0

First install of ICP CE 2.1.0 on Ubuntu 16.04.03 VM running on ESXi5.5. The VM has 4vCPU with 16GB ram and 170GB (small I know). The install runs 10 min and fails. I ran the install with the -vvv and it's doesn't really provide any significant insights. 

TASK [master : Waiting for Etcd to start] **************************************
task path: /installer/playbook/roles/master/tasks/kube-service.yaml:6
Using module file /installer/playbook/library/cfc_wait_for.py
<9.29.100.159> ESTABLISH SSH CONNECTION FOR USER: root
<9.29.100.159> SSH: EXEC ssh -C -o CheckHostIP=no -o LogLevel=ERROR -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o 'IdentityFile="cluster/ssh_key"' -o KbdInteractiveAuthentication=no -o PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o User=root -o ConnectTimeout=10 9.29.100.159 '/bin/bash -c '"'"'echo ~ && sleep 0'"'"''
<9.29.100.159> (0, '/root\n', '')
<9.29.100.159> ESTABLISH SSH CONNECTION FOR USER: root
<9.29.100.159> SSH: EXEC ssh -C -o CheckHostIP=no -o LogLevel=ERROR -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o 'IdentityFile="cluster/ssh_key"' -o KbdInteractiveAuthentication=no -o PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o User=root -o ConnectTimeout=10 9.29.100.159 '/bin/bash -c '"'"'( umask 77 && mkdir -p "` echo /root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067 `" && echo ansible-tmp-1511385912.24-67181235419067="` echo /root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067 `" ) && sleep 0'"'"''
<9.29.100.159> (0, 'ansible-tmp-1511385912.24-67181235419067=/root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067\n', '')
<9.29.100.159> PUT /tmp/tmp_LQQz6 TO /root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067/cfc_wait_for.py
<9.29.100.159> SSH: EXEC sftp -b - -C -o CheckHostIP=no -o LogLevel=ERROR -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o 'IdentityFile="cluster/ssh_key"' -o KbdInteractiveAuthentication=no -o PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o User=root -o ConnectTimeout=10 '[9.29.100.159]'
<9.29.100.159> (0, 'sftp> put /tmp/tmp_LQQz6 /root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067/cfc_wait_for.py\n', '')
<9.29.100.159> ESTABLISH SSH CONNECTION FOR USER: root
<9.29.100.159> SSH: EXEC ssh -C -o CheckHostIP=no -o LogLevel=ERROR -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o 'IdentityFile="cluster/ssh_key"' -o KbdInteractiveAuthentication=no -o PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o User=root -o ConnectTimeout=10 9.29.100.159 '/bin/bash -c '"'"'chmod u+x /root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067/ /root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067/cfc_wait_for.py && sleep 0'"'"''
<9.29.100.159> (0, '', '')
<9.29.100.159> ESTABLISH SSH CONNECTION FOR USER: root
<9.29.100.159> SSH: EXEC ssh -C -o CheckHostIP=no -o LogLevel=ERROR -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o 'IdentityFile="cluster/ssh_key"' -o KbdInteractiveAuthentication=no -o PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o User=root -o ConnectTimeout=10 -tt 9.29.100.159 '/bin/bash -c '"'"'/usr/bin/python /root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067/cfc_wait_for.py; rm -rf "/root/.ansible/tmp/ansible-tmp-1511385912.24-67181235419067/" > /dev/null 2>&1 && sleep 0'"'"''
<9.29.100.159> (0, '\r\n{"msg": "The Etcd component failed to start. For more details, see https://ibm.biz/etcd-fails.", "failed": true, "elapsed": 1965, "invocation": {"module_args": {"active_connection_states": ["ESTABLISHED", "SYN_SENT", "SYN_RECV", "FIN_WAIT1", "FIN_WAIT2", "TIME_WAIT"], "state": "started", "port": 4001, "delay": 0, "msg": "The Etcd component failed to start. For more details, see https://ibm.biz/etcd-fails.", "host": "9.29.100.159", "sleep": 1, "timeout": 600, "exclude_hosts": null, "search_regex": null, "path": null, "connect_timeout": 5}}}\r\n', 'Connection to 9.29.100.159 closed.\r\n')
fatal: [9.29.100.159] => The Etcd component failed to start. For more details, see https://ibm.biz/etcd-fails.

The link https://ibm.biz/etcd-fails takes you to a 1.2.0 Knowledge Center entry about flannel fails to start on worker node.

Whats odd is a docker ps shows that etcd is running

root@sysicpce:~# docker ps
CONTAINER ID        IMAGE               COMMAND                  CREATED             STATUS              PORTS               NAMES
652aab0c1cee        ibmcom/mariadb      "start.sh docker-e..."   17 hours ago        Up 17 hours                             k8s_mariadb_k8s-mariadb-9.29.100.159_kube-system_3b21d2ed8c3e2047c0e457af0e948b97_0
80201425a077        ibmcom/etcd         "etcd --name=etcd0..."   17 hours ago        Up 17 hours                             k8s_etcd_k8s-etcd-9.29.100.159_kube-system_b674f0dc7c07780868387aaea0ba7acc_0
a5be8a1e0c25        ibmcom/pause:3.0    "/pause"                 17 hours ago        Up 17 hours                             k8s_POD_k8s-mariadb-9.29.100.159_kube-system_3b21d2ed8c3e2047c0e457af0e948b97_0
d82b0c6e5fa0        ibmcom/pause:3.0    "/pause"                 17 hours ago        Up 17 hours                             k8s_POD_k8s-etcd-9.29.100.159_kube-system_b674f0dc7c07780868387aaea0ba7acc_0
6574c3760499        ibmcom/kubernetes   "/hyperkube proxy ..."   18 hours ago        Up 18 hours                             k8s_proxy_k8s-proxy-9.29.100.159_kube-system_708dfdafb2a5d66e99356e10e609f6b1_0
3b4621d57fef        ibmcom/pause:3.0    "/pause"                 18 hours ago        Up 18 hours                             k8s_POD_k8s-proxy-9.29.100.159_kube-system_708dfdafb2a5d66e99356e10e609f6b1_0
root@sysicpce:~#

How can I resolve this? Where can/should I look next?

4

2 回答 2

1

根据安装要求，如果您在单主机集群中运行所有管理服务，则至少需要 8 个核心 CPU。如果你有少于这个，你可以禁用管理服务。例如计量和监控，如 config.yaml 文件中的 disabled_management_services: ["metering", "monitoring"]。由于您有 4 核 CPU，您可以在 config.ymal 中禁用这些服务，然后再次尝试安装。

于 2017-11-23T18:13:43.030 回答
0

我遇到了同样的问题，在 Ubuntu 16.04、KVM/OpenStack 上使用 ICP CE 2.1.0。相同的消息：“Etcd 组件无法启动”

当我添加一个规则以允许从 127.0.0.1 访问 ICP 机器上的端口 4001 时，问题就消失了。

于 2017-11-28T17:45:40.067 回答