recent past to be relocated to higher performing, perhaps more failure tolerant storage, ACCAGE is too coarse a filter.
Dynamic Storage Tiering Calculating I/O temperature and access temperature is the interval between the time at which the fsppadm enforce command was issued and that time minus the largest interval value specified in any element in the active policy.
172 Dynamic Storage Tiering Calculating I/O temperature and access temperature transfer activity, which is the the sum of bytes read and bytes written, should be used in the computation. For example, a 50 megabyte file that experienced less than 150 megabytes of data transfer over the 4-day period immediately preceding the fsppadm enforce scan would be a candidate for relocation. VxFS considers files that experience no activity over the period of interest to have an I/O temperature of zero.
Dynamic Storage Tiering Multiple criteria in file placement policy rule statements period are to be relocated to tier1 volumes. Bytes written to the file during the period of interest are not part of this calculation. Using I/O temperature rather than a binary indicator of activity as a criterion for file relocation gives administrators a granular level of control over automated file relocation that can be used to attune policies to application requirements.
174 Dynamic Storage Tiering Multiple criteria in file placement policy rule statements In the following example, a file must reside in one of db/datafiles, db/indexes, or db/logs and be owned by one of DBA_Manager, MFG_DBA, or HR_DBA to be designated for possible action: * tier2 other_statements
