大量频繁发起连接导致瀚高数据库进程被信号6杀死

文章目录环境症状问题原因解决方案环境系统平台银河麒麟 龙芯版本4.5.1症状数据库版本瀚高安全版V4.5.1.2介质 hgdb-see-4.5.1.2-ee53424.loongarch64.rpm信号6的报错信息2025-04-18 08:58:11.266 CST,用户名,数据库名,3171421,27.57.186.3:57716,6801a31e.30645d,2,authentication,2025-04-18 08:55:58 CST,15/269,0,PANIC,XX000,stuck spinlock detected at rdLock, auth.c:298,,,,,,,,, 2025-04-18 08:58:13.061 CST,用户名,数据库名,3171419,27.57.186.3:2867,6801a31e.30645b,2,authentication,2025-04-18 08:55:58 CST,12/243489,0,PANIC,XX000,stuck spinlock detected at rdUnLock, auth.c:328,,,,,,,,, 2025-04-18 08:58:13.101 CST,,,2185685,,67fcd5ba.2159d5,59,,2025-04-14 17:30:34 CST,,0,LOG,00000,server process (PID 3171419) was terminated by signal 6: Aborted,,,,,,,,, 2025-04-18 08:58:13.101 CST,,,2185685,,67fcd5ba.2159d5,60,,2025-04-14 17:30:34 CST,,0,LOG,00000,terminating any other active server processes,,,,,,,,,问题原因应用连接瀚高数据库需要做身份鉴别给临界资源添加自旋锁锁竞争导致等待瀚高数据库自旋锁等待数量限制是1000当短时间发起连接超过1000时某个连接自旋锁等待次数可能大于1000当大于1000时报这个错误。解决方案1、永久解决方案应用短时间内发起大量连接需要应用调整设置如使用连接池连接减少频繁发起连接的次数。2、临时方案申请补丁包加大自旋锁等待超时次数设置.减少触发几率补丁包超时次数修改为2000P001-hgdb-see-4.5.1.2-4345-45d048420250513.loongarch64.rpm