ID: 8309
Title: Fix authorization settings for seeing service in the user interface
Component: The Check_MK Micro Core
Level: 2
Class: Bug Fix
Version: 1.2.7i4
The behaviour of the authorization setting <i>loose</i> for hosts has been
changed to be compatible with the behaviour Nagios has. If host authorization
is set to <i>loose</i> in the global settings of the Check_MK Micro Core, then
it is now suffient to be the contact of a host for seeing all of its service
- even if those services do have explicit <i>contacts</i> assigned. Formery
you needed to be an explicit contact if one had been set. In the case that
you just have contacts assigned to hosts nothing has changed.
The behaviour of the authorization setting <i>strict</i> has also changed.
If you are the explicit contact of service, but not of the host, then you
are allowed to see the service.
<b>Note</b>: This all has <b>no</b> impact on notifications. Nothing has
changed there.
ID: 8310
Title: Fix exception in report schedule in case of reports with special contexts
Component: Reporting
Level: 2
Class: Bug Fix
Version: 1.2.7i4
If you have a report of context type "single Event Console Event" and then in the
scheduler add an event id to the context of the report you would run into an
exception saying <tt>object of type 'int' has no len()</tt>. This has been fixed.
ID: 8342
Title: Dashboard: Fixed new style graph rendering glitches
Component: Metrics System
Level: 1
Class: Bug Fix
Version: 1.2.7i4
When moving new style graphs on the dashboard, they were scaled in a strange
way during movement. Only trying to resize the dashlet after move fixed this.
When having multiple new style graphs on a single dashboard, the contents
of single dashlets could be destroyed by interacting with another dashlet.
ID: 8341
Title: Closing SNMP socket file handles during bulk actions
Component: Inline-SNMP
Level: 1
Class: Bug Fix
Version: 1.2.7i4
e.g. when doing "cmk -II" for a larger number of hosts, it could happen
that the process runs out of file descriptors when having a lot of SNMP
devices. This was caused by the fact that all SNMP sessions were kept
open during this bulk actions. The sessions are closed now after each
host during most common bulk actions (inventory, discovery, snmpwalk).