In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-01-19 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Servers >
Share
Shulou(Shulou.com)06/01 Report--
This article shows you what the RAC refactoring type is like, the content is concise and easy to understand, it will definitely brighten your eyes. I hope you can get something through the detailed introduction of this article.
RAC isolation system can be divided into three layers: DB layer, clusterware layer and OS layer.
Each layer provides a mechanism to ensure that RAC works properly, and each layer can also trigger RAC refactoring
DB layer
Maintained by the LMON process, which provides two services, CGS and NM, with CGS on top of NM
NM: all processes in each instance are treated as a group registered with clusterware, where LMON is the main process; it is the communication channel between DB and clusterware, and is responsible for regular communication with other node NM
When Instance starts or shuts down normally, its NM registers or cancels registration with clusterware, notifies the NM of other nodes, and then NM notifies CGS to be in charge of refactoring
Startup/shutdown normal-> local NM-- > remote NM-- > remote CGS-- > GCS/GES reconfiguration
If the instance shuts down abnormally or the LMON/LMS/LMD communication times out, the CGS needs to IMR, record the result of the refactoring to the CFVRR field (control file vote result record) of the control file, and each instance reads the field to know whether it has been expelled.
Shutdown abort/ipc timeout-- > IMR-- > control file
From the above, the refactoring caused by timeout in the DB layer can be divided into two types:
LMON communication timeout: default is 300s, caused by abnormal LMON/LMS/LMD communication
NAME VALUE ISDEFAULT ISMOD ISADJ
_ cgs_send_timeout 300 TRUE FALSE FALSE
Control file timeout: by default, ckpt updates the control file every 3 seconds. If there is no corresponding control file beyond the specified time, IMR will be triggered.
NAME VALUE ISDEFAULT ISMOD ISADJ
_ controlfile_enqueue_timeout 900 TRUE FALSE FALSE
According to What is Split Brain in Oracle Clusterware and Real Application Cluster [ID 1425586.1]
Controlfile is used similarly to voting disk in clusterware layer to determine which instances survive and which instances evict. The voting result is similar to clusterware voting result
Control files act as voting disk in the RAC layer
Clusterware layer
OCSSD is responsible for providing two heartbeat mechanisms, network heartbeat and disk heartbeat, whose parameters are as follows
[oracle@ ~] $crsctl get css disktimeout
CRS-4678: Successful get disktimeout 200 for Cluster Synchronization Services.
[oracle@ ~] $crsctl get css misscount
CRS-4678: Successful get misscount 30 for Cluster Synchronization Services.
Network heartbeat timeout will cause the cluster to be divided into n subclusters, the subcluster with the largest number of nodes will eventually gain control, and the rest of the nodes will be removed and restarted (11R2 introduces a new mechanism to replace fast reboot)
The group with lower node member in case of same number of node (s) available in each group
If the two subset groups have the same number of nodes, the subset group used for the lower node number will eventually survive.
Disk heartbeat timeout: if ocssd does not update voting disk within the specified time, oracle determines its offline, and if the number of offline is greater than the number of online, the node will be expelled
OS layer
Hangcheck-Timer: used to detect whether linux is hang
Oprocd:10204 introduces this process from linux and performs a check every 1 second. If it fails, sleep will continue to try after 500ms; 11R2 will be replaced by cssdagent.
The above is what the RAC refactoring type looks like. Have you learned any knowledge or skills? If you want to learn more skills or enrich your knowledge reserve, you are welcome to follow the industry information channel.
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
Views: 0
*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.
Continue with the installation of the previous hadoop.First, install zookooper1. Decompress zookoope
"Every 5-10 years, there's a rare product, a really special, very unusual product that's the most un
© 2024 shulou.com SLNews company. All rights reserved.