Merge pull request #85 from rug-cit-hpc/develop

Merge of Develop into Master

.gitignore

@@ -4,9 +4,10 @@
 *.md.html
 *.vagrant
 *.pydevproject
+*.pyc
 *.retry
 *.swp
-.vault_pass.txt
+.vault_pass.txt*
 documentation/.~lock.UMCG Research IT HPC cluster technical design.docx#
 promtools/results/*
 roles/hpc-cloud

callback_plugins/homsaplog.py

@@ -0,0 +1,69 @@
+'''
+callback: homsaplog
+type: stdout
+short_description: Homo sapiens friendly formatted output.
+description: Use this callback to sort though extensive debug output.
+'''
+
+from __future__ import (absolute_import, division, print_function)
+from ansible.plugins.callback.default import CallbackModule as CallbackModule_default
+from ansible.plugins.callback import CallbackBase
+try:
+    # Ansible 2.3
+    from ansible.vars import strip_internal_keys
+except ImportError:
+    try:
+        # Anisble2.4
+        from ansible.vars.manager import strip_internal_keys
+    except ImportError:
+        # Ansible 2.5
+        from ansible.vars.clean import strip_internal_keys
+try:
+    import simplejson as json
+except ImportError:
+    import json
+import sys
+reload(sys).setdefaultencoding('utf-8')
+
+__metaclass__ = type
+
+class CallbackModule(CallbackModule_default):  # pylint: disable=too-few-public-methods,no-init
+    '''
+    Override for the default callback module.
+
+    Render std err/out outside of the rest of the result which it prints with
+    indentation.
+    '''
+    CALLBACK_VERSION = 2.0
+    CALLBACK_TYPE = 'stdout'
+    CALLBACK_NAME = 'homsaplog'
+
+    def _dump_results(self, result, indent=4, sort_keys=True, keep_invocation=False):
+        '''Return the text to output for a result.'''
+
+        if result.get('_ansible_no_log', False):
+            return json.dumps(dict(censored=
+                "The output has been hidden due to the fact that 'no_log: true' was specified for this result."))
+
+        # All result keys starting with _ansible_ are for internal use only,
+        # so remove them from the result before we output anything.
+        reformatted_result = strip_internal_keys(result)
+
+        # remove invocation unless specifically wanting it
+        if not keep_invocation and self._display.verbosity < 3 and 'invocation' in result:
+            del reformatted_result['invocation']
+
+        # remove diff information from screen output
+        if self._display.verbosity < 3 and 'diff' in result:
+            del reformatted_result['diff']
+
+        # remove exception from screen output
+        if 'exception' in reformatted_result:
+            del reformatted_result['exception']
+
+        output = json.dumps(reformatted_result, indent=indent, ensure_ascii=False, sort_keys=sort_keys)
+        output = output.replace('\\r\\n\",', '",')
+        output = output.replace('\\r\\n', "\n\t")
+        output = output.replace('\\n\",', '",')
+        output = output.replace('\\n', "\n\t")
+        return output

cluster.yml

@@ -1,9 +1,41 @@
 ---
-- name: Sign host keys of all cluster hosts.
+- name: Sanity checks before we start.
+  hosts: all
+  pre_tasks:
+    - name: Verify Ansible version meets requirements.
+      assert:
+        that: "ansible_version.full | version_compare('2.4', '>=')"
+        msg: 'You must update Ansible to at least 2.4.x to use this playbook.'
+
+- import_playbook: local_admin_users.yml
+
+- name: Install roles needed for all virtual cluster components.
   hosts: all
   roles:
+    - logins
     - ssh_host_signer
     - ssh_known_hosts
+  tasks:
+    - cron:
+        #
+        # Silly workaround for bug in interaction dbus <-> logind
+        # Need DBus 1.11.10 for a fix, but CentOS 7.6 is stuck on dbus 1.10.24.
+        #
+        name: Restart systemd-logind
+        minute: "/10"
+        user: root
+        job: /bin/systemctl restart systemd-logind
+        cron_file: restart_logind
+      become: true
+
+- name: Mount the ceph volume
+  hosts:
+     - compute-vm
+     - sys-admin-interface
+     - deploy-admin-interface
+  become: True
+  roles:
+     - mount-volume
 
 - name: Install roles needed for all virtual cluster components except jumphosts.
   hosts: cluster
@@ -14,6 +46,8 @@
      - ldap
      - node_exporter
      - cluster
+     - resolver
+     - shared_storage
 
 - name: Install ansible on admin interfaces (DAI & SAI).
   hosts:
@@ -46,18 +80,21 @@
 - hosts: slurm
   become: true
   roles:
+     - slurm
      - prom_server
      - cadvisor
-     - slurm
+  vars:
+     # These variables are needed by the mariadb role.
+     # Which is a depencency of the slurm role.
+     # See roles/slurm/meta/main.yml
+     hostname_node0: "{{ ansible_hostname }}"
+     ip_node0: "{{ ansible_default_ipv4['address'] }}"
 
 - name: Install virtual compute nodes
   hosts: compute-vm
   become: true
   tasks:
   roles:
-     - compute-vm
-     - isilon
-     - datahandling
      - slurm-client
 
 - name: Install User Interface (UI)
@@ -66,19 +103,5 @@
   tasks:
   roles:
      - slurm_exporter
-     - user-interface
-     - datahandling
-     - isilon
      - slurm-client
-
-- name: Export /home on NFS server.
-  hosts: user-interface:&talos-cluster
-  roles:
-     - nfs_home_server
-
-- name: Mount /home on NFS clients.
-  hosts: compute-vm&talos-cluster
-  roles:
-     - nfs_home_client
-
-- import_playbook: users.yml
+...

documentation/Gearshift_technical_design.md

@@ -103,7 +103,7 @@ Figure 4. Network design for gs-compute[0-9] node
 
 Figure 5. Network design for gs-vcompute[0-9] virtual compute node
 
- ![](./media/media/image6.jpg)
+ ![](./media/media/image6a.jpg)
 
 ### Compute cluster design
 

gearshift_hosts.ini

@@ -1,43 +1,3 @@
-[databases]
-gs-openstack
-gs-compute10
-gs-compute11
-
-[keystone]
-gs-openstack
-
-[glance-controller]
-gs-openstack
-
-[heat]
-gs-openstack
-
-[horizon]
-gs-openstack
-
-[rabbitmq]
-gs-openstack
-gs-compute10
-gs-compute11
-
-[memcached]
-gs-openstack
-
-[neutron-controller]
-gs-openstack physical_interface_mappings=provider:enp130s0f0
-
-[nova-controller]
-gs-openstack
-
-[cinder-controller]
-gs-openstack
-
-[cinder-storage]
-gs-compute[01:11] storage_volume=/dev/sdb
-
-[nova-compute]
-gs-compute[01:11] physical_interface_mappings=provider:enp130s0f0
-
 [jumphost]
 airlock
 
@@ -68,7 +28,3 @@ administration
 [gearshift-cluster:children]
 cluster
 jumphost
-
-[metal]
-gs-openstack
-gs-compute[01:11]

group_vars/administration.yml

@@ -0,0 +1,3 @@
+---
+
+volume_mount_point: "/apps"