net_kernel {error, closed}

Vithya Shanmugam -X (vitshanm - SCARLET WIRELESS INDIA PRIVATE LIMITED at Cisco)

2021-04-20 09:14:12 UTC

Hi All,

Follow-up on the previous question, On Tracing the pid from INFO REPORT,
=INFO REPORT==== 20-Apr-2021::07:35:52.946312 ===
{net_kernel,{'EXIT',<0.20724.1>,{error,closed}}}

(<8970.67.0>) call net_kernel:handle_info({'EXIT',<8970.20724.1>,{error,closed}},{state, ...
(<8970.67.0>) call net_kernel:accept_exit(<8970.20724.1>,...

Tracing the calls, observed that ssl handshake is failing with reason {error,closed}

(<8970.20724.1>) call ssl_connection:handshake({sslsocket,{gen_tcp,#Port<8970.1336>,tls_connection,
[{session_id_tracker,<8970.23171.1>}]},
[<8970.23173.1>,<8970.23172.1>]},7000)
(<8970.20724.1>) returned from ssl_connection:handshake/2 -> {error,closed}
(<8970.20724.1>) returned from ssl_connection:handshake/7 -> {error,closed}
(<8970.20724.1>) returned from ssl:handshake/3 -> {error,closed}
(<8970.20724.1>) call inet_tls_dist:trace({error,closed})
(<8970.20724.1>) returned from inet_tls_dist:trace/1 -> {error,closed}

which then calls 'gen_tcp:close(Socket)' ..... 'prim_inet:close(

(<8970.20724.1>) call gen_tcp:close(#Port<8970.1336>)
(<8970.20724.1>) call prim_inet:close(#Port<8970.1336>)
(<8970.20724.1>) call prim_inet:getopts(#Port<8970.1336>,[linger])
(<8970.20724.1>) returned from prim_inet:getopts/2 -> {error,einval}
(<8970.20724.1>) returned from prim_inet:close/1 -> ok
(<8970.20724.1>) returned from gen_tcp:close/1 -> ok

I'm having trouble finding out why this is happening. Is {error, inval} an expected behaviour in this scenario and probable reason for ssl:handshake failure? OTP_VERSION is 23.1.5. Any suggestions would be helpful. Thanks in advance.

From: Vithya Shanmugam -X (vitshanm - SCARLET WIRELESS INDIA PRIVATE LIMITED at Cisco)
Sent: 01 April 2021 11:12
To: erlang-***@erlang.org
Cc: Salil G K (sagk) <***@cisco.com>; Panduranga Rao Sadhu (pansadhu) <***@cisco.com>
Subject: net_kernel {error, closed}

Hello,

I'm trying to understand what this INFO REPORT means,
=INFO REPORT==== 1-Apr-2021::05:28:41.384237 ===
{net_kernel,{'EXIT',<0.11838.0>,{error,closed}}}

This happens in a cluster of nodes(cluster of 2 to 6 nodes) and OTP_VERSION 23. Is it saying there's a connection issue between the local and remote node ?
But, when we try writing data in one node, it gets replicated to the other node and no other issues(no node up and node down messages). It's just that the INFO_REPORT is getting dumped in erlang.log and is getting piled up for a long time every 5 minutes and there is no error in corresponding dev logs. And the synchronization between the nodes also seem to be fine...

Thanks,
Vithya