Skip to content

Temporary repo to reproduce an issue with Ignite and Docker

Notifications You must be signed in to change notification settings

skarpushin/issue20181122

Repository files navigation

This is temp repo just to reproduce issue I encountered with Ignite, SSL and Docker. Here is a related question on stack overflow:

Assumptions:
  1. You're running *nix OS
  2. You have maven and jdk 1.8+
  3. You have docker-compose 3.7+
  4. docker and docker-compose require root to run (thus *.sh scripts I've created will do sudo and ask for sudo password
  5. Your non-loopback network device's name is enp0s3 (if it isn't update variable ETH_DEV in both *.sh files)
  6. None of your network devices have network 10.9.9.0/24 (if they do, change sh scripts and docker-compose to use other IP address to create and bind to)
In order to reproduce the issue:
  1. Clone this repo
  2. Run same-time-ssl-issue.sh script (this script will simultaneously bring up both web0 and web1 containers)
  3. You'll see exceptions just like described in aforementioned stackoverflow question, context initialization failed and application will never recover from this error
  4. Press ctrl+c (this will stop and remove current containers)
  5. Run sequential-no-issue.shscript (this will bring up web0 and then after a 20 second delay, web1). No issue and you can see it shows 2 nodes in cluster (in logs and if you do wget http://10.9.9.1:8080/count -q -O -)
Notes:
  1. I wasn't able to reproduce the issue outside of docker environment
  2. I was able to get rid of this issue by removing SSL configuration from Ignite

About

Temporary repo to reproduce an issue with Ignite and Docker

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published