Failed to run atomically, multiple threads updating balancer simultaneously

[zjlooojoe]

https://user-images.githubusercontent.com/41610809/130401736-c7f2c71a-aee3-46f2-aeb8-ee3cfae22edb.png

2021/08/23 06:44:51 [error] 1827#0: *18477198099 lua entry thread aborted: runtime error: /usr/share/lua/5.1/resty/dns/balancer/base.lua:389: Failed to run atomically, multiple threads updating balancer simultaneously stack traceback: coroutine 0: [C]: in function 'assert' /usr/share/lua/5.1/resty/dns/balancer/base.lua:389: in function 'assert_atomicity' /usr/share/lua/5.1/resty/dns/balancer/base.lua:573: in function 'queryDns' /usr/share/lua/5.1/resty/dns/balancer/base.lua:1206: in function 'cb_expire' /usr/share/lua/5.1/resty/timer.lua:87: in function </usr/share/lua/5.1/resty/timer.lua:51>, context: ngx.timer

Has anyone ever encountered this problem?
I am using 1.5.0 kong.It seem report every moment, but kong still work fine.

[zjlooojoe]

In our work, we use kong as k8s ingress, so target's object will change every moment as long as pod is born and dispatched. Is that mean v1.5.0's resty-dns-balancer.lua cannot handle the above-situation because target and balancer is changed frequently? And it will result in failure to create/update the balancer? Create A new balancer may be a best choice, using the balancer:upstreams:xx cluster_events?