Hi Patrick,
I've set debug=True in /etc/ipsilon/idp/ipsilon.conf and don't see any Ipsilon-generated debug messages in the Apache error logs on restart.
$ head -n 3 /etc/ipsilon/idp/ipsilon.conf
[global]
debug = True
tools.log_request_response.on = False
$ sudo systemctl restart httpd
$ tail -f /etc/httpd/logs/error_log
[Sun Feb 21 18:20:40.036535 2016] [mpm_prefork:notice] [pid 1662] AH00170: caught SIGWINCH, shutting down gracefully
[Sun Feb 21 18:20:41.140473 2016] [suexec:notice] [pid 32618] AH01232: suEXEC mechanism enabled (wrapper: /usr/sbin/suexec)
[Sun Feb 21 18:20:41.159765 2016] [auth_digest:notice] [pid 32618] AH01757: generating secret for digest authentication ...
[Sun Feb 21 18:20:41.160488 2016] [lbmethod_heartbeat:notice] [pid 32618] AH02282: No slotmem from mod_heartmonitor
[Sun Feb 21 18:20:41.167307 2016] [mpm_prefork:notice] [pid 32618] AH00163: Apache/2.4.16 (Fedora) OpenSSL/1.0.1k-fips mod_auth_gssapi/1.3.0 mod_wsgi/4.4.1 Python/2.7.8 configured -- resuming normal operations
[Sun Feb 21 18:20:41.167392 2016] [core:notice] [pid 32618] AH00094: Command line: '/usr/sbin/httpd -D FOREGROUND'
That log output includes the time during which I was logging in and out via the web interface. I did see the full request/response bodies logged to that file when I had tools.log_request_response.on set to True, so I'm pretty sure I'm looking at the right file.
We're using SQLite. Should have mentioned before, I believe cleanup has not run not just because dbinfo doesn't have the expected information, but because we have sessions in the database with expiration dates going back to when we first started using this Ipsilon instance in production this past December.