|
Hi,
We are using LVS in LVS-TUN mode to load balance between six Apache
servers. Our performance tests indicated that the system could manage
55000 concurrent users, but in production response times are getting
very high at 25000 concurrent users. CPU load and network traffic are
getting nowhere near the values observed during performance tests, which
makes me wonder if the LVS (or possibly some network component) has a
limit regarding the number of IP addresses served. Allthough we did
maintain a greater number of connections and a larger network load
during the tests, we did not generate traffic from 25000 different IP
addresses.
My question is: What is the largest number of concurrent users you have
successfully served through LVS? What settings should be checked with
respect to this?
Regards,
Eirik Gjesteland
This message is for the designated recipient only and may contain privileged,
proprietary, or otherwise private information. If you have received it in
error, please notify the sender immediately and delete the original. Any other
use of the email by you is prohibited.
|
