Forum Discussion

Dazzla_20011's avatar
Dazzla_20011
Icon for Nimbostratus rankNimbostratus
Nov 29, 2012

Changing Management IP of Active BIG-IP caused both Big-IPs to become Ative

Hi,

 

 

I need to change the management ip of our Active LTM which I tried this evening but this caused both LTMs to become active so I reversed the change which fixed it but not sure why this happened.

 

Each LTM is in a different DC connected by dark fibre and we have a dedicated interfaces for the heartbeart link. I thought the management interface was as it says just used for management but it seems not. When both went Active obviously not good and caused outages to services.

 

 

Any ideas?

 

 

Thanks

 

Darren

 

8 Replies

  • did you also change peer mgmt address under network failover setting?
  • Hi,

     

    Thanks, yes the Standby LTM has the peer management address set to the managment IP address of the Active LTM but the Active LTM has the peer management address set to the ip address of the dedicated interface we use for the network failover.

     

    The settings are as follows.

     

    ACTIVE LTM

     

    Peer Management Address: 192.168.254.3 ( (Dedicated hearbeat interface of Standby LTM)

     

    Configuration Identifier: f5-int-ltm-sync-unicast

     

    Local Address: 192.168.254.2 (Dedicated hearbeat interface)

     

    Remote Address: 192.168.254.3 (Dedicated hearbeat interface

     

    Port: 1026

     

    STANDBY LTM

     

    Peer Management Address: 10.128.81.43 (Management IP Address of Active LTM)

     

    Configuration Identifier: f5-int-ltm-sync-unicast

     

    Local Address: 192.168.254.3 (Dedicated hearbeat interface)

     

    Remote Address: 192.168.254.2 (Dedicated hearbeat interface)

     

    Port: 1026

     

    So there appears to be some difference in the configuration of each LTM. Would the preference be to change the Peer Management of the Standby LTM to the Dedicated hearbeat interface of the Active LTM and when i do this should I force the Standby LTM offline while I make these changes? What I would also like to do is add the management interfaces as a secondary interface for network failover monitoring so we have two just so any problem affecting the decicated hearbeat interfaces doesn't cause a failover.

     

    Does this sound OK?

     

    Thanks for your help

     

     

     

  • the Active LTM has the peer management address set to the ip address of the dedicated interface we use for the network failover.i understand it must be peer mgmt address (not peer tmm address).

     

     

    What I would also like to do is add the management interfaces as a secondary interface for network failover monitoring this is the best practice when configuring network failover.
  • Hi,

     

     

    When you refer to the peer tmm address do you mean the heartbeat interface we have configured?

     

     

    Thanks

     

    Darren
  • He does. A TMM interface is any interface that isn't the management interface (i.e. a switch port) and thus a TMM IP address is any IP address not assigned to the management interface (i.e. configured under Network not Platform). Apologies if this sounds a bit snotty, I can't think of a better way of putting it without losing the detail. =]

     

  • When you refer to the peer tmm address do you mean the heartbeat interface we have configured? i mean this configuration.

     

     

    Peer Management Address: 192.168.254.3 ( (Dedicated hearbeat interface of Standby LTM)

     

     

    i think it must be mgmt address (not tmm/selfip address) of peer unit.
  •  

    I need to configure the management IP of the Active LTM to 10.136.64.24 and then change the peer management address on the Active LTM to the management address of the Standby LTM which is 10.184.128.6. What is the best way to do this to ensure both LTM's don't become Active again. Should I force the Standby LTM offline while I make the changes?

     

     

    Many Thanks
  • Should I force the Standby LTM offline while I make the changes?i never tested but i think it could help.

     

     

    another way on top of my head is to disconnect one unit from network (not sure if it can be called a proper way :p).