check_multi stable 0.20 released

I confess - this release should have been launched much earlier. I know the OSS mantra: release early, release often. But there were so many enhancements, redesigns, fixes in a row that I really missed to shift the trunk version into a new stable release .

You can download it like always from here.

So these are the new features:

• statusdat [ TAG ] = host:service
  gather states and output from existing servicechecks and integrate this seamlessly into your existing checks. This is a good means to build Business Process Views using check_multi without the need to reexecute existing service checks.
  
  
• One more for this new statusdat function:
  When you specify wildcards for hosts and services, check_multi will automatically expand this to additional child checks.
  And the data gathering from status.dat is done efficiently with a caching mechanism.
  
  
• Support for passive feeding: there are several ways now to feed check_multi results directly into Nagios:
  1. via check_result files (direct and very fast)
  2. via send_nsca (needs nsca daemon on Nagios side)
  3. as a chain of commands: one check_multi sends, the other check_multi receives and inserts all child checks into Nagios queue.
     
     
• eval is not counted any more for the number of plugins
  eeval is visible and therefore counted, eval was counted but not visible, and this confused some people. Now it's not counted any more.
  
  
• COUNT(ALL) keywork added to state evaluations, now it's possible to specify:
  state [ CRITICAL ] = COUNT(CRITICAL) < COUNT(ALL)-1
  This feature is useful for cluster monitoring.
  
  
• Some nifty improvements for the command line and the command files:
  1. the report options -r can be specified with a chain of plus-separated numbers instead of a sum:
     -r 1+2+4+8 is better readable than -r 15, isn't it?.
  2. eval and eeval perl snippets don't need to be written with trailing \.
     This allows comment lines within the code as well as direct copy and paste from perl scripts.
     Nevertheless: the old trailing \ is also valid, so nobody needs to rewrite his command files.
     
     

• At last: configure based installation, added tests (make test) and consolidated directory structure.
  By the way: if you have a complicated configure line and deleted your config.log, no problem: call check_multi -V, and it will print the complete configure line:

  nagios~> check_multi -V
  check_multi: v$Revision: 272 $ $Date: 2009-10-15 08:22:59 +0200 (Thu, 15 Oct 2009) $ $Author: flackem $
  configure  '--prefix=/opt/nagios'

  
  

This version is properly working on 1.000 european servers in the data centers of the telecommunication company I work for. If you find oddities or bugs anyway, please report them in the German Nagios-Portal or send me a mail.

Cheers,
-Matthias