Using MemSQL Ops

MemSQL Ops provides a real-time, dashboard overview of your MemSQL deployment. When used in conjunction with collectd (see collectd Integration), MemSQL Ops provides live system information on all of the nodes in a MemSQL cluster.

Setting Up MemSQL Ops

This section describes how to install, configure, and launch MemSQL Ops on a machine that has access to a MemSQL instance. For information about the most recent version of MemSQL Ops, see http://versions.memsql.com/ops/latest.

Prerequisites

MemSQL Ops requires the following:

  • Access to a MemSQL distributed system from the machine on which MemSQL Ops is installed.
  • Access to a MemSQL single-node instance. This instance can run on the same machine as the one on which MemSQL Ops is installed.

Note

MemSQL Ops can be run on an aggregator of the cluster it is monitoring, or on a separate single box instance of MemSQL. Ops requires relatively little memory and CPU and so, in most cases, can be run on an aggregator without issue. However, note that it could be problematic if Ops runs on the cluster it monitors (suppose the aggregator on which Ops is running goes down). Note these consideration when choosing where to run MemSQL Ops.

Installing MemSQL Ops

The license confirmation email contains download links and brief instructions on how to begin installing MemSQL Ops. You have two installation options:

Note

In general, RPM installation is recommended. However, it does not work on Ubuntu.

Installing MemSQL Ops from an RPM

  1. Using the information provided in the license confirmation email, run wget to download the RPM onto the server.
$ wget http://download.memsql.com/ops-[version]/memsql-ops-[version].x86_64.rpm
  1. Run the RPM Package Manager.
$ rpm -ivh memsql-ops-[version].x86_64.rpm

Installing MemSQL Ops from a Direct Binary

  1. Using the information provided in the license confirmation email, run wget to download the tar.gz file onto the server.
$ wget http://download.memsql.com/ops-[version]/memsql-ops-[version].tar.gz
  1. Expand the downloaded tar.gz archive.
$ tar -xzf memsql-ops-[version].tar.gz
  1. Optionally, move the resulting memsql-ops directory to the desired location. This directory contains the dashboard.cnf configuration file and the launch.sh script.

Configuring MemSQL Ops

Once installed, modify settings in the MemSQL configuration file (dashboard.cnf), which is located in:

  • RPM installation: /var/lib/memsql-ops/dashboard.cnf
  • Direct binary installation: [installation path]/memsql-ops/dashboard.cnf

Open dashboard.cnf and make the following changes:

  • [cluster] - Specify the host and port so that the dashboard can connect to the Master Aggregator.
  • [dashboard] - Specify the host, port, user, password, and database to use as the standalone MemSQL instance.

Here is an example configuration:

[cluster]
; connection information for the MemSQL Master Aggregator
db_host: [MASTER AGGREGATOR IP OR HOSTNAME]
db_port: 3306
db_user: root
db_pass:

[dashboard]
; connection information for the dashboard database
db_host: [DASHBOARD DB IP OR HOSTNAME]
db_port: 3306
db_name: dashboard
db_user: root
db_pass:

[analytics]
; how many hours of analytics data should be kept?
retention: 6
; how many hours of event data should be kept?
events_retention: 6
; how many hours of facts data should be kept?
facts_retention: 6

[http]
host: 0.0.0.0
port: 9000

[global]
debug: false
verbose_errors: false

Note

MemSQL Ops allows you to configure where it stores cluster monitoring data. The data can be stored in a separate single box instance of MemSQL, or on one of the aggregators of the cluster.

Launching MemSQL Ops

Once installed and configured, launch MemSQL Ops using the applicable instructions.

To start an RPM installation of MemSQL Ops, run the following command.

sudo service memsql-ops start

To start a direct binary installation of MemSQL Ops, run the following script (from the dashboard directory):

./launch.sh

The server is split into two parts: cluster_monitor and frontend. Both parts need to be running in order for the dashboard to be operational.

Note

launch.sh is only a wrapper script to get up and running quickly, and will run in the current shell. For production deployments, write an init script to manage the dashboard process.

Viewing MemSQL Ops

Once MemSQL Ops is running on the server, you can access it through a web browser. Note that the dashboard has been tested most heavily using Google Chrome, and will run best in that browser.

To view the dashboard, navigate to the IP address or hostname of the server running MemSQL Ops. If you specified an alternate HTTP port in the configuration, specify that port in the web browser.

collectd Integration

collectd system monitoring daemon is a system monitoring daemon that runs as a client service on every machine in a cluster. MemSQL Ops is fully integrated with collectd. To enable this functionality, install and configure collectd on each machine in your cluster, as well as the official MemSQL collectd plugin.

Installing collectd

collectd is distributed as a binary package for most Linux distributions. Read the official collectd install/download guide to determine the best installation method for your platform.

Note

We recommend collectd version 5.0 and above.

Installing the collectd plugin

MemSQL Ops provides a plugin for collectd in the form of a python package. There are two ways to install it, via the Python package index or from a download.

The following are dependencies for the Python package:

  1. Python 2.6 or 2.7, along with headers
  2. Python distribute version 0.6.28 or higher
  3. MySQL headers

To upgrade setuptools, you can run the following command as root:

$ wget https://bitbucket.org/pypa/setuptools/raw/bootstrap/ez_setup.py -O - | python

Once you have the required dependencies, install the MemSQL Ops collected plugin:

$ easy_install memsql_collectd

If you have issues installing the plugin, please contact support.

Configuring collectd

collectd reads its configuration from /etc/collectd.conf or /etc/collectd/collectd.conf depending on your distribution.

The minimum recommended configuration is the following:

Interval 1
LoadPlugin syslog

LoadPlugin cpu
LoadPlugin memory
LoadPlugin interface
LoadPlugin df

<LoadPlugin python>
    Globals true
</LoadPlugin>

<Plugin python>
    Import "memsql_collectd.plugin"
    <Module "memsql_collectd.plugin">
        TypesDB "/usr/share/collectd/types.db"
        Host "[DASHBOARD DB IP OR HOSTNAME]"
        Port "3306"
    </Module>
</Plugin>

Troubleshooting

Stats are showing up on all the nodes except for the Master Aggregator.
Ensure that you are referencing the IP address of the Master Aggregator (not a hostname pointing to it) in the dashboard.cnf file.
How do I install python/mysql headers?
On Centos the packages you need are mysql-devel and python-devel. On Ubuntu the packages are python-dev and libmysqlclient-dev.

If you have other questions, please contact support.

::...
免责声明:
当前网页内容, 由 大妈 ZoomQuiet 使用工具: ScrapBook :: Firefox Extension 人工从互联网中收集并分享;
内容版权归原作者所有;
本人对内容的有效性/合法性不承担任何强制性责任.
若有不妥, 欢迎评注提醒:

或是邮件反馈可也:
askdama[AT]googlegroups.com



自怼圈/年番新

DU21.7
关于 ~ DebugUself with DAMA ;-)


关注公众号, 持续获得相关各种嗯哼:
zoomquiet


粤ICP备18025058号-1
公安备案号: 44049002000656 ...::