Alex Zhang

> I added an optional parameter to configure extra initContainers > > > > the purpose of this option is to give user an option to do more customization, in...

The PR is still on the going with some conflicts.

> > > 我们也发现了这个问题，之前以为是代码的问题，但是发现nginx进程实际占用的内存是非常小的，后面排查是因为写日志文件accesss.log/error.log过大，然后导致page cache会很大，因为k8s的内存统计是包括rss + buffer/cache的。 在执行执行echo "" > access.log之后发现内存是直线下降的。不知道这个问题是不是也是这个原因导致的。 > > >  > > > 解决办法：目前是通过log-rotate插件进行日志的滚动切割定期删除历史文件。 > > > > > > 我们这边没有开文件日志，用的kafka-logger插件，也有泄漏问题。 > > 这个问题，有解决吗？ This...

Hi, When handling the chunked response. `r:iter_content` doesn't return content according to the chunk boundary. Instead, it tries to read data until the size is equal to the one you...

I'm afraid this failure is not related to APISIX. It should be an exception of the bitnami/etcd (charts or image).

@rthamrin Did you solve this? Did you check the so file mounting in the container?

> Finally, I found cause that I can't bind ports 80 and 443. Release 3.1.0 is the first release that contain this change, right? This affect helm chart too. Have...

Signal 9 means the worker process was killed with `SIGKILL`. Could you checkout the `dmesg` output? Maybe this process was killed by the OOM killer.

> @tokers do you agree with @spacewander ? > > ``` > Since the original code doesn't use buffer (it flushes the buffer immediately), I think we can focus on...

> > @tokers do you agree with @spacewander ? > > ``` > > Since the original code doesn't use buffer (it flushes the buffer immediately), I think we can...