Forum Discussion

ShriRam_373553's avatar
ShriRam_373553
Icon for Nimbostratus rankNimbostratus
Oct 03, 2018

F5 Load balancer LVM Monitoring scenario - assistance

Dear Folks,

 

I'm from monitoring background and here to understand the possibility of getting a situation monitored.

 

Consider I have F5 loadbalancer pair (Active-Standby).

 

Appears there exist a SYNC in place between these two devices to ensure they are on same configuration.

 

Let's say

 

X (Active device) Y (standby one)

 

Consider X went on hung state (for any reason) with not letting the monitoring tool understand that it's broken (because the device is recheable as per monitoring tool - since it's hung that the tool did not know)

 

Upon checking we could understand that the SYNC status says "disconnected" too.

 

Questions: We have couple of MIBs loaded in place and actively monitored.

 

-> Would like to know is there a specific OID that shall be monitored for SYNC status (or) Is there a trap that shall be triggered from the device end when the SYNC breaks.

 

-> How F5 loadbalancer hung situation handled as per monitoring standards? Would like to understand in depth on the possible breaks when a device turns hung to ensure we shall enable live monitoring feed for notification.

 

How customers handles such hung state situation for notification - question comes up since the device will not go down according to SNMP manager since it will be reacheable . Only connections drop/threshold drops.

 

Only option to think is to couple connections and thresholds monitoring. Would like to understand this state in depth - like if it could update a log or if it could break updates in a log - or wateva clue shall be of help.

 

I'm fond of some interesting answers to know from this technical specialist group :) Looking forward.

 

Kind Regards, Lakshmi.

 

application delivery
BIG-IP
LTM

5 Replies

Sort By
  • R__Hickman_1172's avatar
    R__Hickman_1172
    Icon for Nimbostratus rankNimbostratus
    Oct 03, 2018

    The sync between the active and the standby is just to sync any changes in configuration between the active and the standby unit. If it shows "disconnected" it can be a couple of reasons and would be independent of the active unit being unavailable. The sync should be configured to use a different network than the fail over network. When there's an issue with the active unit - the F5 will issue a communications error alert via SNMP to whatever SNMP IP's you have configured. I'd be less concerned with the sync in a failover scenario - though if your F5 isn't set to auto sync, there is always a chance in a failover situation the standby unit won't have the right running config.

     

    • ShriRam_373553's avatar
      ShriRam_373553
      Icon for Nimbostratus rankNimbostratus
      Oct 04, 2018

      Hickman,

       

      Thanks , Would you help me with the questions asked for please.

       

      -> Would like to know is there a specific OID that shall be monitored for SYNC status (or) Is there a trap that shall be triggered from the device end when the SYNC breaks.

       

      -> How F5 loadbalancer hung situation handled as per monitoring standards? Would like to understand in depth on the possible breaks when a device turns hung to ensure we shall enable live monitoring feed for notification.

       

      How customers handles such hung state situation for notification - question comes up since the device will not go down according to SNMP manager since it will be reacheable . Only connections drop/threshold drops.

       

      As you havementioned in the post, there should be a trigger if it's hung. Interested to know such trigger.

       

      Let me know if you have any questions on my look out please.

       

      -Lakshmi.

       

    • ShriRam_373553's avatar
      ShriRam_373553
      Icon for Nimbostratus rankNimbostratus
      Oct 05, 2018

      Would be glad if anyone could be of help. This answer will definitely be a PI for any customer to ensure service availability based on what F5 delivers. This approach will definitely help customers achieve dynamic alerts when the device sets HUNG state.

       

      Any advise based on this content will help.

       

      Anyways I'm planning to run snmpwalk through the device to know the OIDs captured/or shall-be in.

       

      Will keep you guys posted from my end too, if I could be of some help.

       

      -Lakshmi