cancel
Showing results for 
Search instead for 
Did you mean: 

lltstat reporting errors and duplicate MAC addresses?

dbsanders
Level 3
I have a new two node cluster, and lltstat -nvv shows duplicate mac addresses. How can this happen? We have all heartbeats going to two separate switches, but all in the same VLAN. Other clusters seem happy with this configuration. If I plumb the interfaces in Solaris and look at the MAC addresses, they are unique as expected, because use-local-mac-address? is set to true.

lltstat -nvv from both cluster members:

LLT node information:
    Node                 State    Link  Status  Address
     0 buton             OPEN
                                  ce2   UP      00:14:4F:68:4D:16
                                  ce6   UP      00:14:4F:68:4D:16
                                  ce0   UP      00:14:4F:68:4D:64
   * 1 strommer          OPEN
                                  ce2   UP      00:14:4F:68:4D:0A
                                  ce6   UP      00:14:4F:68:4E:26
                                  ce0   UP      00:14:4F:68:4D:08

LLT node information:
    Node                 State    Link  Status  Address
   * 0 buton            OPEN
                                  ce2   UP      00:14:4F:68:4D:66
                                  ce6   UP      00:14:4F:68:4D:16
                                  ce0   UP      00:14:4F:68:4D:64
     1 strommer          OPEN
                                  ce2   UP      00:14:4F:68:4E:26
                                  ce6   UP      00:14:4F:68:4E:26
                                  ce0   UP      00:14:4F:68:4D:08

buton ifconfig:
ce2: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 6
        inet 0.0.0.0 netmask 0
        ether 0:14:4f:68:4d:66
ce6: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 7
        inet 0.0.0.0 netmask 0
        ether 0:14:4f:68:4d:16

strommer ifconfig:

ce2: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 7
        inet 0.0.0.0 netmask 0
        ether 0:14:4f:68:4d:a
ce6: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 6
        inet 0.0.0.0 netmask 0
        ether 0:14:4f:68:4e:26


Error message (only appears on strommer /var/adm/messages)
Dec 12 12:04:31 strommer llt: [ID 525299 kern.notice] LLT WARNING V-14-1-10497 crossed links? link 0 and link 1 of node 0 on the same network



Message Edited by dbsanders on 12-12-2007 12:06 PM
5 REPLIES 5

RiaanBadenhorst
Moderator
Moderator
Partner    VIP    Accredited Certified
can you post your llttab files?

dbsanders
Level 3
Sure. From strommer:
 
set-node strommer
set-cluster 91
link ce2 /dev/ce:2 - ether - -
link ce6 /dev/ce:6 - ether - -
link-lowpri ce0 /dev/ce:0 - ether - -

 
and from buton:
 
set-node buton
set-cluster 91
link ce2 /dev/ce:2 - ether - -
link ce6 /dev/ce:6 - ether - -
link-lowpri ce0 /dev/ce:0 - ether - -


Message Edited by dbsanders on 12-12-2007 01:03 PM

Message Edited by dbsanders on 12-12-2007 01:04 PM

RiaanBadenhorst
Moderator
Moderator
Partner    VIP    Accredited Certified
nothing wrong there. Sorry, cant think of anything else.

Jim_Munroe
Not applicable
Is ce2 on node 0 on the same or different  switch as ce2 on node 1?   We had a similar issue and it turned out to be a cabling issue.  I would double check your connections.

dbsanders
Level 3
I believe this is the issue. Some heartbeats are sharing a switch and are seeing each other in the same VLAN. Although I cannot find any documentation from Symantec that specifically requires separate VLANs, it is certainly suggested. We are creating a new VLAN so the two heartbeats will be more isolated from each other.