GWME-7.2.1-00 4118 Rollup Patch Details

Overview

The 7.2.1 Rollup Patch installer includes fixes for issues in several areas, including the GroundWork Foundation database, the downtime scheduler, NoMa notification, NOCBoard dashboard, LDAP integration, and also the SLA Dashboard and reporting tools.

Important Directives

Because this rollup patch does affect configuration files, in some cases you must make additions or edits in order to benefit from the new features and fixes in the patch. The following is a list of such edits you may want to make, depending upon how you use GroundWork, and whether you make use of a particular feature.

NOCBoard

1. GWMON-13406 Window for NocBoard Query speed up
   File to edit:

   /usr/local/groundwork/config/foundation.properties
   

   These parameters denote the acceptable size of the search window in time, and affect how much data is cached.
   If you select a bigger period in NocBoard than you have set here, performance will suffer. Make sure the window size is right for your installation.

   ############################################################################
   # LogMessage Window Service Configuration
   ############################################################################
   logmessage.window.enabled=true
   logmessage.window.size.hours=24
   logmessage.window.update.interval.seconds=5
   

Changes to this file are effective after you restart GroundWork Services.

/etc/init.d/groundwork restart gwservices

LDAP Users

1. GWMON-13447 Map LDAP groups to GW Roles (fixes for LDAP mapping issues)
   File to edit:

   /usr/local/groundwork/config/foundation.properties
   

Adding the ldap mapping enable parameter and the "|" or function on the roles container search term extends the places that GroundWork will look for users and role mappings. This can add significant redundancy to your authentication/authorization setup with LDAP or AD.

# Whether LDAP Mapper should be enabled
core.security.ldap.mapping_enabled = true
...
core.security.ldap.config.windows2012.roles_ctx_dn = ou=GWRoles,dc=company,dc=com|cn=other,dc=company,dc=com

File to edit:

/usr/local/groundwork/config/ldap-mapping-directives.properties

You can add the mapping of LDAP only groups to GW portal only groups, for example: "aga", "another aga", "ago", "agu", below.
Notice that a space in an LDAP group name requires the replacement of that space with the string "\u0020"

#Portal admin group mapping
GWRoot=gw-portal-administrator
#groundwork admin group mapping
GWAdmin=gw-monitoring-administrator
#Portal operator group mapping
GWOperator=gw-monitoring-operator
#Portal user group mapping
GWUser=gw-portal-user
agr=GWRoot
another\u0020aga=GWAdmin
ago=GWOperator
agu=GWUser

Changes to these files are effective after you restart GroundWork Services.

/etc/init.d/groundwork restart gwservices

1. GWMON-13501 LDAP Group to Role equivalent recursive tour de mort fix.

You may use a notation for ldap mapping that sets GROUPNAME=GROUPNAME (where the ldap group and the GW role are identically spelled).

The following is a description of the main issues that are addressed by applying the Rollup Patch installer version 7.2.2-gw4118. Note that this list will change as we release new versions. Each patch installer creates a backup directory that contains all the changed files for that patch. You may view the changed files on your system at

 /usr/local/groundwork/backup-gwNNNN/files

where NNNN is the number of the patch prior to the one currently installed. Thus the files changed by patch 4114, for example, will appear when you install patch 4118 in

 /usr/local/groundwork/backup-gw4114/files

Changes from default install made by prior versions

Status Viewer Spinner Changes

• GWMON-13511
  GroundWork Status Viewer screens have a "spinner", or an indication that the web application is updating data. This is now optional. You may set the spinner inactive in status-viewer.properties:

  showConnectionStatus=false
  

  The default will be disabled when you apply 7.2.2-gw4114.

Path Issues with RStools

• GWMON-13512: Can not delete downtimes
  Some scheduled downtime items were still are shown as "Active" even when they are not "recurring". This patch allows you to delete a downtime in any state, at any time.

LogMessageWindow changes for NOCBoard Initialization

• GROUNDWORK-331
  These changes to the sliding window algorithm were needed to allow the NOCBoard to initialize efficiently. You may notice better performance of the NOCBoard after this patch is applied.

Feeder Changes

GroundWork Nagios and Cacti feeder programs have been updated in this patch to fix a few minor issues. Please see: https://kb.groundworkopensource.com/display/DOC721/GWME-7.2.1-02+Nagios+Feeders for a description. The fact that the feeders are patched in the roll-up means that it’s not necessary to apply both the rollup-patch and the TB 7.2.1-02. Here is a list of the changes made:

• GWMON-13428 logging improvements recommended for the event feeder
• GWMON-13429 certain Nagios events are not recorded in Foundation
• GWMON-13419 status feeder dies due to empty timestamp from Foundation
• GWMON-13412 status feeder sometimes sends current timestamp to parent as check time

SLA Dashboard Fixes

• GWMON-13452: Inappropriate formatting of BSM service-check text (minor typo)
• GWMON-13450: SLA Dashboards and the availability graphs show pending for services that are actually OK (major fix for an issue that was apparent on newly added BSM services)
• RSTOOL-431 - Create command "yiic maintenance cleanpastaudittrail --days=$audit_trail_retention_days" (this is the pre-requisite for TB7.2.1-04, and the reason you need the roll-up patch installed before that TB)
• RSTOOL-428 - When Contract has non existing Service then the Event Tab Crashes for all Contracts and the specific Contract
• RSTOOL-430 - Add SafeDown to migration m180711_170923_reprocessflagPrimeryKey
• RSTOOL-388 - Add check box on Dashboard Create/Update screen: Show on dashboard carousel
• RSTOOL-433 - generated TAR contains no directory structure
• RSTOOL-434 - On Configuration Reports/List commands externals are not found when entering search patterns for services or hosts
• RSTOOL-399 - Calendar update/create will not close the dialog after save on empty calendar
• RSTOOL-401 - Slash in the servicename causes trouble with right click on widgets
• RSTOOL-403 - The frame of the widgets must not be shown if its a box, image, text or status widget
• RSTOOL-404 - Add Downtime - Calendar, autocomplete (host and service, hostgroup, servicegroup) and Repeat - JavaScript not working on IE 11

New features on SLA dashboards

• RSTOOL-350 - grouping of widget with right click

Improvements to SLA dashboards

• RSTOOL-391 - Purge Job for downtimeschedule table
• RSTOOL-395 - Remove default Calendars for new Installations - Only "Default" Calendar (without holidays) will be created.
• RSTOOL-408 - State priorities like nagvis
• RSTOOL-411 - Host and hostgroup including service status summary should sort hosts/services in popover also for service status
• RSTOOL-383 - The Cronjob writes to the command pipe (External Command) of nagios even when nagios is not started - This results in downtimes not starting
• RSTOOL-396 - When the cronjob for downtimes can't start (REST API or Database down) downtimes will not be correctly scheduled.
• RSTOOL-397 - Maintenance - Remove Unused Service is throwing an error when the entered Search string starts with a number
• RSTOOL-398 - Maintenance - Service Cleanup, after selecting some services and clicking on "Delete selected" there is an error displayed. The services are not deleted.

NOCBoard Fixes

The relatively new NOCBoard dashboard is documented here: https://kb.groundworkopensource.com/display/DOC721/NOCBoard
This development is still in progress, so we didn’t release 7.2.1 with the NOCBoard enabled. If you do set it up, though, you will appreciate the fixes that went in since release, and are included in the roll-up patch installer:

• GWMON-13383 Ensured default value for availability window is within availability range (minor fix to default value that was not obviously wrong)
• GWMON-13406 Added “window” capability to define scope of improved state transition queries against logmessage records in Foundation

API/Foundation Fixes

These issues were resolved in most cases to enhance performance, but also to support features for acknowledging and commenting on hosts and services from more places in the user interface:

• GWMON-13465 RAPID - Add GW8 nagios_performance_info() API Support
• GWMON-13464 RAPID - Support GW8 Foundation API Conventions
• GWMON-13463 Version REST API Returning Incorrect JSON Payload
• GWMON-13461 hostgroup membership is sometimes not processed correctly by Foundation API
• GWMON-13460 REST call fetch of servicegroup returns no info on a host for each included service
• GWMON-13448 Hostnames Containing Colon Breaking Service Caching

Status Viewer Fixes

Some minor issues with the Status view were addressed:

• GWMON-13442 Availability graphs sometimes show no data
• GWMON-13441 Activity Graphs in Status Viewer Broken When LogMessage Window Enabled
• GWMON-13447 Map LDAP groups to GW Roles (fixes for LDAP mapping issues)
• GWMON-13477 Correct inaccurate % uptime reporting in NOCBoard dashboard

Cloud Hub Fixes

We made some additional features available in the Cloud Hub that the roll-up patch includes. There are some fixes there as well.

• GROUNDWORK-9 Amazon connector: Added support for Elastic Load Balancers

LDAP fixes

• GWMON-13501 LDAP Aggregator Logic fix to allow group name and role name to be identical, when mapping is performed only in the ldap-mapping-directives.properties file.

Monarch fixes

• GWMON-13509 Monarch logic fixes to interpret Service Template and Instance check commands and arguments correctly, as well as to provide proper inheritance/local instance assignment of the template details. Previous code allowed a non check command to be passed to Nagios which would then fail to run anything. Also documented the need to have a "time_period" for the service check other than "None" which by itself disables running Freshness check. This patch includes migration scripting to clear out any improper instance arguments and the like.

Iframe Portlet Issue

• GWMON-13506 The GroundWork Iframe Portlet operation is repaired to address two issues, a message to screen about loading page error and the underlying write method which prompts the error message. Changing to use innerHTML instead of document.write() makes the page write less costly and less likely to hit the error message. Suppressing the error message in the nonsensical case (a very complex page may take a long time to paint) completes the fix.

Changes between 4114 and 4118

Aliasing Changes

• GWMON- The aliasing feature is used when combining services on hosts from both Cloud Hub and Nagios/GDMA sources. This feature could sometimes fail to properly account for merged host names, but now does so. If you ever got unexpected results in the use of aliases, it's likely this issue was the cause.

Unified Monitoring Fixes

• GWMON- In the case of merging of services on hosts with the same name across multiple monitoring app types (or aliasing), there is a possible conflicting update. While rare, if this occurs it generates potentially many error events in the event console, similar to:

  Cannot update service from different application: host001:service_aaa owned by NAGIOS update ignored from VEMA
  

  This was due to an issue with merging hosts incompletely, and a more complete approach is now used. These messages will no longer occur after this patch is applied.

Jolokia Support

• GWMON- GroundWork can now monitor java JMX counters with a nagios plugin called check_jmx4perl, which uses the jolokia agent installed on a running java app server (e.g. Tomcat). This patch includes support for the plugin and its dependencies on the GroundWork server. Note that GDMA support is also pending in the next iteration of GDMA. When this is done, the plugin can be run from GDMA installs in both Windows and Linux. This work is an enhancement of the basic plugin, check_jmx, which has more limited capabilities. We are using jolokia/check_jmx4perl as a replacement for the deprecated JDMA agent.

Deletion of Downtime

• GWMON-13512 Downtimes created via the downtime user interface could not, in some cases, be removed. This is now fixed. RSTools 1.13.1.25 contains the fix.

BSM Message is now configurable.

• GROUNDWORK-492 You can now build your own notification messages from BSM services. To format you can add HTML (inline CSS and JS allowed). You can use these Tags:
  [DISPLAY], [DESCRIPTION], [NOTE], [INFOURL], [INFOTEXT], [OUTPUT], [LONGOUTPUT] AND Status.

Performance Enhancements for SLA Dashboards

• RSTOOLS-452, 462 These enhancements add performance to the SLA dashboard and downtime processing.