| *Release Date:* | 04/18/2018 |
| *Build Date:* | 03/13/2018 |
| *Linux Manager/Agent Build Version:* | 6.6.2 build 57 |
| *Windows Agent Build Version:* | 6.6.2 build 57 \\ |
----
{toc:location=top|maxLevel=1|minLevel=1|type=flat|separator=pipe|style=border:1}
----
{note:title=Important Upgrade Notice}
Customers upgrading from Server Backup Manager 6.2.2\+ can upgrade as normal, otherwise please upgrade to 6.2.2 first.
{note}
h1. New Features
* SBM will alert if the protected machine is running Windows with 4K native drives and the agent version is not at least 6.6.2
h1. Fixed Issues
h5. Server Backup Manager
* Agent upgrade now works from SBM installed on CentOS 6.9 and earlier.
h5. Server Backup Agent
* Resolved an issue with change block tracking on Windows with 4K native hard drives.
h1. Upgrade Notes
You must have an active maintenance account to upgrade. If you have a perpetual (non-heartbeat) license, it will be converted to a heartbeat license automatically upon upgrade. Make sure your SBM has _https_ connectivity to complete this conversion. During the upgrade, you will be notified of the change and will be given the option to revert back to the previous version you were running.
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}
h1. Known Issues
{color:#000000}{*}Common Problems and Issues{*}{color}
* 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.
* 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.
* 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.
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..
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.
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 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.
h5.
| *Build Date:* | 03/13/2018 |
| *Linux Manager/Agent Build Version:* | 6.6.2 build 57 |
| *Windows Agent Build Version:* | 6.6.2 build 57 \\ |
----
{toc:location=top|maxLevel=1|minLevel=1|type=flat|separator=pipe|style=border:1}
----
{note:title=Important Upgrade Notice}
Customers upgrading from Server Backup Manager 6.2.2\+ can upgrade as normal, otherwise please upgrade to 6.2.2 first.
{note}
h1. New Features
* SBM will alert if the protected machine is running Windows with 4K native drives and the agent version is not at least 6.6.2
h1. Fixed Issues
h5. Server Backup Manager
* Agent upgrade now works from SBM installed on CentOS 6.9 and earlier.
h5. Server Backup Agent
* Resolved an issue with change block tracking on Windows with 4K native hard drives.
h1. Upgrade Notes
You must have an active maintenance account to upgrade. If you have a perpetual (non-heartbeat) license, it will be converted to a heartbeat license automatically upon upgrade. Make sure your SBM has _https_ connectivity to complete this conversion. During the upgrade, you will be notified of the change and will be given the option to revert back to the previous version you were running.
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}
h1. Known Issues
{color:#000000}{*}Common Problems and Issues{*}{color}
* 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.
* 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.
* 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.
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..
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.
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 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.
h5.