New features in General Purpose and work-optimized Cloud Servers
This article describes some of the major features and benefits of the General Purpose and work-optimized flavors (configurations) of Cloud Servers compared to the Standard server class. Changes have been implemented with General Purpose and work-optimized servers to align with industry standards, improve server imaging, and consistently allocate adequate disk resources for each server.
The General Purpose and work-optimized flavors use faster solid state drives (SSD) and separate the system disk from the data disk, with both disks equally RAID 10 protected. With your operating system on a separate disk from your data, you can more easily create an image of the system disk because it is a fixed size and doesn’t scale up as other resources increase. For more information about data disk imaging limitations, see the “Images capture the system disk only” section later in this article or, for the full procedure, see the article Create an image of a server and restore a server from a saved image.
You can back up the data on your data disk or disks by leveraging either Rackspace Cloud Backup or Rackspace Cloud Block Storage (an option that can also be used to increase the storage capacity of your server, if needed). For a comparison of the two data disk backup options, see Best Practices for Backing Up Your Data: Cloud Block Storage versus Cloud Backup.
Faster server provisioning
The following features noticeably reduce the time to provision General Purpose and work-optimized servers:
- Dedicated 10-gigabit Ethernet (10 GigE) for networking on the host computer
- No requirement to format the entire disk (system disk only) when building your servers
Higher memory availability
The General Purpose and work-optimized flavors are available with RAM selection of up to 120 GB per server (up from 30 GB in the Standard server class).
The General Purpose and work-optimized flavors provide more processing power: up to 32 vCPUs or virtual cores, compared to a maximum 8 vCPUs on the largest-size server in the Standard server class. For information, see Work-optimized server types.
General Purpose and work-optimized servers provide more network bandwidth: 40 gigabit Ethernet (40 GigE) to each host server (the physical machine that houses your virtual server). Each host server gets 20 GigE for management and Cloud Block Storage, and 20 GigE for the server’s public network, ServiceNet, and network traffic. This bandwidth dramatically improves the interoperability of servers with complementary services such as Cloud Block Storage over the Standard server class.
Higher maximum IOPS
Separating the operating system from the data provides higher maximum input/output operations per second (IOPS). For example, the 1 GB General Purpose flavor with 20 GB of system disk and no data disk can process about 20k IOPS. The 120 GB I/O-optimized flavor with a 40 GB system disk and 1200 (or 4x300) GB data disks can perform about 80k IOPS. Compare this with the Standard server class for Cloud Servers, which can process approximately only 2k IOPS.
Work-optimized server types
**I/O-optimized **servers are assigned networking resources and use local high-speed SSD drives for storage. I/O-optimized servers work best for applications that require frequent or sustained disk access, like databases.
Compute-optimized servers have a high CPU allocation to optimize the server for applications with high CPU demands, like web servers and application servers. All disk storage for Compute-optimized servers is on Cloud Block Storage.
Memory-optimized servers have larger allocations of low-latency RAM for memory-intensive applications like caching servers, in-memory analytics, and search indexes. All disk storage for Memory-optimized servers is on Cloud Block Storage.
Standard servers can resize up or down, but General Purpose (formerly Performance 1) servers can only resize up, and work-optimized servers cannot dynamically resize at all.
Along with the rest of the OpenStack community, Rackspace is reducing support for this feature because it does not align with the industry-standard method of scaling. Rather than resize one server for vertical scaling, we instead recommend employing horizontal scaling (adding or removing the number of servers managed by a load balancer) to manage your available resources to suit your needs.
For information about reducing the size of your General Purpose server or changing the size of your work-optimized server, see Upgrading resources for General Purpose or I/O optimized Cloud Servers.
Images capture the system disk only
Images of your system capture only your operating system’s configuration (your system disk). This makes the imaging process run more quickly and reliably without placing undue strain on your server. To retain the information stored on attached data disks for General Purpose and I/O optimized flavors, you can use Cloud Block Storage or Cloud Backup to save only the files and directories that you need. For a comparison of the two options, see Best Practices for Backing Up Your Data: Cloud Block Storage versus Cloud Backup. To learn more about Cloud Block Storage snapshots, which are useful for diskless flavors like Compute and Memory, see Create and Use Cloud Block Storage Snapshots.
Preparing the data disk for use
As a result of having a separate system disk and data disk for I/O optimized servers, you must prepare your data disk by formatting and mounting it to your server before you can use it. Follow the instructions in these articles to prepare your data disk for use:** **
No 512 MB RAM servers
The 512 MB RAM cloud server is not available for General Purpose and work-optimized flavors. As a server that uses shared CPU and networking resources, maintaining the 512 MB size would place too much stress on the host because of the bursting capabilities, and potentially negatively impact other servers on the host computer.
Continue the conversation in the Rackspace Community.
©2016 Rackspace US, Inc.
Except where otherwise noted, content on this site is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License