如何减少Exadata计算节点CPU的Core数量
Posted
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了如何减少Exadata计算节点CPU的Core数量相关的知识,希望对你有一定的参考价值。
最近为某客户做一个Exadata的PoC测试,要求是X8 1/8 rack配置,目前机器是1/4 rack的硬件。
OEDA配置时只选择了1/8 rack选项,其他都没有配置。但是在一键刷机时会发现跳过了计算节点的1/8 rack配置,只对存储节点进行了1/8 rack配置,如下:
Initializing
Disabling Exadata AIDE on [dbm08celadm01.oscbj.com, dbm08celadm02.oscbj.com, dbm08celadm03.oscbj.com, dbm08dbadm01.oscbj.com, dbm08dbadm02.oscbj.com]
Executing Update Nodes for Eighth Rack
Drop cell disks for Eighth Rack setup in [dbm08celadm02.oscbj.com]
Drop cell disks for Eighth Rack setup in [dbm08celadm01.oscbj.com]
Drop cell disks for Eighth Rack setup in [dbm08celadm03.oscbj.com]
Skip Eighth rack configuration in compute node dbm08dbadm02.oscbj.com
running setup on: dbm08celadm03
running setup on: dbm08celadm02
running setup on: dbm08celadm01
dbm08celadm02 needs total CPU cores set from 64 to 32
dbm08celadm03 needs total CPU cores set from 64 to 32
dbm08celadm01 needs total CPU cores set from 64 to 32
Skip Eighth rack configuration in compute node dbm08dbadm01.oscbj.com
Successfully completed execution of step Update Nodes for Eighth Rack [elapsed Time [Elapsed = 58142 mS [0.0 minutes] Fri Jun 04 14:21:26 CST 2021]]
那么现在已经刷机完成,此时需要对计算节点也限制cpu核数,匹配 1/8 rack配置,该如何操作呢?
可以使用dbmcli工具来配置:
[root@dbm08dbadm01 ~]# dbmcli
DBMCLI: 发行版 - 发行时间 Sun Jun 06 11:05:15 CST 2021
版权所有 (c) 2007, 2016, Oracle 和/或其关联公司。保留所有权利。
DBMCLI> LIST DBSERVER
dbm08dbadm01 online
DBMCLI> LIST DBSERVER DETAIL
name: dbm08dbadm01
bbuStatus: normal
coreCount: 48/48
cpuCount: 96/96
diagHistoryDays: 7
fanCount: 16/16
fanStatus: normal
httpsAccess: ALL
id: 1909XLB03N
interconnectCount: 2
interconnect1: ib0
interconnect2: ib1
ipaddress1: 192.168.118.1/22
ipaddress2: 192.168.118.2/22
kernelVersion: 4.14.35-1902.306.2.1.el7uek.x86_64
locatorLEDStatus: off
makeModel: Oracle Corporation ORACLE SERVER X8-2
metricHistoryDays: 7
msVersion: OSS_20.1.3.0.0_LINUX.X64_201023
powerCount: 2/2
powerStatus: normal
releaseImageStatus: success
releaseVersion: 20.1.3.0.0.201023
releaseTrackingBug: 31970366
status: online
temperatureReading: 23.0
temperatureStatus: normal
upTime: 2 days, 2:20
msStatus: running
rsStatus: running
DBMCLI> LIST DBSERVER ATTRIBUTES coreCount,cpuCount
48/48 96/96
DBMCLI> ALTER DBSERVER pendingCoreCount=24
DBM-10004: Decreasing the number of active cores is not supported.
DBMCLI> ALTER DBSERVER pendingCoreCount=24 force
DBServer dbm08dbadm01 successfully altered. Please reboot the system to make the new pendingCoreCount effective.
DBMCLI> LIST DBSERVER ATTRIBUTES coreCount,cpuCount
48/48 96/96
DBMCLI> exit
退出
[root@dbm08dbadm01 ~]# mpstat
Linux 4.14.35-1902.306.2.1.el7uek.x86_64 (dbm08dbadm01.oscbj.com) 2021年06月06日 _x86_64_ (96 CPU)
11时12分04秒 CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
11时12分04秒 all 0.28 0.00 0.28 0.00 0.00 0.00 0.00 0.00 0.00 99.43
[root@dbm08dbadm01 ~]#
[root@dbm08dbadm01 ~]# reboot
可以看到,需要加force参数才可以修改,而且修改后需要重启机器生效,我们按要求重启后再次查看计算节点CPU情况:
[root@dbm08dbadm01 ~]# mpstat
Linux 4.14.35-1902.306.2.1.el7uek.x86_64 (dbm08dbadm01.oscbj.com) 2021年06月06日 _x86_64_ (48 CPU)
11时21分33秒 CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
11时21分33秒 all 2.63 0.03 2.44 1.18 0.00 0.01 0.00 0.00 0.00 93.71
DBMCLI> list DBSERVER detail
name: dbm08dbadm01
bbuStatus: normal
coreCount: 24/48
cpuCount: 48/48
diagHistoryDays: 7
fanCount: 16/16
fanStatus: normal
httpsAccess: ALL
id: 1909XLB03N
interconnectCount: 2
interconnect1: ib0
interconnect2: ib1
ipaddress1: 192.168.118.1/22
ipaddress2: 192.168.118.2/22
kernelVersion: 4.14.35-1902.306.2.1.el7uek.x86_64
locatorLEDStatus: off
makeModel: Oracle Corporation ORACLE SERVER X8-2
metricHistoryDays: 7
msVersion: OSS_20.1.3.0.0_LINUX.X64_201023
powerCount: 2/2
powerStatus: normal
releaseImageStatus: success
releaseVersion: 20.1.3.0.0.201023
releaseTrackingBug: 31970366
status: online
temperatureReading: 23.0
temperatureStatus: normal
upTime: 0 days, 0:09
msStatus: running
rsStatus: running
DBMCLI> list DBSERVER attributes coreCount, cpuCount
24/48 48/48
此时可以看到已经成功减半了CPU的核数,匹配 1/8 rack配置,另外的计算节点同样操作即可,不再赘述。
以上是关于如何减少Exadata计算节点CPU的Core数量的主要内容,如果未能解决你的问题,请参考以下文章
Exadata Cell 节点 griddisk容量计算小节
如何减少 Tensorflow/Keras 使用的 CPU 数量?