You can create a multi-host or single-host deployment of Control Center,
on real or virtual hosts.
Zenoss recommends that all production deployments include one Control
Center master host and two or more delegate hosts. In this
configuration, the master host runs in its own, separate resource pool,
and runs only Control Center internal services. Delegate hosts run
application services in other resource pools.
For multi-host deployments, the master host typically requires 4
real or virtual CPU cores and 16GB RAM. Very large multi-host
deployments may require 8 CPU cores, but should not require additional RAM.
For single-host deployments, the master host requires a minimum of 8
real or virtual CPU cores and 64GB RAM. Actual CPU and RAM
requirements depend on application load.
Warning
An under-resourced master host cannot function properly. Deploy Control
Center and Resource Manager on a master host that meets or exceeds the
minimum requirements.
Master host storage areas
A Control Center master host requires the following storage areas:
Root
Size: Depends on configuration
Type: XFS file system (typically)
Mount point: /
Resource: One or more block devices or partitions, or one or more
LVM physical volumes.
Preparation: None. The file system is created when the operating
system is installed.
Docker temporary
Size: 10GB
Type: XFS file system (typically)
Mount point: /tmp
Resource: One or more block devices or partitions, or one or more
LVM physical volumes.
Preparation: Depends on configuration. Typically, additional space
is allocated for the root filesystem, but a separate real or logical
partition may be used. The Control Center installation procedures
include instructions to link the Docker temporary directory to /tmp.
Swap
Size: 16GB
Type: swap
Mount point: None
Resource: One or more block devices or partitions, one or more LVM
physical volumes, or a special file on the root filesystem.
Preparation: Depends on configuration.
Recommended swap size
For RHEL 9, Red Hat recommends that RAM and swap sizes match so that the
full contents of RAM can be paged to disk. The kernel may prevent some services
from starting if RAM and swap are not equal.
For a stable installation of Resource Manager, swapping should be avoided
and 16GB of swap should be adequate. To override the requirement for matching
swap to RAM, the vm.overcommit_memory=1
parameter will be added to /etc/sysctl.conf as an installation step.
Docker data
Size: 50GB
Type: LVM thin pool
Mount point: None
Resource: One or more block devices or partitions, or one or more
LVM physical volumes.
Preparation: None. The installation procedures include steps for
creating the thin pool.
Control Center internal services data
Size: 50GB
Type: XFS file system
Mount point: /opt/serviced/var/isvcs
Resource: One or more block devices or partitions, or one or more
LVM physical volumes.
Preparation: None. The installation procedures include steps for
formatting the resource.
Control Center audit logging
Size: 10GB (default)
Type: XFS file system
Mount point: /var/log/serviced
Resource: One or more block devices or partitions, one or more LVM
physical volumes, or a remote file server that is compatible with Linux.
Preparation: Depends on configuration. Typically, additional space
is allocated for the root filesystem. For more information, see
Control Center audit logging.
Application data
Size: 200GB suggested. The size should be twice as large as the base
device, which determines the size of tenant virtual devices.
Warning
The ratio of application data stored in the thin pool to the total
amount of storage in the pool should never exceed 50%. When the amount
of available space is less than half the amount of total storage,
the snapshot feature is prone to failure.
Type: LVM thin pool
Mount point: None
Resource: One or more block devices or partitions, or one or more
LVM physical volumes.
Preparation: None. The installation procedures include steps for
creating the thin pool.
Application data backups
Size: 150GB suggested. The size should be at least 150% of the size
of the base device, or a minimum of 150GB, whichever is greater.
Note
For large environments, this size should be much greater.
Individual backup files can be 100GB each, or more.
Type: XFS file system, or a Linux-compatible file server
Mount point: /opt/serviced/var/backups
Resource: One or more block devices or partitions, one or more LVM
physical volumes, or a remote file server that is compatible with
Linux.
Note
When the storage for backups and Control Center
internal services data use the same physical device, resource
contention can cause failures during backup and restore operations.
For optimal results, use separate physical devices.
Preparation: None. The installation procedures include steps for
formatting or mounting the resource.
The following example shows the disk configuration of a Control Center
master host.