ATS transmit timeout often when using cluster pattern

[咸鱼 <hd...@gmail.com>]

Hi,

I was trying to test cluster pattern of ats and I use two servers to build
ats cluster. Refer to the official website, ats  configuration was  in
records.config as follows:

LOCAL proxy.local.cluster.type INT 1
CONFIG proxy.config.cluster.ethernet_interface STRING Bond0
CONFIG proxy.config.cluster.rsport INT 8088
CONFIG proxy.config.cluster.threads INT 10
CONFIG proxy.config.http.cache.cluster_cache_local INT 0

I respectively using jtest on two servers as the source station and client.
When I saw the ats log information found a lot of timeout message , and the
length of timeout was  in accordance with another configuration in
records.config as "proxy.config.http.transaction_active_timeout_in", also
the http stutas was 000.
At first I thought that the problem was related to the configuration of
jtest, so I tried a variety of configuration combinations, but the problem
still occurs frequently.
In desperation I turned off a ats server and test again. Amazing is that
the problem goes away
What could I be doing something wrong with the configuration of cluster?
Thanks.
柴云成 Yuncheng Chai

Mail：hddygs@gmail.com