Server Backup 5.14.3 Release Notes

| *Release Date:* | 06/21/2016 |
| *Build Date:* | 06/15/2016 |
| *Linux Manager/Agent Build Version:* | 5.14.3 build 430 |
| *Windows Agent Build Version:* | 5.14.3 build 430 \\ |

The Disk Safe conversion process is automatic upon upgrade. During the upgrade process, Disk Safes will remain offline until they are individually converted. Once converted, the Disk Safe will be automatically re-opened for normal operations. Overall SBM performance will be significantly impacted while the Disk Safes are upgraded. Do not stop or restart any of the R1Soft processes during this upgrade unless specifically advised to do so by R1Soft Support.

*Please ensure that your Disk Safe storage devices have sufficient free space before proceeding with the upgrade. A minimum of 10GB plus any additional free space needed for normal operations is required.*

This release also introduces a new docstore process for SBM servers. This service is required for the successful operation of the SBM server.

See the [Upgrade Notes|http://wiki.r1soft.com/display/ServerBackup/Server+Backup+5.14.3+Release+Notes#ServerBackup5.14.3ReleaseNotes-UpgradeNotes] section below for additional information.
{note}

{info:title=SBM 5.14.3 build 428 Removed from Publication}
On June 14, 2016, we released Server Backup Manager 5.14 for General Availability to all R1Soft partners. We were made aware of a significant issue which impacted Policies utilizing the MySQL Add-on. The SBM 5.14.3 build 430 was released on June 21, 2016, to address this issue.

If your systems were impacted by this issue, please contact [R1Soft Support|mailto:support@r1soft.com] for remediation assistance.
{info}

h1. New Features

h5. Server Backup Manager

* Now supporting XFS filesystem file and folder restores in addition to Bare Metal Restores with 100% compatibility with previously used Disk Safes.
* Updated Disk Safe format to improve reliability and performance of access to recovery point meta data.
* In preparation for more extensive UI improvements, the options in the _Actions_ column in the _Extended Server Edit_ window have been updated to a more convenient dropdown menu.
* Optimized the Policies user interface to improve load times in large environments.
* Optimized the Recovery Points user interface to improve load times for Disk Safes with a large number of recovery points.

h1. Changes

h5. Server Backup Manager

* Validation of archive point limits within a Policy is now performed while retaining the configuration of the _Edit Policy_ window.
* When editing a Disk Safe or Volume where Recovery Point Archiving was disabled, the Recovery Point Archiving status is correctly maintained.
* Resolved an issue where an unsigned tinyint MySQL data type may not be properly restored.
* A closed Disk Safe will now report its proper status in the Statistics section of the Server Details.
* Resolved an issue where editing the API Server Options may generate an unhandled error.
* Properly handle attempts to import partial or malformed Disk Safes when importing Volumes.
* Addressed a Disk Safe replication issue where the resulting Disk Safe would contain significantly more data than required.
* Improved access control to undocumented API functions to minimize security exposure.

h5. Backup Agent

* Addressed NTFS write errors on Microsoft Windows protected machines.

h5. Data Center Console

* Resolved issue that incorrectly disabled the *Archive Now* button on the _Policies_ screen.
* Addressed a defect in getRegisteredCDPs SOAP API where disabled SBM servers would raise an exception.

h1. Upgrade Notes

h5. Server Backup Manager

This release introduces significant changes to the Disk Safe backup storage system. Following the software update, Disk Safes will be converted automatically upon startup or when imported into the SBM. This may result in longer than expected SBM startup times immediately following the update. Additionally, *Disk Safes will be closed until they are individually converted.* Once converted each Disk Safe will automatically re-open and resume normal operations. *Please allow this process to complete without interruption. Restarting the SBM during this process will significantly delay the overall transition.* Additionally:
* A single active conversion process may require up to 1 GB of free storage space. *Disk Safe conversions are processed in batches of 10 at a time.*
* Disk Safes running under 5.14 are not backwards compatible with previous SBM versions.
* During the Disk Safe update, the ratio of deltas to on-disk size may be temporarily elevated, but will normalize after a few minutes.

{info:title=Disk Safe "docstore" Process}
The updated Disk Safe format includes a new system service called the "docstore". This process is integral to all Disk Safe operations (backups, restores, etc.) and should be allowed to run uninterrupted.

For more information about docstore, [click here|ServerBackup:What Is Docstore?].
{info}

{info:title=APT Package Manager (Debian and Ubuntu Distributions)}
This release introduces the new package dependency of "r1soft-docstore". You will be unable to use "apt-get upgrade" to upgrade from an earlier installed version. Upgrades can be initiated using "apt-get install serverbackup-enterprise serverbackup-manager".
{info}

h5. Data Center Console

After completing the software updates in environments using Data Center Console, R1Soft recommends refreshing the associated SBM server data for proper syncing.

h5. Backup Agent

Backup Agent software requires a glibc version of 2.5 or greater.

{info:title=GNU C Library Compatibility}
Backup Agents on Linux distributions with older glibc libraries should remain on version 5.10 or earlier versions of the Backup Agent software. Refer to the following Knowledge Base article for more information: [Error - Linux cdp agent fails with "Floating Point Exception"|kb3:Error - Linux cdp agent fails with "Floating Point Exception"].
{info}


h2. KNOWN ISSUES

h5. Common Problems and Issues

* Some XFS volume groups are not recognized by SBM in Storage Configuration. This is due to a LVM static binary issue that should be fixed in a future product release.
* Site-to-Site (remote replication) Disk Safe replication tasks for Windows protected machines using Dynamic Disks may fail with protocol errors when replicating partition tables.
* Site-to-Site replication between a 5.12\+ SBM server and an older server is not possible due to SSL library changes.
* Multiple simultaneous Vacuum operations will significantly impact the overall performance of the SBM. R1Soft recommends running one Vacuum operation at a time.
* Multipath I/O storage configurations with duplicate device UUIDs are not supported and have been found to not function in some configurations.
* Partition table information is not gathered on KVM systems using the virtio driver. In order to perform a bare metal restore on these systems, the partition table must be manually recreated prior to the restore.
* When upgrading DCC and SBM from 5.6.2 to 5.12+, associated SBM servers may appear "offline". To resolve this issue, remove and re-add the SBM servers.
* The merge task may fail when you attempt to merge a recovery point that was interrupted due to CLOB/BLOB issues. If the merge fails due to this issue, merge the points one by one. After the merge fails, merge each recovery point one by one until the task is successful.
* Users who have Debian installations should note that a Debian-based machine that has both RPM and DEB packages installed is recognized as RPM-based only by Server Backup Manager.
* Users cannot set heap memory for more than 1024 MB using *serverbackup-setup \-m*. Even if there is more than enough space, Server Backup Manager displays a message indicating that memory specified must be smaller than 1024 MB. Edit server.conf to update memory settings.
* Some users are unable to create backup reports and receive the message, "Unknown exception:java.lang.NullPointerException." A workaround is to add a *FROM*: address to the *Set Report Options*.
* When Server Backup Manager is in Maintenance Mode, the policies can still be driven by DCC.
* During agent deployment,&nbsp;Server Backup displays an incorrect error message when there is not enough space available in {{/tmp}}. The current message, “Path not found /tmp,” is not correct. The message should read, “Not enough usable space in {{<tempdir>}} on host {{<hostname>}}.”
* Some users may experience an unknown exception error when adding a server and attempting to display disk information. This issue results when Server Backup runs as a local system account which does not have rights to read the disk. You can resolve this issue by changing the process to run as an administrator.
* Users who have more than 1,500 mount points per device may experience performance issues during a file restore or when attempting to browse recovery and archive points.
* When using the API you must designate the type of user on the "user=" line as either _\--user=dcc/someuser_ or _\--user=sbm/someuser_. Otherwise, the user defaults to local users.
* *{_}If your Windows environment contains an MSSQL or Exchange database on two different partitions{_}*, you should enable multi-volume snapshot (MVS) in the policy settings.

h5. Installation issues

* Server Backup 5.2.x and later do not support remote deployment to a Windows 2003 server because installing drivers requires user interaction.
* Users who attempt to install Server Backup Manager on a Microsoft Windows 2003 operating system must note that .NET 2.x or 3.5 is required. .NET 4.0 does not satisfy this requirement. When installing SBM on Windows 2003, the installer fails if the proper .NET version is not installed. Note that the installer does not prompt you to install the correct .NET version before failing.
* The temporary (.tmp) folder is not deleted after completing a silent install of Server Backup Manager or Backup Agent using the default values. The folder is empty, but still exists on the server.

h5. Restore issues

* Users who use Centos LiveCD on any bare-metal restores that have LVM may receive error messages that they failed to restore deltas to a certain device number. This issue results from the fact that because the Centos LiveCD uses its own LVM to manage its file systems, dm-0 and dm-1 are always in use. To avoid this issue, choose a target that is two LVM devices higher than the original. For example, if /tmp is originally on _/dev/dm-1_, you should choose _/dev/dm-3_ as your target.
* RecoveryCD versions 4.2.0 and 5.0.0 do not work when performing a bare-metal restore to a target that includes mdadm and LVM devices.
* Some users may receive an exception error message when attempting to exclude a large group of files in a directory when performing a restore.
* Microsoft Exchange Server 2007 users cannot restore a mailbox database to an alternate location.
* Some users may experience Java heap space out-of-memory error messages when attempting to restore a database with large LongBlob or LongText types.
* Disk safe verification verifies only those recovery points with an available status. DSV does not verify recovery points with a locked or incomplete status.

h5. Data Center Console issues

* When editing a Policy in DCC, the Recovery Point limit set at the Volume level can be exceeded. Volume objects must be refreshed in order to see the limit changes. This is done by refreshing SBM or by editing and re-saving the Volume.
* The Data Center Console does not automatically restart after a new Windows installation.
* The Data Center Console displays incorrect values in the *Compression Type* section of the Edit Disk Safe window after editing the values.
* Internet Explorer users who enable the Compatibility view may notice graphical display issues when attempting to view some Data Center Console pages, such as the Disk Usage, Volumes, and Policies pages. This issue does not occur in other Web browsers.

h5. Parallels Virtuozzo user issues

* Not all files are downloaded for a container on Virtuozzo and system displays the error message, {{ERROR - Failed to download file}}, in the log file.
* Parallels Cloud Server 6.0 users may notice that Virtuozzo virtual machines are not detected as devices in the Disk Safe wizard. The VMs are backed up. If you need granular file restore for a VM, you can back up the PCS VM by installing a Server Backup Agent in the VM guest operating system.
* Virtuozzo users may notice that when some containers are downloaded, the compressed file does not contain the first and last container listed in the recovery points.