Forum Discussion

fat_138651's avatar
fat_138651
Icon for Nimbostratus rankNimbostratus
Oct 20, 2014

create new device group: which device will become standby?

Hi,

 

I'm currently setting up HA for two Viprion-Guests running 11.4.1 HF3. The devices are "In Sync" already and the next step would be to set up a device group for sync-failover.

 

My problem is, that my first device is already running with clients connected to several virtual servers. The second device has only basic configuration. If I configure a device group, one device will switch to standby mode.

 

My question is: Which device?

 

I would like to keep the first device active, so that the clients are not affected. After that I would sync the config from device one to the new device group.

 

How can i make sure, that the second device will become standby?

 

Thanks in advance

 

Kind Regards,

 

Thorsten

 

13 Replies

  • Device Management > Traffic Groups > Your-Traffic-Group > Default Device

    A default device is the device on which you want a traffic group to run if possible. You actively specify a default device for a traffic group when you create the traffic group. When a traffic group floats due to failover, the BIG-IP system selects the default device as the target failover device whenever that device has the same or more availability as other group members to process traffic. For example, suppose that during traffic group creation you designated Device B to be the default device. If failover occurs and Device B and Device C have the same availability, the traffic group will fail over to Device B. The default device designation is a user-modifiable property of a traffic group.
    

    For additional security you may disable the uplink network interface on the node you wish to become standby. Don't forget to re-enable it once "Traffic groups" and "Device Groups" sections are configured.

    • nathe's avatar
      nathe
      Icon for Cirrocumulus rankCirrocumulus
      Hannes - is this feature (Default Device) still available in 11.4.1? I can see it's an option in 11.2 but not on my lab 11.4.1 device. Thanks
    • NikhilB's avatar
      NikhilB
      Icon for Employee rankEmployee
      Just to clarify the posting you mentioned; if TG1 is active on device A (and A is also the default device for TG1): if device A reboots and causes a failover to its peer (B), the TG1 will remain on device B unless auto failback is ofcourse configured. Do you see under any circumstances where it would failback to A after a failover event? (minus auto failback)
  • Device Management > Traffic Groups > Your-Traffic-Group > Default Device

    A default device is the device on which you want a traffic group to run if possible. You actively specify a default device for a traffic group when you create the traffic group. When a traffic group floats due to failover, the BIG-IP system selects the default device as the target failover device whenever that device has the same or more availability as other group members to process traffic. For example, suppose that during traffic group creation you designated Device B to be the default device. If failover occurs and Device B and Device C have the same availability, the traffic group will fail over to Device B. The default device designation is a user-modifiable property of a traffic group.
    

    For additional security you may disable the uplink network interface on the node you wish to become standby. Don't forget to re-enable it once "Traffic groups" and "Device Groups" sections are configured.

    • nathe's avatar
      nathe
      Icon for Cirrocumulus rankCirrocumulus
      Hannes - is this feature (Default Device) still available in 11.4.1? I can see it's an option in 11.2 but not on my lab 11.4.1 device. Thanks
    • NikhilB's avatar
      NikhilB
      Icon for Employee rankEmployee
      Just to clarify the posting you mentioned; if TG1 is active on device A (and A is also the default device for TG1): if device A reboots and causes a failover to its peer (B), the TG1 will remain on device B unless auto failback is ofcourse configured. Do you see under any circumstances where it would failback to A after a failover event? (minus auto failback)
  • nathe's avatar
    nathe
    Icon for Cirrocumulus rankCirrocumulus

    Thorsten,

     

    Which device? Well, I don't know specifically but recently newer 11.x TMOS versions the BIG-IP that I've configured the Device Group on has become the Standby and the other one becomes active. Which is odd. I'm sure that's changed from one version to another one.

     

    One option could be to ensure all interfaces are disabled (can you do this on a vCMP guest?) and only leave the one with the configsync configured so you can "see" the other guest.

     

    Hope this helps,

     

    N

     

  • Put the standby device into Offline mode prior to doing the peering. Once the devices are peered, release Offline mode. Make sure you do the config sync in the correct direction ( and take a UCS backup before doing anything )

     

  • Hi Fat, The device on which you will have the latest changes will be the one that will acquire the active state and the other will be standby. So make sure the Unit which you want to be up is the one on which you have the latest config done. This way you don't have to worry that it will switch to the standby.

     

    By the way you have a funny name here... "Fat" ... :D ... Just confirm it worked for u ..

     

    Regards,

     

  • I just set up two guests for testing, and it seems to work without problems performing the following steps

     

    • state Active/Active In sync
    • Force the designated standby device offline (Active/Forced offline)
    • create device group on the active-device (changes pending/Forced offline)
    • release offline (wait for some seconds, then Awaiting initial sync)
    • sync Active to standby

    I will let you know how this steps work on my productive system.

     

    Regarding my nickname: My company used to form logins from the first two letters of the surname and the first letter of the first name...in my case that's fat. As I have no weight-problems that's no problem for me. :D

     

    • PeteWhite's avatar
      PeteWhite
      Icon for Employee rankEmployee
      great news, I have done it myself as well and it seems to work just fine. Offline mode survives reboots as well so well worth doing if you're going to have to put the device into a separated mode to avoid split-brain.
  • So, I did the changes on my Main-System and it worked fine. I lost about 8 ping-sequences while creating the device-group, but that's nothing to worry about.

     

    Thanks a lot guys

     

    • fat_138651's avatar
      fat_138651
      Icon for Nimbostratus rankNimbostratus
      Hmmm...it seems that I accientally created two devcentral accounts on my different clients.... I will fix that ;)