Hi, NFSv3 over TCP hangs in case of network failure for a short while. I investigated further, and found that the nfs socket goes into TIME_WAIT state and doesn't come out of it, even after the default timeout of 120 seconds. This blocks all further traffic and any I/O from NFS client hangs. This happens only if tcp_timestamps is disabled, and in 50% of the cases. Kernel Version: 2.6.33.4 Steps to reproduce (at client end only): 1. Disable tcp_timestamps 2. Perform multi-user I/O on the NFS partition 3. Block traffic to NFS server. (I used iptables) 4. After the I/O queues up for some time (approx 3 minutes), re-enable network to the server. netstat -a | grep nfs shows nfs socket in TIME_WAIT state and it does not come out of it. I have the rpc logs (rpc_debug=129) if anyone requires Let me know if you need anything else. -- Goldwyn -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
