本地XMPP连接随机断开伴XMLStreamException问题排查求助

问题背景 #

我们的两个应用通过本地XMPP服务器通信，日常消息交互正常，但连接会随机断开；执行大批次任务时，几乎必然在运行中出现连接断开。报错显示在[39341,25831]位置发生XML解析异常，但我们从未传输过这么大的数据，请求分析故障原因。

错误日志 #

2022-12-22T21:06:49.406-0500 WARN Connection XMPPTCPConnection[user1@localhost/Avert] (0) closed with error
org.jivesoftware.smack.xml.XmlPullParserException: javax.xml.stream.XMLStreamException: ParseError at [row,col]:[39341,25831]
Message: Connection reset
    at org.jivesoftware.smack.xml.stax.StaxXmlPullParser.next(StaxXmlPullParser.java:193) ~[smack-xmlparser-stax-4.4.3.jar!/:4.4.3]
    at org.jivesoftware.smack.tcp.XMPPTCPConnection$PacketReader.parsePackets(XMPPTCPConnection.java:1140) ~[smack-tcp-4.4.3.jar!/:4.4.3]
    at org.jivesoftware.smack.tcp.XMPPTCPConnection$PacketReader.access$700(XMPPTCPConnection.java:913) ~[smack-tcp-4.4.3.jar!/:4.4.3]
    at org.jivesoftware.smack.tcp.XMPPTCPConnection$PacketReader$1.run(XMPPTCPConnection.java:936) ~[smack-tcp-4.4.3.jar!/:4.4.3]
    at java.lang.Thread.run(Unknown Source) [?:?]
Caused by: javax.xml.stream.XMLStreamException: ParseError at [row,col]:[39341,25831]
Message: Connection reset
    at com.sun.org.apache.xerces.internal.impl.XMLStreamReaderImpl.next(Unknown Source) ~[?:?]
    at org.jivesoftware.smack.xml.stax.StaxXmlPullParser.next(StaxXmlPullParser.java:191) ~[smack-xmlparser-stax-4.4.3.jar!/:4.4.3]
    ... 4 more
21:06:49 XMPPConnection closed due to an exception (XMPPTCPConnection[user1@localhost/Avert] (0))
org.jivesoftware.smack.xml.XmlPullParserException: javax.xml.stream.XMLStreamException: ParseError at [row,col]:[39341,25831]
Message: Connection reset
    at org.jivesoftware.smack.xml.stax.StaxXmlPullParser.next(StaxXmlPullParser.java:193)
    at org.jivesoftware.smack.tcp.XMPPTCPConnection$PacketReader.parsePackets(XMPPTCPConnection.java:1140)
    at org.jivesoftware.smack.tcp.XMPPTCPConnection$PacketReader.access$700(XMPPTCPConnection.java:913)
    at org.jivesoftware.smack.tcp.XMPPTCPConnection$PacketReader$1.run(XMPPTCPConnection.java:936)
    at java.base/java.lang.Thread.run(Unknown Source)
Caused by: javax.xml.stream.XMLStreamException: ParseError at [row,col]:[39341,25831]
Message: Connection reset
    at java.xml/com.sun.org.apache.xerces.internal.impl.XMLStreamReaderImpl.next(Unknown Source)
    at org.jivesoftware.smack.xml.stax.StaxXmlPullParser.next(StaxXmlPullParser.java:191)
    ... 4 more

故障原因分析 #

• XML解析异常是连接断开的次生结果：日志核心错误为Connection reset，[39341,25831]的行列号并非真实传输数据的大小，而是XML解析器在连接突然中断时，当前缓冲区的偏移位置标记，并非实际数据包的尺寸。
• 大批次任务触发资源瓶颈：高并发任务场景下，XMPP服务器或客户端的TCP连接缓冲区可能被耗尽，导致操作系统主动重置连接；另外服务器的线程池、内存资源不足，无法处理大量并发数据包，会强制断开连接。
• TCP层隐式问题：本地网络看似稳定，但大流量传输时可能出现丢包、延迟，触发TCP超时重传机制，最终导致连接被重置；防火墙或操作系统的TCP连接超时配置过短，也会在长时间大流量传输时强制断开连接。
• XMPP配置缺失或不合理：使用的Smack客户端（版本4.4.3）若未配置XEP-0199心跳机制，长时间无交互时会被服务器判定为死连接断开；服务器端的最大数据包大小、并发连接数限制过低，大批次任务会触发阈值导致连接中断。

内容的提问来源于stack exchange，提问作者runfastman