Oracle 12.2新特性 | 基于權(quán)重的節(jié)點驅(qū)逐
前言
在Oracle Clusterware需要從集群中逐出特定節(jié)點或一組節(jié)點的情況時，基于服務器權(quán)重的節(jié)點驅(qū)逐可充當斷裂機制，通過服務器上的負載附加信息來識別要驅(qū)逐的節(jié)點或節(jié)點組。其中所有節(jié)點被驅(qū)逐的可能性是一樣的。
本課程視頻，點擊閱讀原文即可下載
基于權(quán)重的集群驅(qū)逐介紹基于權(quán)重的集群驅(qū)逐（Server Weight-Based Node Eviction）是一項Oracle 12.2版本引入的一項新特性。在此之前，Oracle集群在處理腦裂問題時，通過判斷子群(各自獨立)的規(guī)模情況，來決定在腦裂問題發(fā)生時，終止哪個子群，判斷的依據(jù)如下：
-
如果子群規(guī)模（包含節(jié)點數(shù)量）不等，則大的子群獲勝，例如，{1} {2,3,4}后者獲勝，子群{1}被驅(qū)逐；
-
如果子群規(guī)模（包含節(jié)點數(shù)量）相等，則包含最小節(jié)點號的子群獲勝，例如,{1,4}{2,3}, 子群{1,4}獲勝，子群{2,3}被驅(qū)逐。
而在Oracle 12.2版本上引入的這項新特性，一定程度上增加了我們對于集群的控制。這種控制使得我們能避免由于一些規(guī)則原來的限定，而減少對大范圍的應用產(chǎn)生影響。
注意：
以下針對weight以及權(quán)重表述的是同一個意思。
特性使用>>>>使用前提條件
-
權(quán)重(Weight)分配只能在admin-managed的節(jié)點生效
-
可以對于server或者對于注冊到集群的應用進行權(quán)重(Weight)分配
使用說明
-
權(quán)重分配給server
使用crsctl set server css_critical yes命令
-
權(quán)重分配給數(shù)據(jù)庫實例或服務
在srvctl add/modify databse 或 srvctladd/modify service命令中帶上“-css_critical yes”參數(shù)
-
權(quán)重分配給非ora.*資源
在crsctl add/modify resource命令中，加入-attr "CSS_CRITICAL=yes"參數(shù)
注意：
有些權(quán)重的分配需要重啟集群才能生效，而有些資源的分配，則不需要重啟資源。
根據(jù)目前發(fā)現(xiàn)，非ora.*資源并非所有都能直接加入屬性直接修改，但是屬性中可以看到有CSS_CRITICAL屬性，可能是當前版本未開發(fā)修改接口。
特性測試>>>>實驗環(huán)境說明
-
使用軟件環(huán)境說明：
OS:MacOS 10.11.6
VirtualBox:v5.1.30 r118389 (Qt5.6.3)
-
虛擬機環(huán)境說明：
OEL(OracleEnterprise Linux) 6.5 , x86_64
Oracle 12.2.0.1 (2-node RAC)
實驗前的準備工作
VirtualBox虛擬機中的網(wǎng)卡情況如下：
其中網(wǎng)卡1為Public網(wǎng)絡，網(wǎng)卡2為心跳網(wǎng)絡分別對應OEL6.5中的eth0網(wǎng)卡和eth1網(wǎng)卡。
為了模擬心跳網(wǎng)絡中斷，我們在系統(tǒng)中編寫了如下兩個腳本去模擬心跳網(wǎng)絡中斷以及恢復：
Milo-Mac:lab milo$ ls -l
total 16
-rwxr--r-- 1 milo staff 111 1 24 21:38 interconnect_down.sh
-rwxr--r-- 1 milo staff 109 1 24 21:39 interconnect_up.sh
Milo-Mac:lab milo$
#### 模擬心跳網(wǎng)絡中斷：
Milo-Mac:lab milo$ sh interconnect_down.sh
查看虛擬機的網(wǎng)卡的連通情況：
[root@rac122a ~]# mii-tool eth0
eth0: no autonegotiation, 100baseTx-FD, link ok
[root@rac122a ~]# mii-tool eth1
eth1: autonegotiation restarted, no link
[root@rac122b ~]# mii-tool eth0
eth0: no autonegotiation, 100baseTx-FD, link ok
[root@rac122b ~]# mii-tool eth1
eth1: autonegotiation restarted, no link
可以看到在腳本運行后，私網(wǎng)網(wǎng)卡eth1顯示為no link表示網(wǎng)線沒有連接到網(wǎng)卡中，即我們認為心跳網(wǎng)絡出現(xiàn)故障。
Milo-Mac:lab milo$ ls -l
total 16
-rwxr--r-- 1 milo staff 111 1 24 21:38 interconnect_down.sh
-rwxr--r-- 1 milo staff 109 1 24 21:39 interconnect_up.sh
Milo-Mac:lab milo$
#### 恢復心跳網(wǎng)絡：
Milo-Mac:lab milo$ sh interconnect_up.sh查看虛擬機的網(wǎng)卡的連通情況：
[root@rac122a ~]# mii-tool eth0
eth0: no autonegotiation, 100baseTx-FD, link ok
[root@rac122a ~]# mii-tool eth1
eth1: no autonegotiation, 100baseTx-FD, link ok
[root@rac122b ~]# mii-tool eth0
eth0: no autonegotiation, 100baseTx-FD, link ok
[root@rac122b ~]# mii-tool eth1
eth1: no autonegotiation, 100baseTx-FD, link ok可以看到在腳本運行后，私網(wǎng)網(wǎng)卡eth1顯示為link ok表示網(wǎng)線連接到網(wǎng)卡中，即我們認為心跳網(wǎng)絡恢復，當然實際情況是我們還需要重新禁用和啟用eth1網(wǎng)卡，才能使其完全恢復，因此需要在兩個節(jié)點都執(zhí)行以下腳本：
[root@rac122a ~]# sh recover_interconnect.sh
Device state: 3 (disconnected)
Active connection state: activated
Active connection path: /org/freedesktop/NetworkManager/ActiveConnection/2上述相關腳本如下：
Milo-Mac:lab milo$ cat interconnect_down.sh
VBoxManage controlvm "12c_rac_node1" setlinkstate2 off
VBoxManage controlvm "12c_rac_node2" setlinkstate2 off
Milo-Mac:lab milo$ cat interconnect_up.sh
VBoxManage controlvm "12c_rac_node1" setlinkstate2 on
VBoxManage controlvm "12c_rac_node2" setlinkstate2 on
[root@rac122a ~]# cat recover_interconnect.sh
ifdown eth1 && ifup eth1
經(jīng)過上述測試，我們已經(jīng)可以模擬心跳網(wǎng)絡故障以及恢復心跳網(wǎng)絡故障。
以下測試場景不再贅述，上述故障模擬以及故障恢復過程。
>>>>
測試場景
[oracle@rac122a ~]$ crsctl get server css_critical
CRS-5092: Current value of the server attribute CSS_CRITICAL is no.
[oracle@rac122b ~]$ crsctl get server css_critical
CRS-5092: Current value of the server attribute CSS_CRITICAL is no.
當前系統(tǒng)上存在的服務也為設置：
[oracle@rac122b ~]$ crsctl stat res ora.milodb.milodb_srv1.svc -f | grep CSS
CSS_CRITICAL=no
[oracle@rac122b ~]$ crsctl stat res ora.milodb.milodb_srv2.svc -f | grep CSS
CSS_CRITICAL=no
模擬心跳網(wǎng)絡故障
Milo-Mac:lab milo$ sh interconnect_down.sh
節(jié)點1集群的alert日志：
2018-01-25 08:17:33.423 [OCSSD(3896)]CRS-1612: Network communication with node rac122b (2) missing for 50% of timeout interval. Removal of this node from cluster in 14.610 seconds
2018-01-25 08:17:41.435 [OCSSD(3896)]CRS-1611: Network communication with node rac122b (2) missing for 75% of timeout interval. Removal of this node from cluster in 6.600 seconds
2018-01-25 08:17:45.438 [OCSSD(3896)]CRS-1610: Network communication with node rac122b (2) missing for 90% of timeout interval. Removal of this node from cluster in 2.600 seconds
2018-01-25 08:17:49.554 [OCSSD(3896)]CRS-1607: Node rac122b is being evicted in cluster incarnation 412297497; details at (:CSSNM00007:) in /u01/app/grid/diag/crs/rac122a/crs/trace/ocssd.trc.
節(jié)點2集群的alert日志：
2018-01-25 08:17:34.128 [OCSSD(25756)]CRS-1612: Network communication with node rac122a (1) missing for 50% of timeout interval. Removal of this node from cluster in 14.450 seconds
2018-01-25 08:17:41.140 [OCSSD(25756)]CRS-1611: Network communication with node rac122a (1) missing for 75% of timeout interval. Removal of this node from cluster in 7.440 seconds
2018-01-25 08:17:46.193 [OCSSD(25756)]CRS-1610: Network communication with node rac122a (1) missing for 90% of timeout interval. Removal of this node from cluster in 2.380 seconds
…省略部分信息
從上述信息來看，根據(jù)之前的集群腦裂出現(xiàn)的規(guī)則，在同等規(guī)模的子群中，節(jié)點1節(jié)點號小，因而存活，節(jié)點2被驅(qū)逐出集群。
設置節(jié)點2上的sever級別的權(quán)重
將心跳網(wǎng)絡恢復以及集群狀態(tài)恢復正常后，我們將對server級別的權(quán)重進行設置。
[root@rac122a ~]# crsctl get server css_critical
CRS-5092: Current value of the server attribute CSS_CRITICAL is no.
[root@rac122b ~]# crsctl get server css_critical
CRS-5092: Current value of the server attribute CSS_CRITICAL is no.
[root@rac122b ~]# crsctl set server css_critical yes
CRS-4416: Server attribute 'CSS_CRITICAL' successfully changed. Restart Oracle High Availability Services for new value to take effect.
設置完成后，提示需要重啟OHAS才能生效。
$ srvctl stop instance -d milodb -i milodb2
# crsctl stop crs
# crsctl start crs
模擬網(wǎng)路故障：
Milo-Mac:lab milo$ sh interconnect_down.sh
此時，我們來觀察集群的一些日志情況：
節(jié)點1集群的alert日志：
2018-01-25 09:44:03.671 [OCSSD(3717)]CRS-1612: Network communication with node rac122b (2) missing for 50% of timeout interval. Removal of this node from cluster in 14.890 seconds
2018-01-25 09:44:11.731 [OCSSD(3717)]CRS-1611: Network communication with node rac122b (2) missing for 75% of timeout interval. Removal of this node from cluster in 6.830 seconds
2018-01-25 09:44:15.739 [OCSSD(3717)]CRS-1610: Network communication with node rac122b (2) missing for 90% of timeout interval. Removal of this node from cluster in 2.830 seconds
2018-01-25 09:44:18.573 [OCSSD(3717)]CRS-1609: This node is unable to communicate with other nodes in the cluster and is going down to preserve cluster integrity; details at (:CSSNM00008:) in /u01/app/grid/diag/crs/rac122a/crs/trace/ocssd.trc.
2018-01-25 09:44:18.573 [OCSSD(3717)]CRS-1656: The CSS daemon is terminating due to a fatal error; Details at (:CSSSC00012:) in /u01/app/grid/diag/crs/rac122a/crs/trace/ocssd.trc
2018-01-25 09:44:18.616 [OCSSD(3717)]CRS-1652: Starting clean up of CRSD resources.
2018-01-25 09:44:20.608 [OCSSD(3717)]CRS-1608: This node was evicted by node 2, rac122b; details at (:CSSNM00005:) in /u01/app/grid/diag/crs/rac122a/crs/trace/ocssd.trc.
節(jié)點2集群的alert日志：
2018-01-25 09:44:04.327 [OCSSD(8586)]CRS-1612: Network communication with node rac122a (1) missing for 50% of timeout interval. Removal of this node from cluster in 14.020 seconds
2018-01-25 09:44:11.609 [OCSSD(8586)]CRS-1611: Network communication with node rac122a (1) missing for 75% of timeout interval. Removal of this node from cluster in 6.730 seconds
2018-01-25 09:44:15.611 [OCSSD(8586)]CRS-1610: Network communication with node rac122a (1) missing for 90% of timeout interval. Removal of this node from cluster in 2.730 seconds
2018-01-25 09:44:19.777 [OCSSD(8586)]CRS-1607: Node rac122a is being evicted in cluster incarnation 412336033; details at (:CSSNM00007:) in /u01/app/grid/diag/crs/rac122b/crs/trace/ocssd.trc.
此時，我們從集群的alert日志中看到由于節(jié)點1被集群驅(qū)逐了。
通過從節(jié)點2的ocssd.trc日志我們看到如下信息：
[grid@rac122b trace]$ cat ocssd.trc |egrep -i weight| tail -300
……省略部分信息
2018-01-25 09:44:18.347 : CSSD:1494243072: clssnmrCheckNodeWeight: node(1) has weight stamp(0) pebbles (0) goldstars (0) flags (0) SpoolVersion (0)
2018-01-25 09:44:18.347 : CSSD:1494243072: clssnmrCheckNodeWeight: node(2) has weight stamp(412336032) pebbles (0) goldstars (1) flags (3) SpoolVersion (0)
2018-01-25 09:44:18.347 : CSSD:1494243072: clssnmrCheckNodeWeight: Server pool version not consistent
2018-01-25 09:44:18.347 : CSSD:1494243072: clssnmrCheckNodeWeight: stamp(412336032), completed(1/2)
2018-01-25 09:44:18.347 : CSSD:1494243072: clssnmrCheckSplit: Waiting for node weights, stamp(412336032)
2018-01-25 09:44:19.777 : CSSD:1494243072: clssnmrCheckNodeWeight: node(2) has weight stamp(412336032) pebbles (0) goldstars (1) flags (3) SpoolVersion (0)
2018-01-25 09:44:19.777 : CSSD:1494243072: clssnmrCheckNodeWeight: Server pool version not consistent
2018-01-25 09:44:19.777 : CSSD:1494243072: clssnmrCheckNodeWeight: stamp(412336032), completed(1/1)
2018-01-25 09:44:19.777 : CSSD:1494243072: clssnmCompareNodeWeights: Best map is same as the cohort map of the current node
2018-01-25 09:44:19.777 : CSSD:1494243072: clssnmFindBestMap: Using base map(2) of node(1) count(0), low(65535), bestcount(0), best_low(65535), cur_weightpebbles (0) goldstars (0) flags (0) SpoolVersion (0)best_weightpebbles (0) goldstars (0) flags (0) SpoolVersion (0)
2018-01-25 09:44:19.777 : CSSD:1494243072: clssnmCompareNodeWeights: count(1), low(2), bestcount(0), best_low(65535), cur_weight: pebbles(0) goldstars(1) pubnw(1) flexasm(1)best_weight: pebbles(0) goldstars(0)pubnw(0) flexasm(0)
從上述信息來看，CSSD進程會去檢查節(jié)點的權(quán)重情況(CheckNodeWeight)以及比較節(jié)點的權(quán)重（CompareNodeWeights）節(jié)點2的權(quán)重大，此時，節(jié)點2這個子群戰(zhàn)勝了節(jié)點1子群，因而我們看到的情況是節(jié)點1被集群驅(qū)逐。
設置節(jié)點2上的service的權(quán)重
將心跳網(wǎng)絡恢復以及集群狀態(tài)恢復正常后，我們將對service級別的權(quán)重進行設置。
恢復集群server的權(quán)重：
[root@rac122b ~]# crsctl get server css_critical
CRS-5092: Current value of the server attribute CSS_CRITICAL is yes.
[root@rac122b ~]# crsctl set server css_critical no
CRS-4416: Server attribute 'CSS_CRITICAL' successfully changed. Restart Oracle High Availability Services for new value to take effect.
[root@rac122b ~]# crsctl get server css_critical
CRS-5092: Current value of the server attribute CSS_CRITICAL is no.
[root@rac122b ~]#
設置完成后，提示需要重啟OHAS才能生效。
$ srvctl stop instance -d milodb -i milodb2
# crsctl stop crs
# crsctl start crs
這里我們添加兩個專門用于做權(quán)重控制的服務（不對外使用）：
srvctl add service -db milodb -service milodb_wt_srv1 -preferred milodb1
srvctl add service -db milodb -service milodb_wt_srv2 -preferred milodb2
srvctl start service -database milodb -service milodb_wt_srv1
srvctl start service -database milodb -service milodb_wt_srv2
在實例1上的服務不設置權(quán)重：
[oracle@rac122b ~]$ crsctl stat res ora.milodb.milodb_wt_srv1.svc -f | grep CSS
CSS_CRITICAL=no
在實例2上的服務設置權(quán)重：
[oracle@rac122b ~]$ crsctl stat res ora.milodb.milodb_wt_srv2.svc -f | grep CSS
CSS_CRITICAL=no
[oracle@rac122b ~]$ srvctl modify service -db milodb -service milodb_wt_srv2 -css_critical yes
[oracle@rac122b ~]$ crsctl stat res ora.milodb.milodb_wt_srv2.svc -f | grep CSS
CSS_CRITICAL=yes
[oracle@rac122b ~]$
此時，我們來觀察集群的一些日志情況：
節(jié)點1集群的alert日志：
2018-01-25 11:33:30.343 [OCSSD(3652)]CRS-1612: Network communication with node rac122b (2) missing for 50% of timeout interval. Removal of this node from cluster in 14.910 seconds
2018-01-25 11:33:38.351 [OCSSD(3652)]CRS-1611: Network communication with node rac122b (2) missing for 75% of timeout interval. Removal of this node from cluster in 6.900 seconds
2018-01-25 11:33:42.576 [OCSSD(3652)]CRS-1610: Network communication with node rac122b (2) missing for 90% of timeout interval. Removal of this node from cluster in 2.680 seconds
2018-01-25 11:33:45.255 [OCSSD(3652)]CRS-1609: This node is unable to communicate with other nodes in the cluster and is going down to preserve cluster integrity; details at (:CSSNM00008:) in /u01/app/grid/diag/crs/rac122a/crs/trace/ocssd.trc.
2018-01-25 11:33:45.255 [OCSSD(3652)]CRS-1656: The CSS daemon is terminating due to a fatal error; Details at (:CSSSC00012:) in /u01/app/grid/diag/crs/rac122a/crs/trace/ocssd.trc
2018-01-25 11:33:45.269 [OCSSD(3652)]CRS-1652: Starting clean up of CRSD resources.
2018-01-25 11:33:47.289 [OCSSD(3652)]CRS-1608: This node was evicted by node 2, rac122b; details at (:CSSNM00005:) in /u01/app/grid/diag/crs/rac122a/crs/trace/ocssd.trc.
節(jié)點2集群的alert日志：
2018-01-25 11:33:29.844 [OCSSD(3983)]CRS-1612: Network communication with node rac122a (1) missing for 50% of timeout interval. Removal of this node from cluster in 14.900 seconds
2018-01-25 11:33:37.946 [OCSSD(3983)]CRS-1611: Network communication with node rac122a (1) missing for 75% of timeout interval. Removal of this node from cluster in 6.790 seconds
2018-01-25 11:33:41.947 [OCSSD(3983)]CRS-1610: Network communication with node rac122a (1) missing for 90% of timeout interval. Removal of this node from cluster in 2.790 seconds
2018-01-25 11:33:46.684 [OCSSD(3983)]CRS-1607: Node rac122a is being evicted in cluster incarnation 412345406; details at (:CSSNM00007:) in /u01/app/grid/diag/crs/rac122b/crs/trace/ocssd.trc.
2018-01-25 11:33:47.755 [ORAAGENT(5634)]CRS-5818: Aborted command 'check' for resource 'ora.SYSTEMDG.dg'. Details at (:CRSAGF00113:) {0:1:11} in /u01/app/grid/diag/crs/rac122b/crs/trace/crsd_oraagent_grid.trc.
2018-01-25 11:33:47.759 [ORAAGENT(5634)]CRS-5818: Aborted command 'check' for resource 'ora.DATADG.dg'. Details at (:CRSAGF00113:) {0:1:11} in /u01/app/grid/diag/crs/rac122b/crs/trace/crsd_oraagent_grid.trc.
2018-01-25 11:33:49.250 [OCSSD(3983)]CRS-1601: CSSD Reconfiguration complete. Active nodes are rac122b .
2018-01-25 11:33:49.617 [CRSD(5349)]CRS-5504: Node down event reported for node 'rac122a'.
從節(jié)點2的ocssd日志看到，同樣由于節(jié)點2的權(quán)重高，因而最終將節(jié)點1驅(qū)逐除了集群：
后續(xù)測試將兩個服務都設置權(quán)重時，情況與不設置權(quán)重的情況一致，即節(jié)點1驅(qū)逐節(jié)點2。
總結(jié)通過上述場景的測試，我們得知：
-
基于權(quán)重的節(jié)點驅(qū)逐可以在集群出現(xiàn)腦裂時，通過控制減少消除一些影響；
-
配置server方式的權(quán)重，需要修改配置后，重啟crs才能生效；
-
通過服務的方式配置節(jié)點的權(quán)重不需要重啟資源，可以實現(xiàn)動態(tài)控制，更靈活。
當然，由于在該特性是12.2引入的，屬于新特性，可能會存在一些bug，建議在使用前應該經(jīng)過充分測試后使用。
作者介紹：
羅雪原，云和恩墨南區(qū)交付技術顧問。
個人有著8年的Oracle技術支持經(jīng)驗，曾服務過金融、保險、電力、政府、運營商等客戶，有著較豐富的Troubleshooting以及優(yōu)化經(jīng)驗。
