sg管理和维护（SG management and maintenance）.docVIP

sg管理和维护（SG management and maintenance） Management and maintenance of MC/SG Start and stop of MC/SG system In normal use, any node fails, and the backup node can take over the work of the node. But the last time a normal shutdown, which node is the production machine, then the next boot, the node or production machine. If the node does not start properly, the backup machine cannot automatically become a production machine. The startup and shutdown of MC/Services Guard shall comply with the sequence of operations. The order of manual Mc/Service Guard is as follows. 1. boot (production machine) 1) loading operating system 2) start cluster Be careful： Dont boot two machines at once. After the production machine is normal, another machine should be led; Dont execute the cmruncl command when the cluster is already running, which can lead to inconsistent data; - the usual configuration is that package starts normally as cluster starts normally; - when you manually hang package, you want to restart package, and you can manually start package; 2. power off 1) turn off the database management system 2) hang package 3) hang cluster 4) turn off the operating system Be careful： - turn off the backup machine first. If you turn off the production machine first, the backup machine will be switched to the production machine immediately, and the machine will be the first to boot the next time. When a human database is closed normally, the database daemon does not automatically switch to another machine. Management and maintenance of MC/SG Handover of MC/SG system There are two kinds of switching: local switching (Local, Switch) and Switch When the card failure master node (MAC address or Hardware StateDown), will be the first local switch card backup to work; if the backup Nic has fault, the system will switch over, the backup node system. MC switching system occurred in two cases: the first case is due to hardware failure or application failure, the switching is done automatically;