NimbostratusOneConnect Profile behaviour between TMOS versions
I am troubleshooing the behaviour of OneConnect profile and see some differences in two different TMOS versions we use.
On 11.5.4 HF2, it works perfectly during testing. Serverside connections are maintained and idle connections are reused.
Ltm::OneConnect Profile: pr_oneconnect_aap1
Virtual Server Name N/A
Connections Current Idle 3 Maximum 6 Total Reuses 20.9K New 574
On 12.1.3.3, connections are closed immediately after each request is completed. A new serverside connection is created/closed for each request.
Ltm::OneConnect Profile: pr_oneconnect_app1
Virtual Server Name N/A
Connections Current Idle 0 Maximum 3 Total Reuses 4 New 6.5K
The end servers in both environments are exactly the same and respond in the same way to each request so it is hard to point to a misconfiguration on the server-side. On the other hand, I have compared configuration on both LTM instances and dont see any obvious differences. Are there any settings maybe in HTTP profile or TCP profile that I need to pay attention to?
Anyone else ever had a similar experience setting up OneConnect profile with either version.
In troubleshooting , I've noticed the server response does not include a Connection: keep-alive header although they are part of the client request, but this does not seem to be a problem for the 11.5.4 HF2 instance. Perhaps, that header is strictly required in future TMOS versions like 12.1.3.3?
