Forum Discussion
NimbostratusFast failover in GTM.
Hello,
I have a wideip config and its using bigip monitor and default ttl 30.
We need fast switching to second pool member. Any idea to implement fast switching between pool members.
4 Replies
Cory_50405Noctilucent
To detect failed pool members faster, build a custom monitor and set the interval and timeout values very low to detect failures more quickly. Setting an interval of 5 seconds and a timeout of 16 will allow three failed health checks before a pool member is marked down which should give you some robustness against network congestion, but will ensure unreachable pool members will be removed in under 20 seconds.
The numbers can be tailored to whatever works best in your environment. My numbers were arbitrarily chosen.
- Cory_50405Looking at your config again, your two GTM pool members are LTMs I take it based on the bigip monitor? The health checking from your LTMs to the destination servers may be the best place to modify your monitors in this case.
sinu_121261Hello,
Thanks cory for the update, yes we are using ltms under GTM sync group and using bigip iquiry monitors. I have already created custom monitors in LTM.
.
I am checking the possibility for fast switching in GTM end also using bigip monitor, what about ttl change?
- Cory_50405Okay, so this LTM monitor is what is going to ultimately mark your members up or down. Your LTM will inform your GTM right away when pool member status changes. If you want to adjust the timing at which failed members are detected, lowering the values on your LTM monitor will do it. Adjusting DNS TTL could have an impact, though it would depend on how many nodes sit behind each LTM and how often they have issues.
