This documents the major changes between releases 2.14.6 ???? FIXES: Fixed SQL injection in /status/host (OPS-863) Fixed javascript vulnerability with stripping of '<>;' characters in downtime, acknowledgements and submitting results (OPS-864) Escaped < and > characters when submitting wiki information to host and hostgroup comments WARNING: The database upgrade scripts will convert all instances of < and > to < and > in the tables hostinfo and hostgroupinfo 2.14.5 24th September 2009 ENHANCEMENTS: Added ability to exclude service checks at a host level, even if it is in an associated host template WARNING: when upgrading Opsview, make sure you upgrade to 3.3.2+ or 3.4.0+ to retain this enhancement FIXES: Fixed query_host and snmpwalk when over 10000 bytes of data returned Fixed opsview_sync_ldap to encode problem characters from LDAP information (thanks to Nguyen Thai Ha) Fixed a possible import_runtime hang where the downtime end time is incorrectly passed through to ODW, because of duplicated internal_downtime_ids Fixed webexception when deleting a service check still being used as a dependency for a different service check (thanks to Vincent C for reporting) Ensure all opsview database tables are using the InnoDB engine (OPS-426) Fixed Nagios CGI exploit where arbitrary code could be run from statuswml.cgi Enforced model validation for constraint of 63 characters for host and service check names Fixed bug where servicechecks list page would count the number of hosts twice if in multiple templates Fixed bug where list of hosts for a service check were duplicated 2.14.4 15th May 2009 ENHANCEMENTS: An Overview performance graph now respects the end time and hides the start time option Right hand options menu now made to look more like a tab FIXES: Potential fix for timezone calculations, backported from Nagios 3 Fix for IE6/7 where menus on some status pages would appear in the wrong place Redo fix to cater for more instances when notifications.cgi details come up blank Rework popup menus to use x/y coordinates instead of anchoring to a div Fix bug in check_snmp_linkstatus with quotes in interface names Fix some HTML validation errors on the Hostgroup Hierarchy pages 2.14.3 27th March 2009 Enhanced check_snmp_linkstatus to use 64 bit counters when possible Allow option to select SNMPv1 on host page Reduced performance overhead on mysql when gathering a lot of downtime information Enhanced clustered slave takeover so that the failed node's host state is changed by the others in the cluster Fixed bug where non admin users could not change their own password Fixed /viewport summary page to ignore disabled viewports Fixed configuration generation for some check_interval, retry_check_interval and notification_interval values when interval was set to seconds. Especially relevant when using clustered nodes Fixed Query Host and SNMP walk when querying device with strange shell characters in community string on a slave Fixed IE7 flicker when reload times were shown Fixed IE6/7 display of buttons on SNMP Trap Exceptions and SNMP Trap tracing page Fixed IE6/7 when running snmpwalk and json is not returned Fixed IE6 where snmpwalk table was not constrained horizontally Fixed IE6/7 problem where autocomplete for host names not working on /admin/graph page Fixed bug where Template details didn't work for the last item in the list Now shows error on screen if ajax request fails. Recovers on next successful poll If item is already chosen, then selecting again will delete from full list. For parents list, host templates and notification dependencies Helpful error when no host template selected but Template details pressed If template not selected in full list, will see if one exists in the chosen list Added number of service checks on host template list page Changed text for /viewport check_opsview_slave reporting OK when NSCA state was bad - now changed to CRITICAL Show title for notification dependencies select box for long service check names Apply patch to check_reqs for 64 bit linux servers (Thanks to Ian Yates) Allow opsviewd to rotate its log file (OVS-173) Allow more domain names to be used in email address validation for contacts (6 instead of 3) (OVS-38) Fix SQL statements where referencing a temporary table twice breaks in MySQL > 5.0.68 and 5.1.30. Also default to configured engine rather than forcing MyISAM temporary tables Provide links to online help for exceptions, timed exceptions and event handler For slave clusters, increase max check attempts to 2 to cover any temporary problems before invoking the takeover script Added option for diagnostics in rsync_nmis_database Added /graph/options page for configuring graphing parameters. Includes /graph backported from Opsview 3. Now links from HH and Nagios pages Updates to improve working of check_snmp_linkstatus plugin for interface monitoring Create plugin area within var for status data from scripts Include missing bind-utils prereq for builds on RPM systems Added rrd_converter for helpful conversion of rrd files Allow opsviewd to keep running after fork failures have occurred Better error handling in check_snmp_linkstatus Fix for notifications.cgi details coming up blank when updating page Updated tab colour so there is a difference between active/inactive/disabled Fix for lighttpd spinning and soaking up CPU when opsview-web is stopped 2.14.2 21st January 2009 Viewport summary page, with highest state information for each viewport SNMP interfaces with duplicate names are now listed and can be monitored, but with limitations Can filter HH views based on a specific user using the asuser URL parameter (limited to view all roles) Support for shared ODW database for multiple Opsview Master systems Fixed bug where opsview schema version wasn't updated correctly Apply patch to check_opsview_update (thanks to Steve Huff) Fixed app server shutdown if downtime submitted while running in debug mode Viewport pages now reference custom.css for local overrides Fixed installation on Redhat platforms, where a fresh install shows post install errors Removed java dependency for opsview-core, instead relying on opsview-reports to provide If realm != local, then hide the Atom feed details on the contact page as this does not use the realm of the contact to authenticate Fixed bug where some data points may not be updated in RRD if RRD::Simple raises an exception Added parameters to /rrdgraph of title and vertical-title (thanks to Jose Luis Martinez) Added check_nagios_stats for nagios statistical information Host check commands now evaluate Opsview host macros SNMP polling servicechecks now accept SNMPv3 authentication information Fixed bug where servicegroups did not display embeddedservices Increased length for timeperiod definitions Package dependencies streamlined Slight rework on debian packaging Allow the Solaris packages to up upgraded in situ rather than have to be removed and reinstalled. Removed servicegroups from avail.cgi and summary.cgi as these do not match with Opsview concepts Updates to check_snmp_linkstatus to improve efficiency of querying device Amend reverse slave setup to allow for automatically accepting host keys on an initial run (-i option) Move opsview and opsview-slave rpm info from base spec file to core spec file so they become noarch builds Allow stop and start of notifications, active host and service checks and passive host and service checks to be propogated to slaves when used from 'Process Info' page Replace jpg icons usage with pngs (note: jpgs still created but should not be referenced any more) Reapplied multi-valued insert patch as duplicate host contacts are removed in a seperate patch - this will speed up startup time. Allow sound files from /usr/local/nagios/share/media to be accessed without apache proxy Removed a speedup patch that causes incorrect access rights in distributed environments (distprofile removal patch) Allow passive checks to have data graphed (thanks to Eric Sorensen) 2.14.1 17th November 2008 Security patch to stop linefeeds from being entered through Nagios' cmd.cgi Disabling of CHANGE_* Nagios commands, which don't appear to work anyway Amended some page titles from the default 'Opsview' to aid browser history usage Allow use of IPv6 addresses for hosts Update system preferences to be able to select log event handlers All Unhandled view now bunches all services on failed hosts together for a smaller screen When listing services ordered by status, order is now OK, UNKNOWN, WARNING, CRITICAL Added authtkt_default_username option for using a specific contact when a valid authtkt ticket is presented opsview_sync_ldap has troubleshooting options to test a username and password opsview_sync_ldap updated to support OpenLDAP handling of groups (thanks to Steve Huff) Patch to Nagios to do audible alarms through status.cgi for passive checks (thanks to Jason Mogavero and Mike Dorman @ Data393) Amend ownership of /usr/local/opsview-web in debian packages (trac #311) Amend NetScalar SNMP oids Amend perms of slave /usr/local/nagios (trac #324) Added indexes to runtime tables for faster querying of notification information Log system() error when invoking CGIs check_odw_status now goes critical if > 10 hours delay in import_runtime Added upgrade check for import_runtime Fixed possible coredump with check_opsview_slave on RH5 systems when timeout reached Fixed constraints on usernames with @ in the name Amend debian shutdown scripts to stop opsview before mysql Add in hosticon_admin script for adding custom host icons and symbols Removed need for hosticon_admin dependencies unless adding new icons Catch possible null values from runtime for status_update_time in import_runtime Backup now ignores ndologs directory Die if opsview_web.yml file is not parsed correctly URL to create rrdgraphs, with system preference whether to allow public access or not Generalised viewport templates to allow customisation Speedup of nagconfgen when lots of rrdfiles exist Additional macros allowed for managementurls (thanks to Jose Luis Martinez @ CAPSiDE) Updated slave's sidenav.html 2.14.0 6th October 2008 Rebranded for Opsera Fixed bug in IE6 where an invalid field on one tab caused submit to continue Show message in opsview-web if there is an internal server error 500. Possibly due to lack of memory 2.13.0 Not publicly released LDAP integration, for login via LDAP realm Added Log::Dispatch::FileRotate for automatic rotation of Log4perl log files for opsview-web 2.12.10 Not publicly released Fixed finding Nagios::Plugin perl module for check_disk_smb and check_netflow_database 2.12.9.1494 23rd September 2008 Fixed bug where only first line of map.local file was read (Jason Lavoie) Reverted multi valued insert patch which causes problems when a contact is in several similar contactgroups for a host Fixed opsview-agent process status bug (trac #303) 2.12.9.1461 11th September 2008 Fix javascript error with Management URL edit page when no URLs are defined 2.12.9 10th September 2008 Fixed bug where audit_log_retention of 0 would delete data from Runtime DB Hosts and services get a UP/OK state immediately when created. Fixes trac #43 and #241 Add indexes for speeding up import_runtime, especially when services have not had a state set yet (trac #193 and #201). WARNING: upgrade may take some time as these indexes are created Ensure opsview environment is setup when invoking scripts from cron The duration information when a service is in a failed state is now based on (current time - last_state_change). It was previously based on last_state_ok, but this is not relevant for passive checks based on log results. This mimics logic for Nagios' CGIs Allows use of interval length in seconds, but requires manual changing of frequency value A service is considered handled if the host is in period of downtime Update to NRPE 2.12 NRPE patched to allow larger amounts of data to be returned from more capable NRPE daemons WARNING: Viewport JSON output from /viewport/{keyword}?output=json has changed to be more in line with /api/status/service On Solaris, NRPE will not use 256-bit encryption, so removes the dependency on SUNWcry and SUNWcryr Updated nagiosgraph map file so that certain check_ping failures are considered 100% packet loss Stop error from nmis for missing files in distrbuted setups Opsview master housekeep job now handles table trimming instead of ndo2db. Includes an index added to nagios_conninfo Improved performance of validation scripts so that form submission on edit pages is up to 5x faster on IE6 Fixed bug in import_runtime where a configuration change does not lose the previous state in state_history Fix bug with NMIS not finding configuration details where device names have upper case letters Fix slave apache permissions when reading MRTG data Speedup NDO import at Nagios startup time by 90% on large scale systems by: * removing unnecessary name/distprofile contacts * improving writes to ndo2db from file2sock * converting to multi valued inserts Improve nagiosgraph to auto-add new perfdata datasources to existing rrd files WARNING: if new perfdata has much smaller scales (i.e. max of 1,000,000 to max of 100) then data may look 'absent' but after max timeperiod of each graph has passed (i.e. a day, a week, etc) scale should revert to show data correctly Nagios cmd.cgi now stores the username of a command submission in nagios.log When submitting a result, goes to slave if applicable and does not submit to master - otherwise this was causing double results to be received on master Removed option for logging external commands in system preferences as should be done all the time Always log passive check results on slaves Anchor contextual menus to the triggering link rather than appearing where the mouse pointer is. Fixed bug whereby some database names were hardcoded rather than picking up config Add in Management URLs functionality Move all SNMP MIBs into separate repository/package (opsview-mibs) Move some host icons and symbols into separate repository/package (opsview-hosticons) Add in user customisable check_period on servicechecks Fix some glitches within rpms when installing and removing (more work to be done) 2.12.8 25th July 2008 Housekeeping added for MRTG RRD files that have not been updated, similar to nagiosgraph RRDs Opsview backup now misses out MRTG and NMIS' rrd files - see http://docs.opsview.org/doku.php?id=opsview2.12:backups Nightly backup now includes Reports DB and portions of Runtime Nightly housekeeping now prunes ODW data. System preference to support. Note: when upgrading, retention=0, means no pruning Fixed major bug when upgrading from 2.12.5.1185 2.12.7 23rd July 2008 Fixed major bug where viewports page was not updating correctly Fixed bug with mass acknowledgements not actually submitting Reduced font in service status column to allow more space for duration. Made duration non-breaking Cater for spaces in database passwords (trac #243) Allow for local map file etc/map.local for performance parsing (trac #246) Add in 'comment' icon to HH for hosts and services, with links to detail page Add in links for 'acknowledged' and 'flapping' icons in HH to detail page Allow bin/opsview_api to get status information Fix bug where nmis directory isnt created on slave (trac #260) Fix bug where some macros were not translated on timed overrides (trac #262) Fix bug with downtimes not getting propagated to slaves (trac #263) Allow for downtime to be scheduled and propagated on a service on a host Ensure AuditLog entries are created when adding/removing object information Locked down NMIS to only admin users Rework object comments code and add in tests (trac #266) New URL parameter for statusmap.c: onlyincludehostsinlayerlist. Only works for specific parent/child relationships Updated pst3 from latest Nagios Plugins, to work in Solaris 10 zones 2.12.6 2nd July 2008 On an install, opsview.conf will get generated if it doesn't exist, for local changes WARNING: the auth_tkt shared secret will get changed to a randomly generated value if opsview.conf has it set to the default Major enhancements made to /status/service so that you can specify multiple host names, servicechecks, or states. Also allows sorting by various columns. Added last check time, duration since last ok and check attempts Added mass acknowledgements page Added flapping and acknowledged icons to /status/service and /status/host page Extend API to get status information in XML and JSON format Fixed problem where a contact removed while session was still valid would cause Catalyst to return an error Fixed problem where a change of case for a host group name would cause an Opsview error Show tabs, but unselectable, when creating new items Fixed submit button, where the name may not change correctly across requests Fixed problem where drill down from HH not correct where filter was not retained from /status/host to /status/service Fix bug where runtime helper tables are truncated during a reload and HH loses all data until reload completed (trac #240) WARNING: db_mysql has to be run to update database permissions Fix bug where a form parameter of '0' was interpretted as missing data by the controller Fix Solaris bug where altinity_distributed_commands.o wasn't being linked properly Change number of list items on service, servicegroups and hostgroups pages from 200 to 100 Fixed bug with returned data from service and host group contact lists Amend list of contacts on servicegroup and hostgroup pages to be popup with active links to contact edit pages (trac #239) 2.12.5 13th June 2008 Fixed bug where plugin data returning scientific notation was incorrectly parsed The unhandled link for services was showing all services if a host was unhandled. A new parameter (includeunhandledhosts) will be necessary to show hosts that are unhandled and include the failing services on it Fixed bug where creating new user with no role assumes admin role Fix passive check submission user validation Default option now is to retain session data in file so that an opsview-web restart doesn't force users to re-login again Companion plugin - check_opsview_session to get number of active sessions Fix bug with rc.opsview-agent script picking up wrong processes (trac #234) Allow extra Opsview functionality via plugins at specific points Fix performance issue in nagios where lots of notification prep work was performed when no contacts needed notifications (trac #228) Fix RPM dependency issues (trac items #233, #225 and #226) Host edit and system preferences pages now use tabbed interface Added link from status host menus to NMIS Added search host box in sidenav Get hostgroup breadcrumbs when using /status/service?host={name} Fixed auth_tkt so that it creates a session if the username in the ticket exists in the local database Added auth_tkt to include an OPSVIEW_ADMIN token for 3rd party apps Sidenav menu rearranged Added extra runtime tables for pruning in opsview_master_housekeep Fixed nagiosgraph pages being truncated across the bottom Tightened up valid characters for a contact's username 2.12.4 13th May 2008 Fixed bug where a host that had no contactgroups produced a nagios validation error Fixed bug with non-admin object information links pointing to wrong URL (trac #214) Fixed bug with 'Other Hostname/IP' list on host config page (trac #216) 2.12.3 9th May 2008 Fixed bug where downtime of less than 60 seconds can be scheduled (trac #204) Fixed bug where host only downtimes were not listed in hostgroup or host detail pages Fixed bug whereby quoted metrics in perfdata could not be used for Reports graphs (trac #206) Fixed bug whereby metrics containing /'s in perfdata could not be used for Reports graphs (trac #210) Amend SSO authtkt to use username rather than name for uid for consistency between apps Amend SSO so the cookie domain can be set in opsview.conf for sharing between servers Moved opsview-web pid file to /var/run/opsview/ (trac #192) Amend ODW graphing form to be GET rather than POST to allow for bookmarking "Delete downtime" has now been changed to "Cancel downtime" NMIS: Support for Netscreens, Alteons, Foundrys, Ellacoya and Cisco CSS (arrowpoint) NMIS: Disabled all event logging, as not supported in UI (Mark Lovely, Plusnet) NMIS: Fixed pix typos in nmis-cgi.pl and added PIX v7 support (Mark Lovely, Plusnet) Changed nagiosgraph to use 1px line widths instead of 2px Added port 4125 for tunnelling from slave to master - for ActiveMQ messaging WARNING: On debian systems, /etc/cron.d/opsview will be removed as part of the install. If you have any customisations, put into the nagios user's crontab 2.12.2 24th April 2008 Logic for hostgroup/servicegroup intersection tightened. Previous configuration would allow contacts for hosts based on hostgroup, but now will only set the hosts for the hostgroup/servicegroup combination Fixed bug in Catalyst Engine where a single request with no headers sent would block the entire engine Fix bug whereby invalid ODW database is not detected Fix bug with SNMP polling not parsing community strings with '$' in Force populate_db to see perl library directory Amend rpm init.d script permissions and files Fix bug with template error (trac #183) Ensure opsview can serve NMIS static files Check FastCGI can still be used Allow for use of FastCGI on Solaris Show unhandled hosts in Alerts->Unhandled (trac #106) Allow accents on letters for contact/host/service names in GUI (trac #74) Fix javascript form validation to show better error message Add validation to 'new hostgroup' page (trac #100) Fix bug from buffer overfilling with large query_hosts output (trac #198) Fix bug with resubmitting Hostgroup page with no changes causing error (trac #98) Extra columns for ODW's summary table. Keywords stored for each servicecheck 2.12.1 28th March 2008 Opsview Reports added for PDF reporting off the ODW database NDOutils: Fixed two issues - EINTR or EAGAIN received during a read() and will now reap multiple children Default unset $ADDRESSx$ fields to primary ip address Allow lmstat to be found by PATH for nagios plugins Fixed some timezone issues, where DateTime module was using floating timezone, instead of local. Affected display of timezones and scheduling of downtime Fixed bug where downtimes could not be deleted by id Fixed missing XML::LibXML::Common perl module on redhat systems Fixed SQL error when listing downtimes as non-admin user Fixed error where dummy emails maybe getting sent due to poor contacts configuration Fixed error where contact with only unknowns for notification could raise a Nagios config verification error Code changes for Solaris 10 Set nmis to run in multi-threaded mode when collecting data Removed deletion of downtimes by id in Nagios CGIs Deletion of an snmptrap in the unique view will now delete all similar traps NRPE: Remove " from list of nasty metacharacters NRPE: Use ssldir autodiscovered via ./configure Use /usr/local/nagios/tmp for temporary files Add in 'Parent Tree' on sidenav to show parent relationships Fix debian upgrade scripts when some databases exist on different hosts Added web pages for Opsview reports New default configs file: etc/opsview.defaults. opsview.conf overrides the default values Fix bug with timed overrides in distributed systems Check interfaces via SNMP on host edit page now supports SNMPv3 2.12.0 27th February 2008 Single Sign-On feature, using Apache::AuthTkt as authentication ticket mechanism (Jose Luis Martinez - CAPSiDE) Custom SMS notification methods (Jose Luis Martinez - CAPSiDE) Allow email and sms contact information on master server, even if email and sms notification methods not used (Jose Luis Martinez - CAPSiDE) Scriptaculous updated to 1.8.1 Updates: Nagios 2.10, Nagios Plugins 1.4.11 and NRPE 2.11 Fix for Solaris 10 in ndo2db where EINTR is received on an accept Provide more detail in warnings on reload when hosts appear to share IP addresses 2.11.1 23rd January 2008 Fixed bug with send2slaves overwriting NMIS configuration Fixed restart problem where Opsview was restarted before Nagios completed stopped Fixed automatic RRD generation where rrd label was too long with warn and crit appended NMIS: master not monitoring a node due to duplicated node information NMIS: extra traffic graphs NMIS: Juniper M10 device support NMIS: can search interface names from master NMIS: Reports now generated on master 2.11.0 7th January 2008 NMIS integrated for interface statistics Event handler functionality on a per host per service basis (Josep Ruano - Capside) 2.10.4 Unreleased Update runtime db after every host failure Passive checks are enabled for all hosts by default Amend default %files permissions in RPM spec files IE6 now renders in standards mode Speedup of opsview_host_services helper table Speedup of edit host page, for loading and submitting Changed default colours for handled column back to same as unhandled column. Timings shown in reload page are now based on server's time, rather than UTC Change default host check command from check_ping to check_icmp Remove length limit on command argument strings Log SNMP rule evaluation errors to opsviewd.log Remove duplicate host entries from 'Network Traffic' page (trac #146) Ensure log config changes are picked up by opsviewd and import_ndologd without restart Speedup of upgradedb_odw.pl for notifications table during upgrade (about 98% increase) Take account of timezones better for hostgroup downtime and reload page 2.10.3 5th February 2008 Host check commands now user configurable (Jose Luis Martinez - CAPSiDE) WARNING: If you have custom host check commands, goto /admin/hostcheckcommands to re-configure Include building of opsview-agent in rpm spec file with supporting code changes Backup of opsview-web in rc.opsview backup Include state and output information of notifications in ODW Speedup of edit host page, for loading and submitting Default colours states of unhandled column will now be a pastel shade. Alter custom.css if you want to restore. Only for new installs Menu icon updated Fixed bug in 2.10.2 where new installs would not generate configuration correctly Fixed bug where a Debian upgrade with a database on a different server would be overwritten Re-factored rc.opsview to have a consistent stop/start of applications Reload image overridable in opsview_web_local.yml (Jose Luis Martinez - CAPSiDE) Amend "rc.opsview cron_daily" to make daily backups configurable in opsview.conf Minor RPM tidyups 2.10.2 21st January 2008 Fixed major bug where attempting to restore from audit logs page would leave db in inconsistent state Fixed major bug where mrtg-rrd graphs are not shown if no host parameter provided Fixed major bug with Hostgroup Downtime in contextual menus Added in Host Downtime in contextual menus Fixed bug where snmp agent terminating snmpwalk early didn't show any output Fixed bug where an SNMP rule with no name cannot be editted Fixed bug where submitted passive check message shown inappropriately Force 'Receive from alert' on contacts page to be 1 or more for consistency 2.10.1.703 Fixed issue if import_runtime is run when the servicecheck results have expired from Runtime database Add in host and hostgroup information URL overrides in system preferences Allow overrides in nagios.cfg and cgi.cfg if the parameter has been commented out Improve speed of HH database downtime queries Fixed regression in send2slaves [-p | -s] Fixed issue if NSCA receives a result from slave before Nagios creates nagios.cmd file 2.10.1 17th December 2007 Fixed bug with asynchronous reload errors not being caught Fixed bug with nagios.log not being archived correctly on slaves Amend ndo syslog logging of DB connections and disconnections from INFO to DEBUG Fixed bug where port entry in nagios .ssh/config file was being ignored when using master initiated ssh connections Allow addresses in Other Hostnames/IPs to be able to be used for snmptrap processing lookups Fixed bug where contacts with view some, change some role got DB error Provide access control to MRTG interface Fixed values in ODW so no NULL values in service_availability_hourly_summary Fixed bug with some options in system preferences not being turned off Add in submission of check results, propogated to slaves Updated Opsview XML schema to include new host attributes 2.10.0 27th November 2007 Ability to schedule downtime anywhere in the hostgroup hierarchy Opsview status pages now show downtime/scheduled downtime status for drill down Audit logs available for most changes to database Automatic backups of configuration database after every reload Restoration of backups from audit screen ODW now records all acknowledgements and downtimes, and includes statistics in the summary data SNMP traps now work for hostnames with multiple IP addresses - will push to Nagios if trap received from any of the IPs Cache IP lookups for snmptrap hosts Warning message in reload if an IP address is used more than once for an snmptrap host SNMP traps has a summary page for starters. Able to debug a saved traced trap Opsview status pages include summary of states Asynchronous updates to NDO Improvements in housekeeping to include slaves Default backup directory is now /usr/local/nagios/var/backups Fixed reload via API, reporting failed if warning messages appear Allow limits on simulataneous configuration tasks at once Fixed Nagios bug where some state changes not propagated to database correctly Fixed bug with install_slaves preventing slave installs/updates Implement cached and parallelised Nagios commands when sending to slaves Fixed bug with incorrect file test state during web reload 2.8.8 Unreleased Amend running order of setuid changes when installing plugins Ensure notifications are always enabled on the master server to allow correct use of RSS feeds Fixed bug with rc.opsview restart not working correctly Fixed bug where host notifications are not sent via RSS feed Fixed bug where RSS feed notifications are disabled in distributed setups Removed duplicated content-type tags Fixed some validation issues using SGML parser 2.8.7 9th November 2007 Fixed bug where "view all, change none" contacts still had change permissions in Nagios CGI based on hostgroup/servicegroup authorization User preference page only displays configured hostgroups and servicegroups for non-admin users Fixed bug with removing object information Fixed bug with query_hosts Fixed bug with host icon when querying services based on servicecheck name Fixed bug with trimming snmptrapexceptions table on remote db Include parse_nagios_cfg in installation package Fixed bug with incorrect link from nagios status pages to object information Fix bug with query_hosts Capture ssh errors on reloads sent to slaves Fixed bug with opsviewd killing all nagios processes when a tunnel inactive Correct reason texts for SNMP trap exceptions 2.8.6 24th October 2007 Fix collation issue for pre Mysql 4.1.12 Amend HHG page for IE7 usage with menu icons Fix Odw::Host error on web startup Truncate runtime tables that are no longer used 2.8.5 17th October 2007 New popup menu on config and status pages to link between the two Add object information link to hostgroups Updated check_opsview_slave to check connection to slaves and NSCA status is okay Fix runtime db schema problems when upgrading from 2.7.X versions Fix for corrupted perfdata in passive checks where only plugin output is '|' Fix for cases where snmp interface name has illegal characters (as decided by Nagios) Fix search in hostgroups page when using mysql5 Fix for occasional mysql error during config generation Made search consistent - using space in search box is an OR of the fields Can now search based on trap data in snmptrapexceptions Put navigation links on snmptrap exceptions page for hosts, and servicechecks that fail to match Allow monitoring of non-duplicate SNMP interfaces rather than disallow monitoring of the device Add in 'check_odw_status' HH graphing page now uses id's when deleting rather than name Session access time updated on every page access to stop expiring before time Add in 'additional_freshness_latency' option Add flapping options to services, hosts and contact pages Amend logging of opsviewd process Allow a restart time in import_runtime if a whole period of data is lost NDOutils now shows the mysql error in syslog Slave communications made consistent, with batchmode set for ssh Fixed upgrades from Opsview 2.8.2 where some databases were at the wrong versions Fixed bug with keywords where extra services were listed Validation fails if no cluster nodes selected for a monitoringserver Standardised versioning of all 3 databases Improve checks on ssh tunnels to slaves 2.8.4 11th September 2007 Configuration generation parallelised across all monitoringservers Additional macros for INTERFACES and SNMPv3 credentials custom.css file available to override standard CSS settings Check for duplicate SNMP interface names when querying Fixed bug in UI where you could not deactive an snmp interface once already activated Cater for using databases on remote hosts better Fixed bug where new hosts would not be seen in HH until 2nd reload Removed runtime.cfg references Fixed bug in API where you could not clone a host that had snmpinterfaces Fixed IE7 cell text aligns for Nagios screens Correct flaw in promoting MIBs and also improvements to trap process Provide updated information during MIB promotion 2.8.3 24th August 2007 Optimised the population of opsview_viewports helper table on large scale systems Fixed bug in db_opsview db_install script 2.8.2 21st August 2007 Fixed problem with stale results if the normal interval was a large period, such as a day Optimisation to HH pages Optimisation to Nagios startup time by removing initial host/service state data to db Some unnecessary tables are disabled for population into runtime db SVN repository moved to svn.opsview.org Reimplement nagcmd on slave servers (only) NSCA tuning - removed dump file option and set aggregate writes Fixed some validation issues for sms phone number and new hostgroup names AQL servers can now be contacted via a proxy server Fixed problem where MRTG graphs enabled but no snmp community string could cause Opsview to appear as if it is still reloading 2.8.1 10th August 2007 API available to run a subset of commands to Opsview via XML User Preferences screen available to change own password and contact information Add in further options to System Preferences Fixed Network Map so that changes to view are saved across a refresh Amend Atom feeds such that status going from OK to non-OK is a new item and add test suite for feeds import_runtime now runs continuously until up to date Tidy up host check commands list on Hosts page Cater for non-fatal errors when generating configuration Enable Orphaned Services checks Add more validation to SNMP fields Remove nagcmd group and use nagios group instead Remove unnecessary restarts from some slave updates Add in more help links to wiki Create ODW DB upon upgrade if it doesn't already exist Corrected opsview-web path to Log4Perl config file Amend SNMP interface processes to take into account 63 character description limit. Fixed if two reloads were initiated in close succession 2.8 18th July 2007 Now integrated with AQL's SMS messaging system, for alerts via AQL Added servicecheck exceptions based on timeperiods Exposed timeperiods editting in UI Update to Nagios 2.9 and NDOutils 1.4b3 Include Log4Perl functionality for logging Improve passive caching mechanism for slaves Added icons to Icon dropdown on host page for Firefox Bugfix install_slave for when tar files are not compressed Communications to slaves can now be initiated from the slave Included validation.js functionality for form validation Funny characters in the snmp community name are now escaped correctly 2.7.8 18th July 2007 Fixed problem where a timeout of send_nsca killed ocsp_command script Reduced number of buffer slots in Nagios to limit memory used for each nagios process Amend atom to remove need of config file but use available info instead Amend rpm .spec files to include all missing dependancies Improve atom code to remove potential bugs Remove brackets and quotes from SNMP description strings Amend snmpwalk to try version 2c, then fall back to version 1 Display average reload time on S&R page before button is pressed Change 'Promote MIB' to work asynchronously Input validation on warning/critical fields in host/SNMP/query/interfaces Amend user profile to enable coredumps Improve PATH checks in nagios user profile Correct title in atom feeds 2.7.7 22nd June 2007 Add in "Alert every failure" to service checks Speedups to nagconfgen for hosts and hostgroups Host notifications in a distributed environment fixed Deprecate "Log Every Message" in servicechecks ODW schema changes to save time of services in warning/critical/unknown states Reduce stale results from passive/snmptrap checks on slave when renotify is set Catch snmpwalk executable unavailable Only load SNMP at nagconfgen time when necessary Search in hosts list now also searches in Network Address column During upgrade, catch ODW DB not existing and create it Send nrpe.cfg to slaves; slaves use it if they havnt already got it 2.7.6 13th June 2007 Fix patch to nagios config reloading and freshness checking Fixed mail command for notifications Added patch from Nagios 2.9 re: signal handling - catches SIGSEGV correctly now Fixed segfault on startup with relative config dir patch on 64bit debian Increased maximum size of large fact tables in ODW Now save state changes in ODW. Some fixes if a service is checked less than once an hour and if a service is added/deleted during an hour period Enforce maximum name length of hosts and services to match Nagios 2.7.5 1st June 2007 Fixed IE6 problem with exceptions always being saved when not selected in host/hosttemplate edit pages - will self correct on current systems Stop SNMP errors from failing the submit Fix IE display problems with select boxes Some contention when opsview restarted - nsca may not die correctly and hang Patch nagios to take config reload time into consideration when checking result freshness Fix snmp community names which have $ in it Disable snmptrap debugging on slaves SNMP polling interfaces now works correctly in slave environment Can reference relative nagios config files Stop freshness checks expiring before the initial check spread period is over Speed enhancements to snmptrapexceptions page Only show a subset of page links at bottom Form validation was broken on IE6 Pages now reload when Back is pressed (except Safari). Form pages are excluded otherwise data is lost Disable spaces in Host Title in UI 2.7.4 21st May 2007 UI does not allow form to be submitted if errors are visible on page Fixed problem in class if notification_interval was set to 0 /usr/bin/mail used for mail command - fixes RH issue Removed #---next trap---# from display for snmptrap exceptions from slaves Re-added unique indexes to NDO tables Fixed notification_interval when set to 0, so inherited correctly Runtime table nagios_externalcommands trimmed by ndo2db Fixes "Undefined identifier: snmp" error when restarting snmpd (from net-snmp) for CISCO-GENERAL-TRAPS.my Correctly populates initial configuration database when server does not have DNS setup Fix links in ATOM feed Hide database errors, but allow them to be displayed if requested Fix window width on reload if errors occured Fixed redundant opsview_web_server processes if browser had connection failure 2.7.3 30th April 2007 Fixed UI bug where if the information line for hosttemplates and dependent services is selected, integrity failure occurs Removed population of agent information from populate_db.pl - will be re-introduced later Nagios user is now created with a home dir of /var/log/nagios Fixed collation errors on mysql < 4.1.12 for keywords If config generation error, server light goes red. Now stays yellow, and configuration status goes red Bumped up Nagios buffer slots for all implementations New indexes for some historical tables which were taking too long when deleting old rows 2.7.2 23rd April 2007 Changed warn/crit levels for interfaces to bits/s, rather than bytes/s. Need to manually change host configurations Fixed UI bug where if the information parent line is selected, host not created correctly HH now respects the contact's role, so a view all contact will see all host/services Added 'rename' functionality to HH pages 2.7.1 12th April 2007 Fixes to db_mysql servicedependencies fixed on slave servers Stop ndo2db starting up on slaves Slaves do not require mysql Fixed error when reload duration is 0 Fixed installing slaves from send2slaves Fixed SNMP interfaces so only active ones are checked Started internationalisation Upgrade Nagios Plugins to 1.4.8 2.7 5th April 2007 NDOUtils updated to 1.4b2. Status pages will temporarily fail as database tables are renamed. Will need to restart opsview-web to pick up new SQL commands. Will need to reload Opsview to get helper tables created Nagios updated to 2.8 Supports Mysql 4.1 & 5.0 Massive speedup to Hostgroup Hierarchy Added host meta information Fixed IE bug in HH graphing Viewports, aka Keywords Easy configuration of SNMP interfaces for polling New servicecheck type: SNMP polling - need to manually define servicegroup/notifications via SNMP Interface Polling link Service dependencies can be defined, at the servicecheck level Parallelise sending and restarting of configuration to master and slaves, to speed up reload Some speedups to /admin/status/opsview Removed old performance scripts and scan_snmp_devices.sh Two types of database scripts: db_mysql (for grants - requires root) and db_{dbname} (no prompting required) All databases upgraded from installer/upgradedb.pl Move notificationperiods details into database FastCGI fixes - use $c->stash rather than altering $self->config during runtime Run ssh tunnels with TCPKeepAlive send2slaves uses compressed tar file to send data to slaves Next page for Advanced SNMP servicechecks is the edit rules page SNMPtraps with linefeeds are now probably parsed (but ignored) Switched from RSS feeds to Atom feeds instead Refresh rate for Opsview status screens now configurable in opsview.conf NOTE: Need to install new perl modules NOTE: Config needs to be regenerated and restart of Opsview required NOTE: Need to add a grant to opsview user in opsview database to grant CREATE TEMPORARY TABLES (covered in db_mysql) NOTE: New atom config file instead of RSS file 2.6.5 20th March 2007 Upgraded to Nagios 2.8 that fixes a scheduled downtime problem Monitoring servers page now only lists hosts that are not currently used as monitoringserver Fixed initial create of agent Added import_agentinfo, missing from filelist status.dat not deleted during a reload sidenav.html not generated at build time - needs to be generated during development Block / in monitoringserver name Fixed a double free in Nagios 2.5 if a reload occurs during host unreachability logic checks Recommended Apache configuration has log level set to normal Respects logo_path in opsview.conf in sidenav and login page Fixed error when editing an SNMP Trap Rule Fixed problem with performance graphing not working Fixed problem where host was being incorrectly reported as UP if no host check_command was defined NSCA uses --single Added YAML::Syck to perl-install NOTE: use_datawarehouse must be switched off! Otherwise, scheduled downtime will cause Nagios to die 2.6.4 29th January 2007 Fixed NSCA 2.6's --single mode 2.6.3 24th January 2007 Check for uniqueness of names in /admin/host,servicecheck,contact,hosttemplates,hostgroups,servicegroups Stalking options exposed Support for IE7 Flap detection option now exposed in servicecheck edit pages Only show performance graphs on master server 40% speed up in time taken to generate configuration Fix for customers with spaces SQLite perl module now included from perl-install - plugins can use this for cross-invocation data retention 2.6.2 21st December 2006 NSCA updated to 2.6 Various packaging fixes for redhat and install permissions Optimised SQL calls for /admin/status/opsview page 2.6.1 14th December 2006 Fixed so hostgroup id 1 (customer name) does not show when no other hostgroups have been created Master broker modules only set on master server - ndomod.o is a problem on slaves If drag/drop disabled, then the objects become links to their edit page Can now goto /admin/host?name=hostname and will redirect to an edit page if only one host is found, otherwise searches for that hostname. Also works on other list pages 2.6 11th December 2006 Multiple host templates allowed per host Ability to specify an agent to run a plugin Auto calculation of best monitoring server based on network address on host edit page Hierarchical hostgroups now respects the role setting for a contact Better parent selection box Session expiry increased from 2 hours to 24 hours Search boxes for all list pages Fixed bug where a new hostgroup in host/edit page was not being saved correctly Major speedups to service list page in host/hosttemplates where there is a large number of servicechecks MIBs individually saved in version control. Updated PowerNET mib Fixed bug with MRTG option not sticking in hosttemplate page (was okay for hosts) Major speedups to pages with draggable things - only sets dragability when it is mouseovered Allow disabling of drag drops Add delete options for hostgroups and servicegroups Tested parent selection box + agent lists on IE6. Now uses JSON When running mrtgconfgen.pl, will only ask the first node in a cluster to query mrtg Only show exceptions + help for active checks Fix db_runtime changing opsview database on slaves Move opsview_sidenav.js into opsview-core for slaves to work NOTE: Need to install new perl modules NOTE: Need to update opsview database schema 2.5.4 14th November 2006 Fixed bug with MRTG option not sticking in host/hosttemplate pages Removed basic SNMP traps Hierarchical hostgroups available, but not in sidenav NOTE: must update existing runtime databases with db_runtime 2.5.3 19th October 2006 Fixed bug with servicechecks with args containing double quotes NRPE gracefully dies if SSL connection takes too long Uses perl libraries in /usr/local/nagios/perl Added facilities for modules Nagios CGIs now compiled against /cgi-bin, NOT /nagios/cgi-bin. Apache configuration may need updating 2.5.2 2nd October 2006 Packaged as RPMs Allow renotify and stalking options for passive checks Simple escalations for contacts 2.5.1 September 2006 Advanced SNMP Traps has option to continue notification in failed states and log every error message snmptrapexceptions table moved from opsview to runtime. All data will be lost 2.5 September 2006 New Catalyst framework: stateful web sessions, inline updates to CGIs, paging in admin pages, status bar Advanced SNMP trap processing, based on rules Clustered slave servers Subset of Passive commands on master propagated to slaves Simple escalations via contact definition check_snmp_apcups added with Fahrenheit option