Red Hat Gluster Storage native client logs show 'bailing out frame type' and 'Transport endpoint is not connected' when server node fails

Issue

When one node of a volume replica peer becomes unresponsive or latent due to hardware or software failure, the native clients where the volume is mounted will show in the log file errors and warnings like below.

[2014-01-22 06:23:46.727937] E [rpc-clnt.c:207:call_bail] 0-volume1-client-14: bailing out frame type(GlusterFS 3.3) op(LOOKUP(27)) xid = 0x1970025x sent = 2014-01-22 05:53:44.592273. timeout = 1800
[2014-01-22 06:23:46.728034] W [client-rpc-fops.c:2604:client3_3_lookup_cbk] 0-volume1-client-14: remote operation failed: Transport endpoint is not connected. Path: / (00000000-0000-0000-0000-000000000001)

Note that the op(LOOKUP(27)) portion of the error message may vary.