Recovering lost studies

Running the tool

Run the tool in each mode, in the order listed.

To recover an archive volume

1. Run the utility in mark_affected_locations mode.

2. Run the utility in pre_recovery_assessment mode.

3. Run the utility in recovery_attempt mode.

4. Run the utility in post_recovery_assessment mode.

5. Run the utility in clean_up mode.

1. Run the tool in mark_affected_locations mode

Command

In a Windows command prompt, change to the C:\mvf\bin\ directory and type:

mvf_archive_recovery mark_affected_locations -v volume_ref [-b begin_ordinal[,end_ordinal]]

In a Solaris a terminal window, type:

mvf-archive-recovery mark_affected_locations -v volume_ref [-b begin_ordinal[,end_ordinal]]

Description of mode

This mode uses the list of unreadable block addresses or ordinals for the recovered volume to determine which objects are no longer retrievable from this volume. In the database, the locations for these objects (mostly images) are marked as in recovery by setting their visible attribute to R.

If no begin and end ordinals are specified, the entire volume is marked as in recovery. If only the beginning ordinal is specified, objects from the begin ordinal to the end of the volume are marked as in recovery.

The object count for the studies containing objects in recovery is updated to delete their archive locations on this volume; the system no longer considers these objects to be stored on this volume. This is important for any of these objects that may be in cache. After their archive locations on this volume have been deleted, they are protected against deletion by the autopilot until they have been re-archived.

Tip: The mark_affected_locations command can be run several times to cover multiple ordinal ranges.

2. Run the tool in pre_recovery_assessment mode

Command

In a Windows command prompt, change to the C:\mvf\bin\ directory and type:

mvf_archive_recovery pre_recovery_assessment -v volume_ref -r report_file_name

In a Solaris terminal window, type:

mvf-archive-recovery pre_recovery_assessment -v volume_ref -r report_file_name

Description of mode

This function writes a pre-recovery assessment report into the specified report filename. If the report filename already exists, the date and time are appended to the report filename (for example, pre_report.20010725_142535). The pre-recovery assessment report generates a list of all object locations for the studies being recovered for the specified archive volume_ref. It lists studies being recovered and for each study, the number of objects in the study, the number of recoverable objects, the number of unrecoverable objects, and the locations of all the study's objects.

To indicate which carriers must be online so studies can be recovered, a list of carrier IDs is generated for the system administrator.

Example of Pre-recovery assessment for library volume_ref 20

3. Run the tool in recovery_attempt mode

Command

In a Windows command prompt, change to the C:\mvf\bin\ directory and type:

mvf_archive_recovery recovery_attempt -v volume_ref -n maximum_number_of_jobs

In a Solaris terminal window, type:

mvf-archive-recovery recovery_attempt -v volume_ref -n maximum_number_of_jobs

Description of mode

The recovered volume must be imported before performing this step.

This function tries to create maximum_number_of_jobs SEND jobs, one for each study to be recovered. The SEND jobs gather all of these studies into the cache on the server where the volume recovery tool is run so that the studies can be re-archived on the same archive that owns the lost or recovered volume.

If a SEND job is successfully created, the visible attribute for the objects is changed from R to S to indicate that a SEND job has been submitted. The protected flag of the entry in the dosr_study_location table for the study on the local machine's cache is set to T so that the study cannot be prematurely deleted.

When specifying the maximum number of SEND jobs, select a reasonable number such that the system is not overloaded with too many SEND jobs. The number of recoverable studies is mentioned in the pre-recovery assessment report.

Tip: If many studies must be recovered, the recovery_attempt command can be run multiple times to break up the work into small batches to limit the load on the system.

4. Run the tool in post_recovery_attempt mode

Command

In a Windows command prompt, change to the C:\mvf\bin\ directory and type:

mvf_archive_recovery post_recovery_assessment -v volume_ref -r report_file_name

In a Solaris terminal window, type:

mvf-archive-recovery post_recovery_assessment -v volume_ref -r report_file_name

Description of mode

This function writes a post-recovery assessment report into the specified report filename. If the report filename already exists, the date and time are appended to the report filename. This report shows the recovery status of all studies being recovered. It lists all the studies in recovery and, for each study, lists:

• The status of the recovery attempt (fully recovered, partially recovered, and unrecoverable)

• The study_ref

• The number of objects in the study

• The number of recoverable objects

• The number of objects that have been recovered so far

• The study uid

• The accession number

• The patient name

• The patient ID

• The modality

• The station name

If the status is blank, the study has not been processed yet.

At the end of the report, a list of statistics is provided showing the total number of studies that have been processed so far, the number of studies that were fully recovered, the number of studies that were partially recovered, and the number of unrecoverable studies.

Tip: For a better view of the post-recovery assessment report, import the file into a spreadsheet.

Example of post_recovery_assessment report for library volume_ref 21

5. Run the tool in clean_up mode

Command

In a Windows command prompt, change to the C:\mvf\bin\ directory and type:

mvf_archive_recovery clean_up -v volume_ref

In Solaris, at a terminal window, type:

mvf-archive-recovery clean_up -v volume_ref

Description of mode

Note: To perform the STORE jobs generated by this step, the archive requires a fresh or open volume, even if new studies are not being stored by this archive.

This function cleans up the osr_location table for all recovered studies. After a study has been recovered, the visible attribute for the object locations on the recovered volume is changed to F. When the autopilot wakes up, it deletes these records from the osr_location table. Next, the study is unprotected. An entry is created in the dosr_study_destination table to ensure that the study is archived more quickly.

If the entire volume has been recovered so that the volume_ref is no longer being referenced, the volume recovery tool first deletes these entries from the osr_location table and then removes the volume_ref from the database.

Note: Because the objects are no longer referenced in the osr_location table as being archived, Autopilot cannot delete them from cache before they are re-archived.