Forum Discussion

danielpenna's avatar
danielpenna
Icon for Cirrus rankCirrus
Apr 30, 2015

iRule to delay traffic going to pool2 after pool1

Hi Team, Have been asked to provide a solution to assist with a DR scenario that I am scratching my head on.

 

The scenario is that in a DR event, a team needs 10 minutes to get a backend database cutover to DR and they want the F5 to hold the failover event for 10 minutes before using the DR server.

 

So in normal day, traffic all goes to pool1. pool1 fails, and before traffic goes to pool2 there is a delay of 10 minutes ( and that delay only occurs once ).

 

I was thinking about setting the healthchecks on the server/pool level for that 10 minute mark, but the problem there is traffic would still be sent to dead primary server versus a nicer reset from the F5.

 

Any thoughts ?

 

application delivery
availability
design
devops
iRules
LTM

2 Replies

Sort By
  • Ed_Summers's avatar
    Ed_Summers
    Icon for Nimbostratus rankNimbostratus
    Apr 30, 2015

    Are you using priority group activation or another method to switch between pools?

     

    Could you simply monitor the status of pool2 using a health monitor that gets an object the pool member would need to 'fetch' from the back-end database server? Assuming that pool2 only has database connectivity after this 10-minute event occurs, the pool would be down until the database team finishes their work (regardless of the time it takes them).

     

    Let me know if I mis-read your scenario. -Ed

     

  • danielpenna's avatar
    danielpenna
    Icon for Cirrus rankCirrus
    May 01, 2015

    Hi Ed, I also contacted my local SE as well and he put forward the scenario of a pool that monitors the backend SQL box to be up before sending traffic to Pool2. Think thats the best option and working on that solution. Sometimes it just requires some lateral thinking :).

     

    Appreciate your help.