So I was having an issue with rsyslog in one of my centos 6.6 hosts:
[root@scan ~]# /etc/init.d/rsyslog start Starting system logger: *** glibc detected *** /sbin/rsyslogd: double free or corruption (fasttop): 0x00007f80cc3da880 *** ======= Backtrace: ========= /lib64/libc.so.6(+0x75e66)[0x7f80c9210e66] /usr/lib64/librelp.so.0(relpTcpDestruct+0x5f)[0x7f80c7f1a9bf] /usr/lib64/librelp.so.0(relpSrvRun+0x4c)[0x7f80c7f1d08c] /usr/lib64/librelp.so.0(relpEngineListnerConstructFinalize+0x26)[0x7f80c7f16ac6] /usr/lib64/librelp.so.0(relpEngineAddListner+0x91)[0x7f80c7f16c61] /lib64/rsyslog/imrelp.so(+0xd4c)[0x7f80c8122d4c] [...] 7fff6c305000-7fff6c31c000 rw-p 00000000 00:00 0 [stack] 7fff6c3ff000-7fff6c400000 r-xp 00000000 00:00 0 [vdso] ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0 [vsyscall]
^C [root@scan ~]#
Since this is the only machine I put relp in (I was playing with it), and all the others are working, I decided to remove rsyslog-relp-5.8.10-10.el6_6.x86_64 to see if that solves the issue. So, I
yum remove rsyslog-relp
and restarted rsyslog and it works now. Now that is fixed, what would cause rsyslog-relp to crash rsyslog?