How To speed up case backups

book

Article ID: 100038397

calendar_today

Updated On:

Description

INTRODUCTION

Although the case backup time is proportional to the number of unique documents in eDiscovery, the presence of any large production exports has a much bigger impact on the case backup time. An image production export for example, can consist of thousands, sometimes even millions, of small TIF files. Load File Imports also generate an XML file for each document family. In addition, use of Load File Import custom fields mapped with the data type "Long Text" also generate a separate file for each custom field value. Since a case backup is heavily disk I/O intensive, the presence of many small files can increase the case backup time considerably.  
 

HOW TO DETERMINE IF A CASE IS LIKELY TO CONTAIN LARGE EXPORTS

One way to determine if a case backup is likely to be including one or more large exports, is to examine the file times of the files created in the case backup. 

You can get a listing of the file times in a case backup as follows:

  • Change to the directory containing the case backup. For example:
    d:\>cd CW\V##\caseBackups\cases\0.6.7.346939725-SECvTamas\BACKUP_Mon

  • Generate a file listing and redirect it to an output file for review:
    dir /s /t:w > D:\case_backup_listing.log
     

The following example shows the file times of the top-level directories in an example case backup:

Directory of D:\CW\V##\caseBackups\cases\0.6.7.346939725-SECvTamas\BACKUP_Mon\cluster
04/05/2014 04:23 AM .
04/05/2014 04:23 AM ..
04/05/2014 02:35 AM _lds_case_a82593ltl3
04/05/2014 04:23 AM _lds_case_appliance_a82593ltl5
04/05/2014 04:23 AM _lds_case_group_a82593ltl4
04/05/2014 04:23 AM _lds_email_locator_a82593ltl7
04/05/2014 04:22 AM _lds_index_a82593ltla
04/05/2014 04:23 AM _lds_merge_a82593ltlb

These times indicate that the backup finished around 04:23 AM, and started no later than 02:35 AM.


Looking at the times of the files inside the_lds_case_directory:

Directory of D:\CW\V##\caseBackups\cases\0.6.7.346939725-SECvTamas\BACKUP_Mon\cluster\_lds_case_a82593ltl3
04/05/2014 02:35 AM .
04/05/2014 02:35 AM ..
04/05/2014 04:21 AM data
04/05/2014 02:34 AM 428 dataStoreMeta.properties
04/05/2014 02:35 AM 24,504,164 dsdbdump_lds_case_a82593ltl3.sql
2 File(s) 24,504,592 bytes

Notice how the MySQL part of the backup (filedsdbdump_lds_case_a82593ltl3.sql) finished at 02:35 AM, but the “data” directory has a last modified time much later at 04:21 AM. The “data” directory includes all the files from any production exports (or print jobs). The last modified time of this “data” directory indicates that the majority of the case backup time was spent creating (and compressing) the backup files in the “data” directory. This would suggest that there may be many files there from a production export.

 

HOW TO REMOVE UNWANTED EXPORTS FROM A CASE

An easy way to reduce the case backup time is to remove any unwanted production exports from the case.

Use the following steps to identify and remove any unwanted production exports:

  • Go to the “System > Jobs” page and filter by job type “Export” (Figure 1):

    Figure 1.

  • Click on the yellow icon to the left of “Success” in the “Status” column to see the job log (statusLog.txt) file.
    Scroll to the bottom to see where the export files are located (Figure 2):

    Figure 2.

    This shows that the files can be found under the following directory:
    D:\CW\V##\data\esadb\dataStore_case_aigo6swdfr_13100101\fileManager\0.14.7.1\jobRun_350

    If we look in this directory then we see the following (Figure 3):

    Figure 3.

    The XML_EXPORT_CONTENT sub-directory contains all of our TIF images (Figure 4):

    Figure 4.

  • Backup the D:\CW\V##\data\esadb\dataStore_case_aigo6swdfr_13100101\fileManager\0.14.7.1\jobRun_350 directory (if you want to keep the export).

  • Delete the export from the “System > Jobs” page, by clicking on the trashcan icon (Figure 5):

    Figure 5.

    Once this has been deleted, the directory D:\CW\V##\data\esadb\dataStore_case_aigo6swdfr_13100101\fileManager\0.14.7.1\jobRun_350 will no longer exist, and the case backup will be smaller and therefore quicker.
     

 

Issue/Introduction

How To speed up case backups