Summary
This article outlines recommended methods for performing new DPX block base backup jobs.
Step By Step
When running a new DPX block base backup, Catalogic Software recommends that you consider the following:
1. Define impact.
- Determine the number of DPX Block level for Snapvault primary and DPX Block Level machines that require a new base backup.
- Evaluate the retention date and schedule for the Block Level backups and plan for the best implementation time.
2. Apply the latest updates that are found on Catalogic Software Online Customer Support ( mysupport.catalogicsoftware.com ) to each DPX Block Level node.
- Updates that include a different version of the snap driver will require you to reboot the DPX Block level clients.
- Newly applied modules will ensure the integrity of the future backup jobs.
- We recommend using the bexchkdsk.exe utility on all DPX Block level for Snapvault primary nodes to establish the integrity of the hard disk that will be snap shot and to enforce base backups. This utility can be found on the site mysupport.catalogicsoftware.com.
NOTE: The bexchkdsk.exe utility described in Step 2 will reset change journaling for its appropriate disk. After using this utility, the next backup will run as a base and step 3 will become unnecessary for these disks.
3. Force base backup jobs using one of the two methods described below. For each method the appropriate amount of disk space must be prepared on the Netapp Secondary(7-mode or CDOT) or the DPX OSS Servers in question.
DPX Block Level backup jobs track disk changes on the block level. Thus, forcing base backup jobs to run for a given disk entails resetting Change Journaling functionality that tracks the disk's changed blocks.
a. Use the BEXSNAPMGR utility to reset Change Journaling. The BEXSNAPMGR utility should have been installed on your Windows machine as part of Step 2, described above. NOTE: The following steps solely describe the "Change Journaling Configuration" aspect of running the BEXSNAPMGR utility. Other capabilities of this utility should not be used.
- Open the DPX Command Prompt on the DPX Block level for Snapvault Primary node by selecting Start-> Programs ->DPX -> DPX Command Prompt. ( right-click run as administrator)
- Run bexsnapmgr.bat.
- Use the DPX management console to Disable Change Journaling on the given node from the "Change Journal Configuration" tab
- Use the DPX management console to Enable Change Journaling on the given node from the "Change Journal Configuration" tab.
c. Create or schedule a new job name. Since the job name is one of the main elements establishing the relationship between the DPX Block level for Snapvault primary and DPX Block level for Snapvault secondary as well as the DPX Block level and DPX OSS Server, defining a new job name will establish a new relationship and all prior jobs--even those that back up the same data--will be unaffected. These prior jobs will remain in the DPX catalog until they expire.
Remember to delete old jobs from the schedule so they do not conflict with the new jobs. In order to create a new job you can load an existing job and "Save As" a new job name.
After all old jobs are expired, the Qtree must be manually deleted to free up space on the filer. For more information about how to delete the Qtree.
4. Ongoing maintenance.
a. Catalogic Software recommends regularly using the Instant Availability utility for all DPX Block level backup(This utility can be found at the following location on the Master server in DPX/bin directory: or setting up simple volume verification How to enable simple/light verification for Block level jobs for additional validation.
b. Validate jobs when DPX Block level jobs are not being run. This will prevent backup jobs from running while the LUN is being mapped and can circumvent issues described in knowledge base article Cannot Delete Storage System Snapshot: Understanding the LUN-Busy Snapshot Condition
5. Customers that require a seeding process for the initial base backup are advised to contact Catalogic Technical Support for further assistance.
Running an initial Block level base backup that contains a large amount of data or utilizes a low communication link, such as a dial-up telephone line or fractional T1, may be impractical. The process of creating an initial Block Level base backup in these situations is sometimes referred to as seeding.
6. Customers that need to perform a new base on a Windows cluster are advised to contact Catalogic Technical Support for further assistance.