Shinken v0.8 Release Notes
Release Date: 2011-10-17 // over 12 years ago-
CORE ENHANCEMENTS * Add: WebUI broker module. More info in the UI changelog part. * Add: (by: Vincent Riquer) no_event_handlers_during_downtimes parameter to choose between nagios and shinken behavior to (not) run event handlers during a downtime. * Add: added RUN var relocation in shinken default file * Add: support for serviceextinfo * Add: support for hostextinfo * Add: a workdir parameter in the nagios.cfg file for set the arbiter working directory * Add: modules for modules. Will be useful for the global configuration in the WebUI (it's a module, but will need additonnal modules) * Add: alternative installation method in contrib * Add: shinken-admin command that can connect to an arbiter and ask daemons status. * Add: discovery runners timeout parameter. Set by default at one hour. * Add: rename criticity into business_impact for hosts and services. Criticity is automatically mapped into business_impact in the configuration reading, and rename min_criticity into min_business_impact for contacts and notifways. And keep a mapping for old conf. * Add: if set to 0, min_workers will set the pollers/reactionners to use ALL cpus of the server. * Add: business modulation objects. * Add: new tests for A,B,Cof: rules and a fix for the default return logic. * Add: arbiter ask one a check_interval the managed conf for satellites. Makes for smoother decisions. CORE FIXES * Fix: BugFix Pyro4 compatibility * fix: try to fix bad notification output make scehduler crash. * Fix: (Seb-Solon) catch incomplet service group members * Fix: bad message for arbiter in HA without host_name value. * Fix: (reported by: fraggod): inline comment in boolean cause an arbiter crash when lookup conf. * Fix: (reported by titilambert) bug about linkify and templates, so need to clean search lists after remove templates. * Fix: (reported by Vincent Riquer ) Pyro got problem in connect() and so need a low socket timeout by default. * Fix: (reported by: grim) manage flap_detection_enabled and enable_flap_detection * Fix: when disable active check was done, the pending check was still executed, so the effect was not immediate. Now we change it as a dumym check with curretn value the value of the host/service. * Fix: external_command.py where the config setting log_external_commands was ignored * Fix: (reported by: obiouane) missing object value is now reported correctly. * Fix: (Etienne Obriot) multiple = in macos values * Fix: a bug in config.py where resource macros containing a "=" were not correctly processed. Thanks Etienne * Fix: (reported by Seb-Solon) custome macros get problems in notifications. * Fix: (reported by: Michael Grundmann) a service for a void hostgroup and a host exclusion was reporting a bad conf instead if forgot this service. * Fix: (Reported by: Dirk Hallanzy) was not launching under XP due to some pid file creation bad mode. * Fix: (reported by: Vincent Riquer) now a spare broker can be send to sleep again by the arbiter. It will clean its queues and stop modules. * Fix: arbiter crash with a spare and the new smoothing wat i managed behavior. * Fix: Change default check/eventhandler/notification timeouts and flap thresholds to make nagios more compatible to shinken. * fix: Better handling of check timeouts (there was an empty output) * fix: Log notification and eventhandler timeouts * Fix: (reported by: Denis GERMAIN) service with no description cause a host problem in linking phase. * Fix: (reported by: Seb-Solon) manage utf8 in host and service perfdata. * Fix: protect broker agains module taht ask for full instance init about unknown scheduler. * Fix: (reported by Markus Elger) add next schedule date in the retention data. * Fix: (reported by: Olivier Hanesse) bad realm configuration was not detected for satellites. * Fix: (reported by:Markus Elger) bad host count in realms. * Fix: (reported by:Markus Elger) bad host reaml conf were not detected. MODULES ENHANCEMENTS * Add: module for autotagging poller_tag from a custom macro. * Add: create the simple-log archive dir automatically if it doesn't exist * Add: Initial import of the thrift broker with a sample client in python * Add: Initial import of TSCA, Thrift Service Check Acceptor * Add: Module arbiter load all configuration from GLPI (plugin monitoring) bywebservice * Add: a new module to import configuration data (hosts/services) from a MySQL database * NDO/MySQL: * Add: current_notification_number in the ndo module. * Add: last_notification in ndo module. * Add: last_state_change to the ndo module. * Add: add the is_flapping data to broks and update teh ndo module so it update the value. * Add: percent_state_change as check_result brok and in the ndo module. * LIVESTATUS: * Add: Derive the archives filenames from the main livestatus db * Add: Rewrite the livestatus sqlite code. Now there is one database file with logging events for each day. *!!: If you have a huge livestatus.db it is recommended to stop shinken and run the script contrib/livestatus/splitlivelogs.py * Add: the livestatus sqlite db will now be opened in do_start() instead of init(). Now only the livestatus process has an open filehandle. (the main broker process had one too) * Add: Make the livestatus socket world read/writable * Add: the REUSEADDR flag to livestatus's socket. Thanks Olivier Hanesse * Add: performance improvement for livestatus queries with Limit: MODULES FIXES * Fix: important bugfix in ncpdmod * Fix: Livestatus path database * Fix: string comparison for livestatus. None is equivalent to empty string * Fix: collumn name change in merlin ds. * Fix: (reported by: Seb-Solon) bad contact group insert in objects in NDO * Fix: (reported by: seb-solon) bad contact group id in NDO * Fix: (seb-solon) ndo contacts were not registered with the good id. * Fix: add the flap_detection_enabled in the ndo status tables too. (I love duplicate data...) * Fix: the ndo module with is_flapping value for update states. * Fix: (reported by: Ronny Lindner & Markus Elger) LiveStatus crash in log parsing. * Fix: (reported by: Ronny Lindner) fix status.dat module with utf8 characters * Fix: (reported by: Olivier Hanesse) nrpe module did not tag the check_time value for checks, and so pnp and other tools were not happy. BP rules were alike. * Fix: (reported by Markus Elger) LiveStatus module did not clean previously added elements before recreating them. UI * Add: Andreas to main contributors :) * Add: welcome text and error text in the login page. * Add: make appears on the dep graph only interesting elements, so near the root or hosts or service in real problems. Far more redable with huge conf. * Add: default http backend auto, so if a better one (paste, cehrrypy) is founded, use it instead of the simple swsgi. * Add: less info in the detail page, with a moee button for rarely used infos. * Add: dock effect for actions buttons on /problems. * Add: /system page with satellites info in it. * Add: search form in the problems page, with autocompletion. Search problems and impacts names too. * Add: new way of handling cookies. Now it's a crypted way, and so stateless. Needto remove all session file thing so. * Add: in cfg file password for users, or AD or apache passwd file * Add: AD module that will get user photos from Ldap. * Add: icon sets: network_service, servers., disk, database * Add: gesture management for the detail page. * Add: parallax effect in the top right banner to see huge impacts * Add: make known bad nodes shown in the dep tree, so the admin don't even have to expand and find the bad thing that cause him a problem. Lasy admin mdoe on :) * Add: swtich buttons for ths hostdetail page for enable/disable things. * Add: if you change host topology (new dependency like fror VMotion), the UI is updated too.