PlateSpin Protect 10.3 Release Notes

September 28, 2012

Version 10.3 provides new features, enhancements, and bug fixes.

For Release Notes documents that accompanied previous 10.x releases, visit the PlateSpin Protect 10 Documentation Web Site and go to Previous Releases at the bottom of the main TOC page.

1.0 About This Release

1.1 New Features

  • Support for vSphere 5.1: This release provides support for the VMware vSphere 5.1 platform.

    You can now select a vSphere 5.1 DRS Cluster or an ESXi 5.1 Server as your protection container or your failback target.

    See the “Supported Configurations” section in your User Guide.

  • Support for new workloads: You can now protect workloads running Red Hat Enterprise Linux (RHEL) 5.6-5.8, 6.0-6.2, and 6.3. This also covers the Oracle Enterprise Linux (OEL) distributions that are based on these Linux versions.

    See the Supported Configurations section in your User Guide.

  • Localization: Product documentation and the integrated WebHelp system accompanying this release have been localized for the following languages: Chinese Simplified, Chinese Tranditional, Japanese, German, and French.

    Note that the English version of product documentation at the PlateSpin Protect 10 Documentation Web Site, which is being updated more frequently than the localized versions, should be considered the most current.

1.2 Discontinued Features

  • Discontinued storage management feature: (Linux) Logical Volume Manager (LVM) 1 is no longer supported. Consider upgrading your workloads to LVM 2.

2.0 Bug Fixes

This release addresses the following bugs:

  • 770964 (Windows) Problem running custom configuration script: An issue with how the product handled batch files prevented certain custom configuration scripts from running properly.

  • 753157 Replication reports by e-mail not functioning properly: In some situations the removal of an email account that was listed as a recipient for PlateSpin Protect e-mail notifications might cause erratic behavior, such as ‘flooding’.

  • 753449 (Windows) Workload hostname failing to change as required: In some cases the system might fail to assign a new hostname to the failover VM of a protected Windows Server 2008 workload when it was configured to join a domain.

  • 770996 Wrong user in Events report: In Events reports, all Add Workload jobs were erroneously shown as initiated by system, instead of the actual username.

  • 762850 (Linux) Unable to use non-default shells: PlateSpin Protect failed to protect Linux workloads that had a command line interpreter other than the Bash shell, which PlateSpin Protect Server uses by default. You can now override the default shell used by PlateSpin Protect Server to execute commands on a Linux workload.See KB Article 7010676.

  • 756871 (Linux) Incorrect sequence of 2 NICs on target after failover: In some cases an issue with target NIC mapping caused networking problems, such as the Novell eDirectory service binding to the wrong NIC.

  • 773097 (Windows XP) Incorrect SCSI controller type on failover VM: VM replicas of Windows XP workloads were being assigned BusLogic SCSI controllers (instead of LSI SCSI controllers), which negatively impacted failover functionality.

  • 768137 (Windows) Registry hives not replicating correctly during incrementals: In some cases an issue with how Windows Registry changes are handled in Windows Server 2003 and Windows XP might result in a mismatch between the Registries of a protected workload and its VM replica.

  • 734525 (Linux) Unable to connect to port 3725: An issue with how communication with a Linux workload with two NICs was being managed might occasionally cause problems in connectivity.

  • 744867 Problem inventorying NLB cluster hosts: In some cases an issue with the collection and processing of Windows Network Load Balancing (NLB) Cluster hosts might cause configuration problems on the VM replica.

3.0 Known Issues

  • Support for the GUID Partition Table (GPT) standard: PlateSpin Protect supports the protection of workloads that use the GPT disk partition layout standard. However, targets are always configured to boot from BIOS using an MBR (Master Boot Record). This limitation has the following implications:

    • Max 2 TB per volume: The maximum size of a protected workload’s volumes is restricted to 2.19 terabytes, the maximum for a partition allowed by MBR.

    • Physical targets for failback must boot from BIOS: Most hardware vendors provide support for multiple disk partitioning standards; for information on how to configure a physical target to boot from BIOS, or to reconfigure GPT hardware to operate in “legacy mode” (with support for BIOS), see your hardware vendor documentation.

    See also KB Article 7005452.

  • 781217 (SLES 9) Issue with volumes mounted using UUIDs: An issue with how mount points on SLES 9 workloads are looked up and how PlateSpin Protect handles Linux volumes might negatively impact the protection of SLES 9 workloads with volumes that are mounted by UUIDs. This issue is being investigated.

    Workaround: Modify the workload’s /etc/fstab configuration file to use device names instead of UUIDs for storage devices and partitions. See KB Article 7010812.

  • 737715 Unable relocate failover VM using Storage vMotion: In some circumstances, where your protection container is a VMware DRS Cluster in vSphere 5 and the initial replica of the workload is created incrementally, Storage vMotion might be unable to relocate the failover VM’s disk files across shared storage locations.

    Workarounds: To work around the issue, use one of the following:

    • Use the VMware vSphere Client to unregister and re-register the failover VM, then attempt to relocate the VM using Storage vMotion.

      - OR -

    • Apply the VMware ESXi 5.0 Patch ESXi500-201109401-BG, which addresses an underlying issue. Reboot the host (required), then attempt to relocate the VM using Storage vMotion.

      For further technical information about the issue, see VMware Knowledge Base article 2005740.

  • 737057, 736959 (vSphere 5 Storage DRS) Missing target datastore or inaccurate volume-to-datastore map: When viewing or editing a workload protection contract, the volume-to-datastore map might be inaccurate. The target datastore might be reported as missing in the PlateSpin Protect Web Interface, which might in turn report a validation error. In most cases this is the result of moving a VM to a different datastore in the vSphere 5 Storage DRS (Distributed Resource Scheduler) system of the vSphere 5 platform.

    Workaround: In most cases, the issue has no functional impact on the protection contract. However, manually renaming a datastore before the first successful replication might render the protection contract unusable. Avoid manipulating the failover VM or its underlying DRS structure to preserve the integrity of your protection contract.

  • 686911 Problems with file downloads from or uploads to datastore: Under certain conditions, where the protection target is a VMware DRS Cluster, the system might fail to upload or download a file, such as a boot ISO image. This negatively impacts a protection contract.

    See KB Article 7008306.

  • 756454 (vSphere 5) Recovery points are stored in the same datastore as the target virtual machines’ VMDK files: When protecting a workload to a vSphere 5 DRS Cluster or ESXi Server container, indicating a Configuration File Datastore location only determines the storage location of the failover VMs’ .VMX file, but not the storage location of Recovery Point snapshots. This might result in inaccurate free space calculation, impacting validation.

    For further information, see KB Article 7005494.

  • 595490 Preserving boot partition on failback causes failback to stall: In some failback scenarios, the system improperly allows you to preserve an active (or boot) partition on the target, preventing the target from booting properly. This issue is under investigation.

    Workaround: In Failback Details, do not opt to preserve any boot partitions on the target.

  • 702152 Protection over a WAN takes a long time if VM container has a large number of datastores: Under some circumstances the process of locating the appropriate ISO image required for booting the target might take longer than expected. This might happen when your PlateSpin Protect Server is connected to the VM container over a WAN and your VM container has a large number of datastores. This issue is under investigation.

  • 698611 Full cluster replication failure under certain circumstances: If a Windows 2008 R2 Cluster protection contract is set up through the sync to an existing VM method, and if the active cluster node flips prior to the full replication, the full replication job fails.

    See KB Article 7008771.

  • 655828 Failure to mount NSS volumes: Upon failover or test failover, NSS volumes with snapshots enabled are not automatically mounted as expected.

    See KB Article 7008773.

  • 638392 ESX 4.1: Direct host discovery results in missing VM port groups if dvSwitch port groups share the same name.

    Workaround: Ensure that port group names are unique.

  • 680259 (VMware 4.1) Poor networking performance by traffic-forwarding VMs: In some scenarios, the replica of a workload that is forwarding network traffic (for example, if the workload’s purpose is to serve as a network bridge for NAT, VPN, or a firewall) might show significant network performance degradation. This is related to a problem with VMXNET 2 and VMXNET 3 adapters that have LRO (large receive offload) enabled.

    Workaround: Disable LRO on the virtual network adapter. For details, see the VMware vSphere 4.1 Release Notes. Sscroll down to the bulleted item Poor TCP performance....

  • No software RAID support for Linux workloads: PlateSpin Protect does not support Linux workloads with volumes on software RAID.

  • 590635 Inconsistent failover results after upgrading: Following an upgrade to PlateSpin Protect, a failover operation might fail to complete or might not apply the correct failover parameters, such as the proper hostname and workgroup settings.

    Workaround: Before performing a failover, run a replication.

  • 581860 Browser exception in the Chinese edition of the product: Attempting to connect to the PlateSpin Protect Server with a browser that does not have a specific version of Chinese added might result in Web server errors. For correct operation, use your browser’s configuration settings to add a specific Chinese language (for example, Chinese [zh-cn] or Chinese [zh-tw]). Do not use the culture-neutral Chinese [zh] language.

  • 610918 Unresponsive Expand and Collapse icons in integrated help: On some systems with enhanced browser security settings (such as Internet Explorer 8 on Windows Server 2008), the Expand and Collapse icons (+ and -) in the Table of Contents might fail to work. To fix the issue, enable JavaScript in your browser:

    • Internet Explorer: Click Tools > Internet Options > Security tab > Internet zone > Custom level, then select the Enable option for the Active Scripting feature.

    • Firefox: Click Tools > Options > Content tab, then select the Enable JavaScript option.

  • 558937 Failure of block-level replications that use VSS (Windows): If you are using third-party VSS-based backup software, block-level replications might occasionally fail.

    Workaround: Use blackout windows (see “Protection Tiers” in your User Guide).