Ssh: Could not resolve hostname node1: Name or service not known


#1

Hi,

I tried to run a build with 3 instances and got error messages below on the primary node.

Syncing CircleCI nodes...
Waiting for slave nodes to be ready for download...
ssh: Could not resolve hostname node1: Name or service not known

ssh: Could not resolve hostname node2: Name or service not known

ssh: Could not resolve hostname node1: Name or service not known

ssh: Could not resolve hostname node2: Name or service not known

...

These Could not resolve hostname continues until the build was cancelled by timeout.


SSH between machines
#2

SSH between containers is not yet supported.


#3

I have these steps and it seems the primary node tried to connect via ssh automatically at the end of the steps… I didn’t see the log on other nodes. We are using 3 containers per build. Does 2.0 beta supports the settings which we use for 1.0 ?

 steps:
      - checkout
      - restore_cache:
          key: usr-local-bin-phantomjs-2.1.1
      - restore_cache:
          key: gemfile-{{ checksum "Gemfile.lock" }}
      - run: apt-get update -qq && apt-get install -y build-essential nodejs
      - run: curl --output /usr/local/bin/phantomjs https://s3.amazonaws.com/circle-downloads/phantomjs-2.1.1
      - run: chmod +x /usr/local/bin/phantomjs
      - save_cache:
          key: usr-local-bin-phantomjs-2.1.1
          paths:
            - "/usr/local/bin/phantomjs"
      - run: ./bin/set_github_status_as_pending.sh
      - run: bundle install --path vendor/bundle
      - save_cache:
          key: gemfile-{{ checksum "Gemfile.lock" }}
          paths:
            - "vendor/bundle"
      - run: DISABLE_SPRING=1 bundle exec rake db:create db:schema:load --trace
      - run: echo $PATH
      - run: PATH=$PATH:/usr/local/bin DISABLE_SPRING=1 bundle exec rake db:migrate
      - run: mkdir -p ${CIRCLE_TEST_REPORTS}/rspec
      - run: bin/rspec --fail-fast --color --require spec_helper --format=progress --format=Nc --format=RspecJunitFormatter --out ${CIRCLE_TEST_REPORTS}/rspec/rspec.xml --order rand `circleci tests glob "spec/**/*_spec.rb" | circleci tests split --split-by=filesize`

#4

What settings are you referring to?


#5

What settings are you referring to?

I meant “Adjust Parallelism” in the settings page. We set parallelism to 3x. I expected the steps in the previous posts make our tests run parallel automatically and it worked. But I saw the error messages after on the primary node after the tests finished. I didn’t write any steps to use ssh. I think the error was caused by something related to collect test results from other nodes.


#6

Yes, parallelism is the same, but you can’t SSH between the nodes.

How are you collecting the test results?


#7

The steps in the previous post was whole steps. I haven’t added any steps to collect test results with 2.0. That’s why I am wondering why SSH connections to other nodes happened automatically. Here is a example of the errors happened.

https://circleci.com/gh/ClinicalPlatform/Qualizm/15840


#8

Because your rspec config is telling it to do that. You will need to change the config in question to prevent SSHing.


#9

I looked spec config files but couldn’t find any settings to connect other containers with SSH.

I suppose the collection of test metadata is not working correctly with 2.0. We didn’t add any settings for test metadata with 1.0 because the metadata was collected automatically as described in the help page.

https://circleci.com/docs/1.0/test-metadata/


SSH access to other nodes
#10

Oh, I found the culprit in the gems we are using as you said… I disabled the gem and will try another way to merge test metadata.

Thank you very much for taking your time for us.


#11

This topic was automatically closed 41 days after the last reply. New replies are no longer allowed.