When I am testing app failover its taking >15 min before the system responds that it has failed over to the active server.
I cannot find anything in the logs and maybe the webservers are the culprit where its lagging on failing over. I did a wireshark trace which shows the failover happening in 14 min. Would https have anything to do with it being slow?
4.2 failover in about 3 minutes.
I think there is missing info how is done the app failover? by what product - tomcat cluster or any loadbalancer?
4.3 SP2 Patch 11 Loadbalancer is F5 and tomcat cluster is running on Oracle Linux
I dont know the F5 - but another one where are defined a service monitors - what to check, how many times, what should be the answer etc. and by combination of those failed monitors the app failover is set to switch to another node…
So I would start to check the settings for that app in F5…
e.g. if the monitor check just a response http/200 OK - then even the SIA is dead Tomcat still returns 200. then I would define another monitor to check e.g. status of the SIA service and bind it to the app…
regarding your question - I am not able to simply answer - perhaps as the BOBJ becomes more and more complex the shutdown process takes longer and longer… so the time when https returns something else than 200 - (404 or so…) is longer as well