Selecting a STORE Error Recovery Method

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 errors always cause STORE to terminate:

A command syntax error
An error in the file system directory
An error opening the tape file or an indirect file (a text file containing STORE command and options; refer to "Using Indirect Files" in the previous section)

Disk read errors

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 native mode 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 onto the tape. The STORE operation does not terminate. Since the files that contain disk errors are not stored, you will not be able to restore them.

Tape error recovery

The STORE command's ONERROR parameter lets you specify a tape error recovery procedure. Your options are ONERROR=QUIT and ONERROR=REDO.

QUIT instructs STORE to abort upon encountering a tape I/O error.
REDO instructs STORE to continue after encountering an I/O error. REDO is the default option.

If you specify REDO and the STORE command encounters a tape I/O error on a device, STORE sends an informational message to your terminal (or the file or device to which you have assigned the file SYSLIST). It 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.

To display native mode messages

STORE displays the following messages during tape error recovery:

The user sees the message:

   STORE ENCOUNTERED MEDIA WRITE ERROR ON LDEV #

The operator sees messages similar to the following:

   STORE IS MARKING REEL BAD ON LDEV #



   MOUNT NEXT REEL FOR STORE ON LDEV #

To check file listings for errors

Use the SHOW parameter of the STORE command to check file listings for errors. As STORE copies files to tape, it will list them on the SYSLIST provided you use the SHOW parameter. For example, if you enter this:

   STORE T@.SREXP.SYS;*T;SHOW

The following might be listed on the SYSLIST:



STORE/RESTORE VERSION A.21.01  (C) 1986 HEWLETT-PACKARD

CO. TUE, JAN 6, 1990,  2:57 PM



FILENAME GROUP  ACCOUNT  VOLUME RESTRICTIONS       SECTORS CODE   REEL

TDORSERR.SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S    32          1

TSRBUGFX.SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S   160          1

TSRDOC  .SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S   672          1

TSRERR  .SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S   240          1

TSRGLOB .SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S    64          1

TSRINCL1.SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S   656          1

TSRINCL2.SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S   480          1

TSRINCL3.SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S   272          1

TSRTAPE .SREXP .SYS      MPEXL_SYSTEM_VOLUME_SET  :S   384          1



FILES STORED:  9

When STORE finishes, you see the total number of files stored and if there is an error, the number not stored. Use the listing to verify that the correct files were indeed stored and that none were omitted.

To verify that STORE files are error free

After using the STORE program to write files to tape, you may want to check to be sure that no errors occurred in the process. The VSTORE command allows you to verify that any or all the files were correctly stored. VSTORE can also be used to verify the file system directory or to list the date that a given file was stored.

To verify one or more files on a STORE tape, mount the tape on your backup device, as described in chapter 8. (The write ring should be removed as a protection.) Put the device online.

As with the STORE and RESTORE commands, you should precede the VSTORE command with a FILE command, assigning a file name to the tape drive. If you do not assign a file name, a default name is utilized based on your username.

For example, to verify the files on a tape enter:

   FILE T;DEV=TAPE

Now issue a VSTORE command that backreferences the tape file and includes the file or files you want to verify. For example,

   VSTORE *T;/

If you have omitted the FILE command, you would enter

   VSTORE ;/

In addition to the tape file name and the names of the files to be verified, four parameters are available: SHOW, ONERR[OR], LOCAL, and DIRECTORY. The following is a brief description of each of these parameters:

Use the SHOW parameter to list the names of each file successfully verified. (If you do not specify this parameter, the total number of files verified is displayed.) Additional options used with the SHOW parameter include short and long form descriptions of files, creation, last access, and last modification date, security status, and an offline printing option.
Use the ONERR[OR] parameter to inform VSTORE what to do upon encountering an error within a file. In addition to the QUIT option, which causes the verification process to quit when an error is encountered, you have a second option, SKIP, that allows verification to proceed after skipping a faulty file. SKIP is the default.
Use the DIRECTORY parameter to verify the file system directory on the tape. Use of this parameter requires SM or OP capability.

The following are examples of the above parameters:

   VSTORE *T;/;SHOW

   VSTORE *T;FILE1.MANAGER.SYS;SHOW;ONERR=SKIP

   VSTORE *T;/;DIRECTORY

The VSTORE command is not valid for a STORE tape created with the TRANSPORT parameter.

For a full discussion of all VSTORE command options, refer to the MPE/iX Commands Reference Manual (32650-90003).

To retain the STORE tapes

Unload each tape and check for correct labeling. If you used the SHOW=OFFLINE option of the STORE command, the system has printed a list of the files you stored. Retrieve the report from your printer and either file it with your records or attach it directly to one or more tape reels. Place each tape in your tape library.

To allow users back on the system

If you prevented users from accessing the system during your backup, as discussed in chapter 8, "Backing Up the System," you need to allow users to access the system again. Here are the steps to follow:

Reset the job and session limits to their original values. Enter:
LIMIT nn,nn
To check that you've used the correct numbers, enter:
SHOWJOB STATUS
The last line should report the correct job limit (JLIMIT) and session limit (SLIMIT) for your computer. If the values are incorrect, enter the LIMIT command again with the corrected values.
Reset the system jobfence to its original value by entering:
JOBFENCE nn
To check that the jobfence is correct, enter:
SHOWJOB STATUS

Restart any suspended jobs. To determine whether any jobs had been temporarily stopped, enter:

   SHOWJOB

The system lists your session and any suspended jobs in the following way:



JOBNUM   STATE   IPRI JIN     JLIST     INTRODUCED   JOB NAME

#S185    EXEC          20     20        FRI  1:03P   OPERATOR.SYS

#J17     SUSP          10S    PP        FRI  5:02P   AJOB,MRS.T

2 JOBS:

   0 INTRO; 0 SCHEDULED

   0 WAIT; INCL 0 DEFERRED

   1 EXEC; INCL 1 SESSIONS

   1 SUSP



JOBFENCE= nn; JLIMIT= nn; SLIMIT= nn

The example shows one suspended job (#J17). To restart that job, enter:

   RESUMEJOB #J17

Repeat the RESUMEJOB command for each job listed as SUSP. When you have done this for each suspended job, check that all jobs have been restarted by entering:

   SHOWJOB

The list should look nearly identical, except that jobs that were suspended before should now be listed as executing or EXEC:



JOBNUM   STATE   IPRI JIN     JLIST     INTRODUCED   JOB NAME

#S185    EXEC          20     20        FRI  1:03P   OPERATOR.SYS

#J17     EXEC          10S    PP        FRI  5:02P   AJOB,MRS.T



2 JOBS:

   0 INTRO; 0 SCHEDULED

   0 WAIT; INCL 0 DEFERRED

   2 EXEC; INCL 1 SESSIONS

   0 SUSP

JOBFENCE= nn; JLIMIT= nn; SLIMIT= nn

When you have returned the system to its normal status and restarted suspended jobs, the backup is complete, and users can begin using the system again.