RocketMQ 4.9.7 线程卡死情况 Thread Got Stuck with RocketMQ v4.9.7

[williams-voon]

This issue happened in March 2025, and I had no ideal about the root cause。

In CommitLog.java, the following code prints log:

        if (elapsedTimeInLock > 500) {
            log.warn("[NOTIFYME]putMessage in lock cost time(ms)={}, bodyLength={} AppendMessageResult={}", elapsedTimeInLock, msg.getBody().length, result);
        }

It says "in lock cost time(ms)={some 6000}", and also prints the variable 'result'. Sorry the log has been lost.

I dig into the code and found the thread was stuck at the line: byteBuffer.put(preEncodeBuffer);

            final long beginTimeMills = CommitLog.this.defaultMessageStore.now();
            // Write messages to the queue buffer
            byteBuffer.put(preEncodeBuffer);
            msgInner.setEncodedBuff(null);
            AppendMessageResult result = new AppendMessageResult(AppendMessageStatus.PUT_OK, wroteOffset, msgLen, msgIdSupplier,
                msgInner.getStoreTimestamp(), queueOffset, CommitLog.this.defaultMessageStore.now() - beginTimeMills);

What is more, the WATERMARK log thread was stuck for 3 seconds as well in the mean time.

I am not sure if the issue was due to a bug in RocketMQ, could someone give me some suggestions？

问题是几个月前发现的，当时已经定位到某一行代码卡住了6秒，这是个 byteBuffer.put JDK 调用，且同时watermark的日志也有大概3秒没有打印。当时也查看了CPU、IO、网络的记录，没有异常。就这样放下了。
现在想起来始终没有合理答案，也不确定是不是 RocketMQ的bug，所以发出来问问，希望有大牛帮忙指导一下。