- INSTALLATION AND DEPLOYMENT
- FILE SYSTEMS
- ACCURATE TIME SYNCHRONIZATION
- TCP ENHANCEMENTS
- HIGH AVAILABILITY
INSTALLATION AND DEPLOYMENT
Red Hat Enterprise Linux 7 provides support that simplifies the task of performing in-place upgrades.
A pre-upgrade assistant package is provided in the Red Hat Enterprise Linux 6.5 zstream, which
reports what can be upgraded in-place and what will have to be done manually.
The report describes the issues and links to knowledgebase articles available in the Red Hat
The report includes information on configuration files that will be modified and identifies existing
user-modified configuration files, recommending some to be manually checked. At that point, the
administrator can decide if the end result of an in-place upgrade is sufficient for their needs. Upon
executing the in-place upgrade, the administrator can then inspect the final results and decide to
complete the upgrade.
redhat.com TECHNOLOGY OVERVIEW What’s new in Red Hat Enterprise Linux 7 5
PARTITIONING DEFAULTS FOR ROLLBACK
The ability to revert to a known, good system configuration is crucial in a production environment.
Using LVM snapshots with ext4 and XFS (or the integrated snapshotting feature in Btrfs described
in the “Snapper” section) an administrator can capture the state of a system and preserve it for
future use. An example use case would involve an in-place upgrade that does not present a desired
outcome and an administrator who wants to restore the original configuration.
ANACONDA KICKSTART FOR ACTIVE DIRECTORY INTEGRATION
A system administrator can now create kickstart installation files that do not require administrative
credentials. The installed system can then join an Active Directory domain with a one-time password.
This new feature eliminates the need for writing and maintaining large blocks of interdependent
code in two domains.
CREATING INSTALLATION MEDIA
Red Hat Enterprise Linux 7 introduces Live Media Creator for creating customized installation
media from a kickstart file for a range of deployment use cases. Media can then be used to deploy
standardized images whether on standardized corporate desktops, standardized servers, virtual
machines, or hyperscale deployments. Live Media Creator, especially when used with templates,
provides a way to control and manage configurations across the enterprise.
SERVER PROFILE TEMPLATES
Red Hat Enterprise Linux 7 features the ability to use installation templates to create servers
for common workloads. These templates can simplify and speed creating and deploying Red Hat
Enterprise Linux servers, even for those with little or no experience with Linux.
Red Hat Enterprise Linux 7 includes three desktops to match different work styles and
preferences: GNOME 3, GNOME Classic, and KDE.
GNOME 3 provides a focused working environment that encourages productivity. A powerful
search feature lets you access all your work from one place. Side-by-side windows make it easy
to view several documents at the same time, and you can turn off notifications when you need to
concentrate on the task in hand.
Every part of GNOME 3 has been designed with simplicity and ease-of-use in mind. Activities
Overview gives an easy way to access all your basic tasks. A press of a button is all it takes to view
your open windows, launch applications, or check if you have new messages
GNOME Classic combines old and new; it keeps the familiar look and feel of GNOME 2, but adds the
powerful new features and 3-D capabilities of GNOME Shell.
GNOME 3 and GNOME Classic integrate well with online document-storage services, calendars, and
contact lists, so all your data can be accessed from the same place.
In addition to GNOME 3 and GNOME Classic, Red Hat Enterprise Linux 7 offers the latest stable
version of the popular KDE desktop.
Red Hat Enterprise Linux 7 includes systemd, a system and service manager. It combines compatibility
for most SysV and LSB init scripts as well as the following new capabilities:
• Provides aggressive parallelization capabilities.
• Uses socket and D-Bus activation for starting services.
• Offers on-demand starting of daemons.
• Keeps track of processes using Linux cgroups.
• Supports creating snapshots and restoring system state.
• Maintains mount and automount points.
• Implements fine-grained transactional, dependency-based, service control logic.
The OpenLMI project provides a common infrastructure for the remote management of Linux
systems. Capabilities include configuration, management, and monitoring of hardware, operating
systems, and system services. OpenLMI includes a set of services that can be accessed both
locally and remotely, multiple language bindings, standard APIs, and standard scripting interfaces.
It enables system administrators to manage more systems, automate management operations, and
manage both physical and virtual servers. The standardized tool interface shortens the learning
curve for new administrators and the standard APIs make it easier to build custom tools.
Storage management capabilities simplify configuring and managing storage, especially on systems
with multiple drives. A traditional issue on Linux systems is that volume labels can change when
hardware is reconfigured. OpenLMI avoids this problem by allowing you to address volumes by
volume label, UUID, or Device ID. The combination of a standardized API and persistent device
names makes it easy to keep storage consistent, even when hardware and software change.
OpenLMI enables remote network management by providing a standardized API to query and configure
the network hardware. In addition to standard network configuration, it supports configuring
network bridging and bonding and provides notification of changes in network configuration.
A system administrator can use the OpenLMI Software Provider to remotely add or remove
services, and the Service Provider to determine the state of the service (started, running, stopped,
failed), enable, start, or restart a service
• Red Hat Enterprise Linux now includes XFS as the default file system, providing scalability
enhancements and additional file system choice. Previously, XFS was available only with the
Scalable File System Add-On. XFS supports file system sizes up to 500TB.
• Ext4 supports a file system size of 50TB, up from 16TB.
• Btrfs is a relatively young file system especially useful for local, large-scale use cases. Btrfs
includes basic volume management, snapshot support, and full data and metadata checksum
integrity checksumming, and a command-line interface that makes these advanced features easier
to use than in other large-scale file systems. Btrfs is available as a Technology Preview.
• The CIFS networking file system with server message block (SMB) protocol updates will provide
better performance, security, and more features than were available with previous protocols.
• GFS2 commands now more accurately deal with RAID stripe alignment and placement of critical
elements such as journals and resource groups. This increases the scalability and performance of
GFS2 when the file system is being created and when it is used.
ISCSI AND FCOE TARGETS
Red Hat Enterprise Linux 7 includes a new software implementation of the iSCSI (RFC-3720 mode)
and Fibre Channel over Ethernet (FCoE) targets in the kernel, as opposed to the prior userspace
implementation. This new implementation strengthens the ability to replace expensive shared
storage arrays with Linux-based storage appliances built on commodity hardware.
DYNAMIC LUN DETECTION
Logical units (LUN) can now be dynamically recognized by the operating system with no manual
intervention, resulting in fewer reboots and less downtime.
Snapper is a new utility that creates, deletes, labels, and organizes snapshots of both Btrfs file
systems and LVM logical volumes. The additional information and tooling give system administrators
more control over their backup environment.
The new dynamic firewall service (firewalld) provides greater flexibility over traditional Linux iptables
enabling the unified management of IPv4, IPv6, and Ethernet Bridge rule sets. It allows administrators
to quickly respond to new threats by activating new rules without restarting, thus eliminating
service disruptions. In addition to dynamic configuration capabilities, firewalld supports a powerful
rules language that simplifies firewall configuration and includes nearly 50 pre-defined settings for
many commonly used system services.
Information stored in the system log files can now be accessed in a structured form, making automated
log analysis tools more powerful and effective. The default log file structure is not changed,
ensuring that your existing tools and processes continue to work without requiring modifications.
SELINUX OVER NFS
Labeled NFS allows customers to take advantage of the full power of SELinux and deploy more
secure environments, including secure virtual machine home directories stored on NFS servers.
Images in a Red Hat Enterprise Virtualization storage domain can have labels conveniently assigned
and issued by the Red Hat Enterprise Virtualization Manager.
Many types of attacks on a system can be prevented by providing fine-grained control of who
can access system resources. SELinux protection is now available when using NFS, simplifying the
development of secure applications. The Linux kernel has enhanced support for passing SELinux
labels between a client and server using NFS.
Two new NetworkManager user interfaces are now available in Red Hat Enterprise Linux 7 and are
aimed at administrators who prefer or may require command-line access for managing networking
connections and services:
• Nmcli uses a command-line interface to provide network administration of local, remote, and
headless systems. It can also be used for network management scripting purposes.
• Nmtui is a curses-based, menu driven text user interface designed to replace system-confignetwork-tui
(in Red Hat Enterprise Linux 6) and simplifies configuring many common network
settings, eliminating the need for administrators to manually edit network interface
ACCURATE TIME SYNCHRONIZATION
Chrony is a different implementation of the network time protocol (NTP) than the network time
protocol daemon (ntpd) that is able to synchronize the system clock faster and with better accuracy
than ntpd. Note that ntpd is still included for customers who need to run an NTP service.
Benefits of Chrony include:
• Faster synchronization requiring only minutes instead of hours to minimize the time and
frequency error, which is useful on desktops or systems not running 24 hours a day.
• Better response to rapid changes in the clock frequency, which is useful for virtual machines
that have unstable clocks or for power-saving technologies that don’t keep the clock
• After the initial synchronization, it never steps the clock so as not to affect applications needing
system time to be monotonic.
• Better stability when dealing with temporary asymmetric delays, for example when the link is
saturated by a large download.
• Periodic polling of servers is not required, so systems with intermittent network connections can
still quickly synchronize clocks.
PRECISION TIME PROTOCOL
Red Hat Enterprise Linux 7 supports the IEEE 1588, Precision Time Protocol version 2 (PTPv2)
standard, a method for accurately synchronizing distributed clocks over an Ethernet network. While
similar to Network Time Protocol (NTP), one of the main advantages of PTP is the hardware support
found in various network interface cards (NICs) and network switches. When used in conjunction
with the appropriate hardware, enabled devices, and network drivers, it is capable of achieving clock
accuracy in the sub-microsecond range, which is far more accurate than is typically obtainable with
NTP. This feature is particularly important for applications in the financial services and tradingrelated
industries, where application latency is measured in microseconds.
TEAM DRIVER LINK AGGREGATION
The Team Driver project is new for Red Hat Enterprise 7 and provides a mechanism to bond multiple
network devices (ports) into a single logical interface at the data link layer (OSI Layer 2). This mechanism
is typically used to increase the maximum bandwidth and provide redundancy for links.
Team Driver moves most of the work and logic into a user-space daemon by implementing only the
necessary data fast-path parts in the kernel. This approach provides several advantages over traditional
bonding such as more stability, easier debugging, and making it simpler to extend while providing
equal or better performance.
Various improvements to transmission control protocol (TCP) aim to reduce application latency for
connection-oriented services such as web servers built on Red Hat Enterprise Linux.
• Fast Open is an experimental TCP extension designed to reduce overhead when establishing a
TCP connection by eliminating one round time trip (RTT) from certain kinds of TCP conversations.
Fast Open could result in speed increases of between 4% and 41% in page-load times for busy
• Tail loss probe (TLP), an experimental algorithm, improves the efficiency of how the TCP
networking stack deals with lost packets at the end of a TCP transaction. For short transactions,
TLP could reduce re-transmission timeouts by 15% and shorten HTTP response times by
an average of 6%.
• Early Retransmit (RFC 5827) allows the transport to use fast retransmits to recover segment
losses that would otherwise require a lengthy re-transmission timeout. Connections can recover
from lost packets faster, decreasing overall latency.
• Proportional Rate Reduction (PRR) is an experimental algorithm designed to return to the
maximum transfer rate quickly. It can potentially reduce HTTP response times by 3-10%.
40G ETHERNET LINK SPEED
Red Hat Enterprise Linux 7 supports 40G Ethernet link speeds on the appropriate hardware, which
enables faster network communication for systems and applications.
Low-latency sockets are a software implementation that reduces networking latency and jitter
within the kernel. This implementation makes it easy for applications to poll for new packets directly
in the network driver which speeds up packets moving into the network stack. Applications that are
sensitive to unpredictable latency benefit from the top-down, busy-wait polling method that replaces
interrupts for incoming packets.
ENHANCED CLUSTER RESOURCE MANAGER
Cluster resource management has been enhanced through several additions:
• Simplified administrative procedures reduce the amount of effort it takes to monitor and manage
• Finer-grained monitoring of every component in the cluster stack provides more awareness and
control over applications running in high-availability environments. Resources can have multiple
states associated with them and can be managed on a schedule basis or manually. An important
new feature is the ability to create user-defined actions.
• Resource cloning allows a single command to be replicated across multiple nodes in the cluster.
For example, by using cloned resources, issuing a single command can cause a GFS2 filesystem to
be mounted on all nodes within the cluster.
• The new cluster resource manager has both a graphical and a command-line interface. The
new resource manager provides a single environment for managing clusters running Red Hat
Enterprise Linux 6 and 7.
PACEMAKER POLICY ENGINE
The pacemaker remote capabilities now apply to virtual machines within a cluster. Now with
Red Hat Enterprise Linux 7, users can run pacemaker from within a virtual machine and to control
resources and applications running in other virtual machines in the cluster.
For more detailed technical information and instructions for using the features and capabilities
described in this document, see the Red Hat Enterprise Linux 7 documentation.3
The capabilities previously available in the High Performance Networking Add-On, the Load
Balancing Add-On, and the Scalable File Sy