 |
» |
|
|
|
Use the
following list of items to prepare for the system configuration phase: Know where you stored the XC.lic license key file because you will copy or move it
to the head node. Be prepared to supply the
information listed in Table 3-1, which is required by the cluster_prep command. Table 3-1 Information Required by the cluster_prep Command | Item | Description
and User Action |
|---|
Node name
prefix | During the system discovery process,
each node is automatically assigned an internal name. This name is
based on a prefix defined by you. The default node prefix is the letter n. All node names consist of the prefix and a number based
on the node's topographical location in the system. You have
the option to change this default, perhaps to differentiate the nodes
from another system or to customize node names according to your corporate
policies. Choose a node prefix that adheres to these guidelines: The node name consists of a maximum of six alphanumeric
characters: The lowercase letters a through z The uppercase letters A through Z The numbers 0 (zero) through 9
The first and last characters in the prefix must be
alphabetic. Hyphens (-) and periods (.) are not permitted.
By default, nh is assigned
as the head node alias, regardless of the node prefix. | Maximum number of nodes in the system | Enter the exact number of nodes
in the hardware configuration or enter the maximum number of nodes
allowed by the interconnect. For example, if the interconnect
allows a maximum of 128 nodes, but the hardware configuration contains
only 64 nodes, it is acceptable to enter the number 128 when prompted. This number affects only the internal node naming
of the head node. If you run the cluster_prep command with the --enclosurebased option, you
are not prompted to supply the maximum number of nodes. | Database administrator's password | Define and set a password
to protect the configuration and management database. Make this password
different from the root password. You are prompted for
this password any time you log into the configuration and management
database. See “Password Guidelines” if you need recommendations on choosing
strong passwords. | External Ethernet connection
on the head node | You are prompted to configure an external Ethernet device on the
head node. Doing so connects the node directly to the public network.
The most likely Ethernet device to use as the external network connection
is presented to you by default. You can use this value or provide
a different value. If you need more information to verify
that you are selecting the correct Ethernet device on the head node,
open a new terminal window and use the ethtool ethernet-device-name command to provide more
information about the Ethernet devices on the node. | IP address for the external
network connection | Provide the IP address for the head
node's Ethernet connection to the external network. This is the
address assigned by your site's network administrator. The IP
address must be in your site's Domain Name Service (DNS) configuration. Enter the address in dotted decimal format n.n.n.n, where n is the decimal value in the
range 0 to 255 and is the decimal value of an 8-bit segment of the
address. | Network mask | Provide the network mask (netmask) for the external network
on the head node. The network mask is used for each of
the addresses in the address range and is specified in dotted-decimal
format (for example, 255.255.252.0). | IPv6 address | Provide
the IPv6 address of the head node's Ethernet connection to the
external network, if applicable. Specifying this address is optional
and is intended for sites that use IPv6 addresses for the rest of
the network. Enter the IPv6 address in the format x:x:x:x:x:x:x:x/n where x consists of one to four hexadecimal digits that represent a 16-bit
segment of the address, and n represents
the prefix length (for example, 2001:0db8:0000:0000:0000::1428:57ab/32). The prefix length represents the netmask parameter equivalent
and can be in the range of 1 to 128. You can omit leading
zeros within each hexadecimal field and use two colons (::) to indicate
multiple 16-bit groups of zeros. If a value is already
present in this field, enter left and right brackets [
] to erase it | Gateway IP address | Provide the network gateway IP address for your site. The network gateway is the next-hop router the system uses
to direct any traffic to a host that is not directly on your site's
local network. This address is specified as a dotted-decimal address
(for example, 192.0.2.51) rather than a name. | MTU (jumbo frames) | Enter 9000 to enable jumbo frames
on external Ethernet interfaces or leave this field blank to use the
system default (1500). Jumbo frames sets the maximum
transmission unit (MTU) to 9000 bytes for improved efficiency and
performance with bulk data transfer. The MTU must match on the other
end of the connection. If a value is already present in
this field, enter left and right brackets [ ] to erase it | IP address for the DNS name
server | Provide the IP address for
your site's DNS server. This address is specified in dotted-decimal
format (for example, 192.0.2.50). | Search path | Provide a list of name server
domains to search to resolve host names when the following prompt
appears: Provide one or more DNS domains to use for search paths
or press the Enter key to accept the default response.
Enter one domain name on a line, and after the last domain
name, enter a period (.) on a line by itself and press the
Enter key.
Search path: |
The domain name is typically
in the form my-domain.com (assuming that the head node is intended to be HeadNodeName.my-domain.com). Enter the search entry in the following format: DomainName.something[.something....] The search entry can have from one to six DomainName variables. The DomainName variable is the name of a domain to be included in the search list.
The first DomainName variable is interpreted
as the default domain name. |
Be prepared to supply the information listed in Table 3-2, which is required
by the discover command. You can omit some questions
in the discovery process if you provide the related keyword and value
on the discover command line. Table 3-2 Information Required by the discover Command | Item | Description
and User Action |
|---|
Total number of nodes in this cluster | Enter the total number of nodes
in the system configuration that are to be discovered at this time.
Make sure the number you enter includes the head node and all compute
nodes. You are not prompted for this information if you
are discovering a multi-region, large-scale system. If
you specify the --enclosurebased option on the discover command line, you are not prompted for this information. You can include the nodes= keyword and value
on the discover command line to omit this question
during the discovery process. | Number of
nodes that are workstations | Enter the number of workstation nodes in the hardware
configuration. Workstations do not have console ports. There is no default response, and a response is required. Enter 0 (zero) if the system does not contain workstations. You are not be prompted for this information if you are discovering
a multi-region, large-scale system. If you specify the --enclosurebased option on the discover command line, you are not prompted for this information. You can include the ws= keyword on the discover command line to omit this question during the
discovery process. | MAC address of the root administration switch | Look on
the back panel of the ProCurve Root Administration Switch near the
9-pin console port and write down the MAC address. Use the switch
management serial console with a terminal connected if you cannot
read the MAC address. See “ProCurve Switches Do Not Obtain Their IP Addresses”, which describes how you can identify
which switch is the Root Administration Switch. You can
include the switchmac= keyword on the discover command line to omit this question during the
discovery process. | MAC address of the Onboard Administrator | When only one HP BladeSystem c3000
or c7000 enclosure is present in a hardware configuration, you must
include the --single option on the discover command line. The discover command prompts you
to supply the MAC address of the Onboard Administrator because hardware
configurations with one enclosure do not have any external ProCurve
switches. The Onboard Administrator is the small LED
panel near the bottom of the enclosure and you can obtain its MAC
address directly from it. | Administrator password for the ProCurve switch | Supply the common administrator
password for the ProCurve switches that you defined when you prepared
the hardware. Press Enter if you did
not set this password when you prepared the hardware. You can include the switchpw="-" keyword and value
on the discover command line to omit this question
during the discovery process. | User name and password for the console port management devices | Supply the common user name
and password that you set for the console port management devices
(that is, the MP, iLO and LO-100i devices) when you prepared the hardware. The HP XC Hardware Preparation Guide instructs you to configure all console management
ports on the system with the same user name and password; otherwise,
the discovery process will not succeed. If the hardware
configuration contains HP server blades and enclosures, the password
for the Onboard Administrator must match the password for the console
management device. | | | |  | NOTE: The HP XC Hardware Preparation Guide instructs you to set the password and user
name as a hardware preparation task on all server models except the
HP ProLiant DL140 G2 and DL145 G2 servers. The procedure to change
the default user name and password for those hardware models is described
in the appropriate place in this document (after the discover command discovers the IP addresses of the console ports). | | | | |
You can include the cpuser= and consolepw= keywords and values on the discover command line to omit these questions during the discovery process. | Number of nodes plugged into the Root Administration Switch | If you
are required to use the --oldmp option on the discover command line for HP XC systems with an
HP Integrity head node, you are prompted to supply the number of nodes
that are plugged into the Root Administration Switch. If you specify the --enclosurebased option on
the discover command line, you are not prompted
for this information. | Number of nodes plugged into application cabinets | If you are required to
use the --oldmp option on the discover command line for HP XC systems with an HP Integrity head
node, you are prompted to supply the number of nodes plugged into
application cabinets. If you specify the --enclosurebased option on the discover command line, you are
not prompted for this information. |
Be prepared to supply the information listed in Table 3-3, which is required
by the cluster_config utility. Table 3-3 Information Required by the cluster_config Utility | Item | Description
and User Action |
|---|
Availability sets | You are prompted to configure availability
sets for improved availability of services if you have installed and
configured an availability tool (such as HP Serviceguard) as described
in “Task 9: Plan a Service Availability Strategy”. If you are not prompted to configure availability
sets and you expected to configure availability sets, return to “Task 9: Plan a Service Availability Strategy” and
ensure you followed all instructions. | Node role assignments and Ethernet connections | A command-line menu appears
which enables you to change the default node role assignments and
to add or remove an external Ethernet connection to any node. See Appendix F for node role definitions and for more information about using
the menu. | ssh key
regeneration | If root ssh keys
already exist, you are prompted to regenerate them. If you regenerate
the ssh keys, you cannot run the ssh or pdsh commands to other client nodes until
you reimage them. For this reason, HP recommends that
you regenerate ssh keys so you can determine if
all nodes have successfully booted the new image. | Number of NFS daemons | You are prompted to supply the number of NFS daemons to
be run on the head node and on any other NFS server within the system
to support the number of NFS clients in the system. A default is provided
based on the number of nodes in the hardware configuration. | Quorum server or lock LUN
configuration for Serviceguard clusters | If you have configured improved availability with HP Serviceguard,
you are prompted to supply the name of a quorum server or the full
path to the device name of the lock LUN. | NTP server configuration | One node in an HP XC system acts as an NTP server
for all other nodes. By default, this is the head node. All other
nodes are NTP clients of this server, which is known as the internal
NTP server. All nodes in the system are synchronized to the head node. You have the option to specify up to four external time servers
or to use the default NTP server on the head node. | Improved availability of
the database server | If you assigned the avail_node_management role to a node in an availability
set with the head node, you are prompted to specify how you want to
handle improved availability for the dbserver service.
You can choose between no improved availability or enabling improved
availability through an availability tool. | QsNetII network type | For systems with a QsNetII interconnect, you are asked to supply the type of network.
The network type reflects the maximum number of ports the fabric topology
can support, and it is used to create the qsnet diagnostics database. See Appendix H for information about how to determine
the QsNetII network type. | Number of QsNetII node-level and top-level switches | For systems with a QsNetII interconnect,
you are asked to supply the number of node-level and top-level switches
in the configuration. Enter the number of node level switches in your configuration [1-32]:
Enter the number of top level switches in your configuration [0-32]: |
| LVS configuration | If you
modified the default role assignments and assigned a login role to one or more nodes, you are prompted to enter an LVS alias.
This is the host name that users will use to log in to the HP XC system. You have the option to decide whether you want the LVS
director to act as a real server, that is, a node that accepts login
sessions: Do you want the LVS director to act as a real server? (y/n) |
If you answer yes, the LVS
director is configured to act as a login session server in addition
to arbitrating and dispersing the login session connections. If you answer no, the LVS director does
not participate as a login session server; its only function is to
arbitrate and disperse login sessions to other nodes. This gives you
the flexibility to place the login role on the
head node yet keep the head node load to a minimum because login sessions
are not being spawned. The default is to configure the LVS director as a real server if LVS is not configured
with improved availability. If LVS is configured with improved availability,
the default is to arbitrate and disperse only. If you
assign the login role to nodes in an availability
set, you are prompted to specify how you want to handle improved availability
for the lvs director service. You can choose between
no improved availability or enabling improved availability through
an availability tool. | Nagios configuration | You are prompted to enable web access
to the Nagios monitoring application. HP recommends that you enable
web access because it is the only mechanism with which you can view
the data collected by Nagios. You will have to supply
a password for the Nagios administration user. This password does
not have to match any other password you previously provided. If you assigned the management_server role
to the nodes in an availability set, you are prompted to specify how
you want to handle improved availability for the Nagios master service.
You can choose between no improved availability or enabling improved
availability through an availability tool. | NAT configuration | If you assigned the external role to
the nodes in an availability set, you are prompted to specify how
you want to handle improved availability for the nat service. You can choose between no improved availability or enabling
improved availability through an availability tool. You
are also prompted to enter an additional external IP address to use
as an external alias. | snmptrapd service configuration | You are prompted to configure
the snmptrapd service if the hardware configuration
includes HP server blade enclosures or if you have defined an mcs.ini file to configure MCS devices. Configuring
the snmptrapd service enables Nagios to generate
alerts for critical SNMP traps sent from either an enclosure or an
MCS device. | SVA and remote
graphics software configuration | If you installed SVA or optional remote graphics
software,[1] you are prompted to supply the following information: Whether the visualization nodes (the workstations)
have a KVM attached The host names for display nodes, that is, the nodes
that have monitors connected to them Remote graphics software configuration information: The host names of the remote graphics nodes. You must
configure external connections on these nodes. Whether you want to use RGS sender on the head node.
Doing so enables the head node to display remote nodes (for HP RGS
only). Whether you want to use RGS Receiver Version 3.0 to
connect the system (for HP RGS only). Answering yes means that multiple ports are opened on the firewall.
Answering no opens a single port on the firewall. The current version of HP RGS is Version 4.n. Version 4.n receivers require a single
port opened in the firewall. If you choose to continue using previously
installed HP RGS Version 3.0 receivers, multiple ports are opened
in the firewall. You can use only Version 4.n receivers or a mix of Version 4.n and
Version 3.0 receivers. For more information about HP RGS and its use of firewalls,
see the SVA System Administration Guide .
| Apache self-signed certificate | You are prompted to configure a
self-signed certificate for the Apache server so that all HP XC systems
do not have the same credentials. If you do not configure a self-signed
certificate, the certificate that is shipped with the Linux base operating
system is used instead, which prevents some web browsers from being
able to connect to more than one HP XC system during the same
session. | NIS configuration | If you modified the default
role assignments and assigned a nis_server role
to configure one or more nodes as a NIS slave server, you are prompted
to enter the name of the NIS master server or its IP address as well
as the NIS domain name. | SLURM configuration | You are prompted for the following information about the
SLURM configuration: Whether you want to configure SLURM. You are not required
to configure SLURM, however, SLURM is required by LSF-HPC with SLURM and SVA. A SLURM user name. A default value is provided, but
you can specify your own SLURM user name. The node on which to configure the primary SLURM controller The node on which to configure the backup SLURM controller Whether to configure user-restricted access on the
compute nodes
| LSF configuration | You are prompted for the following
LSF configuration information (default responses are provided for
each item): Whether you want to configure LSF. You are not required
to use LSF as the job management software. What type of LSF to install: LSF-HPC with SLURM or standard LSF A user name for the primary LSF administrator A password for the LSF administrator if it is a new
local account An LSF cluster name (the default is hptclsf)
|
|
Printable version
