cherry-pick from master to 3.0

excessive_rolling_update.yml

@@ -38,6 +38,22 @@
   roles:
     - check_config_dynamic
 
+- name: Pre-check for rolling update
+  hosts: tidb_servers
+  any_errors_fatal: true
+  serial: 1
+  tags:
+    - always
+  tasks:
+    - shell: "{{ deploy_dir }}/bin/tidb-server -V"
+      register: current_version
+
+    - name: Check whether can perform rolling update
+      fail:
+        msg: "Rolling update from {{ current_version.stdout_lines[0].replace(' ','').split(':')[1] }} to {{ tidb_version }} is forbidden"
+      when:
+        - current_version.stdout_lines[0].replace(' ','').split(':')[1] <= "v2.0.1"
+        - tidb_version >= "v2.1.0" or tidb_version == "latest"
 
 - name: rolling update PD cluster
   hosts: pd_servers
@@ -199,7 +215,7 @@
       until: tikv_http_result.status == 200
       retries: 12
       delay: 5
-      when: not enable_tls|default(false) and (tikv_metric_method is defined and tikv_metric_method != "push")
+      when: not enable_tls|default(false)
 
     - name: wait until the TiKV status page is available when enable_tls
       uri:
@@ -212,13 +228,7 @@
       until: tikv_https_result.status == 200
       retries: 10
       delay: 5
-      when: enable_tls|default(false) and (tikv_metric_method is defined and tikv_metric_method != "push")
-
-    - name: wait until TiKV process is up
-      wait_for_pid: |
-        pid_file={{ deploy_dir }}/status/tikv.pid timeout=300
-        thread_name_regex='tikv-server' state=present
-      when: deployment_method == 'binary' and (tikv_metric_method is defined and tikv_metric_method == "push")
+      when: enable_tls|default(false)
 
     - command: cat {{ deploy_dir }}/status/tikv.pid
       register: new_tikv_pid

group_vars/all.yml

@@ -40,9 +40,6 @@ dev_mode: False
 # systemd: Specifies whether to send SIGKILL to remaining processes after a timeout.
 disable_send_sigkill: False
 
-# The way to collect tikv monitor data: "pull" or "push".
-tikv_metric_method: "pull"
-
 # pump
 pump_port: 8250
 pump_data_dir: "{{ deploy_dir }}/data.pump"

migrate_monitor.yml

@@ -326,7 +326,9 @@
               node: "{{ cluster_name | title }}-Node_exporter"
               pd:   "{{ cluster_name | title }}-PD"
               tidb: "{{ cluster_name | title }}-TiDB"
-              tikv: "{{ cluster_name | title }}-TiKV"
+              tikv_summary: "{{ cluster_name | title }}-TiKV-Summary"
+              tikv_details: "{{ cluster_name | title }}-TiKV-Details"
+              tikv_trouble_shot: "{{ cluster_name | title }}-TiKV-Trouble-Shooting"
               binlog: "{{ cluster_name | title }}-Binlog"
               overview: "{{ cluster_name | title }}-Overview"
               disk_performance: "{{ cluster_name | title }}-Disk-Performance"

requirements.txt

@@ -1,3 +1,3 @@
-ansible>=2.4.2,<2.7.0
+ansible==2.5.14
 jinja2>=2.9.6
 jmespath>=0.9.0

roles/check_system_static/tasks/main.yml

@@ -14,7 +14,10 @@
 - name: Preflight check - Linux OS family and distribution version
   fail:
     msg: "System versions lower than Red Hat Enterprise Linux / CentOS 7.3 have been deprecated. Please use CentOS 7.3 and above. See https://github.com/pingcap/docs/blob/master/op-guide/recommendation.md"
-  when: "ansible_os_family == 'RedHat' and ansible_distribution_version < '7.3'"
+  when:
+    - ansible_os_family == 'RedHat'
+    - ansible_distribution != 'Amazon'
+    - ansible_distribution_version < '7.3'
 
 - name: Deploy epollexclusive script
   copy: src="{{ script_dir }}/check/epollexclusive" dest="{{ deploy_dir }}/epollexclusive" mode=0755

roles/grafana/defaults/main.yml

@@ -18,7 +18,7 @@ grafana_api_keys:
 
 grafana_exec_vars_only: false
 
-grafana_version: 4.6.3
+grafana_version: 5.4.3
 
 # docker settings
-grafana_tag: 4.6.3
+grafana_tag: 5.4.3

roles/local/templates/binary_packages.yml.j2

@@ -2,8 +2,8 @@
 
 third_party_packages:
   - name: prometheus
-    version: 2.2.1
-    url: "https://github.com/prometheus/prometheus/releases/download/v2.2.1/prometheus-2.2.1.linux-amd64.tar.gz"
+    version: 2.6.1
+    url: "https://github.com/prometheus/prometheus/releases/download/v2.6.1/prometheus-2.6.1.linux-amd64.tar.gz"
   - name: alertmanager
     version: 0.14.0
     url: "https://github.com/prometheus/alertmanager/releases/download/v0.14.0/alertmanager-0.14.0.linux-amd64.tar.gz"
@@ -17,14 +17,14 @@ third_party_packages:
     version: 0.4.0
     url: "https://github.com/prometheus/pushgateway/releases/download/v0.4.0/pushgateway-0.4.0.linux-amd64.tar.gz"
   - name: grafana
-    version: 4.6.3
-    url: "https://s3-us-west-2.amazonaws.com/grafana-releases/release/grafana-4.6.3.linux-x64.tar.gz"
+    version: 5.4.3
+    url: "https://dl.grafana.com/oss/release/grafana-5.4.3.linux-amd64.tar.gz"
 
 
 third_party_packages_under_gfw:
   - name: prometheus
-    version: 2.2.1
-    url: "http://download.pingcap.org/prometheus-2.2.1.linux-amd64.tar.gz"
+    version: 2.6.1
+    url: "https://download.pingcap.org/prometheus-2.6.1.linux-amd64.tar.gz"
   - name: alertmanager
     version: 0.14.0
     url: "http://download.pingcap.org/alertmanager-0.14.0.linux-amd64.tar.gz"
@@ -35,8 +35,8 @@ third_party_packages_under_gfw:
     version: 0.4.0
     url: "http://download.pingcap.org/pushgateway-0.4.0.linux-amd64.tar.gz"
   - name: grafana
-    version: 4.6.3
-    url: "http://download.pingcap.org/grafana-4.6.3.linux-x64.tar.gz"
+    version: 5.4.3
+    url: "https://download.pingcap.org/grafana-5.4.3.linux-amd64.tar.gz"
   - name: blackbox_exporter
     version: 0.12.0
     url: "http://download.pingcap.org/blackbox_exporter-0.12.0.linux-amd64.tar.gz"

roles/prometheus/defaults/main.yml

@@ -16,4 +16,4 @@ prometheus_extra_labels:
 
 alert_label: "promethues"
 
-prometheus_tag: v2.2.1
+prometheus_tag: v2.6.1