xCAT/xcat-core
2
0
Fork 0
mirror of https://github.com/xcat2/xcat-core.git synced 2025-05-29 17:23:08 +00:00
Code Issues Releases Wiki Activity

Merge pull request #264 from neo954/task60792doc

Browse Source 
Documentation updating, Hardware management for OpenPower systems
This commit is contained in:
Victor Hu 2015-10-15 09:43:16 -04:00
commit 2295560c50
2 changed files with 114 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/source/guides/admin-guides/manage_clusters/ppc64le/index.rst
View File

@ -9,7 +9,7 @@ For a new user, you are recommended to read this chapter in order since later se
:maxdepth: 2
discovery/index.rst
management/index.rst
management.rst
diskful/index.rst
diskless/index.rst
updatenode.rst

113
docs/source/guides/admin-guides/manage_clusters/ppc64le/management.rst Normal file
View File

@ -0,0 +1,113 @@
Hardware Management
===================
Basic Operation
---------------
The Beacon Light
````````````````
Most of modern enterprise level server machines have LEDs installed on their front panel and/or rear panel, which are called beacon lights. When this light has been turned on, the system administrator can use this light to indicate one physical machine out of a bunch of enclosures in a server frame. It makes life easier.
With xCAT, the end user can turn the beacon light on or off with the commands show below. ::
rbeacon cn1 on
rbeacon cn1 off
Please notice, the current state of the beacon light can not be inquery remotely. As a workaround, one can always use the ``rbeacon`` command to trun all the beacon lights in one frame off, and then turn a particular beancon light on. ::
rbeacon a_group_of_cn off
rbeacon cn5 on
Remote Power Control
````````````````````
The next important thing is to control the power of a remote physical machine. For this purpose, ``rpower`` command is involved. ::
rpower cn1 on
rpower cn1 off
In order to reboot a remote physical machine, run ::
rpower cn1 boot
Or do a hardware reset, run ::
rpower cn1 reset
Get the current rpower state of a machine, please refer to the example below. ::
# rpower cn1 state
cn1: Running
Remote Console
``````````````
Most enterprise level servers do not have video adapters installed with the machine. Meaning, the end user can not connect a monitor to the machine and get display output. In most cases, the console can be viewed using the serial port or LAN port, through Serial-over-LAN. Serial cable or network cable are used to get a command line interface of the machine. From there, the end user can get the basic machine booting information, firmware settings interface, local command line console, etc.
In order to get the command line console remotely. xCAT provides the ``rcons`` command. ::
First of all, make sure the ``conserver`` is configured, if not, configue it with ::
makeconservercf
Then check if the ``conserver`` is up and running ::
ps ax | grep conserver
If the conserver is not running, or you just updated its configuration file, restart the conserver with ::
service conserver restart
In case you have ``systemd`` instead of ``sysvinit``, use the command below instead ::
systemctl restart conserver.service
After that, you can get the command line console for a specific machine with the ``rcons`` command ::
rcons cn1
Advanced operation
------------------
Remote Hardware Inventory
`````````````````````````
When you have a lot of physical machines in one place, the most important thing is identify which is which. Mapping the model type and/or serial number of a machine with its host name. Command ``rinv`` is involved in such a situation. With this command, most of the important information to distinct one machine from all the others can be obtained remotely.
To get all the hardware information, which including the model type, serial number, firmware version, detail configuration, et al. ::
rinv cn1 all
As an example, in order to get only the information of firmware version, the follwing command can be used. ::
rinv cn1 firm
Remote Hardware Vitals
``````````````````````
Collect runtime information from running physical machine is also a big requirement for real life system administrators. This kind of information includes, temperature of CPU, internal voltage of paricular socket, wattage with workload, speed of cooling fan, et al.
In order to get such information, please use ``rvitals`` command. Please also notice, this kind of information various among different model types of the machine. Thus, please check the actual output of the ``rvitals`` command against your machine, to verify which kinds of information can be get. The information may change due to the firmware updating of the machine. ::
rvitals cn1 all
As an example, get only the temperature information of a particular machine. ::
rvitals cn1 temp
Firmware Updating
`````````````````
**TODO**: For OpenPower machine, the firmware updating feature is not implement in the ``rflash`` command. The section should be updated after this feature get implemented.
Configures Nodes' Service Processors
````````````````````````````````````
Here comes the command, ``rspconfig``. It is used to configure the service processor of a phyisical machine. On a OpenPower system, the service processor is the BMC, BAse Motherboard Controller. Various variables can be set through the command. But, please also notice, the actual configuration may change among difference machine model type.
Examples
To turn on SNMP alerts for cn5: ::
rspconfig cn5 alert=on