cbbackup continues when node failover happens in the cluster

Backup continues when a node failover happens. And at the end, we just backup all the items on stable nodes in the cluster.

[chisheng@cen-1715 membase-cli]$ ./cbbackup http://Administrator:password@10.5.2.23:8091 ~/backup
warning: could not import couchstore module; accessing local couchstore files will not work
2012-07-27 20:30:14,838: w0 error: recv exception: [Errno 104] Connection reset by peer
########### 59.5% (555497/934098 msgs)
bucket: default, msgs transferred...
: total | last | per sec
batch : 75147 | 75147 | 157.9
byte : 1727825445 | 1727825445 | 3630453.2
msg : 555497 | 555497 | 1167.2
done

[chisheng@cen-1715 membase-cli]$ ./cbrestore ~/backup http://Administrator:password@10.5.2.22:8091 -x try_xwm=0 -b default

################### 100.0% (555497/555497 msgs)
bucket: default, msgs transferred...
: total | last | per sec
batch : 4305 | 4305 | 36.5
byte : 1727825445 | 1727825445 | 14667323.3
msg : 555497 | 555497 | 4715.6
done