Forum Discussion
NimbostratusIssue in Failover of pool member
Scenario
2 pool members A & B behind ltm. Load balancing method .. Round robin No persistence Monitoring Probe used tcp_8443 Interval set to 2 sec Timeout 7 sec
Issue... Client experiences delay of 10 sec while initiation of Failover of URL from pool member A to pool member B
Could anyone advise how could this delay be reduced?Thanks!
RaghavendraSYAltostratus
Hi,
You configured two poole members with round robin method. First request will go to A and second request will go to B. I suggest you configure persistence based on your application requriement.
Your application is using any cookie method ? if yes you can go with Cookie persistence.
Samit_jadhavHi Raghvendra, Application isn't using any cookie persistence.
If persistence isn't set then what will be default persistence value?
When the Failover of poolmember takes place , do I need to try to open url his different pc ?
RaghavendraSY_7Cumulonimbus
Hi,
You configured two poole members with round robin method. First request will go to A and second request will go to B. I suggest you configure persistence based on your application requriement.
Your application is using any cookie method ? if yes you can go with Cookie persistence.
- Samit_jadhav
Hi Raghvendra, Application isn't using any cookie persistence.
If persistence isn't set then what will be default persistence value?
When the Failover of poolmember takes place , do I need to try to open url his different pc ?
- RaghavendraSY_7
If cookie persistence is not there, then go with source based persistence with default Timeut value (300seconds). Please make sure your TCP profile timout and persistence timout has same values.
- Samit_jadhav
But if the source based persistence is used then it may take 5 min to hit the client traffic on next available poolmember when the 1st pool member goes down and the requirement is clients traffic should switchover to next available pool member within 4 sec whenever the 1st pool member goes down...
- RaghavendraSY_7
If your requirement is failover traffic to pool member B only if pool member A goes down? If Yes to the question then you need to go with Ratio load balance method and no need to use persistence.
Ration method:
Pool A --100% Pool B --0%
Hope this helps.
- Samit_jadhav
Thanks Raghvendra, but after setting lb method to ratio , will the Failover take place in 4 sec
anyways I'll set probe interval timings to 1 sec and probe timeout value to 4 sec , will this help me in achieving my target of Failover in 4 sec?
- RaghavendraSY
If cookie persistence is not there, then go with source based persistence with default Timeut value (300seconds). Please make sure your TCP profile timout and persistence timout has same values.
- Samit_jadhav
But if the source based persistence is used then it may take 5 min to hit the client traffic on next available poolmember when the 1st pool member goes down and the requirement is clients traffic should switchover to next available pool member within 4 sec whenever the 1st pool member goes down...
- RaghavendraSY
If your requirement is failover traffic to pool member B only if pool member A goes down? If Yes to the question then you need to go with Ratio load balance method and no need to use persistence.
Ration method:
Pool A --100% Pool B --0%
Hope this helps.
- Samit_jadhav
Thanks Raghvendra, but after setting lb method to ratio , will the Failover take place in 4 sec
anyways I'll set probe interval timings to 1 sec and probe timeout value to 4 sec , will this help me in achieving my target of Failover in 4 sec?