We have 3 node Oracle 184.108.40.206 RAC setup on SUSE Linux Enterprise Server 10 SP4 (x86_64).
While executing index rebuild script "alter index index_name rebuild online" due to delay in response from “index rebuild” command CTRL+C was executed to terminate session. Once again “index rebuild” was fired and this time completed successfully but due to last CTRL+C command "alter index index_name rebuild online" hang on wait event "enq: MD - contention" . This cause whole DB in block state. "
Since CTRL+C should have killed last process, but it did not and cause system caught in to deadlock situation.
Oracle Support confirm this as an Oracle Bug 12987588 - LNX64-11203-RAC:REBUILD INDEX ONLINE HANG AFTER CTRL-C MANY TIME was hit
I would suggest all to avoid executing CTRL+C commands. Though, this has happened just once and crosscheck for termination of sessions after CTRL+C command.