This document describes enhancements and fixes introduced in the 10.3.2 release. For upgrade instructions and considerations, reference the Upgrade Guide for Recovery Series and Unitrends Backup.
Unitrends 10.3.2 builds on the success of Unitrends 10.3.1 with the following enhancements:
- Appliance – System memory usage has been tuned to boost performance.
- VMware backup – An appliance alert now displays if a VM has been removed from a schedule because it is no longer detected in inventory on the ESXi host.
The table below lists bugs that were resolved in this release. Unless stated otherwise, you can resolve each bug by simply upgrading your appliance.
- Windows image-level backup – Resolved an issue where SQL clusters and availability groups were erroneously listed as eligible for image-level backup.
- Scheduled jobs – Resolved an issue where scheduled jobs could fail with the error message: Error Failed to start schedule: *** glibc detected *** bputil: double free or corruption (!prev): 0x0000000000cbb820 ***.
- Hyper-V backup – Resolved an issue with hardware version 8 VMs that are configured to use cluster shared volumes (CSVs). Backup jobs would complete successfully, but could cause a CSV VSS writer error (Windows Event ID 1544.)
- Hyper-V backup – Resolved an issue where backups of Hyper-V 2008 R2 VMs could fail with the error message: Warning: Network connection stopped (time out).
- File-level backup – Resolved an issue where backups containing many very small files could fail or hang.
- SQL 2008 R2 – Corrected an issue where the system could incorrectly identify SQL 2008 R2 instances as SQL 2008. This could result in existing schedules not running for SQL 2008 R2 instances.
|Backup Copy (Cold)
- Cold copy of bare metal backups – Resolved an issue where cold copy jobs containing bare metal backups could fail on some older appliance releases.
- Importing cold copies of bare metal backups – Resolved an issue where importing cold backup copies of Windows bare metal backups and legacy cold bare metal backups could fail with the error: No full backup copy found for instance <…>.
- "Device is busy" error – Resolved an issue where cold copy jobs could fail due to a "device is busy" condition caused by unclean termination of a previous job.
- Malware scan virus definition updates – Resolved a schedule issue that prevented this weekly job from running.
- ssl_error_log – PHP warnings/notices will no longer display in the ssl_error_log.
- Attempt to recover a running backup – This error message now displays if a user attempts to recover from a running backup: Cannot restore in-progress backups.
- Unified bare metal recovery – Enhanced Windows unified bare metal recovery (UBMR) performance. UBMR speed is now faster than Windows replica restores of the same asset.
- UEFI-based replicas with GPT partitions – Resolved an issue in which replica restores could fail.
- Replica configured with SQL – Resolved an issue where replica restores of SQL databases could fail. This fix requires upgrading the Windows agent to version 10.3.2. The Windows replica issue is resolved automatically once a file-level backup is run by the 10.3.2 agent.
- Accessing a "live" Windows replica that resides on a Recovery-Series appliance – To access an on-appliance replica, you open a VNC viewer and connect by entering the appliance IP and a VNC port number. In prior releases, the VNC port only displayed for replicas running in audit mode. The VNC port now also displays for replicas running in live mode (on the Replicas tab under View Details).
- Cold backup copy reports – Corrected an issue where backup copy size was incorrectly reported in some cases.
- Cold backup copy reports – Reports now show the raw backup size (Original Size) as well as the actual space used on media (Size on Media).
- From address for appliance reports – Fixed an issue where hot backup copy email reports contained the incorrect From address for appliances that are configured to use reports@<ApplianceDomainName>.
- Min-Max Retention report – Copied assets are now included in the report.
- Hot backup copy reports – Resolved an issue where asset names for file-level hot backup copies were not displayed correctly.
- SQL 2008 R2 – Corrected an issue where SQL 2008 R2 assets could display as SQL 2008 in reports.
- Resolved an issue in which the High port security setting did not appear to close all of the correct ports.
- Instant Recovery tab – NAS backups are now greyed out and cannot be selected for instant recovery.
- Windows replicas – The Audit button is now disabled if a replica running in live mode is selected on the Replicas tab.
- Editing image-level assets – Clicking the Save button now closes the Edit Asset dialog.
- Links in error messages – Fixed links to KB articles that were displaying as plain text.
- Instant Recovery Details tab – Fixed a display issue that caused long messages to be illegible.
- Job schedule calendar – Jobs scheduled to run on the first Sunday of every month no longer display on every Sunday in the job calendar.
- SQL Server 2008 R2 – Resolved an issue where the R2 in SQL Server 2008 R2 did not display if it was listed below a SQL Server 2008 instance.
- System Notification report – Added a scroll bar to the Notification Information dialog so that all details for the selected row can now be viewed.
- Oracle and SharePoint recovery – Corrected inaccurate display of the number of concurrent recovery objects for SharePoint and Oracle recovery.