0

I am using log4cxx and recently started working with SocketAppender from here 1. I use the following:

java org.apache.log4j.net.SimpleSocketServer 4712 log4j-server.properties

and am able to write logs on the server. I do also have a local SizeRollover policy setup. This all works fine, until I stop running the SimpleSockerServer. At which point my C++ application crashes, I think after the ReconnectionDelayTime specified in the properties file.

0 0x0000003c5fa0e6fd in write () from /lib64/libpthread.so.0

1 0x00007ffff79eade7 in apr_socket_send () from /usr/lib64/libapr-1.so.0

2 0x00007ffff7d52bb1 in log4cxx::helpers::Socket::write(log4cxx::helpers::ByteBuffer&) () from /usr/lib64/liblog4cxx.so.10

3 0x00007ffff7d595ed in log4cxx::helpers::SocketOutputStream::flush(log4cxx::helpers::Pool&) () from /usr/lib64/liblog4cxx.so.10

4 0x00007ffff7d5468c in log4cxx::net::SocketAppender::append(log4cxx::helpers::ObjectPtrT const&, log4cxx::helpers::Pool&) () from /usr/lib64/liblog4cxx.so.10

5 0x00007ffff7cd7d9d in log4cxx::AppenderSkeleton::doAppend(log4cxx::helpers::ObjectPtrT const&, log4cxx::helpers::Pool&) () from /usr/lib64/liblog4cxx.so.10

6 0x00007ffff7cd50f4 in log4cxx::helpers::AppenderAttachableImpl::appendLoopOnAppenders(log4cxx::helpers::ObjectPtrT const&, log4cxx::helpers::Pool&) () from /usr/lib64/liblog4cxx.so.10

7 0x00007ffff7d1adfd in log4cxx::Logger::callAppenders(log4cxx::helpers::ObjectPtrT const&, log4cxx::helpers::Pool&) const () from /usr/lib64/liblog4cxx.so.10

8 0x00007ffff7d1c4ac in log4cxx::Logger::forcedLog(log4cxx::helpers::ObjectPtrT const&, std::basic_string, std::allocator > const&, log4cxx::spi::LocationInfo const&) const () from /usr/lib64/liblog4cxx.so.10`

My question is, it possible to make log4cxx write logs to a local file in case the server stops? If I start the server again, before the application crashes, SocketAppender does not receive the logs again. How can I get around this problem? Any other suggestions to send logs over the network?

Community
  • 1
  • 1
Raaka
  • 304
  • 1
  • 3
  • 15
  • The best way to "get around" this problem is to find out *why* the crash happens, and try to fix that. You might have found a bug in the framework, in which case you should report it. – Some programmer dude Sep 10 '14 at 17:45
  • I think the reason it crashes is that it is trying to send the logs over to an earlier socket connection, which now doesn't exist. – Raaka Sep 10 '14 at 17:49

0 Answers0