ricci crashing immediately after starting and prints an error and backtrace from glibc if started directly in a RHEL 6 High Availability cluster

Issue

• When I start ricci, it dies immediately.
• After I start ricci, 'service ricci status' says it is dead

# /etc/init.d/ricci stop
Shutting down ricci:
# /etc/init.d/ricci start
Starting ricci:                                            [  OK  ]
# /etc/init.d/ricci status
ricci dead but pid file exists
# ricci
# /etc/init.d/ricci status
ricci dead but pid file exists

• If I start ricci directly outside the service, glibc gives an error and backtrace about an invalid free

# ricci -f -u ricci
*** glibc detected *** ricci: free(): invalid pointer: 0x00007ffe0add3c00 ***
======= Backtrace: =========
/lib64/libc.so.6[0x3428e75f4e]
/lib64/libc.so.6[0x3428e78cad]
ricci[0x41dd1f]
ricci[0x40696e]
/lib64/libc.so.6(__libc_start_main+0xfd)[0x3428e1ed5d]
ricci[0x406399]
======= Memory map: ========
00400000-0043e000 r-xp 00000000 fd:00 665063                             /usr/sbin/ricci
0063d000-0063f000 rw-p 0003d000 fd:00 665063                             /usr/sbin/ricci
01e29000-01e4a000 rw-p 00000000 00:00 0                                  [heap]
3428600000-3428620000 r-xp 00000000 fd:00 391788                         /lib64/ld-2.12.so
342881f000-3428820000 r--p 0001f000 fd:00 391788                         /lib64/ld-2.12.so
[...]
342c600000-342c629000 r-xp 00000000 fd:00 396845                         /lib64/libk5crypto.so.3.1Aborted