now you've probably on XenServer 7 and its cool new features, one of However, this version is is. more than just its feature list - we have also in the fundamentals invested XenServer 7 to make a platform on which we can provide even more innovation, features and improvements over time
Let some one views. the platform and architectural changes we have But before we start with the ceiling of XenServer 7.
let us taken a look at XenServer architecture for a little guidance. The Xen hypervisor project is a "Type 1" hypervisor that runs directly on the hardware. On this we a privileged Linux virtual machine, called "Domain 0" and is based on CentOS running, the tool stack management and API runs, includes drivers for networking and storage I / O, etc., and quite a number of other things.
So what has changed?
Domain 0 Linux version
A significant change in XenServer 7 is the upgrade of the domain 0 around CentOS 5.10 CentOS 07/02/1511 . At the domain 0 Linux environment is NOT provide customers visible features, however, because it most of the server components of XenServer hosts, it is extremely important for the operation of these components, and our ability to add new features that it depend on you. By moving up to date to, but stable and enterprise-class Linux platform we are well XenServer 7 placed to assist in the future, benefit from bug fixes and improvements from the CentOS project view and their upstream provider and Open Source -Communities.
The updated Linux platform also offers opportunities for future features and enhancements that would have been difficult on older CentOS 5 platform or impossible. Coupled with XenServer existing Linux 3.10 kernel, the new platform given us more ways to use Linux filesystems as ext4 as the basis for future XenServer storage functions.
You may wonder why the move to CentOS 7 didn 't automatically mean that XenServer 7 acquired different functions and mechanisms that have happened 7 CentOS as SELinux or XFS file systems. This is because the domain 0 Linux platform a component of the XenServer system (rather than XenServer is an application that runs on Linux), and therefore is closely integrated into the overall system - if we have a upgrade the system components, our first priority is to ensure a like-for-like upgrade that preserves the integrity, functionality, quality, performance and safety of the system. If the new component comes with new mechanisms that did not exist in the older version, we carefully consider whether and how we can these new mechanisms in the XenServer system (eg change from ext3 to XFS would require careful examination for XenServer to integrate, upgrade and use cases rollback and is therefore not a transparent change.) - these additional integrations can be used as the underlying component upgrade in later versions
with the release of XenServer 7.0, we mostly use CentOS 7 as again for-like replacement for CentOS 5 - with new mechanisms in the newer version to be candidates for the integration and use in the following XenServer releases. As with many of the fundamental changes we will talk about in this blog, I would expect to see more of the value of CentOS 7 surface in the customer visible features and improvements over the next few XenServer releases.
Partitioning Layout
Earlier versions of XenServer have a partitioning layout of a primary partition used 4GB consisting (this is the domain 0 Linux environment and XenServer stack hosts which it sits), partition a 4GB backup roll-back after an upgrade, and to allow used as a storage repository for VM disk images the rest of the plate. In preinstalled OEM systems an additional OEM partition can be there.
Although 4GB is more than enough for the XenServer software the same partition for third-party add-on "Extra Packs" is also used, like all host protocols and temporary staging for hotfix files and so on. Even with log rotation and compression are some cases of the file system were to fill because of the volume of log files. Before XenServer 7 there was a mechanism to some mitigations also the total volume of the log files (in addition to the per-file rotation) and for placing the log files on a separate volume in space from the local storage repository loan cap. In XenServer 7 we decided that it was time to move larger partitions and make the log partition a default configuration. This means that we now have:
- 18GB XenServer host control domain (Dom0) partition
- 18GB Backup Partition
- 4GB logged partition
- 1GB swap partition (just in case - we try to avoid using swap)
- 5GB UEFI boot partition
the traditional layout used in cases , 4GB, where the primary hard drive is smaller than the required 46GB can. In cases where older XenServer hosts XenServer update 7, the host will be repartitioned when the local storage repository is empty (as is often the case when shared memory is used). See the installation manual for more details.
Xen hypervisor project
XenServer 7 saw us Xen 4.4 Upgrade (in XenServer 6.5) 4.6 to XEN also contains the new contents of Xen 4.5. This brought a number of useful mechanisms fixes and enhancements, including virtual memory event subsystem and support introspection with Intel EPT and AMD RVI , the part of XenServer of Direct Investigate APIs form lets you agentless antimalware VM introspection. The upgrade provided a new framework for the management, the CPU functions are exposed to VMs - a prerequisite for the VMs allow the benefits of CPUs take instructions ahead as AVX2 while allowing downward leveling of feature sets VMs allow, be moved between differential CPU generations the basics and options.
with the domain 0 CentOS 7 Upgrade There are a number of new mechanisms of the updated Xen hypervisor XenServer that are not used today brings for future features, but offer. Some areas of particular interest are extensions Xen Security Modules (XSM) - mandatory access control framework in the style of SELinux, albeit on the hypervisor instead OS layer; Support for vTPM 2.0 , possibly a chain of trust from hardware TXT / TPM boats permitted right by virtual machines on trusted boot; and PVH virtualization mode - a step towards unifying Xen two modes, PV and HVM in order to reduce the complexity and the advantages of both modes at the same time receive
Xen 4.5 and 4.6 also. brought a number of performance enhancements - to http://xenserver.org/ cause an eye out for a series of blog posts from XenServer Performance Jonathan Davies, who in a number of performance and scalability enhancements in XenServer 7.
digActive Directory integration
Because Active Directory support was introduced in XenServer 5.5, we have the Also used tools with AD server to connect to. In XenServer 7 we update these tools to a newer PowerBroker Identity Services (PBIS) packages that provides better support for complex corporate structures and better credential caching along with a number of other compatibility, performance, scalability and stability improvements , used
using the IOMMU
XenServer physical I / O device driver in the domain 0 kernel and must often allocate memory in DMA transactions to hardware devices such as RAID controller or network interface. adds a layer of indirection in memory allocation (compared to a bare-metal operating system) and thus the domain 0 view of the accumulator to have a hypervisor, is different from the view of the memory of the physical device. For individual memory pages, this is easy with a simple translation but DMAs from more than one side adjacent to the size of the physical storage area and therefore can not be a single translation of the start address, is not possible, treated. The simplest solution for this is to pass all DMA transactions by memory from a memory pool with a picture with the host physical memory 1-to-1. This pool is known as software I / O Translation Lookaside Buffer (SWIOTLB) or bounce buffer . However, this adds a memory copy to each DMA, increasing latency and lowering power; and it risks DMA failures when the SWIOTLB is exhausted or fragmented
XenServer 7 avoids these disadvantages by using (if possible) the hardware of IOMMU -. can access a device, the physical device memory using virtual addresses rather than physical addresses and therefore, the same memory allocation is used as the domain 0th This means that, even if the physical memory for the DMA is not connected, the device on / off her DMA as if it, exactly the same address as it was given by the Domain-0 device driver. This avoids the SWIOTLB memory copy and the pool exhaustion problem leads to better performance and reliability.
virtualized I / O
We have just been changes, both our storage and network I / O data paths to increase performance. See Jonathan Davies' performance blog series for more on this topic.
API extensibility
The XenServer tool stack and API consists mainly of the XAPI daemon. XAPI is the API endpoint and almost all API implementations are handled directly by him. 7 XenServer provides an internal API extensibility framework, the new API makes calls added and their implementation are handled by separate executables within the XenServer Domain 0 environment. The aim is a modular, extensible tool stack to enable, where new functionality can be added with the reconstruction XAPI itself.
And more ...
- The Open Virtual Switch (OVS) has been updated to version 2.3.2 - stable and proven release
- to create to manage the use of Linux cgroups resources 0 within domain better response under load
- many internal refactoring of XAPI tool stack to make it easier to add new features
Finally, there are plenty more cool stuff to the ceiling of XenServer 7, as you might think. This is one of the largest and best XenServer releases we've ever done, and it will form the basis of a stream of new features, improvements and enhancements over many publications. Stay tuned ...
0 Komentar