check_mk arrive en version 1.1.2

Posted by on 23 Jan 2010 in Nagios Plugins | Commentaires fermés sur check_mk arrive en version 1.1.2

Mathias Kettner nous annonce sur la mailing list Betatesters que la version 1.1.2rc1 de check_mk va bientôt sortir. Pour rappel check_mk est constitué de deux principaux projets :

  • check_mk est un plugin de check nagios (et un agent multiplateforme : windows, linux, aix, solaris …) novateur
  • livestatus qui est un module de courtage d’évènements (NEB module) permettant un accès direct aux objets Nagios (hôtes, services, groupes d’hôtes ….) sans passer par le classique status.dat (cas des cgi) ou par un module de courtage écrivant les données dans une base de données type ndo ou merlin. L’accès aux données se fait en écrivant dans un socket unix des requêtes grâce à une syntaxe dédiée (LQL). Actuellement il existe des API en PYTHON et en PERL. L’avantage est des performances bluffantes (j’ai personnellement testé la récupération de 50.000 services en moins de 2 secondes sur un quadcore).

Plusieurs projets commencent à utiliser livestatus :

Le changelog est plutôt dense :

Summary:

  • Lots of new checks
  • MK Livestatus gives transparent access to log files (nagios.log, archive/*.log)
  • Many bug fixes

MK Livestatus:

  • Added new table « log », which gives you transparent access to the Nagios log files!
  • Added some new columns about Nagios status data to stable ‘status’
  • Added new table « comments »
  • Added logic for count of pending service and hosts
  • Added several new columns in table ‘status’
  • Added new columns flap_detection and obsess_over_services in table services
  • Fixed bug for double columns: filter truncated double to int
  • Added new column statusrogram_version, showing the Nagios version
  • Added new column num_services_pending in table hosts
  • Fixed several compile problems on AIX
  • Fixed bug: queries could be garbled after interrupted connection
  • Fixed segfault on downtimes:contacts

Checks & Agents:

  • Check ps: this check now supports inventory in a very flexible way. This simplifies monitoring a great number
  • of slightly different processes such as with ORACLE or SAP.
  • Check ‘md’: Consider status active(auto-read-only) as OK
  • Linux Agent: fix bug in vmware_state
  • New Checks for APC Symmetra USV
  • Linux Agent: made <<<meminfo>>> work on RedHat 3.
  • New check ps.perf: Does the same as ps, but without inventory, but with performance data
  • Check kernel: fixed missing performance data
  • Check kernel: make CPU utilization work on Linux 2.4
  • Solaris agent: don’t use egrep, removed some bashisms, output filesystem type zfs or ufs
  • Linux agent: fixed problem with nfsmount on SuSE 9.3/10.0
  • Check ‘ps’: fix incompability with old agent if process is in brackets
  • Linux agent: ‘ps’ now no longer supresses kernel processes
  • Linux agent: make CPU count work correctly on PPC-Linux
  • Five new checks for monitoring DECRU SANs
  • Some new PNP templates for existing checks that still used the default templates
  • AIX Agent: fix filesystem output
  • Check logwatch: Fix problem occuring at empty log lines
  • New script install_nagios.sh that does the same as install_nagios_on_lenny.sh, but also works on RedHat/CentOS 5.3.
  • New check using the output of ipmi-sensors from freeipmi (Linux)
  • New check for LSI MegaRAID disks and arrays using MegaCli (based on the driver megaraid_sas) (Linux)
  • Added section <<<cpu>>> to AIX and Solaris agents
  • New Check for W&T web thermograph (webthermometer)
  • New Check for output power of APC Symmetra USP
  • New Check for temperature sensors of APC Symmetra WEB/SNMP Management Card.
  • apc_symmetra: add remaining runtime to output
  • New check for UPS’es using the generic UPS-MIB (such as GE SitePro USP)
  • Fix bug in PNP-template for Linux NICs (bytes and megabytes had been mixed up).

Core, Setup, etc.:

  • New SNMP scan feature: -I snmp scans all SNMP checks (currently only very few checks support this, though)
  • make non-bulkwalk a default. Please edit bulkwalk_hosts or non_bulkwalk_hosts to change that
  • Improve setup autodetection on RedHat/CentOS.  Also fix problem with Apache config for Mutliadmin: On RedHat
  • Check_MK’s Apache conf file must be loaded after mod_python and was thus renamed to zzz_check_mk.conf.
  • Fix problem in Agent-RPM: mark xinetd-configfile with %config -> avoid data loss on update
  • Support PNP4Nagios 0.6.2
  • New setup script « install_nagios.sh » for installing Nagios and everything else on SLES11