Forum Discussion

smp_86112's avatar
smp_86112
Icon for Cirrostratus rankCirrostratus
Sep 18, 2014

Immediate failback

I'm new to v11, and having trouble with the failover process on 11.4.1. I have two members (u1 and u2) in a Sync-Failover group. When I force a failover from u2->u1, u1 reports Active briefly, then it immediately fails back. This process generates the following sequence log entries in /var/log/ltm on u1:

 

Sep 18 09:07:16 u1 notice sod[6926]: 010c0019:5: Active

 

Sep 18 09:07:17 u1 err sod[6926]: 010c006a:3: Configuration CRC values disagree amongst peers. Suggest configsync peers.

 

Sep 18 09:07:17 u1 notice logger: /usr/bin/tmipsecd --tmmcount 4 ==> /usr/bin/bigstart start racoon

 

Sep 18 09:07:19 u1 notice sod[6926]: 010c004a:5: Leaving active in favor of active peer.

 

Sep 18 09:07:19 u1 notice sod[6926]: 010c0052:5: Standby for traffic group /Common/traffic-group-1.

 

Sep 18 09:07:19 u1 notice sod[6926]: 010c0018:5: Standby

 

This problem started happening after bringing them both up from a graceful shutdown for power maintenance overnight. The only thing that looks slightly "wrong" is the "Next Active Device" property for the (single) traffic group on both units has a status of "Initializing / Not Synced". Seems like it could be related, but I know this was the same status before the power down, when we had no failover issues.

 

Not being familiar with v11, I'm uncertain where to focus and I haven't found any relevant articles in the KB. Has anyone seen this before?

 

application delivery
availability
LTM
management

4 Replies

Sort By
  • nathe's avatar
    nathe
    Icon for Cirrocumulus rankCirrocumulus
    Sep 18, 2014

    smp

     

    Do you have either Auto Failover on the traffic group enabled, or have you got fast failover/ HA group setup? See System - High Availability - HA group.

     

    N

     

  • afedden_1985's avatar
    afedden_1985
    Icon for Cirrus rankCirrus
    Sep 18, 2014

    we had the same issue after 11.2 to 11.5 upgrade we had to change the Failover Method from HA Order to Load Aware for us it was located Device Management ›› Traffic Groups ›› traffic-group-1 we had to set the Failover Method to Load Aware on both units

     

  • smp_86112's avatar
    smp_86112
    Icon for Cirrostratus rankCirrostratus
    Sep 22, 2014

    Turns out we made the rookie mistake of not plugging in the failover cable after one of the units was recently RMA'd. Now that the units are properly connected with a failover cable, they behave as expect and do not fail back automatically.

     

    Thanks for the assistance.

     

    • nathe's avatar
      nathe
      Icon for Cirrocumulus rankCirrocumulus
      Sep 22, 2014
      Good news. Always the simple things eh!