Evolution of Storage and Volume Management in SUSE Linux Enterprise
Data is our customers' gold and Novell's enterprise Linux strategy recognizes storage as a key component. The storage management capabilities in SUSE® Linux Enterprise are evolving to take advantage of the latest technologies and respond to customer needs. This evolution reflects the following elements:
1. Evolution requires investment and room to grow.
SUSE Linux Enterprise has always been a leader in the area of storage and storage management:
- First enterprise Linux distribution to ship with logical volume management (LVM 1 in SUSE Linux Enterprise Server 7)
- First enterprise Linux distribution to ship with a journaling file system (ReiserFS in SUSE Linux Enterprise Server 7)
- The only enterprise Linux distribution shipping with three enterprise-class journaling file systems (ext3, ReiserFS, XFS—all since SUSE Linux Enterprise Server 8)—addressing customer needs for different workloads and purposes
- First enterprise Linux distribution to ship with two "flavors" of volume management (LVM 1/2 and EVMS 1/2 since SUSE Linux Enterprise Server 8)
- First enterprise Linux distribution to ship with a community-supported and upstream-integrated cluster file system (OCFS2 in SUSE Linux Enterprise Server 9 SP 2)
We are proud of this history and confident that we will maintain our leadership working hand in hand with the community and our partners and customers.
2. Evolution presents choice.
For SUSE Linux Enterprise Server 10, Novell built our High Availability Storage Infrastructure (HASI) on EVMS2, which has been donated to the open source community by IBM. At the time this architecture was designed, no other community-based approach offered the features and stability of EVMS2 with respect to clustering and volume management.
For Open Enterprise Server 2 (which is based on SUSE Linux Enterprise Server 10), Novell has integrated our classical storage infrastructure, inherited from NetWare, with the modern device-mapper (DM) infrastructure of the Linux kernel and EVMS.
EVMS2 consists of a middle layer between "userland" tools and applications, and the device-mapper infrastructure in the kernel.
Additionally, SUSE Linux Enterprise Server 10 and Open Enterprise Server 2 support LVM2 for non-clustered environments. LVM2 shares the device-mapper (DM) infrastructure of the Linux kernel with EVMS2.
In the last three years, the Linux community has made tremendous progress with developing and enhancing the cluster components based on LVM2, called cLVM. Thus there is now a choice of multiple enterprise-class volume management technologies that can be incorporated into Linux distributions.
3. Evolution requires decision.
Currently the open source community is working on the tight integration of several clustering components developed independently, as it strives to build a single Linux cluster stack, consisting of, for example:
- Linux HA + Pacemaker (scalable High Availability cluster resource manager)
- OpenAIS cluster communication layer
- cLVM
- Cluster file system, such as OCFS2 or GFS2
In order to advance industry efforts to develop a single Linux cluster stack, Novell has decided not to continue development on EVMS2 in the future, and we will not include EVMS2 in the next major release (version 11) of SUSE Linux Enterprise Server.
Our enterprise distribution concept guarantees that we will not drop a feature within the lifetime of a product generation. Thus we will continue to fully support EVMS2 for the full lifetime of SUSE Linux Enterprise Server 10 and Open Enterprise Server 2 respectively, and customers are safe to begin and continue deployments using the existing High Availability Storage Infrastructure (based on EVMS2) today.
4. And my gold, the data?
As noted above, both EVMS2 and LVM2 (and thus cLVM) are based on the Linux kernel's device-mapper layer, and EVMS2 and LVM2 reside on top of this layer. The metadata format on the storage media is identical. Even today you can access your EVMS2-created data with LVM2 (in non-clustered environments) and vice versa.
Therefore customers will be able to migrate existing SUSE Linux Enterprise Server 10 systems with EVMS2 or LVM2 to SUSE Linux Enterprise Server 11 seamlessly and without requiring format conversion.
Novell's High Availability team will also provide a set of tools with SUSE Linux Enterprise Server 11 to enable automatic offline migration for existing installations and configurations.
5. Decision enables new advancements.
An industry standard High Availability cluster stack on Linux, an effort in which Novell is a key contributor, is an ambitious goal by itself. However Novell is already working to bring this infrastructure to the next level of integration and data management:
- SUSE Linux Enterprise Server leverages Novell's work on various standards, such as CIM and (for storage specifically) SMI-S.
- Our YaST team is working to enhance the existing "partitioner" module into a storage administration tool, which will:
- Integrate the various layers found in the kernel
- Provide a consistent view of storage resources
- Scale better in data center environments
- Ease the administrator's daily work
- Integration of storage and networks, server virtualization, and storage virtualization are key components of today's data center and will become even more important in the future. These elements are primarily supported today in SUSE Linux Enterprise Server 10 and will be fully supported in SUSE Linux Enterprise Server 11.
- SUSE Linux Enterprise Server is the foundation of tomorrow's policy-driven data center, and will help to speed up the flow of information and protect your data for future advancements.