This article will introduce two tips for troubleshooting network issues on VNX for File. Carefully document the step you take. That way, if a tech support call becomes necessary, at least you’ll have a good idea of what isn’t causing the problem.

How to measure the TCP retransmissions on a Data Mover

If the customer encounters a network issue to cause re-transmit packets on the Data Mover, you can execute the following command from control station to check TCP retransmissions.

[nasadmin@NS480-1 ~]$ server_netstat server_2 -s -p tcp

    tcp:

    ****

    4948183 packets sent

    5 data packets retransmitted 

    0 resets

    3654442 packets received

    7332 connection requests

    41     connections lingered

Please note:

1. The statistics here are cumulative, so you need to execute this command at least two times (subtract the results to get the difference) in order to get the actual value during the time period.

2. The retransmitted value only counts from Data Mover to client and does not reflect the retransmit statistics as seen from client.

3. As a rule of thumb, if the retransmission rate is greater than 0.01% (retransmitted packet / total packet sent * 100), there is a problem on the network.

How to troubleshooting port aggregation issues on a Data Mover

When the customer encounters an Ethernet Channel issue on the Data Mover, you can try these troubleshooting tips in the order they’re listed below.

1.   Check all virtual devices on Data Mover, type:

          $ server_sysconfig server_2 -virtual

              server_2 :

              Virtual devices:   

              trk0_ec     devices=cge2 cge3

              fsn         failsafe nic devices :

             trk         trunking devices : trk0_ec

2.   Check port speed and duplex mode on Data Mover, type:

           $ server_sysconfig server_2 -pci

              server_2 : PCI DEVICES:

              Slot: 1

              Emulex LP9000 Fibre Channel Controller

              0: fcp-0 IRQ: 23 addr: 10000000c92b5a10

              1: fcp-1 IRQ: 24 addr: 10000000c92b5a11

              Slot: 2

              Emulex LP9000 Fibre Channel Controller

              0: fcp-2 IRQ: 22 addr: 10000000c92b514e

              Slot: 4

              Intel 10/100/1K Ethernet Controller

              0: cge0 IRQ: 18

              speed=auto duplex=auto rxflowctl=disable txflowctl=disable

              1: cge1 IRQ: 19

              speed=auto duplex=auto rxflowctl=disable txflowctl=disable

              2: cge2 IRQ: 20

              speed=auto duplex=auto rxflowctl=disable txflowctl=disable

              3: cge3 IRQ: 21

              speed=auto duplex=auto rxflowctl=disable txflowctl=disable

              Slot: 5

              Alteon Tigon-2 Gigabit Ethernet Controller

              0: ace0 IRQ: 25

              linkneg=enable rxflowctl=disable txflowctl=disable

              0: ace0 IRQ: 25

              linkneg=enable rxflowctl=disable txflowctl=disable

3.   Check the network statistics for the Data Mover, type:

$server_netstat server_x –i | -a | -s | -r

Note:

-a: displays the state of all the IP, ICMP, TCP, and UDP sockets.

-i: displays a summary of the state of all physical interfaces including the total amount of data transferred both ways and the number of errors during this process as well as the MTU size and MAC address.

-r: displays the routing table.

-s: displays the statistics of each protocol.

4.   Check the ARP table, type:

           $ server_arp server_2 -all

              server_2 :

              172.24.102.254 at 0:d0:3:f9:37:fc

              172.24.102.20 at 0:d0:b7:82:98:e0

              172.24.102.24 at 0:50:56:8e:1d:5

              128.221.253.100 at 0:4:23:a7:b1:35

5.   Check the route table, type:

           $ server_route server_2 -list

              server_2 :

              net 128.221.253.0 128.221.253.2 255.255.255.0 el31

              net 128.221.252.0 128.221.252.2 255.255.255.0 el30

              net 172.24.102.0 172.24.102.238 255.255.255.0 cge0

              host 127.0.0.1 127.0.0.1255.255.255.255 loop

6.   Verify IP address, network mask and broadcast address, type:

           $ server_ifconfig server_2 -all

              server_2 :

              loop protocol=IP device=loop

              inet=127.0.0.1 netmask=255.0.0.0 broadcast=127.255.255.255

              556 Command Line Interface Reference for File 7.0

              The server Commands

              UP, loopback, mtu=32768, vlan=0, macaddr=0:0:0:0:0:0 netname=localhost

              cge0 protocol=IP device=cge0

              inet=172.24.102.238 netmask=255.255.255.0 broadcast=172.24.102.255

              UP, ethernet, mtu=1500, vlan=0, macaddr=0:60:16:4:29:87

              el31 protocol=IP device=cge6

              inet=128.221.253.2 netmask=255.255.255.0 broadcast=128.221.253.255

              UP, ethernet, mtu=1500, vlan=0, macaddr=0:60:16:4:11:a6 netname=localhost

              el30 protocol=IP device=fxp0

              inet=128.221.252.2 netmask=255.255.255.0 broadcast=128.221.252.255

              UP, ethernet, mtu=1500, vlan=0, macaddr=8:0:1b:43:7e:b8 netname=localhost

7.   Looking for an interface up/down in the log of Data Mover, type: 

$server_log server_x | egrep trk

8.   Use ping command to check connectivity, type:

           $ server_ping server_2 172.24.102.2

              server_2 : 172.24.102.2 is alive, time= 0 ms

9.   Check the fail-safe network status, type:

           $ server_sysconfig server_2 -v -i fsn_test

              server_2 :

              *** FSN fsn_test: Link is Up***

              active=cge-2-1 primary=cge-2-1 standby=cge-2-0

                                                                                                                                              

Author: Jeffey Liu