Release notes

Update information for NConf 1.2.6 to 1.3.0
Please examine the following release notes carefully. They contain important update instructions.

Changes to the NConf configuration

Added new deployment functionality. To migrate your old config deployment or to set up new deployment rules, please refer to the online documentation: Deployment setup

The following configuration options have been removed:

  • CONF_DEPLOY_URL
  • CONF_DEPLOY_USER
  • CONF_DEPLOY_PWD

The following (optional) configuration options have been added:

  • DEBUG_GENERATE [1-5]
  • CHECK_UPDATE [1|0]
  • CHECK_STATIC_SYNTAX [1|0]
  • AUTH_METHOD [login|basic]
  • BASICAUTH_REALM [text]
Additional changes

New files being generated:

  • host_dependencies.cfg
  • service_dependencies.cfg
  • advanced_services.cfg

Files not generated anymore:

  • extended_host_info.cfg
  • extended_service_info.cfg

Security issue with ‘call_ajax.php’:
Please remove this file from your nconf/ directory. A replacement which is more secure has been implemented (‘call_file.php’).

CAUTION 

The following attributes are DEPRECATED and have been removed in NConf 1.3: Attributes under 'timeperiods':

- max check attempts
- check interval
- retry interval
- notification interval
- host notification options
- service notification options

Attributes under 'nagios-collectors / monitors':
- active checking
- passive checking
- notification enabled
- check freshness
- freshness threshold

We are providing a script that will convert timeperiods, nagios-collector and nagios-monitor servers in the following way: If these items use one or more of the attributes listed above, the script will create host- and service-templates based on the content of the attributes and then link the new templates with the timeperiods, nagios-collectors and monitors. This way, the original attributes will still be applied to hosts and services via e.g. timeperiods, but users will have the ability to override inheritance by setting the same attributes on a host or service level directly.

If you are updating using the update procedure in the GUI, then you will be asked if you want to run the conversion script.
If you are updating manually and wish to run the conversion, then you need to run the following script BEFORE DOING THE DATABASE UPDATE:

$> nconf/bin/convert_timeperiods_collectors.pl

The goal is to ensure that there is no data missing after an update to NConf 1.3 and that the config can be generated as usual. If prior to the update you weren’t using any of the attributes listed below or, if you don’t wish for any auto-created templates, then you do not need to run this script.

Limitations of the conversion script:
Each timeperiod is analyzed by the conversion script and if any of the 5 deprecated attributes are in use, host- & service-templates are generated and linked to the original timeperiod. So far so good.

The tricky thing now is that depending if the timeperiod is used as check_period or notification_period, different combinations of the 5 deprecated attributes are statically copied into the hosts / services. This cannot be reflected by host- or service-templates because the same timeperiod might have been used as a check or notification_period and different attributes will be inherited from it.

The bottom line: Check your hosts and services after the update! The attributes that might be affected by this problem are: max_check_attempts check_interval retry_interval notification_interval notification_options

The conversion script cannot work around this, because there is NO WAY of knowning how a timeperiod will be used: as check_ or as notification_period.

Affected by the problem are:

  • 1.2.6 users upgrading to 1.3.0 who are using any of the 5 deprecated timeperiod attributes AND have differing check_ & notification_period settings for hosts & services (most likely everyone)

Not affected are:

  • pre-1.2.6 users upgrading directly to 1.3.0
  • 1.2.6 users who were using the ExportNagios.pm.PATCH from the ADD-ONS directory
  • 1.2.6 users who always have set the check_period + notification_period to be identical for each host or service