VMware vCenter Update Manager provides a patch management framework for VMware vSphere. IT administrators can use it to patch and upgrade VMware ESX/ESXi hosts, apply patches to Windows and certain versions of Linux in virtual machines, upgrade VMware Tools and virtual hardware for virtual machines, and patch and upgrade virtual appliances. A new white paper, VMware vCenter Update Manager Performance and Best Practices, is now available.
In this paper, we discuss VMware vCenter Update Manager 4.1 host deployment, latency, resource consumption, guest OS tuning, VM operations in high latency networks, the impact of on-access virus scanning, and host operations in WAN environments. We also provide performance tips to help customers tune the system for better performance, such as:
- Separate the Update Manager database from the vCenter database when there are 300+ virtual machines or 30+ hosts.
- Separate both the Update Manager server and the Update Manager database from the vCenter Server system and the vCenter Server database when there are 1000+ virtual machines or 100+ hosts.
- Make sure the Update Manager server host has at least 2GB of RAM to cache patch files in memory.
- Allocate separate physical disks for the Update Manager patch store and the Update Manager database.
- Deploy the Update Manager server close to the ESX hosts if possible. This reduces network latency and packet drops.
- On a high-latency network, powered-on virtual machine scans are preferred because they are not sensitive to network latency.
- Host operations in a slow network will take a long time. Refer to the white paper for the maximum time estimation. Don’t interrupt ongoing operations.
Please read the white paper for more performance tips with more details. You can download the full white paper from here.
VMware recently released a
href="http://www.vmware.com/resources/techresources/10145">white paper on
performance and best practices for vCenter Server in VMware vSphere 4.1. This
paper addresses four common areas of concerns regarding vCenter Server
- Performance improvements in vSphere 4.1 compared to vSphere 4.0
- vCenter Server sizing guidelines and software requirements
- Best practices in performance monitoring, tuning and troubleshooting
- Case studies demonstrating performance improvements in vSphere 4.1.
The vCenter Server in vSphere 4.1 supports a larger
inventory in a vSphere environment when compared with that supported in vSphere
4.0, both at the vCenter Server level and at the single cluster level. A table
is provided in the white paper to show the new supported limits. In addition,
in vSphere 4.1, vCenter Server can handle a larger number of concurrent tasks
at a time when compared with vSphere 4.0.
Significant performance improvements have been made in vSphere
4.1 compared to vSphere 4.0. The following list highlights some of the most
important performance improvements:
- Improved performance at higher vCenter Server inventory limits – up to 7 times higher operational throughput and up to 75% reduced operational
- Improved performance at higher cluster inventory limits – up to 3
times higher operational throughput and up to 60% reduced operational latency
- Faster vCenter Server startup – around 5 minutes for maximum
vCenter Server inventory size
- Better vSphere Client responsiveness, quicker user interaction,
and faster user login
- Faster host operations and VM operations on standalone hosts – up
to 60% reduction in latency
- Lower resource usage by vCenter agents by up to 40%
- Reduced VM group power-on latency by up to 25%
- Faster VM recovery with HA – up to 60% reduction in total
recovery time for 1.6 times more VMs
- Better load balancing with improved DRS/DPM algorithm
For more information, please read the full paper:
Server Performance and Best Practices for vSphere 4.1.
Simplify management of VMware vSphere by automating patches and updates. VMware vCenter Update Manager makes it easy to manage tracking and patching of VMware vSphere hosts, as well as selected Windows virtual machines. A new KB article, “VMware vCenter Update Manager host tasks might fail in slow networks", is now available.
If you run VMware vCenter Update Manager over a low-bandwidth, high-latency, or lossy network, host operations might time out or fail. These operations include host patch and extension scanning, staging, and remediation, as well as host upgrade remediation tasks. Timeout might also occur under normal network conditions if the Update Manager operation takes more than 15 minutes to complete.
The task timeout or failure might occur because of the following reasons:
- The connection to the Update Manager server times out.
- The connection to vCenter Server times out.
This article introduces the methods of setting timeouts for various releases of Update Manager, and estimating timeouts for host tasks.
Simplify management of VMware vSphere by automating patches and updates. VMware vCenter Update Manager makes it easy to manage tracking and patching of VMware vSphere hosts, as well as selected Windows virtual machines. A new sizing tool, VMware vCenter Update Manager 4.1 Sizing Estimator, is now available.
The following input parameters are used to estimate database size, patch store disk space, and temporary disk space:
- Feasibility for virtual machine remediation
- Number of ESX and ESXi flavors in the deployment
- Number of hosts, virtual machines, Windows distributions, average number of locales for Windows distribution, average number of different Service Pack levels for Windows distribution
- Patch scan frequency for virtual machines
- VMware Tools upgrade scan frequency for virtual machines
- Virtual machine hardware upgrade scan frequency
- Patch scan frequency for hosts
- Upgrade scan frequency for hosts
- Whether VUM 4.1 is upgraded from 4.0 and old ESXi upgrade bundles are still used
The following are the outputs from the tool:
- VMware vCenter Update Manager 4.1 database deployment model recommendations
- VMware vCenter Update Manager 4.1 server deployment model recommendations
- Initial disk space utilization in MB for database, patch store, and temporary space
- Monthly disk space utilization growth in MB for database and patch store
- The upper and lower bounds on the estimation, assuming a 20% variance
VMware published a paper
which shows that vSphere 4.1, out of the box, easily supports a
demanding SAP ERP installation by allowing administrators to either
scale up or scale out.
Scaling Up By Adding Processors
test scale-up performance, a similarly-configured physical machine and
virtual machine were used to see how many SAP users could be supported
with one, two, four, and eight processors.
Figure 1. Number of users supported while scaling up virtual and physical (native) CPUs
the experiments, a virtual machine with 8 vCPUs was able to support
approximately 1,200 users, achieving 93%-95% of the number of users as a
similar native test.
Scaling Out By Adding Servers
4.1 can be scaled out to support larger user workloads. A series of
tests with multiple virtual machines were run to compare the performance
of different scale-out scenarios with SAP. The following graph shows
the scale-out test results and is highlighted by 16-vCPU tests. These
tests showed that approximately 2,500 SAP users could be supported
using four virtual machines each with 4 vCPUs, or 8 virtual machines
each with 2 vCPUs (for a total of 16 vCPU in both cases).
Figure 2. Scale-out test with 2- and 4-way VMs
test results for both scale-up and scale-out show that vSphere 4.1
provides good performance for SAP in a variety of environments. To read
the full paper, see SAP Performance on vSphere with IBM DB2 and SUSE Linux Enterprise.