Installation Notes

Install Update

Please read the following release notes carefully, before continuing. When installing this update on a Collax V-Bien cluster the following steps have to be performed:

/ Important information

Please have a look at the Collax V-Bien 6.5.28 Release Notes if the nodes have Version 6.5.26 or above and haven’t been updated yet. For this update, it is necessary that both nodes are set in maintenance mode simultaneously.

Procedure

1. Click at High Availability within the Dashboard.
2. Select the other (not the local) and change into the Start Maintenance dialog.
3. Now, put this other node into maintenance mode by confirming the query in the dialog. Wait until the maintenance mode is reached. Depending on the number of virtual machines and hard disk, this can take some time. Virtual machines will be migrated live, which can take a few minutes depending on the size of the RAM. The progress will be displayed in the list of background activities.
4. Close the dialog.
5. In the list of nodes, select again the other node whose software is to be updated by right-clicking it.
6. Change to System Update ….
7. Click Check for Updates. The completion will be displayed as “100%”.
8. Click Download and Install. This function will download the latest software packages from the update server and install them. The completion will be displayed as Install - 100%.
9. Close the dialog and set the node to active.
10. Please repeat this procedure for the local node.

New in this Version

Extended desciption field of a vm

The “comment” field for VMs is beeing extended within this release to a multi-line input field and the complete description can be shown in the tooltip.

Extended virtual hard disc without reboot of vm

The dialog “Virtual Hard Disk” lists all drives that are available for use in virtual machines in the cluster. If a virtual hard disk provides too little space for the application, the hard disk can be enlarged. For the virtual machine to be able to use the enlarged hard-disk space, since now the VM had to be shut down and rebooted. Thereafter, the file system of the VM must be resized if this does not take place automatically. Within this release this is not longer necessary any more. If the hard disk is of the type “Virtio”, it may be sufficient to re-read the volume within the VM, depending of the operating system of the VM.

Mail Server Port adjustable

In this version a port for the mail server can be specified. The adjustment of another mail server port may be needed if the relay-server does not accept E-mails on standard port 25.

Boot System UEFI

In this version you can choose the UEFI mode over the default legacy BIOS for virtual machines. The dialogue is located under the extended settings of the virtual machine.

VESA and QXL graphics card

In this version VESA is used as default graphics card for nely created virtual machines. The dialogue is located under the extended settings of the virtual machine.

New Windows Virtio Driver for virtual Machines

The Virtio drivers for Windows will be renewed with this update. From this update on virtio driver version virtio-win-0.1.126 are available.

Changelog / Windows Virtio Drivers

QEMU Guest Support

With this update you can use a virtio serial device to communicate between the virtual machine and the hyperviros. There are virtioserial drivers and guest-agent tools on the Virtio Driver CD which have to be installed therefore. Systems with running QEMU guest-agent service now support shutdown/reboot via guest-agent per default. Fall back to the usage of ACPI when necessary.

Info

Kernel configuration item CPUFreq governor

Within this release the Linux kernel configuration item CONFIG_CPU_FREQ_DEFAULT_GOV use “performance” as default resulting in lower latency and higher performance.

Configuration Management Details

Changes in the Web interface are not adopted directly in the system. Rather, a configuration must be “activated” explicitly. The performed changes are listed more detailled within this release. In the pop-up under Details, you can see which settings have been changed.

Embedded SAN (DRBD)

Within this release, the replicating block device driver and tools drbd version 8.4 are going to be installed.

Fujitsu ServerView RAID Manager netservice

This update adds the netservice “Fujitsu ServerView” (port tcp/3173) to the list of services.

RDP Access to VM Consoles

With this update a GUI-button was added to the dialog of the virtual machine management to directly establish a connection via the protocols RDP (Remote Desktop Protocol) to the VM Consoles.

Additional hardware support

This update brings support for additional network interface cards based on the Intel xgbe driver v5.1.3 and i40e driver 2.0.26 for the Intel platform. These are X552, X553, X540, X722 und XXV710 (25 Gbps Ethernet) to number a few.

Network drives under new protocol

Backups of VMs can be backed up to network drives. The protocol used for this is updated to version SMB 2.0. In this way, security, stability, and performance are significantly increased.

different timeoption enabled by default for all virtual machines

Due to changed default settings for virtual machines the time drift is lower than before. Within this update these settings apply to all virtual maschines per default.

Issues Fixed in this Version

Preferred Node and Shut Off VM

To achieve optimum utilization of the cluster nodes, you may want to migrate a virtual machine to another cluster node. Due to an erroneous in the programm code, the location constraint was not right in all cases. If the VM is inactive and a preference is set, the VM will be started on the specified node the next time it is started.

Bug in Intel Skylake/Kaby Lake processors

Systems with the Intel processors code-named “Skylake” and “Kaby Lake” could, in some situations, dangerously misbehave. The microcode Update microcode-20170511 fixed this issue with this update for Intel Skylake processors. When using Kaby Lake processors, it’s recommended to disable Hyper-Threading in the BIOS. Also see here :

Stack Clash Attack

A security advisory researched various security flaws. These holes have been published as the “Stack Clash” and are going to be fixed within this release. Various patches for the kernel and the C-library glibc are included with this software update. See here .

Assigned Common Vulnerabilities and Exposures (CVE) numbers:

CVE-2017-1000364 CVE-2017-1000365 CVE-2017-1000366 CVE-2017-1000367 CVE-2017-6891

Patched Kernel 4.4.70

Several flaws regarding “The Stack Clash” were detected in the kernel. This update installs a patched kernel 4.4.70 where these flaws are fixed.

Proactive HA: Failure of LAN interfaces

The Proactive HA ensures that a node that loses a connection to the LAN is evacuated. This behavior has been adjusted. If several LAN connections exist and one LAN connection has failed, the node will not evacuated. Only when all LAN connections have failed, an evacuation occurs.

Error adding virtual harddiscs

If a virtual hard disc in the cluster was being deleted and added again without activating the config, this could lead to a failure. Within this update, this is being fixed. The config needs to be activated before.

Notes

V-Cube+: Windows Setup notes Error Message 0x80300001

If Windows Server 2008 R2 and above is to be installed on Virtio hard disks using the Virtio driver CD for Windows, Windows reports the error code 0x80300001. The error message occurs if the partitions are recognized and the user clicks the button Next. The message means, that the installation CD of Windows needs to be inserted again.

Restore virtual harddiscs

From Collax V-Bien version 6.8.15 on the internal path for virtual harddiscs changes and a restore from previous backups isn’t possible anymore without manual intervention. In case of a baremetal resore please use the previous verion or don’t hesitate to contact support.