Mihail created LOG4J2-2829:
------------------------------
Summary: TcpSocketManager.write swallows IOException when retry =
false or reconnector != null
Key: LOG4J2-2829
URL: https://issues.apache.org/jira/browse/LOG4J2-2829
Project: Log4j 2
Issue Type: Bug
Components: Appenders
Affects Versions: 2.12.1
Reporter: Mihail
Hi, I see very stange behavior when using tcp SocketAppender. My configuration
is:
<Socket name="auditSocket" protocol="TCP" host="10.8.84.194" port="4587"
connectTimeoutMillis="10000" reconnectionDelayMillis="10000"
immediateFail="true" immediateFlush="true" ignoreExceptions="false"
bufferedIO="false">
After application starts and before writing any logs I simulate connection
reset with:
iptables -A OUTPUT -p tcp -d 10.8.84.194 --dport 4587 -j REJECT --reject-with
tcp-reset
Now the strange part:
# First log is written to output stream successfully (but nothing actually
gets sent)
# Second log throws IOException on outputstream write, this gets handled with
a reconnection attempt (which fails) and result is AppenderLoggingException -
this is what I expect because I'm using a failover appender to route the log
into file.
# Third and subsequent log attempts also fail due to IOException on
outputstream write BUT no AppenderLoggingException is raised - reconnector is
not null at this point and IOException gets swallowed up.
I'm using log4j2 2.12.1
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
