To be more accurate, client_idle_timeout_ms settings sometimes works with
grpc version 1.57.0, I tried to use "GRPC_VERBOSITY=debug
GRPC_TRACE=client_idle_filter,channel_stack" to print some logs, below is
the log when it works, and the last line is missing when it didn't work,
seemed like the channel didn't go to idle status correctly.
I1027 11:58:22.187377817 1 channel_stack.cc:116] CHANNEL_STACK: filter
client_idle [promise-capable]
I1027 11:58:22.187806892 1 channel_stack.cc:116] CHANNEL_STACK: filter
client-channel
D1027 11:58:22.238282078 73 rb_channel.c:732] GRPC_RUBY:
run_poll_channels_loop - create connection polling thread
D1027 11:58:22.238531816 73 rb_channel.c:656] GRPC_RUBY:
run_poll_channels_loop_no_gil - begin
I1027 11:58:22.348362088 1 channel_stack.cc:114] CHANNEL_STACK: init
DynamicFilters
I1027 11:58:22.348391301 1 channel_stack.cc:116] CHANNEL_STACK: filter
retry_filter
D1027 11:58:22.371003348 73 tcp_posix.cc:2004] cannot set inq fd=16 errno=92
I1027 11:58:22.819900113 73 channel_stack.cc:114] CHANNEL_STACK: init
subchannel
I1027 11:58:22.822087210 73 channel_stack.cc:116] CHANNEL_STACK: filter
authority [promise-capable]
I1027 11:58:22.822687281 73 channel_stack.cc:116] CHANNEL_STACK: filter
client-auth-filter [promise-capable]
I1027 11:58:22.823268462 73 channel_stack.cc:116] CHANNEL_STACK: filter
message_size [promise-capable]
I1027 11:58:22.823767668 73 channel_stack.cc:116] CHANNEL_STACK: filter
http-client [promise-capable]
I1027 11:58:22.824236760 73 channel_stack.cc:116] CHANNEL_STACK: filter
compression [promise-capable]
I1027 11:58:22.824823023 73 channel_stack.cc:116] CHANNEL_STACK: filter
connected [promise-capable]
*I1027 11:58:22.900787905 1 channel_idle_filter.cc:247] (client idle
filter) timer has started*
On Saturday, October 28, 2023 at 2:25:06 AM UTC+8 anto oant wrote:
>
> by the way, the reason I met this "reset by peer" error is because my
> server is in AWS cloud (don't know why), I don't have this problem when
> running server locally.
> On Saturday, October 28, 2023 at 2:13:01 AM UTC+8 anto oant wrote:
>
>> I am building a ruby GRPC client and for certain reason, I need to close
>> TCP connection after client being inactive(not sending any request) for 5
>> minutes, otherwise, new request send through such connection will end up
>> with "
>> grpc_status:14, grpc_message:"recvmsg:Connection reset by peer" error.
>>
>> With version 1.43.1 x86_64-linux, I ever used channel_args: {
>> "grpc.client_idle_timeout_ms" => 5 * 1000 } to achieve this, but it
>> seems no longer work with version 1.57.0.
>>
>> I wonder whether it is intentional or not, if yes, what should I do to
>> achieve my goal in 1.57.0?
>>
>> any pointer will be really appreciated!
>>
>
--
You received this message because you are subscribed to the Google Groups
"grpc.io" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/grpc-io/83de92e5-38ba-4b83-b1a5-38461893938en%40googlegroups.com.
