Add more details about 2.5.0 release
* add mysql section * add tuning section * update release notes * add links to fixed bugs * add links to documentations Change-Id: I7091ccd4241bfb820804e2b2b9b8786b09fcd663
This commit is contained in:
Swann Croiset
parent
f0d9299a29
commit
fc71bbbbba
Binary file not shown.
|
After Width: | Height: | Size: 75 KiB |
|
|
@ -2,6 +2,55 @@
|
|||
Appendix
|
||||
==================
|
||||
|
||||
.. _tuning:
|
||||
|
||||
Zabbix configuration tuning
|
||||
============================
|
||||
|
||||
*New in version 2.5.0*
|
||||
|
||||
Zabbix server
|
||||
-------------
|
||||
|
||||
To be able to handle large environments, Zabbix server is configured with the following
|
||||
parameters (provided that there is enough memory on the node otherwise the default
|
||||
values are used).
|
||||
|
||||
Memory cache sizes:
|
||||
|
||||
* CacheSize = 32M (default 8M)
|
||||
* HistoryCacheSize = 128M (default 8M)
|
||||
* TrendCacheSize = 512M (default 4M)
|
||||
* HistoryTextCacheSize = 128M (default 16M)
|
||||
|
||||
The process numbers are also increased:
|
||||
|
||||
* StartPollers = 30
|
||||
* StartPollersUnreachable = 30
|
||||
* StartTrappersStartTrappers = 15
|
||||
|
||||
Refer to the `Zabbix server <https://www.zabbix.com/documentation/2.4/manual/appendix/config/zabbix_server>`_ official documentation for further details
|
||||
and this `blog entry <http://blog.zabbix.com/monitoring-how-busy-zabbix-processes-are/457/>`_ can be useful to configure the optimal number of Zabbix processes.
|
||||
|
||||
Zabbix agent
|
||||
------------
|
||||
|
||||
The following parameters are set up:
|
||||
|
||||
* StartAgents = 10 (number of processes used to collect data, default 3)
|
||||
* Timeout = 30 (default 3 seconds)
|
||||
|
||||
Refer to the `Zabbix agent <https://www.zabbix.com/documentation/2.4/manual/appendix/config/zabbix_agentd>`_ official documentation for futher details.
|
||||
|
||||
Kernel
|
||||
------
|
||||
|
||||
Since cache related parameters of Zabbix server daemon are increased, Linux kernel
|
||||
has to be configured accordingly.
|
||||
The plugin also configures the maximum shared memory to 1GB (sysctl kernel.shmmax).
|
||||
|
||||
Refer to the `How to configure shared memory <https://www.zabbix.org/wiki/How_to/configure_shared_memory>`_ for further details.
|
||||
|
||||
Links
|
||||
=========================
|
||||
|
||||
|
|
@ -19,6 +68,7 @@ deb packages::
|
|||
zabbix-frontend-php: GPL-2.0
|
||||
zabbix-server-mysql: GPL-2.0
|
||||
zabbix-agent: GPL-2.0
|
||||
zabbix-sender: GPL-2.0
|
||||
|
||||
rpm packages::
|
||||
|
||||
|
|
|
|||
|
|
@ -19,22 +19,42 @@ Fuel 7.0 and higher
|
|||
Release Notes
|
||||
=============
|
||||
|
||||
**2.5**
|
||||
**2.5.0**
|
||||
|
||||
* Add Ceph metrics
|
||||
* Configuration tunning for server and agents
|
||||
* Check processes nova-conductor and nova-novncproxy
|
||||
* Add MySQL cluster metrics (wsrep global variables)
|
||||
* Embed all package dependencies
|
||||
* Enhance :ref:`Ceph` monitoring
|
||||
* :ref:`tuning` for server and agents
|
||||
* Add :ref:`MySQL` cluster metrics (wsrep global variables)
|
||||
* Embed all package dependencies (bug 1483983_)
|
||||
* Fix HAproxy configuration behind the Zabbix VIP (bug 1510115_)
|
||||
* Compatibility with MOS 7.0 *(follow-up)*
|
||||
|
||||
**2.0**
|
||||
* Fix NTP monitoring on controller nodes (bug 1513454_)
|
||||
* Monitor `cinder-volume` process (instead of the Pacemaker resource which has
|
||||
been removed)
|
||||
* Fix trigger for Neutron DHCP/L3 agents (these agents run now on all controllers)
|
||||
|
||||
* New process checks
|
||||
|
||||
* nova-conductor
|
||||
* nova-novncproxy
|
||||
|
||||
* Generate documentation with `Sphinx <http://sphinx-doc.org/>`_
|
||||
|
||||
.. _1483983: https://bugs.launchpad.net/fuel/7.0.x/+bug/1483983
|
||||
.. _1510115: https://bugs.launchpad.net/fuel/+bug/1510115
|
||||
.. _1513454: https://bugs.launchpad.net/fuel-plugins/+bug/1513454
|
||||
|
||||
**2.0.0**
|
||||
|
||||
* Fix HA issue when scaling down/up a controller (bug 1506767_)
|
||||
* Compatibility with MOS 7.0
|
||||
* Disable user Guest in zabbix
|
||||
* Use HTTPS Zabbix UI
|
||||
* Use dedicated VIP for Zabbix server
|
||||
|
||||
**1.0**
|
||||
.. _1506767: https://bugs.launchpad.net/fuel-plugins/+bug/1506767
|
||||
|
||||
**1.0.0**
|
||||
|
||||
* This is the first release of the plugin.
|
||||
|
||||
|
|
@ -43,10 +63,12 @@ Release Notes
|
|||
Limitations
|
||||
===========
|
||||
|
||||
* Prior to version 2.5, the plugin requires access to distribution repository,
|
||||
* If a base-os role node is deployed within the environment, the plugin
|
||||
installation may fail because the management network is not configured
|
||||
(see bug `1515956 <https://bugs.launchpad.net/fuel-plugins/+bug/1515956>`_).
|
||||
|
||||
* Prior to version 2.5.0, the plugin requires access to distribution repository,
|
||||
external or local mirror, in order to download necessary packages for proper
|
||||
installation.
|
||||
With plugin version 2.5, the `fuel-createmirror` command is supported.
|
||||
Since plugin version 2.5.0, the `fuel-createmirror` command is supported.
|
||||
|
||||
* If a base-os role node is deployed within the environment, the plugin
|
||||
installation may fail because the management network is not configured.
|
||||
|
|
|
|||
|
|
@ -149,24 +149,27 @@ several pages placed under Configuration tab.
|
|||
<https://www.zabbix.com/documentation/2.4/manual/config/notifications>`_
|
||||
upon events chapter in the official Zabbix Documentation.
|
||||
|
||||
.. _Ceph:
|
||||
|
||||
Ceph
|
||||
====
|
||||
|
||||
When Ceph is deployed the plugin configures:
|
||||
|
||||
* A Screen 'Ceph' providing an overview of the cluster with the following metrics:
|
||||
* A check on the general health of the cluster as reported by the command `ceph health` *(version 1.0.0)*
|
||||
* A Screen 'Ceph' providing an overview of the cluster with the following metrics *(version 2.5.0)*:
|
||||
|
||||
.. image:: ../images/ceph_screen.png
|
||||
:width: 100%
|
||||
|
||||
* Zabbix is configured to verify that all Ceph processes are running
|
||||
* Zabbix is configured to verify that all Ceph processes are running *(version 2.5.0)*:
|
||||
|
||||
* ceph-osd
|
||||
* ceph-mon (on controller nodes)
|
||||
|
||||
* A Host 'CephCluster' is configured where are attached all cluster metrics:
|
||||
* A Host 'CephCluster' is configured where are attached all cluster metrics *(version 2.5.0)*:
|
||||
|
||||
* Metrics collected for the cluster:
|
||||
* Metrics collected for the cluster *(version 2.5.0)*:
|
||||
|
||||
- monitor_count
|
||||
- quorum_count
|
||||
|
|
@ -207,7 +210,7 @@ When Ceph is deployed the plugin configures:
|
|||
- osd_count_in
|
||||
- osd_count_out
|
||||
|
||||
* Triggers configured:
|
||||
* Triggers configured *(version 2.5.0)*:
|
||||
|
||||
- An alert is triggered when Zabbix fail to collect Ceph cluster metrics
|
||||
- An alert is triggered when the cluster free capacity is under 10%
|
||||
|
|
@ -215,3 +218,22 @@ When Ceph is deployed the plugin configures:
|
|||
.. image:: ../images/host_ceph.png
|
||||
:width: 100%
|
||||
|
||||
.. _MySQL:
|
||||
|
||||
MySQL
|
||||
=====
|
||||
|
||||
The following metrics are retrieved for each MySQL server (generally running
|
||||
on controller nodes):
|
||||
|
||||
.. image:: ../images/mysql_items.png
|
||||
:width: 80%
|
||||
|
||||
and 4 triggers are configured:
|
||||
|
||||
- a simple `mysqladmin ping` check *(version 1.0.0)*
|
||||
- and 3 cluster related checks *(version 2.5.0)*:
|
||||
|
||||
- the cluster node is ready
|
||||
- the cluster node is connected to the cluster
|
||||
- the cluster node status (Primary, Non-Primary or Disconnected)
|
||||
|
|
|
|||
Loading…
Reference in New Issue