Sophie

Sophie

distrib > Mandriva > 2008.1 > x86_64 > media > contrib-backports > by-pkgid > 4f795a1da021b02603401d4d712e1cc7 > files > 37

nagios-3.0.6-3mdv2008.1.x86_64.rpm

#####################
Nagios 3.x Change Log
#####################


3.0.6 - 12/01/2008
------------------
* Fix for CGI submission of external commands (writing newlines and submitting service comments)
* Fix for Apache group membership in RPM spec file
* Fix for improper notification propagation command processing
* Better out-of-disk-space error handling when writing retention and status files
* Disabled adaptive check and eventhandler commands for security reasons
* Fix for reading output from system commands (event handlers, etc) that have timed out
* Added wildcard host matching in CGIs
* Fixes for playing audio alerts in CGIs
* Fix for incorrect host status links in status CGI when viewing hostgroup summary
* Added support for x509 cert authentication in the CGIs



3.0.5 - 11/04/2008
------------------
* Security fix for Cross Site Request Forgery (CSRF) bug reported by Tim Starling. 
* Sample audio files for CGIs removed from distribution
* Fix for mutliline config file continuation bug
* Minor fix to RPM spec file
* Fix for AIX compiler warnings
* Minor sample config file fix
* Added documentation on CGI security issues



3.0.4 - 10/15/2008
------------------
* Fix for properly terminating plugins when parent processes get killed (e.g. using 'killall nagios' with check_timeout plugins running)
* Fix for event broker callback when service notifications are disabled
* Fix for scheduling scheduling servicegroup downtime with 'hosts too' option in CGIs
* Fix for segfault under Solaris with NULL plugin output
* Fixes for bugs in sample event handlers - stop/start active service checks and enable notifications
* Cosmetic fix for logging of notifications
* Fix for high CPU utilization under OS X
* Fix for host/service name encoding in CGIs (trends and availability reports, etc.)



3.0.3 - 06/25/2008
------------------
* Typo fix in object config routines
* Segfault fix in history CGI
* Fix for bug in retention of contact notification options
* Fix for bug in calculation of negative (month) days in timeperiod definitions
* Fix for debug logging of notifications
* Fix for encoding host and servicegroup names in trends and availability CGIs
* Fix for checking for abnormal termination of host checks
* Fix for spurious orphaned host check errors with non-scheduled hosts
* Fix for url encoding null string in CGIs (caused status CGI segfault)
* Fix for encoding URLs in macros
* Fix for bug in handling of contact groups in escalation definitions
* Changes to service check event broker logic (DNX patches)
* Minor doc updates



3.0.2 - 05/19/2008
------------------
* Minor bug fixes in CGIs to ensure extra host/servicegroup url strings are terminated properly
* Minor bug fix in navigation frame link for unhandled service problems
* Better error logging during fork() errors
* Embedded Perl is now disabled by default
* Fixed bug in parsing host dependencies
* Added note to Makefile about 'make install-webconf' option
* Fixed bug in config CGI where active host check attributes were not displayed properly
* Fixed bug in status CGI where sounds were not played for passive service problems
* Fixed sample script for distributed monitoring
* Updated p1.pl to allow for 4KB lines in Perl plugin output under epn
* Fixed bug in command for disabling contact notifications
* Fix for bugs in host and service orphan check logic
* Fix for 'make install' functionality for contrib directory
* Fix for host problem links in CGI status summary view
* Fix for properly escaping macros containing URLs
* Patches for possible XSS vulnerability in CGIs (CVE-2007-5803) - Florian Weimer & SUSE Linux team



3.0.1 - 04/01/2008
-------------------
* Fixed bug in trends CGI with user not being authorized for hosts/services
* Fixed bug in status CGI with page layout messing up when sounds enabled



3.0 - 03/13/2008
-------------------
* Faster program startup times (especially in large installs) with new object lookup code
* Fix for special contact inheritance rules in host and service escalations
* Fix for threading-related bugs that resulted in checks failing under NetBSD
* Configure script fix
* Fix for bug in processing $CONTACTADDRESSx$ macros
* Documentation fixes



3.0rc3 - 02/26/2008
-------------------
* Fix for passive service check stats bugs in nagiostats
* Nagios no longer warns or errors when encountering host/contact/service groups with empty memberships
* Debugging info fix in notification logic
* Fix for multiline support in config files (again)
* Fix for stripping illegal chars in output, perfdata macros
* Patch to allow non-ASCII characters in notifications, etc.
* Fix for bug in on-demand service macros with blank/null host names
* Fix for cleaning up zombie child processes if large installation tweaks disabled and fork twice option disabled
* Fix for incorrect total checks counts in nagiostats
* Fix for segfault if embedded Perl support compiled in, but disabled by config file option
* Smarter host check logic to reduce (uncached) on-demand host checks during "stable" service problems
* Fix for recursive subgroup membership for host/service/contact groups
* Fix for renaming/moving files on network file systems
* New flap detection startup logic speedups when large installations tweaks enabled
* Speed improvements during startup routines when using precached config
* Speed improvements in reading retention data during startup
* Safer loading mechanism for NEB modules to prevent segfaults in Nagios
* Fix for segfault in CGIs when converting object names with extended ASCII characters to URLs



3.0rc2 - 01/29/2008
-------------------
* Changed embedded audio MIME types in CGIs to "application/wav"
* Fixed bug in sample helloworld NEB module
* Modified p1.pl to allow for multiline-line output of Perl plugins under embedded Perl interpreter (epn)
* Fix for incorrect environment variables names for custom host and contact macros
* Fix for external command file thread polling code
* Fix for cfg_dir directive not working on Solaris
* Fixed segfault in extinfo CGI when passed an invalid service description/name
* Fixed bug in summary CGI with reports on specific hosts
* Fix for writing check result files under Cygwin
* Fix for not building sample helloworld.o event broker module if event broker is disabled by configure script
* Fix to sample event handler scripts
* Fix for handling plugin output with escaped newlines and backslashes
* Fix for bug in service and host event scheduling logic
* Fix for reversed object sort order when using fast startup options
* Fix for bug with notification commands not being run if notifications were not logged



3.0rc1 - 12/17/2007
-------------------
* Fixed bug where status log was left after a restart with faulty config
* Fix for compilation bug on systems that don't support setenv() (e.g. Solaris)
* Support for line continuation/breaks in config files - end lines with one backslash (\) to continue on next line
* Fixed bug with not deleting old check result files that contained results for invalid host/service
* Fixed bug with init script option to check Nagios config 
* Auto-filled comment/author fields for acknowledging problems through WAP interface
* Fixed bug with processing of CONTACTGROUPNAMES, NOTES, NOTESURL, and ACTIONURL macros
* Doc fix regarding soft state dependencies
* Fix for segfault in event broker module code



3.0b7 - 11/23/2007
------------------
* Memory leak fix in macro code
* Added use_timezone option to allow Nagios instances to run in non-native timezones
* Fix for encoding HTML characters in CGIs
* Fix for unimplemented $LASTHOSTSTATE$, $LASTHOSTSTATEID$, $LASTSERVICESTATE$, and $LASTSERVICESTATEID$ macros
* Fixes for memory leaks with passive check, performance data routines
* Makefile fixes
* Code cleanup, fixes for compiler warnings
* Fix to prevent FLAPPINGSTART notifications from incorrectly being resent after program restarts
* Added free_child_process_memory and child_processes_fork_twice options for performance tweaks
* Fix for bug in processing of on-demand service macros



3.0b6 - 10/31/2007
------------------
* Logging API revamp, cleanup
* Misc internal performance improvements
* Bug fix with attempting to access an uninitalized mutex if external commands are disabled
* Updates to .cvsignore files
* Fix for embedded Perl interpreter to recache scripts when they're modified
* Fix for cancelling inheritance with 'null' in various object directives
* Additional functionality of additive inheritance feature
* Fixed bug where missing host/servicegroups in object definitions didn't always generate errors
* Increased max plugin output length cap from 4kb to 8kb
* Caching of Perl scripts now enabled by default if embedded Perl interpreter is compiled in
* Fix for segfault when contact definitions had no notification timeperiods defined
* Retention of host/service check scheduling options
* Fix for race condition when freshening host and service check results
* Added null -b option to install-sh for Solaris so install scripts don't break
* Added .gitignore files for git users
* Added new external commands to change notification period for host, services, and contact on-the-fly
* Added new external commands to change modified object attributes (affecting data retention logic)
* Added $ISVALIDTIME:$ and $NEXTVALIDTIME:$ on-demand macros
* Added enable_environment_macros option to determine whether or not macros are set as environment vars
* Major overhaul (read: complete rewrite) of macro code, so macros are now only computed when found
* Summary macros are now available as regular (non-environment) macros if requested when large installation tweaks are enabled
* Bug fix for scheduling hostgroup host downtime through web interface
* Better error logging when failing to rename/move files
* Added $LASTHOSTSTATE$, $LASTHOSTSTATEID$, $LASTSERVICESTATE$, and $LASTSERVICESTATEID$ macros
* Addition of object pointers to event broker module data structures (may require event broker module rebuilds)
* Spec file changes to better support RPM builds on 64-bit systems
* Possible fix for Unicode html encoding errors in CGIs



3.0b5 - 10/08/2007
------------------
* Minor bug fix in CSS files
* Bug fix in sample config files
* Fix for recovery notifications not being sent out when first_notification_delay option used in host/service definitions
* Fix for ochp commands not running properly
* Special additive inheritance rule: escalation contact(group)(s) that begin with '+' will  now inherit from associated host/service definition if not inheriting from another escalation template
* Status file is no longer deleted during restarts due to a SIGHUP or external command
* Improvement in service flap detection logic
* Added additional_freshness_latency config file directive
* Improvements in freshness check logic
* Temporarily removed $SERVICEGROUPMEMBERS$ macro, as it was causing crashes for some users (this need further investigation)
* Added $EVENTSTARTTIME$ macro to indicate time Nagios started processing events (checks, etc.)
* Added use_pending_states option to CGI config file to determine displayed state of hosts/services that have not been checked
* Workaround to prevent browsers from incorrectly using cached statusmap image on page refresh
* Fix for incorrect latency calculations for passive host/service checks
* Added passive check latency stats to nagiostats
* Added support for relative paths in config_file and config_dir directives
* Fix for compile error under Solaris



3.0b4 - 09/27/2007
------------------
* New macros: $HOSTGROUPMEMBERS$, $HOSTGROUPNOTES$, $HOSTGROUPNOTESURL$, $HOSTGROUPACTIONURL$, $SERVICEGROUPMEMBERS$, $SERVICEGROUPNOTES$, $SERVICEGROUPNOTESURL$, $SERVICEGROUPACTIONURL$, $CONTACTGROUPALIAS$, $CONTACTGROUPMEMBERS$, $NOTIFICATIONRECIPIENTS$, $NOTIFICATIONAUTHOR$, $NOTIFICATIONAUTHORNAME$, $NOTIFICATIONAUTHORALIAS$, $NOTIFICATIONCOMMENT$
* Removed host/service downtime author/comment macros introduced in 3.0b2 in favor of more generic $NOTIFICATION...$ macros
* Fix for segfault when cancelling active scheduled host downtime
* Macro code cleanup 
* Added on-demand contact and contactgroup macro support
* More complete (but still partial) support for macros in CGIs ($xNOTES$, $xNOTESURL$, and $xACTIONURL$ macros)
* Fixed bug in config CGI with displaying incorrect notification interval for escalations
* Added new 'check' option to init script to verify configuration
* Added custom host and service notifications, with option to force the notifications and broadcast them to all contacts
* Fix for on-demand/cached host check statistics
* Fixed bug where null host check command would cause high CPU utilization
* Alias, if not specified, now defaults to object name in host, hostgroup, servicegroup, contact, and contactgroup definitions
* Fixed bug with excluding hosts, services, and contacts with bang (!) in object definitions
* Fixed bug in nagiostats with NULL mrtg data arguments printing bogus string
* Added custom delimiter option in nagiostats output (useful for CSV output)



3.0b3 - 08/30/2007
------------------
* Minor bug fix for debug logging of macros
* Version number is now printed in CGI pages
* HTML documentation cleanup (HTML Tidy, link checking, etc.) 
* Fixed bug where notifications would not be sent out host/service contact group members



3.0b2 - 08/29/2007
------------------
* Fix for some sample config files getting installed with .sample extension
* Improvements to the host check logic performance (more use of cached and parallel checks)
* Minor bug fix with notification timeout error messages
* Fix bug with logging of passive host and service check results
* Fixed bug with warning about no contactgroups being defined
* Internal data structure cleanups
* New macros: $SERVICEISVOLATILE$, $TOTALHOSTSERVICES$, $TOTALHOSTSERVICESOK$, $TOTALHOSTSERVICESWARNING$, $TOTALHOSTSERVICESUNKNOWN$, $TOTALHOSTSERVICESCRITICAL$, $HOSTDOWNTIMEAUTHOR$, $HOSTDOWNTIMEAUTHORNAME$, $HOSTDOWNTIMEAUTHORALIAS$, $HOSTDOWNTIMECOMMENT$, $SERVICEDOWNTIMEAUTHOR$, $SERVICEDOWNTIMEAUTHORNAME$, $SERVICEDOWNTIMEAUTHORALIAS$, $SERVICEDOWNTIMECOMMENT$
* Added 'lock_author_names' option to CGI config file to prevent alteration of author names
  when submitting comments, scheduled downtime, etc.
* Fix for concatentation of multiline plugin perfdata
* Added status CGI host/service property filters for hard and soft states



3.0b1 - 07/31/2007
------------------
* Fixed bug with processing epn directives in Perl plugins
* Fixed bug with check_result_path config option being ignored
* Added $MAXHOSTATTEMPTS$ and $MAXSERVICEATTEMPTS$ macros
* Fixed bug with incorrect output returned from embedded Perl plugins
* Fixed bug where status data file was not read by CGIs using mmap()
* Fix for CGI segfault
* Program status now updated at least every 5 seconds for addons that watch NDOUtils DB
* Added escape_html_tags option to CGI config file to escape HTML tags in plugin output
* Added optional integration with Splunk into the CGIs
* Added new action and notes URL target frame options to CGI config file
* Added new 'exclude' option to timeperiod definitions for easy on-call rotation definitions



3.0a5 - 06/19/2007
------------------
* Fixed problem with multiline long plugin output and Perl plugin output
* Fixed compilation problem with embedded Perl
* More debug/trace log functionality
* Added new 'passive_host_checks_are_soft' config option
* Fixed bug with host notifications occurring during soft host states
* Fixed bug in processing multiple check results contained in a single file
* Expanded timeperiod definitions to allow for day/date exceptions
* Changes to sample config files and installation location
* Changed debug file format to include current pid
* Added 'initial_state' option to host and service definitions to allow for non-UP/OK initial states
* Minor changes to freshness threshold calculation
* Documentation updates



3.0a4 - 05/08/2007
------------------
* Fix for segfault fix when low priority event queue's only item is removed
* Added test for writeability of temp directory during configuration test
* Fix for unescaping backslashes in plugin output in the CGIs
* Removed check_result_buffer_slots option from main configuration file
* New IPC method for passing host/service check results back to main daemon
* Added check_result_path option to set path used to store check results
* Added max_check_result_file_age option to control processing of check results in older files
* Added new --with-temp-dir option to configure script
* Removed legacy 2.x host check logic and use_old_host_check_logic option
* Removed DEBUGx options from configure script
* Added new debug/trace log functionaltiy (only partially complete)
* Fixed compilation error under OSX
* Fix for SIGTERMs being seen as SIGEXITs, non-logging of SIGTERMs/shutdowns



3.0a3 - 04/10/2007
------------------
* Configure script fix for no mail program found/installed on system
* Fix for compilation warnings about implicit declaration of round() - may (strangely enough) also
  fix segfaults that were occuring on some systems
* Added detection of Apache conf.d directory and 'make install-webconf' functionality
* Configure script option bug fixes for cygwin and embedded perl
* Added buffer stats and check statistics to performance information in extinfo CGI
* Hostgroup and servicegroup definitions now have notes, notes_url, and action_url directives
* Patch for incorrect time down percentage in availability CGI
* Command definitions and host/service plugin perfdata with HTML should now be escaped in CGIs
* Updated init script to fix a race condition during restarts
* Documentation updates



3.0a2 - 03/27/2007
------------------
* Added warning about aggregate_status_updates directive being deprecated.
* Added new "p" option to host/service_perfdata_file_mode directives for pipes.
* Fix for incorrect performance data file write/append mode options
* Regular expression matching in object config files now also triggered by '+' and '\.'
* Added support for extended regular expression matching in object config files
* Fix for incorrect processing of y_2d coordinates for host objects in statusmap layout
* Fix for current status of hosts with no check command defined
* SIGSEGV signals should now be logged again (broken in 3.0a1)
* Added warning for invalid temp_path directory
* Documentation bug fixes and updates



3.0a1 - 03/06/2007
------------------
Here are all the changes since Nagios 2.x:

* Adaptive monitoring:
     - Check timeperiod can now be modified on-the-fly

* Notification changes:
     - Added first_notification_delay to host and object definitions to delay first notification
     - Notifications are now sent out when flap detection is disabled on a host/service-specific
       or program-wide basis ($NOTIFICATIONTYPE$ macro = "FLAPPINGDISABLED")
     - Notifications can now be sent out when scheduled downtime starts, ends, and is cancelled.  The
       $NOTIFICATIONTYPE$ macro will be set to "DOWNTIMESTART", "DOWNTIMEEND", or "DOWNTIMECANCELLED",
       respectively.  In order to receive downtime notifications, specify 's' or 'downtime' in contact, host,
       and/or service notification options.

* Object changes: 
     - Added ability to use multiple template names (separated by commas) in object definitions
     - Added ability to null out optional character directives in object definitions (using 'null' without quotes)
     - Added hostg/service/contactgroup_members directives to host/service/contactgroup definitions,
       respectively, for allowing including host, service, or contact members from "sub" groups.
     - Added support for custom variables in host, service, and contact definitions
     - Added host_notifications_enabled, service_notifications_enabled, can_submit_commands to contact definitions
     - Added optional display_name directive to host and service definitions
     - Removed serviceextinfo definitions and merged variables with service definitions
     - Removed hostextinfo definitions and merged variables with host definitions
     - Services inherit contactgroups, notification interval, and notification period from associated host if not specified
     - Service escalations inherit contactgroups, notification interval, and escalation period from associated service if not specified
     - Host escalations inherit contactgroups, notification interval, and escalation period from associated host if not specified
     - Host, service, host escalation, and service escalation definitions now support a 'contacts' directive, along with 'contact_groups'
     - Service dependencies with blank dependent host/hostgroup names will create "same host" dependencies


* Performance optimizations:
     - Added ability to precache object config files
     - Added ability to exclude object relationship and circular path
       checks from verification process

* Check output:
     - Multiline output support for host and service checks

* Macro changes:
     - Added $LONGHOSTOUTPUT$ and $LONGSERVICEOUTPUT$ macros
     - Added $TEMPPATH$ macro
     - Removed $COMMENTDATAFILE$ and $DOWNTIMEDATAFILE$ macros
     - Added $HOSTDISPLAYNAME$ and $SERVICEDISPLAYNAME$ macros
     - Custom host/service/contact macros accessible via $_HOST<varname>$, $_SERVICE<varname>, or $_CONTACT<varname>$.
       On-demand host/service macros for custom vars are working.  Custom vars are also set as environment vars.
     - On-demand service macros can contain and empty host name field.  In this case, the name of the current host will be used.
     - $HOSTNOTES$ and $SERVICENOTES$ macros may now contain macros themselves

* Flapping changes:
     - Added flap_detection_options directive to host and service definitions to control which
       states (i.e. OK, WARNING, UNKNOWN, and/or CRITICAL) are used in flap detection logic
     - Percent state change and state history are now retained and recorded even when host/service
       flap detection is disabled.
     - Hosts and services are immediately check for flapping when flap detection is enabled program-wide.
     - Hosts/services that are flapping when flap detection is disabled program-wide are now logged

* External command changes:
     - Added PROCESS_FILE external command to allow processing of external commands found in
       an external (regular) file.  Very useful for passive checks with long output or scripting.
     - Custom commands can now be submitted to Nagios.  Custom command names are prefixed with an underscore
       and are not processed internally by the Nagios daemon.  They may, however, be processed by NEB modules.
     - External commands are now checked by default.  Nagios 2.x and earlier did not check for external commands
       by default.

* Status data changes:
     - Contact status information is now saved (although it is not processed by the old CGIs)

* Retention data changes:
     - Contact status information is now retained across program restarts
     - Comment and downtime IDs are now stored across program restarts, should be unique unless reset
     - Added retained_host/service_attribute_mask variables to control what host/service attribs are retained globally
     - Added retained_process_host/service_attribute_mask variables to control what process attribs are retained
     - Added retained_contact_host/service_attribute_mask variables to control what contact attribs are retained globally

* Downtime changes:
     - Scheduled downtime entries are now stored in the status and retention files

* Comment changes:
     - Comments are now stored in the status and retention files
     - Non-persistent acknowledgement comments are now deleted when the acknowledgement is removed (instead of when Nagios restarts)

* Host check logic changes:
     - Most host checks are now run asynchronously, in parallel!
     - Scheduled host checks now help improve performance, instead of hinder it (if caching is enabled)
     - Added cached_host_check_horizon option for enabling use of cached host check results
     - Added enable_predictive_host_dependency_checks for enabling predictive checks of dependent hosts
     - Added retry_interval to host definitions
     - Added check_for_orphaned_hosts option and support for orphaned host checks
     - Passive host check states can now be translated from their original DOWN/UNREACHABLE state if the
       new translate_passive_host_results option is enabled

* Service check logic changes:
     - 'parallelize' option in service definitions deprecated/no longer used
     - Added cached_service_check_horizon option for enabling use of cached service check results
     - Added enable_predictive_service_dependency_checks for enabling predictive checks of dependent services

* Dependency changes:
     - Host and service dependencies can now have a timeperiod during which they're valid (dependency_period directive)

* Event broker changes:
     - Updated NEB API version
     - Modified adaptive program status callback
     - Added adaptive contact status callback
     - Added host/service precheck callbacks to allow modules to cancel/override internal host/service checks

* Embedded Perl changes:
     - Added 'enable_embedded_perl' option to main config file to control whether epn is enabled/disabled
     - Added support for perl plugins to specify whether or not they should be run under the epn...  The
       second to tenth line of a perl plugin may start with '# nagios: +epn' or '# nagios: -epn' to explicity indicate that it
       should be run under the epn.
     - Added 'use_embedded_perl_implicitly' option to main config file to determine whether or not perl
       plugins will use the epn if they don't explicitly allow/disalow it

* CGI changes:
     - Hostgroup and servicegroup summaries now show important/unimportant problem breakdowns like the TAC CGI
     - Minor layout changes to host and service detail views in extinfo CGI

* Misc changes:
     - More information given when testing scheduling (-s command line option)
     - Removed fixed length restrictions for host names and service descriptions
     - Plugin output length restriction bumped up to 4K
     - Added temp_path directive to main config file for specifying temp directory
     - Multiline output support for system commands via my_system()
     - Added global event_id and notification_id vars that are at least unique during a single run of Nagios
     - Default is now to check for orphaned services, rather than not
     - Renamed service_reaper_frequency to check_result_reaper_frequency
     - Fractional notification and check intervals are now supported (e.g. "3.5" minutes)
     - Backslash chars are now used to escape command arguments that contain \ or ! characters
     - Added 'external_command_buffer_slots' and 'check_result_buffer_slots' variables to specify size of internal buffers
     - Added check statistics to status file, available via nagiostats for graphing in MRTG
     - Added $HOSTGROUPNAMES$, $SERVICEGROUPNAMES$, $HOSTACKAUTHORNAME$, $HOSTACKAUTHORALIAS$, $SERVICEACKAUTHORNAME$,
       and $SERVICEACKAUTHORALIAS$ macros