Optin and FlowVisor flood disk space through logs

[CarolinaFernandez]

There is some I/O error related to FlowVisor that is flooding logs (up to 17G and 22G respectively) in the following files (with sample messages that are repeated):

/var/log/messages

Jan 14 16:44:46 ControlFramework flowvisor: WARN-slicer_SLICE1_ID__exp_i2cat_fp7-ofelia_eu_50_dpid=00:10:00:00:00:00:00:04: got i/o error; tearing down and reconnecting: java.io.IOException: Connection reset by peer

/var/log/syslog

Jan 12 06:27:08 ControlFramework rsyslogd: [origin software="rsyslogd" swVersion="4.6.4" x-pid="666" x-info="http://www.rsyslog.com"] rsyslogd was HUPed, type 'lightweight'.

Possible problems?

• The offending slice contained loops (and -only for the record- VM names with incorrect syntax (slashes)
• Also, the controller URL corresponds to a VM that was not up; therefore a packet could be asking forever for instructions on its handling:
  • Slice: SLICE1 (id=50)
  • tcp:10.216.12.145:6633

[CarolinaFernandez]

Take a look to opennetworkinglab/flowvisor#107.
Newer FlowVisor versions use log4j, so log level can be configured.