突发故障,集群两个节点同时重启
某天深夜突然接到客户电话,集群脑裂导致集群2个节点在3分钟内先后重启,影响了业务。按照正常情况,脑裂只会踢出一个节点,两个节点分别重启的现象倒是比较少见。不由得让我好奇心大起,拿到客户提供的日志立即着手分析。
OSW收集到的数据,故障的时间线大致如下:
21:42:29开始,节点2日志采集中断,OS夯住导致操作系统命令无法正常执行。
21:44:12左右,节点1报出通讯错误,和节点1之间的通讯发生中断。
2024-06-19 21:44:27.744 [OCSSD(26526)]CRS-1612: Network communication with node orcldb02 (2) missing for 50% of timeout interval. Removal of this node from cluster in 14.830 seconds
2024-06-19 21:44:35.745 [OCSSD(26526)]CRS-1611: Network communication with node orcldb02 (2) missing for 75% of timeout interval. Removal of this node from cluster in 6.830 seconds
2024-06-19 21:44:39.745 [OCSSD(26526)]CRS-1610: Network communication with node orcldb02 (2) missing for 90% of timeout interval. Removal of this node from cluster in 2.830 seconds
看到以上信息,有经验的DBA都知道系统发生了脑裂,