Implement a proper WSREP check for galera

The galera cluster rely on WSREP for cluster consistency. While the
default MySQL monitor will allow us to know when the database node is
minimally functional it does not provide the ability to query the node
state allowing loadbalancers, operators, and deployers to know a node
is healthy prior to being allowed to accept connections. This change
implements the checkcluster script as provided by the fine folks at
Percona. The implementation of this check follows the guild-lines noted
here [0]. With this in-place, we'll be able to convert our haproxy check
for the galera cluster nodes to use an HTTP check on port 9200 instead
of the default MySQL login which will provide for a more robust and
fault tolerant cluster.

[0] https://www.percona.com/doc/percona-xtradb-cluster/LATEST/howtos/virt_sandbox.html

Combined backport of:
- https://review.openstack.org/520665
- https://review.openstack.org/523850

Conflicts:
>------handlers/main.yml
>------tests/test-galera-server-functional.yml
>------vars/suse-42.yml

Closes-Bug: #1665667
Change-Id: Ie1b3b9724dd33de1d90634166e585ecceb1f4c96
Signed-off-by: Kevin Carter <kevin.carter@rackspace.com>
(cherry picked from commit ed739a5243)
This commit is contained in:
Kevin Carter 2017-11-16 11:30:55 -06:00 committed by Jesse Pretorius
parent b586b9fa1e
commit 515ab131fa
9 changed files with 191 additions and 0 deletions

View File

@ -69,6 +69,18 @@ galera_running_and_bootstrapped: false
galera_monitoring_user: monitoring
galera_monitoring_user_password: ""
# WARNING: Set this to open xinetd rules for galera monitoring.
# This is REQUIRED to run a working openstack-ansible deployment.
# If it's undefined the galera cluster state can't be reported,
# and haproxy would fail to do proper load balancing on the cluster.
# Because this opens connections to the cluster status, this
# should be restricted, which we do in the integrated build.
# Please override accordingly to your use case.
# This can be replaced with other hostnames, cidr, ips, and ips + wildcards.
#
#galera_monitoring_allowed_source: "0.0.0.0/0"
galera_root_user: root
# WARNING: This option is deprecated and will be removed in v12.0

View File

@ -56,3 +56,8 @@
until: galera_restart_fall_back | success
retries: 3
delay: 5
- name: Restart xinetd
service:
name: xinetd
state: restarted

View File

@ -0,0 +1,7 @@
---
features:
- The galera cluster now supports cluster health checks over HTTP using port
9200. The new cluster check ensures a node is healthy by running a simple
query against the wsrep sync status using monitoring user. This change will
provide for a more robust cluster check ensuring we have the most fault
tolerant galera cluster possible.

View File

@ -0,0 +1,7 @@
---
features:
- |
Galera healthcheck has been improved, and relies on an xinetd service.
By default, the service is unaccessible (filtered with the no_access
directive). You can override the directive by setting any xinetd
valid value to ``galera_monitoring_allowed_source``.

View File

@ -124,3 +124,31 @@
command: "systemctl daemon-reload"
when:
- ansible_service_mgr == 'systemd'
- name: Create clustercheck script
template:
src: "clustercheck.j2"
dest: "/usr/local/bin/clustercheck"
mode: "0755"
tags:
- galera-config
- name: Create mysqlchk config
template:
src: "mysqlchk.j2"
dest: "/etc/xinetd.d/mysqlchk"
mode: "0644"
notify:
- Restart xinetd
tags:
- galera-config
- name: Add galera service check to services
lineinfile:
dest: /etc/services
state: present
regexp: '^mysqlchk'
line: 'mysqlchk 9200/tcp # MySQL check'
backup: yes
tags:
- galera-config

110
templates/clustercheck.j2 Normal file
View File

@ -0,0 +1,110 @@
#!/bin/bash
#
# Script to make a proxy (ie HAProxy) capable of monitoring Percona XtraDB Cluster nodes properly
#
# Author: Olaf van Zandwijk <olaf.vanzandwijk@nedap.com>
# Author: Raghavendra Prabhu <raghavendra.prabhu@percona.com>
#
# Documentation and download: https://github.com/olafz/percona-clustercheck
#
# Based on the original script from Unai Rodriguez
#
# {{ ansible_managed }}
if [[ $1 == '-h' || $1 == '--help' ]];then
echo "Usage: $0 <user> <pass> <available_when_donor=0|1> <log_file> <available_when_readonly=0|1> <defaults_extra_file>"
exit
fi
# if the disabled file is present, return 503. This allows
# admins to manually remove a node from a cluster easily.
if [ -e "/var/tmp/clustercheck.disabled" ]; then
# Shell return-code is 1
echo -en "HTTP/1.1 503 Service Unavailable\r\n"
echo -en "Content-Type: text/plain\r\n"
echo -en "Connection: close\r\n"
echo -en "Content-Length: 51\r\n"
echo -en "\r\n"
echo -en "Percona XtraDB Cluster Node is manually disabled.\r\n"
sleep 0.1
exit 1
fi
MYSQL_USERNAME="${1-{{ galera_monitoring_user }}}"
MYSQL_PASSWORD="${2-{{ galera_monitoring_user_password }}}"
AVAILABLE_WHEN_DONOR=${3:-0}
ERR_FILE="${4:-/dev/null}"
AVAILABLE_WHEN_READONLY=${5:-1}
DEFAULTS_EXTRA_FILE=${6:-/etc/my.cnf}
#Timeout exists for instances where mysqld may be hung
TIMEOUT=10
EXTRA_ARGS=""
if [[ -n "$MYSQL_USERNAME" ]]; then
EXTRA_ARGS="$EXTRA_ARGS --user=${MYSQL_USERNAME}"
fi
if [[ -n "$MYSQL_PASSWORD" ]]; then
EXTRA_ARGS="$EXTRA_ARGS --password=${MYSQL_PASSWORD}"
else
EXTRA_ARGS="$EXTRA_ARGS --password="
fi
if [[ -r $DEFAULTS_EXTRA_FILE ]]; then
MYSQL_CMDLINE="mysql --defaults-extra-file=$DEFAULTS_EXTRA_FILE -nNE --connect-timeout=$TIMEOUT \
${EXTRA_ARGS}"
else
MYSQL_CMDLINE="mysql -nNE --connect-timeout=$TIMEOUT ${EXTRA_ARGS}"
fi
#
# Perform the query to check the wsrep_local_state
#
WSREP_STATUS=$($MYSQL_CMDLINE -e "SHOW STATUS LIKE 'wsrep_local_state';" \
2>${ERR_FILE} | tail -1 2>>${ERR_FILE})
if [[ "${WSREP_STATUS}" == "4" ]] || [[ "${WSREP_STATUS}" == "2" && ${AVAILABLE_WHEN_DONOR} == 1 ]]; then
# Check only when set to 0 to avoid latency in response.
if [[ $AVAILABLE_WHEN_READONLY -eq 0 ]];then
READ_ONLY=$($MYSQL_CMDLINE -e "SHOW GLOBAL VARIABLES LIKE 'read_only';" \
2>${ERR_FILE} | tail -1 2>>${ERR_FILE})
if [[ "${READ_ONLY}" == "ON" ]];then
# Percona XtraDB Cluster node local state is 'Synced', but it is in
# read-only mode. The variable AVAILABLE_WHEN_READONLY is set to 0.
# => return HTTP 503
# Shell return-code is 1
echo -en "HTTP/1.1 503 Service Unavailable\r\n"
echo -en "Content-Type: text/plain\r\n"
echo -en "Connection: close\r\n"
echo -en "Content-Length: 43\r\n"
echo -en "\r\n"
echo -en "Percona XtraDB Cluster Node is read-only.\r\n"
sleep 0.1
exit 1
fi
fi
# Percona XtraDB Cluster node local state is 'Synced' => return HTTP 200
# Shell return-code is 0
echo -en "HTTP/1.1 200 OK\r\n"
echo -en "Content-Type: text/plain\r\n"
echo -en "Connection: close\r\n"
echo -en "Content-Length: 40\r\n"
echo -en "\r\n"
echo -en "Percona XtraDB Cluster Node is synced.\r\n"
sleep 0.1
exit 0
else
# Percona XtraDB Cluster node local state is not 'Synced' => return HTTP 503
# Shell return-code is 1
echo -en "HTTP/1.1 503 Service Unavailable\r\n"
echo -en "Content-Type: text/plain\r\n"
echo -en "Connection: close\r\n"
echo -en "Content-Length: 44\r\n"
echo -en "\r\n"
echo -en "Percona XtraDB Cluster Node is not synced.\r\n"
sleep 0.1
exit 1
fi

20
templates/mysqlchk.j2 Normal file
View File

@ -0,0 +1,20 @@
# default: on
# description: mysqlchk
# {{ ansible_managed }}
service mysqlchk
{
disable = no
flags = REUSE
socket_type = stream
port = 9200
wait = no
user = nobody
server = /usr/local/bin/clustercheck
log_on_failure += USERID
{% if galera_monitoring_allowed_source is defined %}
only_from = {{ galera_monitoring_allowed_source }}
{% else %}
no_access
{% endif %}
per_source = UNLIMITED
}

View File

@ -27,6 +27,7 @@ galera_server_required_distro_packages:
- libgcrypt
- MariaDB-client
- MariaDB-devel
- xinetd
_galera_mariadb_server_package: "MariaDB-server"

View File

@ -38,6 +38,7 @@ galera_server_required_distro_packages:
- libstdc++6
- python-software-properties
- software-properties-common
- xinetd
# The package name for mariaDB is set as a variable
# so that it can be used in debconf later in the