How to debug problem with SEMP timing out in container

Options
olly
olly Member Posts: 4 ✭
in General Discussions #1

Hi all, I am experiencing an issue when a docker container comes up, the exposed ports are all available, but the SEMP port times out when it is sent an HTTP patch to http://localhost:<mapped port>/SEMP/v2/config/msgVpns/default

I have confirmed in Wireshark that the connection has been opened at the TCP level and the http request set, but SEMP never responds. This is with old and latest versions of Solace. The most recent one tested against is solace/solace-pubsub-standard:9.12.0.15

The real mystery is that this seems to fail on some machines and not on others. For example, it works on Docker for Mac, Google's Container Optimised OS, but is failing on latest debian and Arch linux.

What is the best way to proceed with debugging this issue?

Answers

  • brenda
    brenda Member, Employee Posts: 9 Solace Employee
    #2
    Options

    You may need to ensure that /proc/sys/net/ipv4/ip_forward is set to 1 so that the packets will be passed to the container.

  • olly
    olly Member Posts: 4 ✭
    #3
    Options

    Hi, I already have this flag set to 1. I can't see how a successful TCP connection could have been established if the packets weren't being forwarded.

    Also, whilst the HTTP patch is being waited for, I can do a GET via curl on the port and receive an HTML page. This suggests that it is whatever is receiving the PATCH command that can't proceed.

    Any further ideas welcome, including how to set the debug/trace log level via container parameters

Categories

This Month's Leaders

This Week's Leaders