Fix: bootstrap: Detect cluster service on init node before saving the canonical hostname (bsc#1222714)

[liangxin1300]

# When cluster service is down on init node
# crm cluster join -c 15sp5-1 -y
WARNING: chronyd.service is not configured to start at system boot.
ERROR: cluster.join:

See crmsh.log:

Traceback (most recent call last):
  File "/usr/lib/python3.6/site-packages/crmsh/ui_context.py", line 86, in run
    rv = self.execute_command() is not False
  File "/usr/lib/python3.6/site-packages/crmsh/ui_context.py", line 267, in execute_command
    rv = self.command_info.function(*arglist)
  File "/usr/lib/python3.6/site-packages/crmsh/ui_cluster.py", line 543, in do_join
    bootstrap.bootstrap_join(join_context)
  File "/usr/lib/python3.6/site-packages/crmsh/bootstrap.py", line 2246, in bootstrap_join
    join_ssh(cluster_node, remote_user)
  File "/usr/lib/python3.6/site-packages/crmsh/bootstrap.py", line 1596, in join_ssh
    return join_ssh_impl(local_user, seed_host, seed_user, keys)
  File "/usr/lib/python3.6/site-packages/crmsh/bootstrap.py", line 1628, in join_ssh_impl
    user_by_host.add(seed_user, get_node_canonical_hostname(seed_host))
  File "/usr/lib/python3.6/site-packages/crmsh/bootstrap.py", line 468, in get_node_canonical_hostname
    utils.fatal(err)
  File "/usr/lib/python3.6/site-packages/crmsh/utils.py", line 2711, in fatal
    raise ValueError(error_msg)
ValueError

With this fix, when cluster service is down on init node

# crm cluster join -c 15sp5-1 -y
WARNING: chronyd.service is not configured to start at system boot.
INFO: A new ssh keypair is generated for user hacluster.
WARNING: Cluster is inactive on 15sp5-1. Retry in 10 seconds
WARNING: Cluster is inactive on 15sp5-1. Retry in 10 seconds
...

[nicholasyang2022]

This is required by line 1633 when a hostname alias is used.

[liangxin1300]

Then we can try to check init node's cluster service status just after the line 1620

[liangxin1300]

Then the process/order for this part will be:

1. call ping_node to check if the init node is up
2. call join_ssh to setup passwordless with init node
3. inside join_ssh, after swap the key for the first user('hacluster' is the second), and before using utils.HostUserConfig to save the info of user, check the status of cluster service on init node
4. then get the canonical hostname from init node

[codecov]

Codecov Report

Attention: Patch coverage is 70.00000% with 3 lines in your changes are missing coverage. Please review.

Project coverage is 53.27%. Comparing base (09f3ed8) to head (5036bc6).

Files	Patch %	Lines
crmsh/bootstrap.py	70.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1386      +/-   ##
==========================================
- Coverage   53.32%   53.27%   -0.05%     
==========================================
  Files          80       80              
  Lines       23941    23942       +1     
==========================================
- Hits        12766    12755      -11     
- Misses      11175    11187      +12     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.