3 Replies Latest reply on Dec 29, 2009 3:57 PM by tessebie

    Failover assistance, please!

      I am trying to configure fail-over with a pair of SG560's (v3.1.5u1 and v3.1.3, respectively).

       

      I configured the 4-port switch on each router with a separate a port for failover.

      I then set the port (A2) with a LAN IP address and the gateway of the other SG560.

       

      Failover is configured to route out through A2 upon failure.

      Howver, in the failover condition, network traffic gets to the 2nd SG560 but goes no further.

      Am I missing something?

       

      I even tried connecting the primary Internet connection (T-1) to port B and the secondary Internet connection (Comcast) in A2 (with the appropriate IP/Gateway/DNS).

      But the A2 port still doesn't route the traffic when port B is down/disconnected.

       

      Is there something else (routing?) that has to be configured, too?

      Any assistance will be appreciated!

       

      Thanks.

       

      John -Z-

       

       

      Message was edited by: JohnZ on 12/29/09 12:44:34 PM CST
        • 1. Re: Failover assistance, please!

          make sure you have one set to Primary and one set to Secondary in the failover page.

          • 2. Re: Failover assistance, please!

            Mike, Thanks for the response.

             

            We did set one one as primary and the other as secondary.

            Both ports were marked as ENABLED.  Should the primary be REQUIRED?

             

            Thanks!

            • 3. Re: Failover assistance, please!

              > ... v3.1.5u1 / v3.1.3 ...

              Its a good idea to run the latest firmware. Perhaps 4.0.5 is too big a step - but upgrading to 3.2.2

              on both of these is a no-brainer, particularly since there were a number of HA/failover/highavaild

              related fixes between 3.1.5u1 and 3.2.2.

               

              > ... network traffic gets to the 2nd SG560 ....

              presumably you are using tcpdump or the equivalent UI tool to see this happening.

              check the src/dst addr of the packets in question and make sure routing/NAT is set

              up on the 2nd 560 to deal with it. eg. return traffic could be going into a bit-bucket, or

              dst-addr could be being directed to the wrong place. There is a route-table under

              'Diagnostics'. or you can ssh into the cli and type 'route'.

               

              If a 'tcpdump -ni any' doesn't show the packet(s) leaving to anywhere, they are probably

              being dropped by the firewall. Worth comparing 'known good traffic going through 2nd 560'

              tcpdump -ni any behavior. Use a port restrictor to eliminate noise. eg.

                'tcpdump -ni any icmp' or 'tcpdump -ni any port 23'

              then generate test-traffic appropriate for the filter (eg. ping, or telnet some-internal-site

              All these can work from the 'Diagnostics -> Packet Capture' window as well, I just find

              the CLI tcpdump command quicker/easier once I get stuck into a debug cycle.

               

              Check syslog for any firewall rule dropping packets unexpectedly, and/or custom firewall

              rule counters (can be hard to see if other network traffic is going on).

               

              That's probably enough to establish a baseline and eliminate a bunch of potential issues.

               

              Cheers

              tom