ID: 7088
Title: omd update can now be performed without access to source version
Component: Site Management
Level: 2
Class: New feature
Version: 1.6.0i1
The "omd update" procedure, which is used to perform a version update for a
site, always needed access to the previous version, the new version and the
site.
Since this change it is now possible to perform an update without access
to the old version. This is possible becase we copy the information that
are needed to the Check_MK site during site creation.
This new mechanism can only be used when updating FROM a site that already
implements this werk.
Technical detail:
<ul>
<li>The meta files are saved for the first time during "omd create"</li>
<li>The meta files are updated to the new version during each "omd update"</li>
<li>The files are copied to the sites <tt>.version_meta</tt> directory.</li>
<li>The directory <tt>/omd/versions/[version]/skel is copied.</li>
<li>The file <tt>/omd/versions/[version]/share/omd/skel.permissions is copied.</li>
<li>A file <tt><tt>.version_meta/version</tt> is created.</li>
<li>When an update is performed, the meta files are used if they are available
and up-to-date. In case they don't fit these conditiones, the previous version
files need to be available as before this werk.</li>
<li>
ID: 7164
Title: cmk-update-agent: Add timeout handling for connection to server
Component: agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
Previously, the connection from the agent updater to the Check_MK Server did
not handle timeouts. This could lead to situations where the agent updater
would run forever waiting for a response.
This was particularly noticeable when running the agent updater under Windows,
because the agent updater would fail to start over again, showing an error
like <code>copying plugins\cmk-update-agent.exe to temp\cmk-update-agent.exe failed</code>.
The agent updater now will cancel the connection if there is a connection timeout detected
or if the connection freezes. The timeout limit is set to 120 seconds.<br>
This timeout handling does not limit the overall connection time.
ID: 7143
Title: mk_docker.py: A new docker monitoring plugin
Component: Checks & agents
Level: 1
Class: New feature
Version: 1.6.0i1
We introduce a new plugin for docker monitoring, 'mk_docker.py'.
It combines the functionality of the two deprecated plugins
'mk_docker_node' and 'mk_docker_container_piggybacked', while
adding more configuration options.
If you where using both plugins, you can use the new plugin in its
default configuration to replace them. If you have been using only
one of them, you can disable the opposite parts of the new plugin,
either by using the bakery rule "Docker node and containers" or via
the configuration file (please refer to the example files in
agents/cfg_examples fo rmore details).
The previous rule "Piggybacked docker containers" is deprecated -
please set up a corresponding new one.
To deploy this plugin you must have the 'docker-py' python library
installed on the monitored system. You can install it using the
python package manager pip (using the shell command
"pip install docker-py").
ID: 7085
Title: Fixed parsing of special syslog messages which don't contain a host name
Component: Event Console
Level: 1
Class: Bug fix
Version: 1.6.0i1
The Event Console is now able to process syslog messages that don't contain the
host name field. An example for such a message is this one:
C+:
Feb 13 08:41:07 pfsp: The configuration was changed on leader blatldc1-xxx to version 1.1366 by blatldc1-xxx/admin at 2019-02-13 09:41:02 CET</tt>
C-:
In previous versions messages like this resulted in log messages like this in
the event console log (var/log/mkeventd.log):
C+:
2019-02-13 09:41:07,338 [40] [cmk.mkeventd.EventServer] Got non-syslog message "Feb 13 08:41:07 pfsp: The configuration was changed on leader blatldc1-xxx to version 1.1366 by blatldc1-xxx/admin at 2019-02-13 09:41:02 CET" (need more than 1 value to unpack)
Traceback (most recent call last):
File "/omd/sites/ggmcmpp1/lib/python/cmk/ec/main.py", line 2916, in create_event_from_line
event.update(self.parse_syslog_info(rest))
File "/omd/sites/ggmcmpp1/lib/python/cmk/ec/main.py", line 2667, in parse_syslog_info
tag, message = line.split(": ", 1)
ValueError: need more than 1 value to unpack
2019-02-13 09:41:07,338 [20] [cmk.mkeventd.EventServer] Parsed message:
application:
core_host:
facility: 1
host:
host_in_downtime: False
ipaddress: 1.23.45.67
pid: 0
priority: 0
text: Feb 13 08:41:07 pfsp: The configuration was changed on leader blatldc1-xxx to version 1.1366 by blatldc1-xxx/admin at 2019-02-13 09:41
:02 CET
time: 1550047267.34
C-:
A fallback event was created that had no syslog fields set and contained the
whole unparsed syslog message in the text field. If you have EC rules matching
on this fallback event, you will have to change these rules to match the parsed
event fields.
Now that the parsing has been added, events created by such a syslog message now
have the fields set as follows for our example:
C+:
application: pfsp
core_host:
facility: 1
host: 127.0.0.1
host_in_downtime: False
ipaddress: 127.0.0.1'
pid: 0
priority: 5
text: The configuration was changed on leader blatldc1-xxx to version 1.1366 by blatldc1-xxx/admin at 2019-02-13 09:41:02 CET
time: 1550043667.0
C-:
Please note that the EC uses the sender IP addresse of the syslog message to populate the host field.
ID: 6705
Title: agent_netapp: fixed typo in agent which caused the agent to fail on bigger netapp server responses
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
ID: 7154
Title: OpenSSL update on our build platform
Component: Site Management
Level: 1
Class: Security fix
Version: 1.6.0i1
Centos 6, Suse Enterprise 11 SP2 and SP3 updated OpenSSL. Therefore, we updated our build systems and python packages. In case you experience SSL problems on one of these systems, please update your OS.
ID: 7097
Title: diskstat.include: Read/write latency are configurable now
Component: Checks & agents
Level: 1
Class: New feature
Version: 1.6.0i1
Nearly all check plugins which use the diskstat.include in order to check
disk IO provide an average value for the disk latency. Newer checks may
provide values for disk read and write latency.
ID: 6704
Title: Fixed problem where persistent information of hosts was randomly missing, causing services to go stale
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
ID: 6703
Title: Remove SLA painter from default services view
Component: Multisite
Level: 1
Class: New feature
Version: 1.6.0i1
The SLA information is no longer shown per default, since its empty most of the time.
If you want to reenable this painter, just edit the view and add one of the following columns.
<ul>
<li>Host/Services: SLA column specific</li>
<li>Host/Services: SLA service specific</li>
</ul>
Keep in mind that you still have to configure a SLA configuration before this column can display any meaningful data.