ID: 6702
Title: Introduced various performance improvements for cmc config generation (e.g. multiprocessing)
Component: cmc
Level: 2
Class: New feature
Version: 1.6.0i1
Previous versions only used one CPU core for the config generation.
The time to generate the config highly depends on the number of hosts, services and especially rulesets.
Lots of the underlying computation code has been changed, caches and functions were optimized.
As a result Check_MK is now able to distribute the work load of the config generation over several CPUs.
Per default, this feature is activated. Unless configured otherwise, it uses up to 75% of the available
CPUs during the config generation, leaving some CPUs for running monitoring core.
A new configuration option <tt>Generate monitoring configuration via multiprocessing</tt> has been introduced.
You can either switch of multiprocessing or configure the number of used CPUs manually.
Tests have shown that the performance can increased by a factor of 5-10 on a 8 CPU core setup.
ID: 7055
Title: jolokia_generic: remove duplicate MBean in item
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
Werk 6901 introduced an additional word "MBean" in the item of
jolokia_generic checks. This changed the service description
to "MBean text1 MBean text2". This werk reintroduces the old
service description "MBean text1 text2".
If you have services that contain two times the word "MBean"
a service rediscovery is necessary to obtain the old format.
ID: 6657
Title: Kubernetes checks for node Network and Filesystem usage
Component: Checks & agents
Level: 1
Class: New feature
Version: 1.6.0i1
Kubernetes special agent obtains the node statistics of network and filesystem usage for each node. Corresponding checks are available at node or cluster level.
ID: 7075
Title: Fixed possible "omd backup" error when persisted sections change during backup
Component: Site Management
Level: 1
Class: Bug fix
Version: 1.6.0i1
When persisted sections of Check_MK in <tt>var/check_mk/persisted/*</tt> and
<tt>var/check_mk/persisted_sections/*</tt> vanished during the backup, the backup
job could fail in previous versions.
ID: 6935
Title: Improvements in Solaris Agent packages
Component: agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
This werk introduces some improvements in the postinstall- and postuninstall-routines
for Solaris agent packages that come from the agent bakery:
<ul>
<li>Properly remove the installed inetd service. This has been done incompletely
before, which could lead to problems on reinstallation.</li>
<li>simplified syntax in order to avoid errors like <code>/var/sadm/pkg/check-mk-agent/install/postinstall: !: not found</code>
on some Solaris systems.</li>
<li>some improvements in timing and robustness.</li>
</ul>
ID: 6939
Title: Improve agent updater start script compatibility on Solaris
Component: agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
The agent updater start script (located at /usr/bin/cmk-update-agent)
is now compatible to older shells like the default sh on Solaris 10.
Please note that you still need a python 2.7 installation and corresponding
packages "requests" and "pyOpenSSL" in order to run the agent updater
on Solaris. You can obtain these software packages e.g. via the
"OpenCSW" project using the "pkutil" tool.
ID: 7054
Title: agent_vsphere, esx_vsphere_hostsystem.multipath: use new API endpoint
Component: Checks & agents
Level: 1
Class: New feature
Version: 1.6.0i1
For ESX versions above or equal to 6.5 the API endpoint "multipathState"
does not deliver sufficient data for the subcheck esx_vsphere_hostsystem.multipath
which may lead to a wrong output of the subcheck. Therefore, the subcheck now
uses the "multipathInfo" API endpoint instead.
For ESX versions below 6.5 which already use the ID of a LUN as an item
no further actions are necessary. For new ESX version which use an item
like e.g. "L0 physical" a service rediscovery is necessary. After the
rediscovery new ESX versions will use the ID of a LUN as an item as well.
ID: 7131
Title: win.mem: Fix incorrect representation of absolute memory levels in graphs
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
Previously, a warn or crit level of less than e.g. 20 GB free space would be displayed
in the graph as a warn or crit level of more than 20 GB used space.
The problem only affected the levels passed to the metrics system for representation,
not the actual levels used for evaluation.
ID: 6756
Title: oracle_instance: Fixed false-positive if the output shows the creation time of instance
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
Previously the categorized as stopped or login failed or invalid data if
the output showed 12 elements per instance. The last element has been the
creation time of the database, which is currently not used in the created
service. This problem is solved know.