Veritas Storage Foundation 5.1 SP1 Cluster File System Administrator"s Guide (5900-1738, April 2011)

Table 2-2
I/O fencing scenarios (continued)
Operator actionNode B: What
happens?
Node A: What
happens?
Event
Restart Node B after
private networks are
restored.
Node B has crashed.
It cannot start the
database since it is
unable to write to the
data disks.
Node A continues to
work.
Both private
networks function
again after event
above.
Repair private
network. After
network is repaired,
both nodes
automatically use it.
Node B prints
message about an
IOFENCE on the
console but
continues.
Node A prints
message about an
IOFENCE on the
console but
continues.
One private network
fails.
Verify private
networks function
and restart Node A.
Node B loses
heartbeats with Node
A, and races for a
majority of
coordinator disks.
Node B wins race for
coordinator disks and
ejects Node A from
shared data disks.
Node A is extremely
busy for some reason
or is in the kernel
debugger.
When Node A is no
longer hung or in the
kernel debugger, any
queued writes to the
data disks fail
because Node A is
ejected. When Node
A receives message
from GAB about
being ejected, it
panics and removes
itself from the
cluster.
Node A hangs.
Storage Foundation Cluster File System architecture
About I/O fencing
40