Forum Discussion

MichaelM's avatar
MichaelM
Icon for Altostratus rankAltostratus
Jan 22, 2016

How to verify GTM VS failover was successful?

Hi all,

I was wondering if you could give me an answer (or a hint) to a F5 GTM related question.

We use a GTM device to switch between two virtual servers as automated failover of our inbound networks (Primary and Backup). The Wide IP pool consists of two virtual servers with different IPs (related to network - Primary or Backup). On the LTM these two virtual server use the same pool to the backend servers.

The GTM monitors the availability of the Primary (inbound) network by pinging a specific router. If the Primary network becomes unavailable, the Backup (inbound) network will be used. GTM switches from Primary-VS to Backup-VS.

Last night the Primary network was down for 10 minutes (due to planned work on the Primary router). The GTM reported Primary-VS down and "should have used" the Backup network connection incl. Backup-VS.

Jan 11 02:04:53 myGTM01 alert gtmd[32568]: 011ae0f2:1: Monitor instance /Common/bigip_link 192.168.1.41:0 UP --> DOWN from 192.168.1.46 (DOWN)
Jan 11 02:04:53 myGTM01 alert gtmd[32568]: 011a2003:1: SNMP_TRAP: Link /Common/Primary-uplink (ip=192.168.1.41) state change green --> red ( Monitor /Common/bigip_link from 192.168.1.46 : DOWN)
Jan 11 02:04:53 myGTM01 alert gtmd[32568]: 011a6006:1: SNMP_TRAP: VS prod-Primary (ip:port=10.10.1.29:80) (Server /Common/dc1-ltm) state change green --> red ( Inherited red from parent link /Common/Primary-uplink.)
Jan 11 02:04:53 myGTM01 alert gtmd[32568]: 011a4003:1: SNMP_TRAP: Pool /Common/prod-pool member prod-Primary (ip:port=10.10.1.29:80) state change green --> red ( Inherited red from parent link /Common/Primary-uplink.)
Jan 11 02:14:43 myGTM01 alert gtmd[32568]: 011ae0f2:1: Monitor instance /Common/bigip_link 192.168.1.41:0 DOWN --> UP from 192.168.1.46 (UP)
Jan 11 02:14:43 myGTM01 alert gtmd[32568]: 011a2002:1: SNMP_TRAP: Link /Common/Primary-uplink (ip=192.168.1.41) state change red --> green
Jan 11 02:14:43 myGTM01 alert gtmd[32568]: 011a6005:1: SNMP_TRAP: VS prod-Primary (ip:port=10.10.1.29:80) (Server /Common/dc1-ltm) state change red --> green
Jan 11 02:14:43 myGTM01 alert gtmd[32568]: 011a4002:1: SNMP_TRAP: Pool /Common/prod-pool member prod-Primary (ip:port=10.10.1.29:80) state change red --> green

Note:
192.168.1.41 = router in Primary (inbound) network
192.168.1.46 = F5 GTM Self-IP (non-floating)
  10.10.1.29 = F5 LTM Primary-VS

How can I check/verify that this really happened and the solution works?

I can’t find any evidence that the switch happened and Backup-VS was used. Nothing in gtm/ltm log on the GTM device or ltm log on the LTM device. I can see in the VS stats on the LTM that there was traffic on the Backup-VS but can’t tell when.

Any hint is much appreciated. cheers

application delivery
BIG-IP
BIG-IP DNS

1 Reply

Sort By
  • IanB's avatar
    IanB
    Icon for Employee rankEmployee
    Jan 27, 2016

    What you're asking for is data about what answers were given in response to DNS queries. In GTM 11.5.x and later, there's logging you can enable logging on a listener, or record the decision log on a wide-ip basis, which will write data to the logs for every query coming in.

     

    Failing that, you could create some sort of monitoring script on your network monitoring server, to periodically make a request to the GTM for that wideip, and record the result.

     

    You could also look at the traffic patterns that hit the pool members themselves, though this would only be useful if the backup VS doesn't normally receive any traffic.