|
|
STORE and TurboSTORE/iX Products Manual: 900 Series HP 3000 Computer Systems > Chapter 6 STORE and TurboSTORE/iX Command OptionsSTORE Options |
|
Whenever you store a set of files, the system displays the total number of files stored at your terminal. If there was an error, it also displays the number and names of files not stored and the reason each was not stored. Use the SHOW parameter and its options to display the names and additional information about the files stored and to list them on your system printer as well as at your terminal. Internally, the STORE program writes information about the files it stores to a file with the formal file designator SYSLIST. It prints file information at your terminal by equating SYSLIST with the standard listing device for your session (your terminal). You can redirect the file information to another file or device by issuing a FILE command that assigns SYSLIST to that device or file. For example, you might want to redirect the file information to a disk file so that you can keep it online for your records. To write STORE information to the disk file, SHOWFILE, enter:
To print information on the files you stored at the system printer as well as at your terminal, use the SHOW=OFFLINE option:
To redirect the OFFLINE listing to another file or device, issue the following FILE command, it will send a copy of the STORE listing to the disk file SHOWOFFL:
To generate a store listing to the terminal as well as to a disk file, enter the previous FILE command followed by the STORE command:
The STORE command gives you two options for displaying additional information about the files stored. Using the SHOW parameter, you can choose to display file information in either long or short form. A short file display prints the file's full name, volume restrictions, number of sectors, file code, and media number for each file you store. In addition to there being a long and a short SHOW listing, there are also two different versions of each - MPE and HFS. The type of listing printed depends on the input fileset list. You will get a HFS style listing if:
The HFS style listing contains the same information as the MPE style listing. The only difference is that the filename is printed as an absolute pathname at the end of the line. Since HFS pathnames can be of variable length, and can be quite long, they are placed at the end of the line so that they will not affect the spacing of the other columns. If the name does not fit on one line, it will be wrapped to the next line. An asterisk (*) will be placed at the last column of the line to indicate that the rest of the filename will continue on the next line.
To display the list of files stored using the short form in MPE format, use the SHOW=SHORT option in your STORE command. For example:
Figure 6-1 “Short File Display (MPE Format)” illustrates a short file display in MPE format using the above example. Figure 6-1 Short File Display (MPE Format)
This example uses the same fileset as the first example, except that it is specified in HFS syntax.
Figure 6-2 “Short File Display (HFS Format)” illustrates a short file display in HFS format using the above example. Figure 6-2 Short File Display (HFS Format)
This example shows how the HFS syntax names will be wrapped, if they cause the line to exceed 80 characters:
Figure 6-3 “Short File Display with Wrapped HFS Names” illustrates a short file display in HFS format with line wrap using the above example. Figure 6-3 Short File Display with Wrapped HFS Names
A long file display contains the same information as a short file display plus each file's ending media number, record size, blocking factor, maximum number of extents allowed, end-of-file, and file record limit. Once again, there are two formats for the long listing -- MPE and HFS. The HFS format contains the same information as the MPE format, except that the variable-length HFS filename is at the end of the line. It will be wrapped in the same was as the short listing if the filename length exceeds the line length. To display the list of files stored using the long form in MPE format, use the SHOW=LONG option. For example:
Figure 6-4 “Long File Display (MPE Format)” illustrates a long file display from the above example. Figure 6-4 Long File Display (MPE Format)
Figure 6-5 “Long File Display (HFS Format)” illustrates the same example as above, except specifying the filenames in HFS syntax: Figure 6-5 Long File Display (HFS Format)
If you use the SHOW parameter, but do not specify either LONG or SHORT, the system displays files in short form when the record size of the output device or file (SYSLIST) contains fewer than 132 characters in native mode and 114 characters in transport mode. The system displays files in long form when the record size is equal to or greater than 132 characters in native mode and equal to or greater than 114 characters in transport mode. You can combine either LONG or SHORT with any of the other SHOW options, except NAMESONLY, but you cannot specify both LONG and SHORT at the same time within a STORE command. To display each file's creation date, last access date, and last modification date for the files stored, use the SHOW=DATES option. For example:
Figure 6-6 “File Date Information (MPE Format)” illustrates a file listing created from the above example. Figure 6-6 File Date Information (MPE Format)
When the output listing is in HFS format, an additional field is added to the dates display. The STATE_CHANGE field indicates the last date that a file attribute was changed. An example listing is: Figure 6-7 “File Date Information (HFS Format)” illustrates a file listing created from the above example in HFS format. Figure 6-7 File Date Information (HFS Format)
To display file security information for the files stored, use the SHOW=SECURITY option. The listing includes the file creator's user and account name, the security matrix, and a flag indicating the presence or absence of an ACD. The HFS format displays the same security information. For example:
Figure 6-8 “File Security Information” illustrates a file listing with security information from the above example. Figure 6-8 File Security Information
The SHOW=NAMESONLY allows an abbreviated listing to be displayed. The only fields shown are the starting and ending media number, and the name of the file. This option is most useful when displaying HFS syntax files. Since they can be longer than MPE syntax names, it is sometimes useful to allow more room on the line to display the full pathname. NAMESONLY cannot be used with the SHORT or LONG options to SHOW. It has a slightly different format for MPE and HFS. The HFS style has the filename as the last field, while the MPE style has the filename as the first field. Figure 6-9 “Names Only Format” is an example for HFS format. If SHOW=PATH is specified, the output STORE listing will be in HFS format, even if no HFS syntax files are being stored. You may want to use this so that all of your listings appear in the same format. When creating backups using the ONLINE=START or ONLINE=END options of the TurboSTORE/iX 7x24 True-Online Backup product, additional information is shown in the STORE listing. This additional information is a single character, immediately following the volume restrictions field. The values and meanings for this character are as follows:
An example of these flags follows. Note that FOO5 was modified during the backup, FOO6's file label information was updated, FOO9 was removed from the backup, and FOO1 was added to the backup.
When STORE encounters an error, it either automatically recovers or terminates, depending upon the nature of the error and the error recovery method you choose. The following unrecoverable errors always cause STORE to terminate:
If STORE encounters a disk read error while storing a file, it skips the remainder of that file and sends an error message to the standard listing device describing the offset from the start of the file and the number of bytes that have been lost. STORE then continues to copy the remainder of the files. The STORE operation does not terminate. Since the files that contain disk errors are not stored, you will not be able to restore them. When a media error occurs while writing to a output device, the behavior of STORE depends on the value of the ONERROR option, as explained below. The STORE command's ONERROR parameter lets you specify a backup error recovery procedure. Your options are ONERROR=QUIT and ONERROR=REDO. REDO is the default. QUIT instructs STORE to abort upon encountering a backup I/O error while REDO instructs STORE to continue after encountering an I/O error. If you specify REDO and the STORE command encounters a backup I/O error on any device, STORE sends a message to your terminal (or the file or device to which you have assigned the file SYSLIST). This message indicates where the error occurred. For example, STORESET=(*T1,*T2,*T3) is requested and a backup error occurs on the device referenced by *T2 while it is creating reel 2. STORE rewinds the reel to the load point, marks the reel as bad, and asks you to mount another reel. After you mount the new reel, STORE automatically stores all the files from the point where the bad reel began. The message is similar to the following:
If an optical disk write error is encountered, STORE displays the same error messages that it would display for tape backup devices. You can determine which piece of media is has been marked bad by using the STATUS command of the MOUTIL program. The status display indicates which piece of media is bad by showing "THE MEDIA IS BAD" in the media name field. If a disk write error occurs while storing to a disk file, STORE aborts, irrespective of what ONERROR option was specified. Once a piece of media has been marked bad, it should not be used for future STORE operations. If a media that has been marked bad is mounted for use in a STORE operation, STORE will issue the following message to the console:
If you reply YES, then STORE will write over this media. If you reply NO, the media will be unloaded, and you should mount another piece of media.
The DATE option specifies which files are stored based on two possibilities:
You can use the DATE parameter to store only files that were modified or their states were changed since a specified date. (The state change date and time are updated any time the file is renamed or an ACD is changed. However, this information is not automatically shown. The state change date and time can be different from the last modified date displayed with the LISTFILE filename,3 command. See the note below for information on displaying the state changes.) The DATE parameter enables you to do, among other things, a partial system backup. For example, suppose a full system backup was performed on September 30, 1993. The following STORE command would copy all system and user files modified since that date:
The date must be entered in either the above format (mm/dd/yy) or as mm/dd/yyyy (for example, 09/30/1993). The files stored include any modified on the same date as the date entered. The files stored also include those modified since the specified date on nonsystem volume sets that are online. Use this form of the DATE parameter when you perform regular backups of a fileset. Store the entire set, for example, once a week and store only those files that were modified or their states were changed since the full backup the remaining six days.
You can also use the DATE parameter to store only files that have not been accessed since a certain date. For example:
The year can be entered as two digits or four digits (93 or 1993). The above command stores any files in the OPERATOR group of the SYS account that have not been accessed since June 10, 1993. You may want to use this form of the DATE parameter with the PURGE parameter to remove out-of-date files from your system. Refer to the next section. If your system contains infrequently used or out-of-date files, you can store them onto backup and permanently remove them from the system disks at the same time. This process, called archiving files, increases available space and provides a backup copy of the files in case you need them in the future. To store disk files and then delete the disk files in a single operation, use the PURGE parameter of the STORE command. For example:
The command above stores all files in the OLD group of the account ACCTG and deletes the files from disk. PURGE deletes the files only after it has successfully stored all of them. Thus, if STORE terminates in the middle of the process, it does not purge any files. You can combine the DATE and PURGE parameters to periodically archive unused files. For example, the following STORE command stores to backup all files in the EMPRECS group of the PERSONEL account that have not been accessed since February 2, 1992 and then deletes the disk files:
You may want to warn users before you archive files and inform them how to request to have archived files restored. If you need to restore purged files to the system, you can find them on the piece of media to which you stored them and on the media from the last full system backup you performed before you purged the files. Use the PROGRESS parameter to display STORE command progress messages at regular intervals. For example, to display progress messages every five minutes, use the following command:
If you use the PROGRESS parameter alone, without specifying an interval, STORE displays status messages every minute. If STORE is being run from a MPE session, the progress messages will be send to the standard list ($STDLIST). Otherwise, the messages will go to the system console. The system directory keeps track of the accounts, groups, users, and files in your system. Nonsystem volume set directories keep track of the contents of nonsystem volumes. Directory information can be stored when you back up system or nonsystem volumes. Storing directory information helps you rebuild your system from backups should it unexpectedly lose information. To store system directory information, use the DIRECTORY parameter. All HFS directories on the system are also stored. The following example will store all files on the system, plus the system volume set accounting information.
To store nonsystem volume set directory information, use DIRECTORY and ONVS in the same command. For example:
In the above example, the ONVS parameter will limit @.@.@ to only files that reside on the specified volume sets. Only the directory information for groups and accounts on the specified volume sets will be stored. The system directory information will NOT be stored. See "Storing Files From Volume Sets" in this chapter for more information. As illustrated above, the DIRECTORY parameter of the STORE command allows you to back up the system and/or nonsystem volume set directories (account structure). You may then reload the account structure in the event of a system disaster by restoring the directory onto the system from the backup media. You have the added flexibility of backing up the directories of nonsystem volumes attached to the system as a part of or as a separate entity from the system. This allows a nonsystem volume (volume set) to become a separate reliable entity. System supervisor (OP) or system manager (SM) capability is required to use the DIRECTORY parameter.
Here is an example of how to back up the system directory and all other files along with a nonsystem volume set directory and its files:
where NV1 is the name of a nonsystem volume set. The TRANSPORT option stores files on media that is compatible with:
Use the TRANSPORT parameter of STORE to create STORE backups that are compatible with MPE V/E. For example, to store all of the files in the PERSONEL account onto an MPE V/E compatible backup, enter:
The backup that is produced with the TRANSPORT parameter uses the same format as MPE V/E STORE backup format. Therefore, files that are not compatible with MPE V/E (due to file/record size limitations or new file types), are not stored if TRANSPORT is specified. The TRANSPORT option cannot be used with the following STORE and TurboSTORE options:
Character set specifications, as listed in "Using Wildcard Characters" in chapter 5, "Storing Files," cannot be used with the TRANSPORT parameter. With the addition of HFS syntax files and other new POSIX-related features, mechanisms exist in STORE to allow backups to be created on a POSIX system and restored on a pre-POSIX system (4.0 or earlier). Pre-POSIX backups are created using the TRANSPORT=MPEXL option. For example:
There are two general scenarios when files would need to be transferred to a pre-POSIX system:
With these two cases in mind, STORE has the following objectives with respect to data transport:
The following sections list the features in STORE which address these objectives. Every POSIX STORE media is created such that any HFS syntax file can be restored to a pre-POSIX system with a unique MPE syntax name. When restoring a backup to a pre-POSIX system, the fileset of @.@.@HFSACCT will contain all HFS syntax files on that media. These files are named with the MPE syntax of:
Where ####### represents numbered files starting at 0000001. A mapping file is also on the media. This file specifies the mapping between the _HFSACCT name and the actual HFS name. The name of this file is HFSMAP._HFSGRP._HFSACCT. Sample entries in the file would look similar to this:
Since _HFSGRP and _HFSACCT are illegal MPE filenames, special steps must be taken when attempting to restore these files on a pre-POSIX system. The LOCAL option can be used to restore the HFS files into your logon group. For example, on a pre-POSIX system, issue:
Alternately, you can specify a specific group and account to place the HFS files, using the CREATE, GROUP, and ACCOUNT options:
This will restore all of the HFS syntax files to the group NEWGRP.MYACCT. The HFSMAP file will allow to you identify the real names of the F####### files. If you want to only restore MPE syntax files from the backup, that can be accomplished by the following command:
An argument of the TRANSPORT option allows POSIX style ACD's to be translated to pre-POSIX systems, while maintaining the correct security. Using TRANSPORT=MPEXL will cause STORE to write out the pre-POSIX format ACD's. If a POSIX media is created without TRANSPORT=MPEXL, and then restored to a pre-POSIX system, the ACD's will not be understood and the system will enforce a "creator only" security. This provides a more restrictive security. Also, all POSIX-related fields of the file labels being stored are cleared on the backup media. The FILES parameter works with the TRANSPORT parameter. By default, MPE/iX stores a maximum of 4000 files at a time. If you are storing more than 4000 files using the TRANSPORT parameter, use the FILES= parameter to specify the maximum number of files to be stored. If you do not use the TRANSPORT parameter, the FILES= parameter is ignored. For example, to set the maximum number of files stored to 6000 for a TRANSPORT backup, you might enter the following:
Use the MPE/iX RESTORE command to restore files from MPE V/E format backup tapes to an MPE/iX system. MPE/iX RESTORE automatically detects that the tapes are in MPE V/E format, and invokes CM(RE)STORE (Compatibility Mode Restore) to restore the files. An additional way to manage the security of files being stored is to use the COPYACD and NOACD options. The default is COPYACD. This means that in addition to the MPE security matrix, a file's ACD (if one exists) is written to backup. This allows you to store the extra security information associated with an ACD. If for some reason you want to store files without their ACD, the NOACD option can be specified. This would remove the extra security associated with ACD's, which may be desired if you are moving files to a system where the current ACD's would be inapplicable or undesired. When using these options, keep in mind that RESTORE also has COPYACD and NOACD options. So, even if you create a STORE media with ACD's, you have the choice at RESTORE time to restore or not restore the ACD information. Use the FCRANGE parameter to store files with certain file codes. MPE XL file codes distinguish different types of files. You can select up to eight file code ranges to store. For example, files with codes 1100, 1101, and 1102 are HPWORD files. Files with codes 1152 and 1153 are SLATE files. To store all HPWORD and SLATE files in the LETTERS group of the PERSONEL account enter:
Files are stored with greater efficiency by using the maximum backup record (buffer) size allocated to a given backup device. Performance increases because data transmission is more efficient and because the amount of media used for interrecord gaps decreases. The maximum backup record size of MPE/iX supported backup devices differs from device to device and is shown in Table 6-1 “Maximum Tape Record Sizes”. Table 6-1 Maximum Tape Record Sizes
MAXTAPEBUF should be used when storing to magnetic tape, DDS, or 3480 cartridges. Currently, it is not necessary to specify MAXTAPEBUF when storing to MO devices or disk files, since the maximum record size of 32K is already being used. Normally when a STORE command is issued, a default record size of 16 kilobytes is used. However, by using the MAXTAPEBUF parameter of the STORE command, the maximum record size of the specified device is automatically used. If multiple backup devices are specified, using the STORESET parameter, in conjunction with MAXTAPEBUF, STORE uses the maximum record size, which is the lowest common denominator of the different devices. This is necessary to ensure that the record sizes are the same on all volumes of the backup set.
To increase record size when storing files, use the MAXTAPEBUF parameter in your STORE command. For example:
During the STORE process, the fileset that you are storing is usually unavailable for modification. Files can be modified again after STORE is completed. If you are doing an online backup using:
The NOTIFY option can be used to help automate the process of performing some action at the time where the files being stored are available for modification, either at the end of the STORE, at the end of the attach phase, or at the end of the sync point. NOTIFY allows STORE to stream a user-specified job file (using the MPE/iX STREAM command) either at the end of the attach phase (while using ONLINE), at the end of a successful STORE operation (normal backup), or at the end of the sync point (ONLINE=START or ONLINE=END). The job file is specified by the user by setting up a formal file designator named NOTIFY, prior to running the store job. This job can contain various actions that notify users that the files are free, start other jobs that modify the files, or allow users back on to the system, for example. If the specified job does not contain passwords, STORE will prompt the user for the passwords if STORE is being run from a session. If STORE is being run from a job and passwords are not present, the job stream will fail. When the job fails to stream, due to lack of passwords, or other reasons, all error messages will be sent to the standard list ($STDLIST). All other output as the result of streaming the job will also be sent to $STDLIST. The following is an example of using the NOTIFY command in conjunction with a user specified job file. First, set up the file equation, using the FILE command:
Where the job stream could contain, for example:
Then, run STORE with the NOTIFY option:
Since this is an online store, the job stream will be executed once the files have been attached, and a message would then be sent to all users telling them that their files are now free to be accessed. In addition to the counts of files stored and not stored that are normally reported at the end of a backup, the STATISTICS option provides extra data about the backup. This extra information includes:
If software compression is enabled with the COMPRESS option, additional data on the amount of compressed data written, and compression ratios and percentages is displayed. If an online or 7x24 true-online backup is performed, then information on the amount of log data written is displayed. All data totals are shown in bytes. To convert these totals to sectors, divide by 256. When displaying the amount of backup data written, up to five columns of data can be displayed, depending on what options were specified for the backup. These columns display the following backup data information:
These data columns are displayed for each piece of media written to during the backup. If a parallel backup is performed (using the STORESET or MOSET options), then the media will be grouped by parallel set. After all parallel sets are displayed, totals for each parallel set, and then for the entire backup are displayed. Look for the "Backup Total" line to determine the total amount of all data written for the backup. The Throughput section follows the data written columns. The throughput for a particular piece of media is computed by dividing the total amount of data written for a piece of media by the time it took to completely write that data. The time required to switch media is NOT accounted for in this computation. Only time spent actually writing the media is used. Also, note that if software compression is being used, then the amount of compressed data written is used to compute the throughput. Your virtual throughput may actually be higher, if software compression is used. The number of hardware retries encountered when writing the data is also displayed here. This only applied to magnetic tape, DDS, and 3480 media. It has no meaning when writing to disk files or magneto-optical devices, and will always be zero. Monitoring the number of retries for a piece of media is important in determining if that media is starting to go bad. An increased number of retries could indicate that the media should no longer be used, or that your devices should be cleaned and/or serviced. After the Throughput section, the total elapsed time for the backup is displayed. This time does include media switches. If software compression is used for the backup (with COMPRESS=LOW or COMPRESS=HIGH), then a section on compression ratios will follow the Throughput section. For each media, or each media in each parallel set, the amount of uncompressed file data written is displayed with the amount of data actually written when that file data is compressed. The ratio between these two numbers and a compression percentage is also displayed. The higher the ratio is, the better the data compressed. Similarly, the higher the percentage, the better the data compressed. A negative percentage is possible, and indicates that the data you were storing did not compress at all. If you see this in your statistics display, or see very low compression ratios, it indicates that you are storing data that does not compress well, and that it might be better not to use software compression, or try a different software compression (LOW instead of HIGH, for example).
Following are examples of output from the STATISTICS options. The following example is of a normal backup to a single device without software compression or online backup.
The following example is of a backup to two pieces of media.
The following example is of a backup to parallel devices with software compression and online backup.
To store the files on a particular volume set, use the ONVS parameter of the STORE command. ONVS lets you store files from up to 20 volume sets at once. For example, to store all files from the nonsystem volume sets named VOL_SET_A and VOL_SET_B, enter:
To store files from system and nonsystem volumes, include the system volume set name within the ONVS parameter. For example:
Use the SPLITVS option to store files that are located on the specified split volume sets. These volumes are also known as mirrored disks. This provides a form of concurrent, or online backup, because you can:
The SPLITVS option requires:
The basic procedure to use split volumes and STORE together is:
For more information on using mirrored disks, see the Mirrored Disk/iX User's Guide (30349-90003).
When using the RENAME option, an original file refers to the original disk file, and a target file indicates the final filename on the backup media. Renaming a file requires the capabilities necessary to manipulate files across groups and accounts. Table 6-2 “Required Capabilities for RENAME” assigns capabilities to the different levels of file renaming. Table 6-2 Required Capabilities for RENAME
The RENAME option has the following syntax: STORE file set [=targetname [:creatorname [.creatoraccount] ]] [,file set[=targetname [:creatorname [.creatoraccount] ] ]] ;*TAPE;RENAME;SHOW The fileset consists of the inclusion fileset and all exclusion filesets. All legal filesets are allowed. The targetname can be any legal MPE or HFS filename with the following restrictions on wildcards:
Each component of the creator name must either be equal to @ or have no wildcards. The account part of the creator name may be omitted. The @ is used to specify that the component in the original filename should be used for that component of the target filename. When specifying HFS pathnames, components will be replaced from the root on down. However, if the last component of the target pathname is an @ the last component of the original pathname will be used. If there is not a corresponding component for an @ in the original pathname the file will not be stored. The following tables illustrate sample translations for the target filenames:
The following table illustrates sample translations for a target file directory when the fileset specification was not a directory:
The following table illustrates sample translations for a target file directory when the fileset specification was a directory:
In a similar manner, the wildcard @ is used to specify the use of the original creator name component in the new creator name. If no creator is specified, the original creator is kept. This is equivalent to specifying @.@ in the creator field. If the creator name is specified without an account part, the account part of the creator name will be made equal to the account of the new file. There is an exception to this rule when the new filename does not have a legal account name in its pathname. A directory off of the root node that has a legal account name will be considered the account for the creator. This is true whether this directory is an account or not. Table 6-3 “Creator File Translations” below illustrates the translations for the creator. All files below in the "sg" group of the "sa" account were created by "CREATOR.CRACCT." Table 6-3 Creator File Translations
If the SHOW option is specified, only the target names are listed. The security option of the SHOW option displays the target creator name if one is specified. The ACD associated with a file is not changed when the file is renamed. As with other STORE operations, the ACD is copied unless the NOACD option is specified. If you need to change the ACD, use the ALTSEC command. You can specify the same target file information for more than one org file. However, when two or more files have the same name on a backup created using the RENAME option, RESTORE will write the second file over the first file. The files excluded may only be specified on the original side of the fileset. As with normal exclusion, there is no limit to the number of excluded filesets. The example below shows an exclusion file specification:
Lockwords are not allowed on target files, and if specified cause STORE to abort. The options that are not supported with the RENAME option and will cause STORE to abort are:
By default, STORE uses the last character of a HFS filename to determine if the specified file or fileset should be scanned recursively to include all files below the specified wildcards or directories. If a filename ends in a /, then it will be scanned recursively to include all files below it in the hierarchical directory. Otherwise, the file or fileset is not scanned recursively, resulting in a horizontal cut at its level in the hierarchical directory. Using TREE and NOTREE can override that default behavior. If TREE is specified, ALL files and filesets are scanned recursively, regardless of their ending character. If NOTREE is specified, then NO files or filesets are recursively scanned, resulting in a horizontal cut in the directory structure. By using the new options PARTIALDB or FULLDB, you can back up either a partial or an entire TurboIMAGE database and/or ALLBASE/SQL DBEnvironment. The type of backup you are performing determines the type of option you would use:
When performing a 7x24 true-online backup using the ONLINE=START or ONLINE=END options, by default, STORE backs up the entire database.
Specification of the TurboIMAGE root file or the ALLBASE/SQL DBE file causes STORE to store the entire database. If a dataset file is specified without the root file, then STORE prints the following warning:
If any of the dataset files are specified in addition to the root file, no warning will be displayed but the individual dataset files will be counted as redundantly specified files. If you want to back up a partial database, you can use the PARTIALDB option. STORE will backup only those dataset files specified in the fileset list.
When performing a non-7x24 true-online backup (where you do not specify the 7x24 true-online backup options, ONLINE=START or ONLINE=END), by default, STORE allows a partial database backup of only those database files included in the fileset list. For example:
If you want to store an entire database, use the new option, FULLDB. To increase the ease and speed of file recovery, all STORE and TurboSTORE products have the ability to put a copy of a backup's STORE label and STORE directory into a disk file. This directory information allows RESTORE to more quickly determine where files are located on the backup, and to prompt users to mount the necessary media. If you want to enable this feature, specify the STOREDIR[ECTORY] option.
The filename used by STORE uniquely identifies the file and associates it with a specific backup. The file is placed in the store_dirs HFS directory in the HPSTORE.SYS group.
The following naming convention is used:
For example:
This name contains the date and time of the backup and the pin of the STORE process. This naming convention allows a LISTFILE,2 to show a chronological list of STORE directory files. You should not rename or purge STORE directory files that you plan on using, since RESTORE looks for a specifically named file. However, if the file does not exist, all files can still be restored from the backup. The size of the directory file is directly proportional to the number of files stored in the backup. You can control how much disk space is used for these directory files by putting a disk file limit on HPSTORE.SYS. The following example limits the space used by the HPSTORE group to 4000 sectors:
You can have STORE associate a shorter, more meaningful name to the disk directory files by specifying an optional filename argument to STOREDIR. You can specify an MPE or HFS filename. STORE then creates a symbolic link with the specified name, which points to the original disk file. For example, to create a full backup on Friday and give it the name FRIFULL, issue the command:
To create a partial backup of all files that were modified since May 4, 1994, and create a disk directory filename, WedPartial, in the current working directory, issue the command:
If you want to create a 7x24 true-online backup with the sync point at the end of the backup, and do not want a disk directory file to be created, specify the NOSTOREDIRECTORY or NOSTOREDIR option on the STORE command line. With STORE, TurboSTORE/iX II, and TurboSTORE/iX 7x24 True-Online Backup, you have the capability of simultaneously reading information from different files on multiple disk drives and storing the accessed information on the backup device in a single operation. This process called interleaving allows you to match slow input devices (such as disk drives), with fast output devices (such as streaming tape devices). This effectively increases the input rate. By varying the number of concurrent input operations, the effective input rate can be matched with the output rate. This allows the backup to execute closer to the speed of the output device. When you STORE files in interleave format, STORE:
Interleave format significantly reduces the time necessary to store a large set of files, such as when performing a full system backup. It merges the stream of data from multiple files on separate disks into a single stream of data sent to the backup device or devices. If necessary, STORE processes files out of sequence to keep backup devices operating efficiently. Interleave format is most efficient when your fileset is evenly spread across two or more system disks. Use sequential devices or parallel devices alone when the set of files to store is smaller and less evenly distributed, such as when doing a partial backup, or if the files are only coming from one disk. To STORE files in interleave format, use the INTER parameter. For example:
When these files are restored, RESTORE will automatically recognize the interleave format and restore the files properly. Using the STORESET and INTER parameters together improves backup time on large systems. For example, the following STORE command uses both parallel device pools and interleaving to minimize backup time:
|
|