Mark110513 edited a comment on issue #3948: URL: https://github.com/apache/rocketmq/issues/3948#issuecomment-1062633416
> waitTimeMillsInSendQueue=400 这个配置的有点小，另外可以同步增加queue size 的大小 warmMapedFileEnable=true 这个参数关掉 useReentrantLockWhenPutMessage 这个可以用回自旋锁，同步降低线程数，可以保持默认 transferMsgByHeap 这个建议保持默认 另外关注下磁盘的表现，stats.log 中的时间分布，丁老师的文章也可以看看： https://www.cnblogs.com/dingwpmz/p/12104702.html 调整后的broker配置： brokerClusterName = RaftCluster brokerName=39.3-broker0 listenPort=30911 namesrvAddr=10.254.39.1:9876;10.254.39.2:9876;10.254.39.3:9876 storePathRootDir=/data/rocketmq/rocketmq-4.9.2/sh/dledger/store/node00 storePathCommitLog=/data/rocketmq/rocketmq-4.9.2/sh/dledger/store/node00/commitlog enableDLegerCommitLog=true dLegerGroup=group-10.254.39.3 dLegerPeers=n30-10.254.39.3:40911;n31-10.254.39.3:40912;n32-10.254.39.3:40913 dLegerSelfId=n30 sendMessageThreadPoolNums=16 brokerIP1=10.254.39.3 deleteWhen=04 fileReservedTime=48 brokerRole=ASYNC_MASTER flushDiskType=ASYNC_FLUSH #发送消息线程等待时间，默认200ms waitTimeMillsInSendQueue=60000 #发送队列线程容量 sendThreadPoolQueueCapacity=100000 #发送消息的最大线程数，默认1 #sendMessageThreadPoolNums=16 #使用可重入锁 #useReentrantLockWhenPutMessage=true #开启临时存储池 transientStorePoolEnable=true #开启Slave读权限（分担master 压力） slaveReadEnable=true #关闭堆内存数据传输 #transferMsgByHeap=false #开启文件预热 #warmMapedFileEnable=true 出现以下报错信息： [PCBUSY_CLEAN_QUEUE]broker busy, start flow control for a while, period in queue: 967ms, size of queue: 19 [REJECTREQUEST]system busy, start flow control for a while -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
