Forum Discussion
NimbostratusRDP Sessions exceeding limit through F5.
Hi I have VIP over the RDP port 3389 to accept RDP connections for 15 terminal servers in backed. In configuration we have connection limit set to 60 for each server accordingly we have limit of 900 total connections on F5.
We are facing issue regarding to maximum connections suddenly getting reached and after that nothing accepted by F5 as which is correct as per design what we have set.
My concern is how to identify what are those sessions affecting this ?Also we have identified duplicate sessions on server from same client IP associated with different servers what is the reason for that.
Do we need to enable persistence profile ? that only one option i am able to see now. Anyone have any thoughts why this duplication of sessions happening on server through F5 ?
RDP session requires connection persistence unless the RDP servers are clustered and mirror session information to one-another. Depending on the exact setup, BigIP may have initiated 1 connection to each server while just one connection carries meaningful RDP session information to a single server. Can you check connection tables or do a measured test during off-peak hours to see how many pool member counters increase when you initiate a RDP session?
For fix. Give a try with following settings and see if the problem persist. Source IP persistence profile (Virtual Server setting), Least Sessions load-balancing algorithm (Pool setting). Source IP persistence can probably be replaced with something better but for initial test it's good enough. As you apply changes, old connections should be killed to see immediate effect, otherwise they will continue to function with previous configuration until timed out. To kill all connections to a particular Virtual Server, use
tmsh delete /sys connection cs-server-addr VS.IP.AD.DR cs-server-port 3389
Hannes_RappRDP session requires connection persistence unless the RDP servers are clustered and mirror session information to one-another. Depending on the exact setup, BigIP may have initiated 1 connection to each server while just one connection carries meaningful RDP session information to a single server. Can you check connection tables or do a measured test during off-peak hours to see how many pool member counters increase when you initiate a RDP session?
For fix. Give a try with following settings and see if the problem persist. Source IP persistence profile (Virtual Server setting), Least Sessions load-balancing algorithm (Pool setting). Source IP persistence can probably be replaced with something better but for initial test it's good enough. As you apply changes, old connections should be killed to see immediate effect, otherwise they will continue to function with previous configuration until timed out. To kill all connections to a particular Virtual Server, use
tmsh delete /sys connection cs-server-addr VS.IP.AD.DR cs-server-port 3389
Gunavant_Patil_Hi Rapp first of all Happy new your to you !!!
Thanks for all explanation,Agreed on persistence profile configuration and least session load-balancing method configuration suggestion.
And answer to your first question i already did close monitoring and seen that Round-Robin working good for initial few connections but after connections start increasing rapidly its seems not working as expected and few servers not seems accepting connections.
Hannes_Rapp_162Nacreous
RDP session requires connection persistence unless the RDP servers are clustered and mirror session information to one-another. Depending on the exact setup, BigIP may have initiated 1 connection to each server while just one connection carries meaningful RDP session information to a single server. Can you check connection tables or do a measured test during off-peak hours to see how many pool member counters increase when you initiate a RDP session?
For fix. Give a try with following settings and see if the problem persist. Source IP persistence profile (Virtual Server setting), Least Sessions load-balancing algorithm (Pool setting). Source IP persistence can probably be replaced with something better but for initial test it's good enough. As you apply changes, old connections should be killed to see immediate effect, otherwise they will continue to function with previous configuration until timed out. To kill all connections to a particular Virtual Server, use
tmsh delete /sys connection cs-server-addr VS.IP.AD.DR cs-server-port 3389- Gunavant_Patil_
Hi Rapp first of all Happy new your to you !!!
Thanks for all explanation,Agreed on persistence profile configuration and least session load-balancing method configuration suggestion.
And answer to your first question i already did close monitoring and seen that Round-Robin working good for initial few connections but after connections start increasing rapidly its seems not working as expected and few servers not seems accepting connections.