Root cause for the node Eviction:-


=== Update ===
Glibc package versions have met pre-requisities for 10.2.0.2 RAC & RDBMS

The reason why server has been evicted from cluster is visible in /var/log/messages:
Jun 27 17:47:40 orinoco1 kernel: bnx2: eth0 NIC Link is Down

OCFS stopped itself and node has been evicted. This is an expected behavior to avoid data corruption.

We cannot say why NIC link went down. It can be a NIC bonding bug in RHEL, network switch outage or cable disconnected.

Oracle RAC, OCFS neither RDBMS is not managing network devices so it is not a Oracle related issue.