Unable to login to the system (eg. with ssh) after rsyslog crashes, it looks frozen

Issue

• System hangs after sending a large amount of syslog messages via TCP
• I'm not able to login through ssh, but the server answers ping requests
• abrt-addon-ccpp user space deadlock when rsyslog daemon crashes
• rsyslogd crashes and no core dump is saved by abrt
• Below are the common call traces found on the logs when the problem happens:

INFO: task rsyslogd:13421 blocked for more than 120 seconds.
      Not tainted 2.6.32-431.29.2.el6.x86_64 #1
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
rsyslogd      D 0000000000000003     0 13421      1 0x00000080
 ffff88204eb15c98 0000000000000082 ffff882053935a28 ffff882053933500
 ffff88204eb15c68 ffffffff810af000 ffff882053935a28 0000000000000000
 ffff882050d105f8 ffff88204eb15fd8 000000000000fbc8 ffff882050d105f8
Call Trace:
 [<ffffffff810af000>] ? exit_robust_list+0x90/0x160
 [<ffffffff810767e5>] exit_mm+0x95/0x180
 [<ffffffff81076c2f>] do_exit+0x15f/0x870
 [<ffffffff81077398>] do_group_exit+0x58/0xd0
 [<ffffffff8108cd46>] get_signal_to_deliver+0x1f6/0x460
 [<ffffffff8100a265>] do_signal+0x75/0x800
 [<ffffffff810149b9>] ? read_tsc+0x9/0x20
 [<ffffffff811a04b8>] ? poll_select_copy_remaining+0xf8/0x150
 [<ffffffff8100aa80>] do_notify_resume+0x90/0xc0
 [<ffffffff8100b341>] int_signal+0x12/0x17