How do I solve the “Waiting for calico node service” error message

问题

Need assistance with ICP install. I have encountered the following error:

TASK [addon : Waiting for calico node service] *********************************
failed: [localhost -> 129.40.227.142] (item=129.40.227.142) => {"elapsed": 600, "failed": true, "item": "129.40.227.142", "msg": "Timeout when waiting for 129.40.227.142:9099"}

PLAY RECAP *********************************************************************
129.40.227.142             : ok=172  changed=66   unreachable=0    failed=0
129.40.227.143             : ok=157  changed=55   unreachable=0    failed=0
129.40.227.144             : ok=116  changed=24   unreachable=0    failed=0
localhost                  : ok=118  changed=52   unreachable=0    failed=1

回答1:

Try explicitly setting the calico_ip_autodetection_method: interface to your interface name in your config.yaml file.

Here's the documentation of all things you can do in that file. https://www.ibm.com/support/knowledgecenter/SSBS6K_2.1.0/installing/config_yaml.html#network_setting.

Kudos to Harry P.

回答2:

We had a similar issue installing.

The solution was the following command on each machine. (Use Ansible to make your life easier)

sysctl -w net.ipv4.conf.all.rp_filtered=1

make sure to add it to the sysctl.conf file so it persists on restart.

echo "net.ipv4.conf.all.rp_filtered=1" | tee -a /etc/sysctl.conf

Hope this helps others.

回答3:

As for me I had similar issue, caused by running out of disk space in the /var folder. Check if you have no error messages in the /var/log/containers folder stating some pods could not be started or similar. The documentation states you will need 40G of space. If you setup docker to place the containers on your data drive, the install will use ~ 500MB in the /opt and ~300MB in the /var.

来源：https://stackoverflow.com/questions/47210290/how-do-i-solve-the-waiting-for-calico-node-service-error-message

标签

ibm-cloud-private